USF Libraries
USF Digital Collections

Investigation of geometrical effects on microneedle reliability for transdermal applications

MISSING IMAGE

Material Information

Title:
Investigation of geometrical effects on microneedle reliability for transdermal applications
Physical Description:
Book
Language:
English
Creator:
Shetty, Smitha
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Skin
Drie
Porous silicon
Penetration force
Fracture force
Dissertations, Academic -- Electrical Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Hollow biocompatible microneedle arrays were designed and fabricated using two different bulk micromachining techniques-Deep Reactive Ion Etching and Coherent Porous Silicon technology to investigate their reliability for transdermal applications. An in-house experimental setup was developed for microneedle fracture and split thickness penetration force measurements. Out of plane needle array configurations (100and#956;m needle length) with intra array geometric variations including needle shape, diameter, intra-array pitch and density (1a 625) were characterized on cadaver skin to predict skin barrier penetration without fracture. Use of microneedle array as transdermal patch necessitates reliable penetration and not just pushing against stratum corneum like a bed of nails. Critical in plane fracture tests were conducted on single microneedle columns with different geometry to validate the failure mechanism with force quantification relations.
Thesis:
Thesis (M.S.E.E.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Smitha Shetty.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 102 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001670380
oclc - 62321499
usfldc doi - E14-SFE0001249
usfldc handle - e14.1249
System ID:
SFS0025570:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001670380
003 fts
005 20051216095050.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 051122s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001249
035
(OCoLC)62321499
SFE0001249
040
FHM
c FHM
049
FHMM
090
TK145 (Online)
1 100
Shetty, Smitha.
0 245
Investigation of geometrical effects on microneedle reliability for transdermal applications
h [electronic resource] /
by Smitha Shetty.
260
[Tampa, Fla.] :
b University of South Florida,
2005.
502
Thesis (M.S.E.E.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 102 pages.
3 520
ABSTRACT: Hollow biocompatible microneedle arrays were designed and fabricated using two different bulk micromachining techniques-Deep Reactive Ion Etching and Coherent Porous Silicon technology to investigate their reliability for transdermal applications. An in-house experimental setup was developed for microneedle fracture and split thickness penetration force measurements. Out of plane needle array configurations (100and#956;m needle length) with intra array geometric variations including needle shape, diameter, intra-array pitch and density (1a 625) were characterized on cadaver skin to predict skin barrier penetration without fracture. Use of microneedle array as transdermal patch necessitates reliable penetration and not just pushing against stratum corneum like a bed of nails. Critical in plane fracture tests were conducted on single microneedle columns with different geometry to validate the failure mechanism with force quantification relations.
590
Adviser: Dr. Shekhar Bhansali.
653
Skin.
Drie.
Porous silicon.
Penetration force.
Fracture force.
690
Dissertations, Academic
z USF
x Electrical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1249



PAGE 1

Investigation of Geometrical Effects on Microneedle Geometry for T ransdermal Applications by Smitha Shetty A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Major Professor: Shekhar Bhansali, Ph.D. Sangchae Kim, Ph.D. William Lee, Ph.D. Thomas Koob, Ph.D. Date of Approval: July 19, 2005 Keywords: Skin, Drie, Porous Silicon, Penetration Force, Fracture Force Copyright 2005 Smitha Shetty

PAGE 2

DEDICATION To My Parents

PAGE 3

ACKNOWLEDGEMENTS This project is supported by the National Science Foundation (NSF) Career Award 0239262. I would like to thank my major advisor, Dr. Shekhar Bhansali for providing me a n opportunity to pursue active research in MEMS. I am indebted to Dr Sangchae Kim for his gui dance, technical expertise and insights on the research which enabled me to hold a firm gras p on mechanical aspect of this project. Special thanks to Dr Tom Koop for his invaluable sug gestions and guidance especially during the microneedle testing and data analysis phase. I would also l ike to thank Douglas Pringle at Shriners hospital for his help with skin and polymer compress ive and indentation tests. I would like to acknowledge Dan Hernandez form Shriners hospit al for his help on frozen section on cryotome for skin imaging. I would like to acknowledge Star C enter, Largo for DRIE processing and provision of motorized micrometer. Thanks to Bill Pic kens at the Skin Science Institute (Cincinnati, Ohio, USA) for help in obtaining the skin spe cimens and information on skin. Jay Beiber, Robert Tufts and Richard Everly from NNR C need to be acknowledged for their help while operating cleanroom and metrology tools during f abrication. Finally, my thanks to Helen for her help with the skin testing and my fellow MEMS teammates for their help in the fabrication process. Last but not the least I am appreciative to Sunny for his motivation, support and guidance all throughout my stay at USF.

PAGE 4

i TABLE OF CONTENT LIST OF TABLES.......................................................................................................................... iv LIST OF FIGURES........................................................................................................................vi ABSTRACT ...............................................................................................................................x CHAPTER 1 INTRODUCTION.................................................................................................1 1.1 Motivation.............................................................................................................................1 1.2 Thesis Overview....................................................................................................................3 1.3 Applications for Microneedle................................................................................................4 1.4 State -of-Art Microneedle Research......................................................................................5 1.4.1 Solid Silicon Microneedle Array by Black Silicon Method...........................................5 1.4.2 Bulk Micromachined Multichannel Silicon Neural Probes............................................5 1.4.3 Surface Micromachined Hollow Metallic Microneedles...............................................6 1.4.4 Hollow Deep Reactive Ion Etching (DRIE) Based Etching Needle Arrays..................7 1.4.5 Polysilicon Molded Microneedle Array.........................................................................8 1.4.6 Silicon Microneedles......................................................................................................9 1.5 Significance of Current Work................................................................................................9 CHAPTER 2 SKIN AND MICRONEEDLE DESIGN.............................................................11 2.1 Understanding Skin Barrier.................................................................................................11 2.1.1 Structure and Significance of Skin Barrier...................................................................13 2.1.2 Biomechanical Properties of Skin................................................................................14

PAGE 5

ii 2.2 Design Consideration for Microneedle Strength.................................................................15 2.2.1 Buckling Analysis........................................................................................................16 2.2.2 Parametric Study for Buckling Analysis......................................................................17 2.2.3 Shear and Bending Failure Analysis............................................................................19 2.2.4 Shear Stress Analysis using ANSYS...........................................................................21 CHAPTER 3 MICRONEEDLE FABRICATION.....................................................................23 3.1 DRIE Based Microneedles..................................................................................................23 3.1.1 Introduction: DRIE.......................................................................................................23 3.1.2 Detailed Process Flow and Results..............................................................................24 3.2 Porous Silicon Based Microneedles....................................................................................34 3.2.1 Introduction to Macroporous Silicon...........................................................................34 3.2.2 Fabrication of Porous Silicon based Microneedles......................................................37 3.3 DRIE Vs. Porous Silicon Based Microneedle Fabrication Process.....................................43 CHAPTER 4 BIO-MECHANICAL CHARACTERISATION.................................................44 4.1 Introduction.........................................................................................................................44 4.2 Experimental Setup..............................................................................................................44 4.3 Calibration of Load Cell......................................................................................................47 4.4 Calibration of Motorized Micrometer.................................................................................50 4.5 Mechanical Tests on Skin like Polymer and Split Skin.......................................................51 4.5.1 Compressive Tests on Polymer and Skin.....................................................................52 4.5.2 Indentation Test on Polymer and Split Thickness Skin...............................................56 CHAPTER 5 FRACTURE AND PENETRATION TESTING.................................................61 5.1 Measurement of Fracture Force...........................................................................................61 5.1.1 Experimental Plan........................................................................................................61 5.1.2 Fracture Results Analysis.............................................................................................62

PAGE 6

iii 5.2 Insertion Testing on Skin like Polymer...............................................................................68 5.3 Penetration Testing on Split Thickness Skin.......................................................................71 5.3.1 Skin Tests Using Manual Translation..........................................................................71 5.3.2 Skin Test Using Motorized Micrometer.......................................................................75 5.4 Confirmation of microneedle penetration as against indentation on split th ickness skin....79 5.5 Penetration Tests on Isolated Stratum Corneum.................................................................81 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK ........84 6.1 Conclusions.........................................................................................................................84 6.2 Recommendations for Future Work....................................................................................86 REFERENCES ............................................................................................................................87

PAGE 7

iv LIST OF TABLES Table 2.1 Microneedle Design Constraints....................................................................................17 Table 2.2 Analytical Buckling Results..........................................................................................18 Table 2.3 Mathematical Analyses for Shear Force........................................................................20 Table 3.1 Design Patterns for Geometrical Investigation..............................................................27 Table 3.2 RIE Process Parameters.................................................................................................31 Table 3.3 Bosch Cycle in DRIE.....................................................................................................32 Table 3.4 Comparison of Porous Silicon and DRIE Processes.....................................................43 Table 4.1 Comparison Between Experimental and Data Sheet Readings......................................47 Table 4.2 Calibration Table for Motorized Micrometer................................................................51 Table 4.3 Polymer Plug Specimen Dimension Chart....................................................................53 Table 4.4 Elastic Modulus of Polymer Plugs at Different Strain Levels .......................................53 Table 4.5 Skin Plugs Dimension Chart..........................................................................................53 Table 4.6 Elastic Modulus of Skin Plugs at Different Strain Levels.............................................55 Table 4.7 Indentation Stiffness at 1mm/sec and 5mm/sec Indentation Rates................................58 Table 4.8 Indentation Stiffness for Cadaver Skin..........................................................................60 Table 5.1 Microneedle Dimensional Variation for Fracture Failure Tes ts....................................62 Table 5.2 Fracture Failure Test Results.........................................................................................64 Table 5.3 Theoretical Fracture Force Values for Short Beam Structu res Derived from Material Strength (Compressive Strength of Thermal SiO 2 =690-1380 MPa)..............................................65 Table 5.4 Penetration Force Summary with Needle Specifications...............................................75

PAGE 8

v Table 5.5 Split Thickness Penetration Results Using Motorized Meter................................ ........78 Table 5.6 Stratum Corneum Penetration Results...........................................................................82

PAGE 9

vi LIST OF FIGURES Figure 1.1 Solid Silicon Microneedle Array Fabricated by Black Silicon Me thod [6]...................5 Figure 1.2 Bulk Micromachined Multichannel Silicon Neural Probes [5]......................................6 Figure 1.3 Surface Micromachined Metallic Array........................................................................7 Figure 1.4 DRIE Based Needles by Gardeneir et. al.......................................................................7 Figure 1.5 Sharp Needle Array by Stoeber et al............................................................................8 Figure 1.6 Pointed, Side Opened Out of Plane DRIE Needles by Griss et. al................................8 Figure 1.7 Microneedles by Two Wafer Micromolding Process.....................................................9 Figure 1.8 Silicon needles by Lin and Pisano..................................................................................9 Figure 2.1 Schematic Illustrating the Layers of Human Skin [14]................................................12 Figure 2.2 Schematic Illustrating Arrangement of Corneocytes in Stratum C orneum [15].........13 Figure 2.3 Elastic Behaviour of Ligament[17].............................................................................14 Figure 2.4 Compressive Behaviour as a Function of Length.........................................................15 Figure 2.5 Buckling Failure Mode.................................................................................................16 Figure 2.6 Cross-sectional Dimensions for Buckling Calculations...............................................16 Figure 2.7 ANSYS Modeling for Microneedle Column...............................................................21 Figure 2.8 Von Mises Simulation Results Using Finite Element Analysis ...................................21 Figure 3.1 DRIE Based Microneedle Process Flow......................................................................25 Figure 3.2 Microneedle Mask Layout in Coventorware................................................................26 Figure 3.3 1x1 Array of Width 40 m and 5x5 Array of Width 60 m...........................................30 Figure 3.4 25x25 Array after Al Etching.......................................................................................31

PAGE 10

vii Figure 3.5 Microloading Effect in Die 34 Illustrating Etch Depth Variat ion with Increasing Window Size on Masking Layer in DRIE.....................................................................................33 Figure 3.6 Scalloping Effect in DRIE Illustrating Surface Roughness.........................................33 Figure 3.7 Final Microneedle Array..............................................................................................34 Figure 3.8 I-V Characteristics Governing Electrochemical Dissolution of Silicon [23]...............35 Figure 3.9 Variation in Pore Cross-section if Doping Density or Bias is I ncreased [25]..............37 Figure 3.10 Process Flow for Porous Silicon Based Microneedles...............................................38 Figure 3.11 Photograph Illustrating the Porous Silicon Etching Setup at USF.............................40 Figure 3.12 Experimental Setup for Macroporous Silicon Etching Setup at USF.........................41 Figure 3.13 SEM Image Illustrating 5x20 m Macropore Array, Height=150 m after Etching for 24 hrs..............................................................................................................................................41 Figure 3.14 20x100 m Array with Branched Pores (Bias 2.5V)...................................................42 Figure 3.15 20x100 m Array with Reduced Branching (Bias 2V)...............................................42 Figure 3.16 Blocked, Out of Plane Microneedles.........................................................................42 Figure 4.1 Testing Module.............................................................................................................45 Figure 4.2(a) Load Cell Block (Top) (b) Microneedle Block ( Bottom).......................................46 Figure 4.3 Calibration Setup..........................................................................................................47 Figure 4.4 Best Fit Graph for Experimental Values Compared with Standard Da ta Values.........48 Figure 4.5 Experimental Calibration Readings..............................................................................48 Figure 4.6 Illustration of Load Cell Output Variation over Time under (a) Nor mal Conditions (b) Partial Shield (c) Complete Shield (d) Variation over 3 hrs Time wit h Metal Fixture..................50 Figure 4.7 Compressive Tests: Stress-Strain Relation for Polymer Pl ugs....................................53 Figure 4.8 Stress-Strain Relation for Skin Plugs on Compressive Loading..................................54 Figure 4.9 Comparative Analysis Between Elastic Modulus of Polymer and S plit Thickness Skin .......................................................................................................................................................55

PAGE 11

viii Figure 4.10 Force-Depth Plots for Indentation Rate of 1mm/sec..................................................57 Figure 4.11 Force-Depth Plots for Indentation Rate of 5mm/sec..................................................57 Figure 4.12 Force Normalisation for Stiffness Calculation at Position 1-10mm........................... 58 Figure 4.13 Relaxation Test Result Illustrating Elastic Behaviour of Pol ymer with Small Viscous Element (Straight Step Plot Represents Instantaneous Step Strain whil e Wavy Curve Indicates Force Values).................................................................................................................................59 Figure 4.14 Indentation Plot for Split Thickness Skin at 1mm/sec................................................59 Figure 5.1 Typical Plot Displaying Fracture Peak for 40 m wide square microneedle of length 125 m............................................................................................................................................63 Figure 5.2 Fracture Test Results as Function of Needle Length and Width..................................65 Figure 5.3 Force-Length Relation for 40 m Wide Square Microneedle.......................................66 Figure 5.4 Force-Length Relation for 40 m Wide Circular Microneedle.....................................66 Figure 5.5 Force-Length Relation for 60 m Wide Square Microneedle......................................67 Figure 5.6 Force-Length Relation for 60 m Wide Circular Microneedle.....................................67 Figure 5.7 Fracture Force vs Width (Length=constant) for Circular Geomet ry............................68 Figure 5.8 Wall Buckling after TMAH Etch (10x-Optical Microscope).......................................68 Figure 5.9 Illustrating the Condition of 5x5 Microneedle Array (Width 40 m, Pitch 150 m and Length 75 m) Before and After Needle Insertion. Needles show Polymer Trace in the Needle Lumen after insertion.....................................................................................................................69 Figure 5.10 Penetration Plot for 5x5 Array with Penetration Discontinuity Obs erved at 0.2gF...70 Figure 5.11 Penetration Plots for 25x25 Array (Width=60 m, Pitch=150 m, Length=75 m) with First Force Peak at 0.8 gF..............................................................................................................70 Figure 5.12 Typical Force-Time Plot for 25x25 Array Marked by a Number of Peaks Due to Non Uniform Insertion Rate..................................................................................................................72

PAGE 12

ix Figure 5.13 Optical Microscope Image Illustrating Black India Ink Stai ns on Cadaver Skin after Insertion for (A) 25x25 Array –Circular Needles (Width=20 m, Pitch=100 m, Height=125 m)72 Figure 5.14 Microneedle While Being Removed from the Skin Sample Indicating P enetration..73 Figure 5.15 SEM Image of 25x25 Array of Circular Needles (Width=20 m, Pitch=100 m, Height=125 m) after Skin Insertion..............................................................................................73 Figure 5.16 SEM Image Illustrating Clogging of Needle with Skin..............................................74 Figure 5.17 SEM Image Showing Effective 4.5 m Clogging of Needle Lumen..........................74 Figure 5.18 Magnified Peak in Skin Penetration Force-Time Plot for 25X25 Arra y (Circular Needle-Width 60 m, Pitch 150 m, Length 105 m) Indicating 4.03gF as Penetration Force.......76 Figure 5.19 Typical Skin Penetration Force-Time Plot for 25X25 Array (Cir cular Needle-Width 60 m, Pitch 150 m, Length 105 m) with Motorized Micrometer...............................................77 Figure 5.20 Plot Illustrating Work Done by 25 X 25 Needle Array (Width 60 m, Pitch150 m, Length 105 m )-Area Between Force Displacement Curves for Chips with and without Ne edles. .......................................................................................................................................................78 Figure 5.21 Optical Microscope Image (4X Magnification) Illustrating Micr oneedle Penetration Marks.............................................................................................................................................80 Figure 5.22 Optical Microscope Image (20X Magnification) Illusrating (A) Spl it Thickness Skin Section Without Penetration (B) Split Thickness Skin Section with 9 0 m Deep,20 m Wide Needle Marks into Epidermis........................................................................................................80 Figure 5.23 Penetration Marks of 25X25 Microneedle Array (Square Needle-Wi dth 40 m, Pitch150 m, Length 100 m) on SC after Insertion Test with Magnified Single Needle Mark....82 Figure 5.24 SC Penetration Plot for 25X25 Microneedle Array ((Square Needle-W idth 40 m, Pitch150 m, Length 100 m).........................................................................................................82

PAGE 13

x INVESTIGATION OF GEOMETRICAL EFFECTS ON MICRONEEDLE RELIA BILITY FOR TRANSDERMAL APPLICATIONS Smitha Shetty ABSTRACT Hollow biocompatible microneedle arrays were designed and fabricated us ing two different bulk micromachining techniques-Deep Reactive Ion Etching and Cohe rent Porous Silicon technology to investigate their reliability for transdermal a pplications. An in-house experimental setup was developed for microneedle fracture and split thic kness penetration force measurements. Out of plane needle array configurations (100 m needle length) with intra array geometric variations including needle shape, diameter, intra-array pitch and density (1~625) were characterized on cadaver skin to predict skin barrier penetration without fracture. Use of microneedle array as transdermal patch necessitates reliable pene tration and not just pushing against stratum corneum like a “bed of nails”. Critical in plane fractu re tests were conducted on single microneedle columns with different geometry to validate the failure mechanism with force quantification relations. Preliminary penetration characterization was performed on skin like polymer followed by direct testing on cryogen preserved cadaver skin. Compress ive and indentation test were performed on both excised skin and polymer to analyze thei r mechanical behavior on loading and establish a mechanical correlation. Finite element modeling using ANSYS was done to examine the effect of shear loading on the needles due to lac k of experimental verification.

PAGE 14

1 CHAPTER 1 INTRODUCTION 1.1 Motivation Rapid advancement in pharmaceutical industry has necessitated dev elopment of physical enhancement techniques for transdermal applications to overcome li mitations of oral drug administration. These include poor absorption in intestine and liv er (first pass metabolism effect) and sensitivity to enzymatic degradation. As compared to oral treat ment, topical route provides large surface absorption area and negligible degradation of dr ug. For effectual delivery, the drug has to diffuse through the outermost barrier of the skin, stratum cor neum, at therapeutic rate and reach the blood vessel located in the dermis. Several methodol ogies [1] are being investigated to increase the permeability of stratum corneum: Electrical based techniques: iontophoresis, electroporation, ultra sound, photomechanical wave Structure based techniques: microneedles Velocity based techniques: jet propulsion. Velocity and electrical based delivery systems tend to be unreliable since they induce skin irritation, burns and shock the cells. Hence there has been an incr eased emphasis on development of structure based techniques. Microneedles are emerging as critical drug delivery and biofluid extraction mechanism owing to advances in microfabrication technology. As compared to comm ercially available hypodermic gauge needles (needle length in mm range and thickness greater than 300 m), they are much smaller (diameter less than 100 m) and exhibit features like minimal pain and tissue

PAGE 15

2 trauma during skin insertion. They also provide increased control over drug dosage, independent of drug composition and concentration. Due to short lengths, microneed les increase the skin permeability without stimulating the nerve endings located in the dermis enabling minimally invasive drug administration. The size and geometry of these mic roneedles are lithographically defined so that they can be fabricated in accordance with the appl ication requirement. In addition, microneedles offer mass production capabilities since fabricat ion processes are conducive for batch production making them cost effective. Furthermore, the out -of-plane structure translates to their scalability into a multi-array configuration. To accelerate their inclusion into mainstream medicine and to a dvance their scope of applicability, microneedles need to conform to requirements of sim ple design, low price and high reliability. Design and cost factors have been sufficientl y optimized in past research; however, performance studies specific for biomedical applications have been inadequate. While exploring the avenue of microneedle array patch for transdermal applicati on, it is critical to ensure that all needles reliably penetrate through the stratum corneum and not just push against it like a “bed of nails”. Studies have been conducted to study the mechanics of singl e needle insertion into skin [2]; however effect of intra-array geometry on safe dermal insertion is still an unreported research domain. Current research focuses on fabrication and bio-mechanical char acterization of out of plane, hollow Silicon dioxide microneedle array chips exhibiting intra-arr ay parametric variation : needle width (5~60 m), cross-section (square and circular), intra-array pitch( 20~200 m) and needle density (1~625). Two different approachesDeep Reactive Ion Etching (DRIE) and Coherent Porous Silicon technology (variation in the technique to obtain bulk micromachined straight wall pores) were implemented to fabricate multi-g eometry microneedle arrays. Critical in-plane buckling and penetration tests were conducted to study ne edle efficacy. Preliminary characterization was performed using artificial skin polymer (palpability similar to skin) before

PAGE 16

3 direct testing on skin. Polymer Compressive and indentation tests were carried out to correlate its properties to excised cadaver skin. Finite element modeling using ANSYS TM was done to corroborate the experimental results obtained after buckling and i nsertion tests. This research would be further extended to develop implantable needle array f or continuous body monitoring for medical conditions. 1.2 Thesis Overview Chapter one briefly discusses the necessity and current appli cations of microneedles in biomedicine. An overview of existing microneedle research is pres ented with an analysis of the functionality and limitations of several proposed schemes in the literature. Chapter two elaborates on heterogeneous skin anatomy discussing the structure and properties of the physical layers. The design consideration for needle array has also been investigated fo r reliable penetration overcoming skin barrier. This chapter presents buckling and she ar stress analysis in a column analogous to the needle structure and analyses results of finite element analysis (ANSYS) used to model the effect of varying needle dimension and shape on buckling. Cha pter three discusses the two fabrication methodologies employed to realize hollow needles, hi ghlighting the rationale behind adoption of the processes for current study. Chapter four pres ents the mechanical characterization of successfully fabricated microneedles, describing the experimental setup in detail. Stressstrain results obtained from compressive tes ts and indentation tests on commercially available polymer employed as skin substitute has been included for correlation studies. This is followed by buckling and real skin testing and discussion. Chapter Five summarizes the work and provide recommendations towards incorpora tion of fabricated needles for realization of implantable sensors for body monitoring.

PAGE 17

4 1.3 Applications for Microneedle Recently, bioinstrumentation research has exhibited a growing i nterest in micro system technologies for development of biodevices owing to miniaturiza tion, increased functionality, bulk fabrication capabilities and reliability. Microneedles ar e finding increasing application in all areas of biomedical instrumentation including drug delivery, diag nostics, neural and minimally invasive surgery. Continuous body fluid monitoring for diseased conditions like diabetes : Microneedles could be used in a feedback mechanism to monitor the blood sugar le vel and administer therapeutics like insulin in precisely controlled amounts [3]. Cellular delivery: Cellular and molecular biology genera lly requires injection of membrane impermeable molecules like DNA, proteins and other ge netic entities into cells. Arrays of solid silicon microprobes and hollow glass capil laries have been successfully utilized for injecting DNA into tobacco cell congl omerates [4]. Further sharpening of these microcapillaries would enable their application for bacteria transfer. Neural stimuli to cortical membrane and electrical signal r ecording: Micromachined silicon neural probes with microchannels developed at Universit y of Michigan [5] have been interfaced to the neurons with minimal tissue disruption for deliv ery of neurosimulating drugs while simultaneously recording electrical si gnals. These probes significantly indicate contribution of MEMS towards neuroscience instrum entation. Antibiotic administration in controlled quantity Intravascular drug delivery for stenotic arteriosclerosis t reatment: The efficacy of antirestenotic drugs prescribed to prevent reblockage of arter ies is hindered due to difficulty in local delivery to the clogged arteries. Coronary stents could be developed with microprobes on the periphery for delivering anti-clogging ag ents [4]. Silicon

PAGE 18

5 microprobes of 140 m length have been successfully pierced into plaqued rabbit iliac arteries in vitro at 500 mm Hg pressure. 1.4 State -of-Art Microneedle Research Many fabrication approaches utilizing different design schemes and materials have been employed for microneedle development. These processes are aimed t owards optimization of geometry and process cost targeting various applications. 1.4.1 Solid Silicon Microneedle Array by Black Silicon Method Early work in the area of transdermal microneedle arrays was reported by Henry et. al [6], testing solid needles to ensure if micro-holes created after ins ertion increase the skin permeability. 20x20 needle arrays (dimension: length 180 m, diameter 50-80 m and tip radius close to 1 m) were fabricated by Black Silicon process (Silicon reactive i on etching using SF6/O2) followed by lateral under etching needles as shown in Figure 1.1. Authors confirmed the microneedle strength to pierce skin without fracture to enable therapeutic diffusi on of calcein in skin by three orders of magnitude. However this solid model design was inefficient for large volume drug delivery and fluid extraction. Figure 1.1 Solid Silicon Microneedle Array Fabricated by Black Silicon Method [6 ] 1.4.2 Bulk Micromachined Multichannel Silicon Neural Probes Biocompatible Neural probes with hollow buried channels developed by Chen, Wise et al [5] contributed significantly to needle research owing to highly loc alized drug delivery and in situ

PAGE 19

6 chemical stimuli monitoring at the cortical cellular level a s illustrated in Figure 1.2. Flow channels with precisely controlled widths (10 m width) and shank length (4mm) were defined by anisotropic wet etching of Silicon with highly boron doped Silicon as ma sk. The channels were sealed by thermal oxide followed by LPCVD dielectrics to house t he electrodes for recording and stimulation on the same chip. Polyimide tubes are accommodated on fl uid port on the rear end. The probes were reported to successfully examine the neural r esponse to local application of specific medication. Figure 1.2 Bulk Micromachined Multichannel Silicon Neural Probes [5] 1.4.3 Surface Micromachined Hollow Metallic Microneedles Hollow fluid coupled Palladium needle arrays with improved functiona lity like mechanical penetration stops, microbarbs and multiple outlet ports was developed [7] as shown in Figure 1.3. This approach marked an important step towards batch de velopment of metallic needles. Palladium layer was electroplated into the patterne d photoresist layer to form the bottom shell. Inner lumen was defined by 40 m thick sacrificial layer photoresist. After sputtering t he seed layer, Palladium is electroplated into the photoresist m icromold to form the top and side walls for the needle. The needles are then released from Silicon surface by etching the base seed layer. The resulting structure was structurally more robust th an previous designs [8] due to inclusion of microrivets between bottom and side walls. These m icromachined metal arrays with

PAGE 20

7 greater taper angle were reported to penetrate skin like photoel astic material with 50% failure rate. They also demonstrated high fluid flow rates of the order of 4000 l/hr when subjected to fluid pressure of 1MPa, enabling them as apt drug delivery microdevices [9]. Figure 1.3 Figure 1.4 Figure 1.3 Surface Micromachined Metallic Array Figure 1.4 DRIE Based Needles by Gardeneir et. al 1.4.4 Hollow Deep Reactive Ion Etching (DRIE) Based Etching Needle Ar rays One of the approach reported by Gardeneirs et. al [10] combines th e anisotropic wet etching and DRIE technique to produce out of plane Silicon needle s with sharp off-center flow channel. They were addressing the perceived disadvantage of fl at hollow needles for transdermal application; they get clogged by skin which tends to block the fluid fl ow during needle insertion. The essential feature includes the shape of the needle w ith a wider base, narrow needle tip and side defined flow channel on the (111) Silicon plane obtained after ani sotropic wet Silicon wet etching as illustrated in Figure 1.4. Another DRIE based technique was suggested by Stoeber et. al [11] utilizing two lithography steps. The process flow deployed anisotropic DRIE t o define backside through opening, protecting side walls by nitride layer followed by isotr opic wet and plasma etching of the patterned front to form pointed out of plane 200 m and 40 m diameter structures. Sharp tips were created due to the offset of the center lines of t he two etch masks as shown in Figure 1.5

PAGE 21

8 This structure improves the mechanical stability of the needl es with superior penetration abilities. However the drawback with this design was clogging of the channel. Griss and Stemme [12] developed a similar procedure to obtain side opene d geometry for Silicon microneedles with high structural stability and minim ized blockage. Flow channel was anisotropically etched in a way similar to the above described process. This is followed by ICP and anisotropic etching to underetch the oxide front mask and form the cross structures respectively without the side opening. A subsequent isotropic e tch step opens up the side walls. Figure 1.6 illustrates the resulting needles. Figure 1.5 Figure 1.6 Figure 1.5 Sharp Needle Array by Stoeber et al Figure 1.6 Pointed, Side Opened Out of Plane DRIE Needles by Griss et. al 1.4.5 Polysilicon Molded Microneedle Array Hollow polysilicon hypodermic microneedles were fabricated using two wafer micromolding process developed by Zahn et al. [13] illustrated in F igure 1.7. The mold wafer is patterned with the needle shape on the front side which is aligned to a through hole on the backside (etched using KOH). This needle mold is then etched by DR IE followed by Phosphosilicate glass (PSG) deposition. This mold is bonded to another P SG coated bare silicon wafer. Subsequent process includes polysilicon deposition onto the mold, a nnealing and wafer release in HF. Since polysilicon is ceramic, cracking is a crucial cause of failure. This was

PAGE 22

9 precluded by the diffusion of phosphorus from PSG into Polysilicon during annealing, as it develops compressive stresses that combat crack propagation. T his structure also supported additional features like microfilters, bends and outlet ports. Figure 1.7 Figure 1.8 Figure 1.7 Microneedles by Two Wafer Micromolding Process Figure 1.8 Silicon needles by Lin and Pisano 1.4.6 Silicon Microneedles Lin and Pisano[14] demonstrated IC compatible fabrication of mi croneedles with facility for on board polysilicon heaters for bubble powered micropumps, ports f or fluid flow and base interface region for incorporating drive electronics and fluidic devices. The hypodermic needles as shown in Figure 1.8 were processed by a combination of surface and bu lk micromachining techniques. The flow channel extends around 1-6 mm long, 9 m in height and 80 m wide. These needles were mechanically more robust due to thicker sidewall s (70 m) and reliably penetrated muscle tissue (steak) without bending/breakage. 1.5 Significance of Current W ork Current research work focuses on design and fabrication of bulk mi cromachined silicon dioxide needle arrays to investigate the effect of geometry for transdermal applications without fracture. This work contributes significantly to the existing techno logies as follows:

PAGE 23

10 Ease in fabrication: In this work, out-of-plan array structures w ere fabricated using DRIE and porous silicon etching methods, both enabling mass production capabilities. Biocompatibility: Silicon dioxide is known to be biocompatible hence these needles can form a part of an implantable device without susceptibility to body corrosion. Multidimensional chips from one Silicon wafer: Since the needl es are lithographically patterned, it is possible to obtain different geometry chips wit h the same photo mask to investigate biomechanical characteristics. Length control: The needle length can be precisely controlled by anis otropic wet etching. Hence one can control the penetration depth into the skin enabli ng local drug administration. Simulation and validation of models with real skin tissue.

PAGE 24

11 CHAPTER 2 SKIN AND MICRONEEDLE DESIGN 2.1 Understanding Skin Barrier Skin is referred as a prototypical “smart” material pro viding a dynamic interface between the environment and the human body. Skin maintains the water homeostasi s of the body and protects the internal organs from damage. This chapter presents an overview of the skin anatomy followed by detailed discussion on the structure and properties of the skin barrier-stratum corneum. The equivalent microneedle model and motivation for proposed design parameters for reliable skin penetration are explained. Skin is heterogeneous in nature, comprising of three layers epidermis, dermis and hypodermis each differing in physiology, thickness and function as illus trated in Figure 2.1. Cellular epidermis forms the upper protective region compose d of stratified epithelial cells called keratinocytes. These cells continuously regenerate as cuboidal c ells that differentiate and migrate from the basal layer (region that separates epidermis from dermis) during the process of desquamation. This 0.1-1mm thick region is further classified in to following sub layers: stratum corneum (outermost region consisting of 10-30 layers of cornified c ells), stratum lucidium (found in thick regions), stratum granulosum (granular cell region r egulating water loss), stratum spinosum (2-7 layered spinous cells) and stratum basale (one c ell thick cuboidal cell layer). Dermis is next “live” region approximately 1-2mm thick and en riched with sensory receptors, blood vessels, hair follicles and integumentary glands. The connect ive tissue content (i.e. elastin

PAGE 25

12 and collagen fibre) in the dermis is responsible for the tensile strength and elasticity of skin layer. Hypodermic needles are invasive since they stimulate the nerv e endings in the dermis region and drugs penetrate into the blood vessels. Hypodermis constitutes the ba sal network of loose connective and adipose tissue that binds the skin to the underlying org ans. This layer plays an important role in metabolism, serving as insulation padding against injur y. Figure 2.1 Schematic Illustrating the Layers of Human Skin [14] The stratum corneum forms the primary region of interest for st udying the permeability and transport mechanism through the skin. It is essential to under stand the structural and functional details of this skin barrier.

PAGE 26

13 2.1.1 Structure and Significance of Skin Barrier Stratum corneum, the topmost layer of the epidermis is esse ntially 5-20 m thick and forms an effective barrier to the milieu owing to its unique wa ter retention ability. Stratum corneum is comprised of matrix of hexagonal flat cells called cor neocytes, embedded in lipid rich intracellular space. Corneocytes are about 30 m in diameter and 0.3 m thick, surrounded by cornified envelope and contain horizontally arranged keratin fibril s which can retain water in the vertical direction. “The corneocyte envelope consists of two parts: a thicker protein envelope (~15nm) composed of cross-linked structural proteins adjacent to the interior cytoplasm and a thinner (~4nm) lipid envelope on the intercellular face of the pr otein.”[15]. These units are interconnected by means of protein rich rivet like structures called corneodesmosomes as illustrated in Figure 2.2. The lipid components incorporate mixture of f atty acids, ceramides, sterols and cholesterol esters arranged in bilayer form. Compounds can penetrate into the skin through the soft intracellular lipids, protected from mechanical abrasi on by corneocytes. Figure 2.2 Schematic Illustrating Arrangement of Corneocytes in Stratum Corneum [ 15] Several models have been proposed to explain the stratum corneum architecture [15], earliest concept being the “brick and mortar model”. This struct ural scheme suggested the two heterogeneous compartment system with corneocytes (protein loade d bricks) arranged in lipid phase mortar (intercellular lipid). A diffusion based model “Doma in mosaic model” was later proposed by Forlind which stated that bulk of stratum corneum lipids are arranged in domains

PAGE 27

14 with a crystalline packing minimizing penetration of water. “ These crystalline domains have fringes with lipids in a liquid crystalline phase, permitting diffusion of wa ter” [15]. Under normal conditions, there would be no penetration through the c orneocytes. The penetrating substances need to penetrate through the lipids in intercell ular space. 2.1.2 Biomechanical Properties of Skin Skin is a viscoelastic material with its properties varying as f unction of its heterogeneous composition, strain rate, hydration content and age. Young’s modulus obta ined through stressstrain characterizes skin elasticity. The elastic modulus of skin varies from 2-12MPa and increases with age [16]. The typical stress strain curve for ligament on tensile loading biomechanism similar to skin is demonstrated in Figure 2.3 Figure 2.3 Elastic Behaviour of Ligament [17]

PAGE 28

15 On loading the skin tissue, initially the strain increases w ith small increase in stress. This is associated to straightening of elastin owing to breaking of cross bond between elastin molecules in the dermis layer. On further loading, the strain va ries linearly with stress indicating elasticity. Even before the yield stress is reached, some of the collagen fibers begin to develop micro-fractures. If the loading is further continued, the tissue permanent ly ruptures. 2.2 Design Consideration for Microneedle Strength For effectual drug delivery, microneedles need to have suffic ient strength and length to penetrate through the lipid layer without fracture, necessit ating a judicious choice of design parameters Failure mode analysis is executed in order to realize the mechanical strength of the needles and set the limiting conditions for the design parameter s. The equivalent model of the needle is established and simulated using Finite element mode ling. This study investigates the vulnerability of the needle to failure due to following mechanis m: buckling and shear. Compressive failure is a measure of slenderness ratio (funct ion of length and radius of gyration) and material property (Young’s Modulus and yield strength) as illust rated in Figure 2.4. Buckling occurs in the case of long needles and for stresses less than yie ld strength of material. Fracture is probable failure mode for short column governed by mechanical property of nee dle. Figure 2.4 Compressive Behavior as a Function of Length

PAGE 29

16 2.2.1 Buckling Analysis Typically microneedles are modeled as long rectangular or cyl indrical columns. During needle penetration into a membrane, if the axial compressive l oads exceed the critical load as defined in Euler’s equation stated below, the needles buckle as show n in Figure 2.5. This loading setup for the microneedle array can be considered similar to l oading of array of hollow beam fixed at base. Euler’s equation gives a mathematical expressio n of the buckling force for a beam of as a function of its length L: 2 2 4 L EI F bucklingp= ………………………………………………………………….. (2.1) Where E=Modulus of Elasticity I = Area moment of Inertia = (b 1 h 1 3 -b 2 h 2 3 )/12 ……………………………………………………………….. (2.2) (For rectangular hollow beam as shown in Figure 2.6) = (d 2 4 -d 1 4 )/64 ……………………………………………………………….. (2.3) (For hollow beam with circular cross section as shown in Figure 2.6) Figure 2.5 Figure 2.6 Figure 2.5 Buckling Failure Mode Figure 2.6 Cross-sectional Dimensions for Buckling Calculations h1 b2 h2 b1 d2 d1

PAGE 30

17 2.2.2 Parametric Study for Buckling Analysis Microneedle design was governed by basic assumption of maximum fr acture force that a microneedle can withstand to be 5gF/needle. From Euler’s equation one can derive an optimized relation between width of square needles (b) and needle length (L) with fixed needle wall thickness ( =1.5 m), restricted by fabrication limitation. From (2.2), I= (b 1 h 1 3 -b 2 h 2 3 )/12= (b+2 ) (h+ 2 ) 3 (b) (h) 3 /12 Neglecting higher order increments, I= h 2 (3b+h)/12= b 3 /3………….. (For square geometry, b= h) From (2.1), F buckling = 2 E b 3 /12L 2 >50 For =1.5 m, b 3 /L 2 >5 This constraint leads to determination of the design chart as illust rated in Table 2.1. Table 2.1 Microneedle Design Constraints DESIGN PARAMETERS TYPE Geometry Square, Circular Needle Width 5 m,10 m,20 m,40 m,60 m Needle Length 25~175 m Wall thickness 1.5 m Intra array needle pitch 20,50,100,150,200 Needle density 1,5x5,25x25 Typical values for elastic modulus and yield strength of bulk SiO 2 is 70GPa and 8.4GPa respectively [18, 19]. The following parameters are varied to study the ir effect on buckling force: Needle shape: Square and circular cross section Needle length: (25 m~175 m) Needle width: (5 m ~ 60 m) The analytical results have been presented in Table 2.2

PAGE 31

18 Table 2.2 Analytical Buckling Results Microneedle Design Parameter Square Geometry Circular Geometry Width ( m) Length ( m) Thickness ( m) Moment of Inertia (m 4 ) Buckling Force (N) Moment of Inertia(m 4 ) Buckling Force(N) 25 3.39e-2 1.99e-2 50 8.48e-3 4.98e-3 75 3.77e-3 2.21e-3 100 2.12e-3 1.24e-3 125 1.36e-3 7.98e-4 150 9.43e-4 5.54e-4 5 175 1.5 1.23e-22 6.92e-4 7.22e-23 4.07e-4 25 2.68e-1 1.57e-1 50 6.69e-2 3.94e-2 75 2.98e-2 1.75e-2 100 1.67e-2 9.84e-3 125 1.07e-2 6.29e-3 150 7.44e-3 4.37e-3 10 175 1.5 9.69e-22 5.46e-3 5.69e-22 3.21e-3 25 2.16e+0 1.27e+0 50 5.41e-1 3.18e-1 75 2.40e-1 1.41e-1 100 1.35e-1 7.95e-2 125 8.65e-2 5.09e-2 150 6.01e-2 3.53e-2 20 175 1.5 7.83e-21 4.42e-2 4.60e-21 2.59e-2 25 1.75e+1 1.03e+1 50 4.37e+0 2.57e+0 75 1.94e+0 1.14e+0 100 1.09e+0 6.42e-1 125 6.98e-1 4.11e-1 150 4.85e-1 2.85e-1 40 175 1.5 6.32e-20 3.57e-1 3.72e-20 2.09e-1 25 5.92e+1 3.48e+1 50 1.48e+1 8.70e+0 75 6.58e+0 3.87e+0 100 3.69e+0 2.18e+0 125 2.37e+0 1.39e+0 150 1.64e+0 9.67e-1 60 175 1.5 2.14e-19 1.21e+0 1.26e-19 7.10e-1

PAGE 32

19 2.2.3 Shear and Bending Failure Analysis Bending failure is commonly observed for free standing microneedle columns as a result of bending moment produced by shear force acting on needle tips. Fl exure formula states the expression for maximum bending stress generated in a beam given as: = My/I where, y = distance from neutral axis to outer edge of the beam M= Bending moment I = Moment of Inertia about centroid of the beam The maximum shear force that the needle of length L can withstand before fracture is given as: F = I/ Yl Assuming the yield strength for thermal SiO 2 as 8.4GPa, the mathematical analysis for the parametric studies have been summarized in Table 2.3

PAGE 33

20 Table 2.3 Mathematical Analyses for Shear Force Design Parameter Square Geometry Circular Geometry Width ( m) Length ( m) Moment of Inertia (m 4 ) Max shear force (N) Bending moment (N-m) Moment of Inertia (m 4 ) Max shear force (N) Bending Moment (N-m) 25 1.65e-2 1.65e-2 50 8.25e-3 4.85e-3 75 5.5e-3 3.24e-3 100 4.13e-3 2.43e-3 125 3.30e-3 1.94e-3 150 2.75e-3 1.62e-3 5 175 1.23e-22 2.36e-3 4.13e-07 7.22e-23 1.39e-3 2.43e-7 25 6.51e-2 3.83e-2 50 3.25e-2 1.94e-2 75 2.17e-2 1.28e-2 100 1.63e-2 9.57e-3 125 1.30e-2 7.66e-3 150 1.09e-2 6.38e-3 10 175 9.69e-22 9.30e-3 1.63e-06 5.7e-22 5.47e-3 9.57e-7 25 2.63e-1 1.55e-1 50 1.32e-1 7.74e-2 75 8.77e-2 5.16e-2 100 6.58e-2 3.87e-2 125 5.26e-2 3.09e-2 150 4.39e-2 2.58e-2 20 175 7.83e-21 3.76e-2 6.58e-06 4.6e-21 2.21e-2 3.87e-6 25 1.06e+0 6.24e-1 50 5.31e-1 3.12e-1 75 3.54e-1 2.08e-1 100 2.66e-1 1.56e-1 125 2.12e-1 1.25e-1 150 1.77e-1 1.04e-1 40 175 6.32e-20 1.52e-1 2.66e-05 3.72e-20 8.92e-2 1.56e-5 25 2.39e+0 1.41e+0 50 1.19e+0 7.05e-1 75 7.99e-1 4.70e-1 100 5.99e-1 3.53e-1 125 4.79e-1 2.82e-1 150 3.99e-1 2.35e-1 60 175 2.14e-19 3.43e-1 5.99e-05 1.26e-19 2.02e-1 3.53e-5

PAGE 34

21 2.2.4 Shear Stress Analysis using ANSYS Finite Element tool ANSYS 8.0 was used to simulate the maximum stress generated due to application of shear force on a silicon dioxide needle column of l ength 200 m. Von Mises analyses was performed to study the stress and out of plane deflection of needle s for forces range 0.1N-80N. The results from the simulation have been presented in Figure 2.7 and 2.8. Figure 2.7 ANSYS Modeling for Microneedle Column Figure 2.8 Von Mises Simulation Results Using Finite Element Analysis

PAGE 35

22 The results from buckling and shear mode analysis confirm the stur diness of the designed oxide needles, though SiO 2 is known as a brittle material. The chosen span of dimensions provides a rich assortment of data sets to perform the experime ntation, thus enabling moderately accurate extrapolation of results The next step involves the fabrication of the microneedle arra y as per the designed dimension.

PAGE 36

23 CHAPTER 3 MICRONEEDLE FABRICATION This chapter presents a detailed discussion on the fabricati on approaches adopted for development of hollow, out of plane silicon dioxide microneedle arr ays. Two approaches differing in the techniques employed to obtain bulk micromachined ani sotropic pores: Deep Reactive Ion etching (dry process) and Porous Silicon etching (we t electrochemical process) were adopted. However, the post pore formation process remains the same for each of these methods. The process flow details have been presented with description of issues encountered in each approach. 3.1 DRIE Based Microneedles 3.1.1 Introduction: DRIE Also referred as Bosch process, DRIE is anisotropic, cry stal-orientation independent dry etch technique capable of producing High Aspect Ratio Structures (HARS). It is used widely for fabrication of micro system components like capacitors, acceleromete rs, optical switches etc. This is a sub class of inductively coupled plasma etch schemes ca pable of generating high density plasma discharge on RF power coupling to a low pressure gas medi um. The etch rate for Silicon substrates ranges from 1-3 m/min and the variation in selectivity to masking is from 50 t o 100:1 for photoresist and from 120 to 200:1 for oxide mask [20]. The process operates on a continuous time multiplexing scheme consisting of alternate etch and polyme r deposition cycles to obtain a nearly vertical side wall (902). The etch cycle exploits t he high etch rates obtained by ion

PAGE 37

24 bombardment of fluorinated discharges (such as SF 6 ) but yields an isotropic geometry. Highly anisotropic profile is obtained by subsequent side wall passivation cycl e (using C 4 F 8 ) via polymer deposition. The polymer is immediately sputtered away on the horiz ontal surfaces only due to the physical nature of the etching, with the sidewalls remaining una ffected. This behavior results from the directional nature of the accelerated ions. Features like anisotropy, etch rate and selectivity can be controlled through proper choice of RF power, gas pressure a nd time. 3.1.2 Detailed Process Flow and Results A similar process flow was implemented at University of Tok yo to obtain hollow microcapillaries for DNA injection [21] .The process flow des ign for fabrication of open ended microneedle array has been outlined in Figure 3.1. Since the current research is dedicated towards mechanical characterization, the later steps of the process flow have been slightly modified to open closed base needles for ease of fabrication. 4” DSP n-type <100> Silicon wafers of thickness 400-500 m were employed for processing. These wafers were subjected to RCA clean prior to de position of masking layer for further bulk micromachining. Silicon dioxide (selectivity 150:1) and Aluminum (selectivity 300:1) were used as masking materials for DRIE etching of S ilicon. 1.5 m thick oxide layer was thermally grown on three wafers at 1050 0 C after subjecting to 8 hrs of wet oxidation. 3000 A of Al was deposited on three other wafers using AJA International Electron Beam Evaporator. A 5 ” dark field Chrome mask was designed in order to study the effect of multidimensional geometric patterns as discussed in Chapter 2. 68 different design as listed in Table 3.1 were finalized for fabrication and were individually implemented in form of 0.9 by 0.9 square cm feature size on photomask.

PAGE 38

25 Figure 3.1 DRIE Based Microneedle Process Flow Each die is labeled using an alphabet (A for square and C for c ircle) and number (3 m feature size) at the lower right corner of the die. The mask layout was implemented using Coventorware TM as illustrated in Figure 3.2. A) Deposition of masking layer (3000 A evapora ted Al or 1.5 m thermally grown oxide) on 4” Si wafer E) Lap the wafer chips to expose the top silicon layer F) Frontside TMAH Si etch to obtain out of plane hollow Silicon dioxide needle array E) Lap the backside of the wafer chips in order to obtain through needles F) Backside TMAH Si etch to obtain out of plane hollow Silicon dioxide needle array Al Open base hollow needle array D) Strip Photo-resist and etch remaining Al This is followed by dicing into 0.9 square die and thermal oxidation to define needle side wall (1.5 m) B) Pattern masking layer using microneedle mask followed by mask layer etching C) DRIE Silicon etching400 Bosch cycles Closed base hollow needle array

PAGE 39

26 Figure 3.2 Microneedle Mask Layout in Coventorware The following steps were carried out to define the microneedle region on the masking layer: Spin primer HMDS and positive photoresist S1827 on the mask lay er at 2500rpm for 30 sec @ 100rpm/sec acceleration using P-6000 resist spin coater. Th is program yields around 3 m thick layer which provides good selectivity to silicon dioxid e during subsequent oxide reactive ion etch. The photo resist thickness i s not a crucial factor with Aluminum as masking layer. Soft bake at 100 C for 15 min in oven Expose the photoresist for 45 sec using soft contact printing on K arl Suss mask aligner (UV light intensity:11mJ/cm 2 ) Develop for 30sec using developer MF319 Hard bake at 110C for 30min in oven Profilometer thickness measurement: 3.3 m Wafer flat aligner 33 Square feature (A01-A33) with dicing marks 33 Circular features (C01-C33) with dicing marks DRIE Microloading study pattern

PAGE 40

27 Table 3.1 Design Patterns for Geometrical Investigation Pattern No Pattern type Pattern type Purpose of pattern Width ( m) Pitch ( m) Array No. of Chips(Ea) 1 Square Circular A01 C01 5 1x1 1 1 2 Square Circular A02 C02 10 1x1 1 1 3 Square circular A03 C03 20 11 1 1 4 Square circular A04 C04 40 11 1 1 5 Square circular A05 C05 Single needle Fracture & (Maybe difficult to find the needle) 60 11 1 1 6 Square circular A06 A06 5 20 5x5 1 1 7 Square circular A07 C07 5 40 5x5 1 1 8 Square Circle A08 C08 5 50 5X5 1 1 9 Square Circle A09 C09 10 20 5X5 1 1 10 Square circular A10 C10 1. fracture & penetration test 2. (5x5) needle array: effect of pitch and width on penetration 10 50 5x5 1 1

PAGE 41

28 Table 3.1 (Continued) 6 Square circular A06 A06 5 20 5x5 1 1 7 Square circular A07 C07 5 40 5x5 1 1 8 Square Circle A08 C08 5 50 5X5 1 1 9 Square Circle A09 C09 10 20 5X5 1 1 10 Square circular A10 C10 10 50 5x5 1 1 11 Square circular A11 C11 10 100 5x5 1 1 12 Square circular A12 C12 20 100 55 1 1 13 Square circular A13 C13 20 150 55 1 1 14 Square circular A14 C14 20 200 55 1 1 15 Square circular A15 C15 40 100 55 1 1 16 Square circular A16 C16 1. fracture & penetration test 2. (5x5) needle array: effect of pitch and width on penetration 40 150 55 1 1

PAGE 42

29 Table 3.1 (Continued) 17 Square circular A17 C17 40 200 55 1 1 18 Square circular A18 C18 60 100 55 1 1 19 Square circular A19 C19 60 150 55 1 1 20 Square circular A20 C20 60 200 55 1 1 21 Square circular A21 C21 5 20 25x25 1 1 22 Square circular A22 C22 5 50 25x25 1 1 23 Square circular A23 C23 10 50 25x25 1 1 24 Square circular A24 C24 10 100 25x25 1 1 25 Square circular A25 C25 20 100 25x25 1 1 26 Square circular A26 C26 20 150 25x25 1 1 27 Square circular A27 C27 20 200 25x25 1 1 28 Square circular A28 C28 1. Fracture test 2. 25x25 needle Array: effect of pitch, and width on penetration 40 100 25x25 1 1

PAGE 43

30 Table 3.1 (Continued) 29 Square circular A29 C29 40 150 25x25 1 1 30 Square circular A30 C30 40 200 25x25 1 1 31 Square circular A31 C31 60 100 25x25 1 1 32 Square circular A32 C32 60 150 25x25 1 1 33 Square circular A33 C33 60 200 25x25 1 1 34 Test Pattern 34 To study microloading effect in DRIE 20x4000,40x4000,60x4000,80x4000, 100X4000,120X4000,140X4000, 160X4000,180X4000,200X4000 2 The microscopic images of the pore patterns with Aluminum as ma sking layer are shown in Figure 3.3. Figure 3.3 1x1 Array of Width 40 m and 5x5 Array of Width 60 m The next step involves etching the mask layer at the lithogra phically exposed region. Wet etch is preferred for wafers with Aluminum as masking layer Aluminum Etchant Type A

PAGE 44

31 Phosphoric-Nitric-Acetic Acid) etches Al at the rate of 100 A /sec at 50C. Figure 3.4 illustrates the exposed Silicon region after Al etching and photoresist strip. Figure 3.4 25x25 Array after Al Etching Dry etch technique is preferred for oxide mask layers since buffered oxide etchant (wet) produces considerable amount of undercutting while etching 1.5 m thick oxide. The oxide wafers are subjected to reactive ion etching in the Unaxis RIE tool using CHF 3 /O 2 etch chemistry with 1:1 oxide to photoresist selectivity. The process parameters fo r oxide etch have been summarized in Table 3.2. Table 3.2 RIE Process Parameters Gases (sccm) Time (in min) RF power (watts) CHF 3 O 2 Pressure ( mtorr) DC bias (V) Etch rate (A/min) 45 200 45 5 40 440V 400-500 The patterned wafers were subjected to 400 Bosch cycles in Unaxis DRIE Tool (Courtesy: Star Centre, Largo) with process parameters as listed in Table 3.3.

PAGE 45

32 Table 3.3 Bosch Cycle in DRIE Process Time (sec) RF power (watts) Gases ( sccm) Pressure ( mtorr) RF 1 RF 2 C 4 F 8 SF 6 Ar He Passivation 5 1.0 825 70 0.5 40 5.18 23.0 Clean 2 9.0 825 0.5 50 40 5.18 23.0 Etch 6 9.0 825 0.5 100 40 5.18 23.0 Etching Issues associated with DRIE process : Microloading: Die 34 in the microneedle mask is dedicated to study the loadi ng effect in DRIE. From the SEM image in Figure 3.5, die 34 manifests an intra -die variation in etch rate due to Aspect Ratio Dependent Effect (ARDE). The etch de pth increases to about 80 m with increase in local opening from 20 m to 200 m. This is due to reduced transport of reactive species in narrow structures. Surface Roughness: Owing to alternate spontaneous etch and deposition processes involved in Bosch cycles, etched high aspect ratio DRIE struct ures exhibit scalloped rough sidewalls as observed in Figure 3.6. This issue can be minimized by opt imizing the operating DRIE parameters. The next step involves sidewall development of silicon oxide micr oneedle. The 4” wafer is diced into individual chips each of 0.9 square cm along dicing mar ks using Kulicke & Soffa Dicing Saw. This process is performed before oxidation since subsequent Silicon wet etch of diced oxidized wafer causes non uniform edges, eating away pattern s on large density needles. Each chip is tracked by a label, visible under 5x optical micro scope. This is followed by dry oxidation (1hr) and wet oxidation (11 hrs) of chips at 1100C in a quartz tube mounted on inverted boat in the furnace tube to obtain 1.5 m thick sidewall.

PAGE 46

33 Figure 3.5 Figure 3.6 Figure 3.5 Microloading Effect in Die 34 Illustrating Etch Depth Variation with Increasing Window Size on Masking Layer in DRIE Figure 3.6 Scalloping Effect in DRIE Illustrating Surface Roughness The top oxide layer has to be removed in order to expose the underlying bare silicon for TMAH etch. This is achieved by lapping chips (wax bonded to 4” glass plate) on Logitech PM5 The lapping parameters employed for fabrication is stated below: Lapping solution: Calcined Alumina powder + water (1:10) Lapping Time: 5 min at rate of ~2-4 m/min The chips are debonded and excessive wax is removed from the die by means of Opticlear wax remover and solvent wash. The final fabrication step involves silicon wet etching in orde r to expose the bulk micromachined needles. Tetramethyl Ammonium hydroxide (25% TMAH, (CH 3 ) 4 NOH) is preferred over other silicon etchants owing to increased selec tivity of Silicon to oxide (TMAH=500:1) [20]. The etching was performed at 85 0 C for 2hrs with uniform stirring. The observed etch rate was found to be around 25~30 m/hr. One of the issues faced with the etching was surface roughness which was overcome by addition of IPA and py razine [20]. Figure 3.7 exhibits the final hollow out of plane microneedle array with closed bas e. Scalloping

PAGE 47

34 Figure 3.7 Final Microneedle Array 3.2 Porous Silicon Based Microneedles 3.2.1 Introduction to Macroporous Silicon Porous Silicon is considered as a therapeutic biomaterial owing to ability to directly interface with human tissue like bones and biological molecules [22]. Porous sil icon etching is formed by anodic dissolution of Silicon in HF electrolyte. The I-V charac teristics governing pore formation has been illustrated in Figure 3.8 with pore formation condition s (highlighted in the dot area) below critical current density JPS. Beyond JPS peak, electro pol ishing occurs. Based on the substrate doping and anodization conditions like HF concentration, cur rent density, potential and illumination intensity, different pore diameter materials c lassified as micropores (<2nm), mesopores (2-50nm) or macropores (>50nm, ranging in microns) are obtained.

PAGE 48

35 Figure 3.8 I-V Characteristics Governing Electrochemical Dissolution o f Silicon [23] Macroporous Silicon formation occurs in n type substrate in presen ce of illumination and aqueous electrolyte during etching process [24]. Backside Illumi nation generates the holes required to promote silicon dissolution, otherwise restricted by holes present as minority carriers in n type substrate. Under anodic bias at constant current density J PS space charge region (SCR) is generated at pore tips and silicon dissolution reaches steady state condition between charge transfer and mass transport. Macropores can be localized by pore initiation etch pits defined by lithography. The value of J PS is anisotropic and is found to be maximum in (100) direction thereby making it the dominant pore growth direction without branc hing. The critical current density J PS (in mA-cm 2 ) is calculated by the Arrhenius expression: ) / exp( 2 /3 kT E Cc J a PS = Where c = electrolyte concentration (in wt% HF) T = absolute temperature (in K) Ea = Activation energy (0.345 eV) C = 3300A/cm2 wt%2/3=constant experimentally determined by Lehmann[23]

PAGE 49

36 Faraday’s law governs the pore growth rate n dependent on current density at pore tip(=JPS), atomic density of silicon NSi (5X1022cm-3), dissolution valence (number of holes consumed for dissolution of one silicon atom = 2.6 (empirical)) and electronic cha rge (1.602x10-19). Si PS N e n J ) ( / =n The etching current can be adjusted by changing light intensity Since current density remains constant at JPS, any change in current reflects a change in pore cross-sectional area. I f I etching represents the etching current and Aetching is the initial sample area (defined by the wafer sealing ring area=d2), then the apparent current density is given as: etching etching app A I J / = If Apores represents sum of cross-sectional area of the pores, fill factor (FF ) or local porosity [25, 26] is defined as PS app etching poresJ J A A FF / / = = For an homogenous and orthogonal square pore pattern of pitch p and width d, 2 2 / p d FF = The total etching current is thus calculated using the relation: 2 2 / p A J d I etching ps etching = The above expressions enable accurate determination of etching condi tions for desired diameter pore formation. Substrate doping density is crucial for pore diameter since it governs the SCR around pore tip, with misalignment leading to branching. The cross ectional shape of the pores varies from circle to branched state as shown in Figure 3.9. Branching of pores can be corrected by increasing the current density. The fill factor al so plays an important role in uniform cross-section of pores at constant current density under homogenous backside il lumination. Stable pore formation is predicted for fill factors ranging from 0.01 to 1[27].

PAGE 50

37 Figure 3.9 Variation in Pore Cross-section if Doping Density or Bias is Incr eased [25] 3.2.2 Fabrication of Porous Silicon based Microneedles This research is focused on development of CPS microneedle arra y with geometric variation. Two different patterns (Array 1: width=5 m, pitch=20 m & Array 2: width=20 m, pitch=100 m) were selected for fabrication. The process flow followed for fabrication of macroporous silicon has been summarized in Figure 3.10. The post pore formation process essentially remains the same as DRIE based method. 2” DSP, n-type (100) substrates with approximate thickness 250 m with two different resistivity (20-25 W -cm & 400-500 W -cm) were chosen as starting substrates. Lehmann empirically showed that wafer resistivity in ohm-cm should approxim ately be the square of the desired pore size in micron [26]. The wafers were subjected to RCA cleans and 4 probe point resistivity measurement before thermal oxidation. 3000A thick oxide is grown, acting as mask layer for contact diffusion.

PAGE 51

38 Figure 3.10 Process Flow for Porous Silicon Based Microneedles The oxide on the backside is patterned (lithography using contact mask followed by Buffered oxide etch) to form the contact region for subsequent n+ diffusion. T he lithography details have been listed below: Spin primer HMDS and positive photoresist S1813 on the mask laye r at 2500 rpm for 30 sec @ 255 acceleration using P-6000 Resist Spin Coater. Soft bake at 90C for 60sec on hotplate Expose the photoresist for 45sec using soft contact printing on Karl Suss mask aligner (UV light intensity:11mJ/cm 2 ) Develop for 30 sec using developer MF319 Hard bake at 110C for 60 sec on hotplate

PAGE 52

39 Frontside protection: spin HMDS and PR 1813 BOE etch for 10 min (observed BOE etch rate =400 m/min) Photoresist strip The wafers are now subjected to solid state Phosphorus diffusion whi ch ensures the ohmic contact essential for electrochemical etching. The n/n+ juncti on creates a built-in field, forcing the holes towards the pores and reduces the recombination at the semiconductor surface. The P 2 O 5 formed after diffusion is deglazed by BOE dip for 5minutes with front side r esist protection. Two 4 ” masks (Mask one: square patterns 5 m wide and 25 m pitch & Mask two: square pattern 20 m wide and 100 m pitch) were designed in Coventorware. Lithography was performed on the front side with similar parameters as discuss ed above in step b to define the pores on the wafer. The exposed oxide is etched using BOE. Pore init iation through wet etching: Anisotropic KOH etching is performed on the front side wafer supported in place in a Teflon jig at 90C for 13 minutes in order to obtain V grooves. These 15 m inverted pyramids act as pore initiation sites. 1000 A of Aluminum is e-beam evaporated on the backside of Si wa fer. The metal is then patterned to form contact electrode (lithography using conta ct mask followed by exposed Al etch using PNA) on the n+ diffused region. The next step involves Macro porous Silicon etching to form micromac hined pores in bulk Silicon. The experimental setup for porous silicon etching has been illustr ated in Figure 3.11. The wafer with pore initiation sites is placed in a cylindrical etching jig containing the electrolyte covered by a metal plate on one side. The plate has an opening to illuminate the wafer backside using the Oriel lamp source as shown in Figure 3.12. The wafer is se aled to the jig using rubber O-rings and screws. The wafer serves as the anode while the platinum electrode placed on top of the jig constitutes the cathode. The etching jig, the power suppl y and the lamp box are interfaced to the computer controlling the etching current via Lab view program. The lamp intensity is varied till the actual etch current follows the calculated value (Changing the lamp intensity varies

PAGE 53

40 the photocurrent and hence the total current through the sample). F ollowing are the etching parameters: Electrolyte: 2%HF + ethanol (10%) + water Bias voltage: 2.5V Etching time: 24 hrs Results for Array 1: The SEM image demonstrating the 5u wide pores is shown in Figure 3.14. The etching was uniform with etch depth of around 150 m. Result for Array 2: SEM images for Array 2 in the figure depict that pores were branched. This may be due to improper bulk substrate doping level or higher bias Figure 3.15 showed reduced branching at low bias (2V) with other etching conditions remaining unchanged. The tape red pores were observed due to drop in illumination intensity after 15 hrs. Figure 3.11 Photograph Illustrating the Porous Silicon Etching Setup at USF

PAGE 54

41 Figure 3.12 Experimental Setup for Macroporous Silicon Etching Setup at USF Figure 3.13 SEM Image Illustrating 5x20 m Macropore Array, Height=150 m after Etching for 24 hrs A H F v Etch rig with 2% HF IR Filter Reflection Mirror Lamp Source Silicon Substrate Lab view control for maintaining current density

PAGE 55

42 Figure 3.14 Figure 3.15 Figure 3.14 20x100 m Array with Branched Pores (Bias 2.5V) Figure 3.15 20x100 m Array with Reduced Branching (Bias 2V) The post pore processing is similar to that adopted in DRIE m ethod. Wafer 1 was diced, oxidized lapped from the reverse surface and subjected to TMAH e tch to form blocked, out of plane needle structures as shown in Figure 3.16. Figure 3.16 Blocked, Out of Plane Microneedles

PAGE 56

43 3.3 DRIE Vs. Porous Silicon Based Microneedle Fabrication Process Table 3.4 Comparison of Porous Silicon and DRIE Processes DRIE based microneedle fabrication Porous Silicon based microneedle fabrica tion Different microneedle design chips can be obtained from one single wafer since they are lithographically defined Single pattern obtained from one wafer since pore dimension is a function of %HF, etching current, substrate doping levels, bias etch Ease in fabrication, single lithography step Complex fabrication process involving three lithography steps Higher etch rate Time consuming process with low etch rates DRIE etch independent of wafer specifications Substrate specification like type, direction and resistivity are crucial for pore formation Pores necessarily occupy entire wafer. No space for additional features. Selective etching is possible in certain areas as per requirement. No disposal of dangerous acids and solvents involved. However process involves use of toxic etch gas composition Hazardous acids and solvents need to be disposed. Loading effect due to selective etching No loading effect, all pores are approximately etched to same depth

PAGE 57

44 CHAPTER 4 BIO-MECHANICAL CHARACTERISATION 4.1 Introduction DRIE based SiO 2 microneedle arrays were employed for experimental testing to e xamine the significance of geometry on transdermal testing. This schem e was adopted due to ease in usage and speedy fabrication of desired patterns from a single w afer as compared to porous silicon based etching. Reliability testing based on measurement of buckling force and skin penetration force was performed before confirming use of the nee dles for applications like drug delivery and fluid extraction. Measurement was carried out by an in-house developed testing module utilizing an accurate load transducer and real time viewi ng capability. Preliminary characterization was established on skin like polymer before real time exper imentation on excised split skin samples. Mechanical tests were also performed on t he polymer to correlate its properties with skin. 4.2 Experimental Setup The experimental setup was designed to house two basic modules: skin l oading block and needle attachment block as shown in Figure 4.1. The skin is mounted on a 2c m x 4cm aluminum block using double sided sticky tape. The penetration force was meas ured using a sensitive tension-compression load cell (LCFA-500gF sensing capacity, Omega C o.) attached to the skin mounting fixture. The internal construction of these load cells consis ts of a full four-arm Wheatstone bridge, capable of producing repeatable measurements.

PAGE 58

45 Figure 4.1 Testing Module The load cell was interfaced to one channel load cell input 16-Bit, RS-485 Data Acquisition Module (Superlogics-8016) to obtain real time results The load cell is in turn screwed to a steel block supported on an L-bracket, mounted on a XY stage with two 1 micrometer resolution (SM-13, Newport Co.). The needle attachment b lock (0.9cm x 0.9cm microneedle chip mounted on steel block) is supported on XYZ stage a nd is slowly driven by 1 um resolution differential micrometers (DM-13) into skin. The des igns for supporting blocks for the needle array and the skin have been illustrated in Figur e 4.2. The manual translation of microneedles by micrometer was replaced by motorized micromet er (Optomike B, Sigma Koki B Inc.) in order to maintain constant insertion speed. Real time ima ge capture is obtained by means of camera mounted on the probe station to view the instant of needle inserti on. Interfaced to computer

PAGE 59

46 (a) Figure 4.2(a) Load Cell Block (Top) (b) Microneedle Block ( Bottom) (b) One through hole -diameter 7 mm hole thru diameter 11 mm hole with depth 7 mm 33 25 60 10 Block with microneedles Front view Top view 16.5 11 7 One through slot Inner diameter 7 mm thru hole diameter 11 mm hole with depth 7.5 mm 11 7 60 60 15 15 7 23 60 60 15 15 7 23 17 10 7 Back side Top view front Front view

PAGE 60

47 4.3 Calibration of Load Cell Figure 4.3 Calibration Setup Experiments were conducted to calibrate the new load cell o f 500 gF capacity. The rated output for load cell = 20.81mV with sensitivity of 2.08mV/V and 10V ex citation input voltage requirement. A fixture was made to mount on the new load cell which would hold the skin samples (Weight of loading plate = 20.7g) as shown in Figure 4.3. Load was varied from 0-500 gF in steps of 100g (including plate weight) and the obtained resul ts were compared with data sheet readings provided by the manufacturers as tabulated in Tabl e 4.1. Best fit linear curve was plotted through the experimental values as shown in Figure 4.4. Table 4.1 Comparison Between Experimental and Data Sheet Readings Point Load (gF) Transducer output (experimental)-mV Transducer output (data sheet)-mV 1 0 0.02 0 2 20.7 1.02 N/A 3 96.9 4.19 4.1(100 g) 4 206.7 8.42 N/A 5 298.2 12.01 12.3(300 g) 6 407.9 15.75 N/A 7 502.7 20.89 20.5 (500g) Loading Plate Load cell (0 500g) Data Acquisition System

PAGE 61

48 Calibration curvesTransducer Output vs Loady = 0.0428x R2 = 0.9949 0 5 10 15 20 25 0100200300400500600 Load (gms)Output voltage(mV) Experimental Calibration Load cell Data Sheet Linear (ExperimentalCalibration ) Figure 4.4 Best Fit Graph for Experimental Values Compared with Standard Data Values Figure 4.5 Experimental Calibration Readings For no load condition With loading plate (20.7 g) Load of 96.95g Load of 206.75 g

PAGE 62

49 Figure 4.5 Experimental Calibration Readings (Continued) Experiments were conducted to determine the variation of the loa d cell output with time. Figure 4.6 (a) illustrates the output of the load cell (in m V) with a plastic fixture and polymer specimen mounted on it. Actual weight of plastic and polymer is 2.368g. The plot measures ~0.145 mV [0.145 20.172 (from data sheet) = 2.926g] after elimination of A /D conversion error through software programming. The variation over period of 300 sec (e stimated as experiment time) was approximately .02mV equivalent to 0.5gF. Figure 4.6(b) and 4.6(c) emphasizes the significance of shielding on output variation with time for t he above fixture setup. The drift observed under partial shield conditions in probe station was 0.025 mV ( 0.5178 gF) as compared to complete isolation value 0.0102 mV (0.2112 gF). Figure 4.6 (d) illustrate s the variation of 1.5 gF over a period of 3hrs (non shield). This variation s attribu ted to device drift as well as dust particle deposition on the load cell. Load of 298.16g Load of 407.8g Load of 502.7 g

PAGE 63

50 (a) (b) (c) (d) Figure 4.6 Illustration of Load Cell Output Variation over Time under (a) Normal Condi tions (b) Partial Shield (c) Complete Shield (d) Variation over 3 hrs Time with Metal Fixt ure 4.4 Calibration of Motorized Micrometer During mechanical testing on skin, there arose a need to maintain a constant in sertion rate as compared to that achieved manually by repeated measurements o f the micrometer. A motorized micrometer (Optomike B, Sigma Koki B Inc.) with two a xis drive remote control controller adjustable to 10 different speed modes (fast and slow at each mode) [28] was employed (Courtesy: Star Center, Largo). This motor, obtained from another device, was not calibrated and lacked an electronic display of traversed distance in the a bsence of feedback through an interface with the PC. Calibration tests were performed in order to determine the exact speed at each of the speed modes as summarized in Table 4.2 and illustrated in Figure 4.6.

PAGE 64

51 Table 4.2 Calibration Table for Motorized Micrometer Mode Front time Back time Average Time Distance Speed (sec) (sec) (sec) (mm) ( m/sec) 9-Fast 17 21 19 5 263.2 9-Slow 65 68 66.5 5 75.2 8-Fast 23 22 22.5 5 222.2 8-Slow 75 75 75 5 66.7 7-Fast 24 24 24 5 208.3 7-Slow 81 82 81.5 5 61.4 6-Fast 30 29 29.5 5 169.5 6-Slow 97 98 97.5 5 51.3 5-Fast 31 33 32 5 156.3 5-Slow 117 114 115.5 5 43.3 4-Fast 38 40 39 5 128.2 4-Slow 140 140 140 5 35.7 3-Fast 47 50 48.5 5 103.1 3-Slow 189 188 188.5 5 26.5 2-Fast 63 67 65 5 76.9 2-Slow 286 285 285.5 5 17.5 1-Fast 144 144 144 5 34.7 1-Slow 560 560 560 5 8.9 Speed calibration_motorised micrometer0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270987654321 Mode Speed (um/sec) Fast mode Slow mode Figure 4.6 Calibration Curves for Fast and Slow Modes at Different Speed Positions for Motorized Micrometer 4.5 Mechanical Tests on Skin like Polymer and Split Skin Compressive and indentation tests were performed on skin like polymer (Chester Chest Difficult Accessing Insert 440, VATA Inc., size=13x13cm) to corr elate its mechanical

PAGE 65

52 properties with skin, followed by identical tests on skin sample plugs. The tissue insert is considered as simulation syringe insertion undertaken during nurse tra ining and has palpability similar to skin. The polymer providers had asserted that no m echanical tests had been conducted to confirm skin similar mechanical properties. For this study, 2x3 cm excised split-thickness skin specimens were obtained from U.S. Cell and Tissue Bank (Cincinnati Ohio). 6 mm dermal punch was used to punch out polymer and skin plugs for compressive loading. T he testing was performed on MTS system (858 Minibionix II, Courtesy: Shriners Hospit al). Factors affecting skin testing are hydration, testing temperature and humidity a nd specimen site location owing to skin heterogeneity. 4.5.1 Compressive Tests on Polymer and Skin Compressive tests were conducted on 5 polymer plugs to determine t he elastic constant of polymer (Strain limit =50%, strain rate=10% and retracte d at same rate). The stress-strain relationship on compressive loading for polymer plugs have been summ arized in Figure 4.7, with plug dimensional details tabulated in Table 4.3. The polymer was obse rved to be isotropic in nature, yielding non linear elastic behavior with small visco us element. This is evident from similar nature of the stress strain curve during compressio n and retraction. The young’s modulus computed from the stress strain plots have been stated in Table 4.4 at differ ent strain levels. Similar tests were performed on skin plugs (Table 4.5) genera ting results as shown in Figure 4.8. These specimen exhibited nonlinear elasticity dominated by large viscous drag. The viscous nature owing to heterogeneous and anisotropic nature of skin i s evident due to dissimilar nature of curves during compression and retraction. For low stres s only the dermis compresses due to low elastic constant. With increase in stress, the curve exhibits a rapid increase indicating the compression of stiffer layers. The Young’s Modulus computed from the stress stain plots have been summarized in Table 4.6.

PAGE 66

53 Table 4.3 Polymer Plug Specimen Dimension Chart Polymer Plug Height (mm) Diameter (mm) Cross-sectional area (mm 2 ) 1 5.41 5.19 21.44 2 5.38 4.96 19.31 3 5.58 5.22 21.39 4 5.49 5.13 20.66 5 5.50 5.30 22.05 Figure 4.7 Compressive Tests: Stress-Strain Relation for Polymer Plugs Table 4.4 Elastic Modulus of Polymer Plugs at Different Strain Levels Polymer Elastic Modulus (MPa) Strain level Polymer plug specimen 0-10% 10-25% 25-40% 40-50% 1 0.15 0.21 0.61 1.58 2 0.17 0.24 0.64 1.49 3 0.10 0.19 0.51 1.61 4 0.13 0.24 0.65 1.71 5 0.11 0.22 0.59 1.49 Table 4.5 Skin Plugs Dimension Chart Skin Plug Height (mm) Diameter (mm) Cross-sectional area (mm 2 ) 1 1.73 4.81 18.16 2 1.28 5.01 19.70 3 1.76 4.92 18.85 4 1.45 5.45 23.32 Compressive test results for polymer plugs-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.250 0.050.090.140.180.230.270.320.360.410.460.490.44 0.4 0.350.310.260.21StrainStress (MPa) Polymer plug1 Polymer plug2 Polymer plug 3 Polymer plug 4 Polymer plug5 Compression Retraction

PAGE 67

54 Skin plug 1Compressive loading-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.40 0.030.060.090.120.150.180.210.240.27 0.3 0.330.360.390.420.450.480.34strainStress (MPa) Skin plug 2 -compressive loading-0.1 0 0.1 0.2 0.3 0.4 0.50 0.030.060.090.120.150.180.210.240.27 0.3 0.330.360.390.420.450.480.29StrainStress (MPa) Skin plug 3-Compressive loading-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.60 0 0 564 9 2 0 1 14 67 9 0.170025 0 2 27 16 31 0 2 8 45784 0 3 424 0 57 0 3 99 11 2 5 0.4542392 0 5 12 68 01 0 5 6 82767 0 6 250 7 27 0 6 8 132 4 4 0.6143 9 77 StrainStress (MPa) Skin plug 4-Compressive loading-0.5 0 0.5 1 1.5 2 2.50 0 .0698 1 38 0 1 38 23 3 8 0.2079138 0 2 75 7 0 3 4 48959 0 4 144 7 59 0 4 82 92 4 8 0.5520945 0 6 21 72 69 0 6 8 94497 0 7 60 51 1 0 8 2 727 5 9 0.7435 8 14 StrainStress(MPa) Figure 4.8 Stress-Strain Relation for Skin Plugs on Compressive Loading

PAGE 68

55 Table 4.6 Elastic Modulus of Skin Plugs at Different Strain Levels Polymer Elastic Modulus (MPa) Strain levels Skin plug specimen 0-10% 10-25% 25-40% 40-50% 1 0.014 0.048 1.680 9.631 2 0.019 0.024 0.558 4.182 3 0.028 0.019 0.022 1.802 4 0.025 0.006 0.148 1.092 The results from compressive tests for polymer and split thickness ski n plugs have been summarized in Figure 4.9. Polymer compressive tests 0 0.5 1 1.5 2 12345Polymer plugsYoungs Modulus (MPa) 0-10%strain 10-25%strain 25-40%strain 40-50%strain Figure 4.9 Comparative Analysis Between Elastic Modulus of Polymer and Split Thickne ss Skin Skin compressive tests 0 2 4 6 8 10 1 2 3 4 Skin plug 0-10% strain 10-25% strain 25-40%strain 40-50%strain Young’s

PAGE 69

56 4.5.2 Indentation Test on Polymer and Split Thickness Skin Indentation tests were performed on polymer and skin using 4mm diameter stainless steel indenter hemisphere. Polymer (size:13cm X 13cm) was indented at tw o different rates set at 1mm/sec and 5mm/sec to verify strain rate dependency (viscoel asticity). Five tests at different positions (10mm, 20mm, 30mm, 40mm & 50mm from polymer edge) with forc e and indenter position measurement at 0.05sec interval were conducted at each s train rate. The results have been summarized in Figure 4.10 and 4.11. With increase in indentation depth, the indenterspecimen interfacial area increases leading to increasing s lopes of force curve deformation. This was accounted by normalizing the force to projected indenter are a at each depth as shown in Figure 4.12 [29]. The indentation stiffness (slope from normalized plots ) was observed to remain constant for all five tests owing to observed linear relations hip between force and indentation depth over the range 0.5 to 1.5mm. Average indentation stiffness for each experiment was computed from the plots tabulated in Table 4.7. The polymer was f ound to be elastic due to fairly similar stiffness coefficients at different strain rates. T his was further confirmed by relaxation test where the polymer was indented by an instantaneous step strain te st for 60sec followed by retraction as shown in Figure 4.13. The results indicate that the stress required in maintaining constant stain doesn’t decrease gradually with time as expected for visc ous solid. This was followed by 6 indentation tests on cryo-preserved cadaver skin pi ece (Size: 12mm x 22mm x 1.4mm) at the rate of 1mm/sec. The force-indent ation plots have been illustrated in Figure 4.14. The average indentation stiffness coeff icients have been stated in Table 4.8. The results indicate variation in indentation stiffness at di fferent regions of the specimen as well as variation at same point. This can be attributed to ani sotropic nature of skin and “creep” (residual strain in viscous material after stress removal).

PAGE 70

57 Force vs indent depth at 1mm/sec (Indentation depth =2mm)0 0.5 1 1.5 2 2.50 0 .20184 0.40239 0.604 67 0.79994 1.005 37 1.202 97 1. 403 1.604 14 1. 80435 1.92234 1.512 49 1.111 0.71149 0.310 86Indent depth (mm)Force (N) Position1-10mm Position2-20mm Position3-30mm Position4-40mm Position5-50mm Compression Retraction Figure 4.10 Force-Depth Plots for Indentation Rate of 1mm/sec Compression Retraction Figure 4.11 Force-Depth Plots for Indentation Rate of 5mm/sec Force vs Indentation at 5mm/sec (Indentation depth= 2mm) 0 0.5 1 1.5 2 2.5 00.250.50.7511.251.51.751.751.651.551.451.35Indent Depth(mm)Indentation Force(N) Position1-15mm Position2-25mm Position3-35mm Position4-45mm Position5-55mm

PAGE 71

58 Normalised force vs depth (indentation at 1mm/sec)0 0.05 0.1 0.15 0.2 0.25 0.3 0.350.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.951.011.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1.45 1.5Indentation depth (mm)Normalised Force (Mpa) Position1-10mmIndentationstiffnes=0.2064 Mpaper mm Increasing depth Figure 4.12 Force Normalization for Stiffness Calcu lation at Position 1-10mm Table 4.7 Indentation Stiffness at 1mm/sec and 5mm/ sec Indentation Rates Indentation rate =5mm/sec Test PositionEdge distance Slope (indent depth=0.5-1.5mm) Position1-15mm 0.232 Position 2-25mm 0.224 Position3-35mm 0.238 Position4-45mm 0.233 Position 5-55mm 0.235 Average 0.232 Indentation rate =1mm/sec Test PositionEdge Distance Slope (indent depth=0.5-1.5mm) Position1-10mm 0.206 Position 2-20mm 0.210 Position3-30mm 0.214 Position4-40mm 0.212 Position 5-50mm 0.210 Average 0.210

PAGE 72

59 Figure 4.13 Relaxation Test Result Illustrating Ela stic Behaviour of Polymer with Small Viscous Elemen t (Straight Step Plot Represents Instantaneous Step S train while Wavy Curve Indicates Force Values) Compression Retraction Figure 4.14 Indentation Plot for Split Thickness Sk in at 1mm/sec

PAGE 73

60 Table 4.8 Indentation Stiffness for Cadaver Skin Mechanical properties of the polymer and cadaver skin were expe rimentally determined on the macro-scale. The results confirm that the polymer does not behave like skin since its properties are isotropic and elastic unlike skin. However thi s material can be used for preliminary characterization of microneedles before real time testing on cadaver sk in. Indentation rate =1mm/sec Test PositionDistance from edge Stiffness -Slope (indent depth=0.5-0.8mm) Stiffness -Slope (indent depth=0.8-1.1mm) Position1-5mm 1.301 6.392 Position 2-10mm 1.632 7.206 Position3-15mm 2.279 9.570 Position4-20mm 0.579 5.289 Position 5-10mm 2.962 11.086 Position 6-15mm 1.836 8.563

PAGE 74

61 CHAPTER 5 FRACTURE AND PENETRATION TESTING This chapter discusses experimental fracture force meas urement and detailed failure analysis for fabricated SiO 2 microneedles. Preliminary test results utilizing skin like polymer to validate the ability of the microneedles to penetrate skin ha ve been presented. The chapter also elaborates on real time insertion tests on split thickness ski n and isolated stratum corneum using both manual and motorized micrometer at constant rate, quantifying the penetra tion force. 5.1 Measurement of Fracture Force 5.1.1 Experimental Plan In plane fracture tests were performed on 1 1 array to observe the effect of needle geometry (shape, width and height) on fracture force. The ax ial force required to fracture silicon dioxide needle column was experimentally measured by compressing the chip (mounted using double sided sticky tape) at rate of 1 m/sec against the load cell mounting screw. The point of fracture was visually captured by the camera (change in int ensity) with the data acquisition system displaying a sharp peak in force-time graph, interprete d as “needle failure point”. Once this peak was achieved, the needles were retracted at the same rate. The microneedle parametric dimensional variation has been tabulated in Table 5.1. Since it was difficult to fabricate and view single needle, 5x5 microneedle array chips were manipulated with probe tips to obtain chips with 13 needles.

PAGE 75

62 Table 5.1 Microneedle Dimensional Variation for Fracture Testing Parameters Variation Needle shape Square, circular Needle width 5 m,10 m,20 m,40 m,60 m Needle height 25~120 m 5.1.2 Fracture Results Analysis Figure 5.1 represents a typical microneedle compressive force vs. t ime plot with the peak representing the maximum force applied on needle before failure (2.33 gF for 40 m wide, 125 m long needle). The sudden drop after the peak indicates that the needle and l oad cell are no longer in contact. The testing results have been summarized in Table 5.2 with frac ture force-length comparison plot for varying widths illustrated in Figure 5.2. The observed readings did not obey the inverse parabolic relation between f orce and length given by analytical values computed in Chapter 2 ( in accordance to Euler’s buckling relation for long slender columns). However as expected, the fracture forc e showed decrease with increasing length. The slenderness ratio (the ratio of second moment of i nertia to radius of gyration for the hollow structure) for the needles was then realized t o be small, imposing needle consideration as short columns leading to inelastic stability as fail ure mechanism on compression. But the readings were found to be smaller than designed compressive force of 5gF, even lower than theoretical fracture force determined from material stre ngth (690-1380 MPa)[30]) tabulated in Table 5.3 .The force-needle length relation for 40 m and 60 m wide needles of both square and circular geometry have been plotted in Figures 5.3-5.6. Owing to lower crosssectional area during compressive failure, circular needle fractured at lower fo rce compared to square geometry. The fracture force also augmented with increasing microneedle diameter (constant length=75 m) for circular geometry as shown in Figure 5.7.

PAGE 76

63 The possible reasons attributed for reduction of compressive streng th of material has been discussed further. Brittle fracture in SiO 2 can be attributed to microscopic crack defects in the structure (Griffith’s stress concentrators [31]) leading to cr ack initiation and propagation. These stress concentrators could have magnified the stresses at the crack tip, thus causing the material to fracture long before it ever reached its theoretical strength. The hollow needle body also could have behaved as stress raiser leading to fracture. Qualit y of needle material depends on oxide wall formation, and release during fabrication process. The needle e dges are non uniform and jagged after mechanical lapping process before needle release. In addition out of plane needle release by silicon etchant (TMAH) produced buckling of oxide sidewal ls as shown in Figure 5.8. This was due to compressive stresses generated by thermal co ntraction of the silicon during oxidation process due to difference in thermal expansion coefficients of silicon and oxide [32]. A04(Width=40um)_BUCK_125um-2 -1.5 -1 -0.5 0 0.5 1 1.5 147101316192225283134374043464952555861646770 time(sec)Force(gF) Figure 5.1 Typical Plot Displaying Fracture Peak for 40 m wide square microneedle of length 125 m

PAGE 77

64 Table 5.2 Fracture Failure Test Results Chip details Test No No of needles-Needle Height Before Test Needle Height After Test Measured Force/needle (gF) 1 Single needle -45 m height Fracture 0.47 2 2 needles 45 m height Fracture 1.05/2=0.52 3 3 needles 78 m height 1needle of height 45 m 0.8/3=0.26 Circular-5 m width 4 4 needle -75 m height 1.6/4=0.40 Circular-10 m width 1 2 needle 25 m height Fracture 0.9/2=0.45 1 Single needle52 m height Fracture 0.28 Square -20 m width 2 2.5 needle-60 m height Fracture 0.73/2=0.37 1 Single needle85 m height 27 m 3.18/2=1.59 3 Single needle 75 m height Fracture 1.23 4 Single needle – 75 m height 1.5 Circular-20 m width 5 2 needles – 37 m height Fracture 0.9/2=0.45 1 Single needle-65 m height Fracture 8.20 2 Single needle-47 m height Only edge dislocation, no length variation 12.05 3 Single needle-above 47 m test continued Fracture 2.05 Square -40 m width 4 Single needle -125 m height Fracture 2.33 1 2 needles -75 m Height 1 needle-52 m height 4.78/2=2.39 2 Single needle -45 m height Fracture 2.48 3 Single needle – 45 m height Fracture 2.33 4 2 needles -75 m Height Fracture 2.15/2=1.07 5 Single needle – 75 m height 1.35 6 Single needle-52 m height Fracture 1.88 7 Single needle-115 m height 95 m height 1.90 Circular-40 m width 8 Single needle-95 m height Fracture 2.05 1 Single needle-100 m height Fracture 1.83 2 Single needle-33 m height Fracture 3.43 3 Single needle -85 m height Fracture 1.85 4 1.5 needle85 m height Single needle -22 m height 1.45 Square-60 m width 5 Single needle-60 m height Fracture 2.05 1 Single needle-40 m height Fracture 2.78 2 1.5 needle-65 m height Fracture 1.28 3 Single needle -75 m height 2.43 4 Single needle -120 m height 80 m height 1.00 Circular-60 m width 5 2 needles-136 m height Fracture =0.50

PAGE 78

65 Fracture Force vs length0 2 4 6 8 10 12 14 050100150 Length (um) Force (gF) 40um width square 40um width circular 60um width square 60um width circular Figure 5.2 Fracture Test Results as Function of Needle Length and Width Table 5.3 Theoretical Fracture Force for Short Beam Structures Derived from Mat erial Strength (Compressive Strength of Thermal SiO 2 =690-1380 MPa) Chip no Wall thickness-1.5 m A01 2.02-4.04 C01 1.59-3.17 A02 4.09-8.18 C02 3.21-6.43 A03 8.23-16.46 C03 6.46-12.93 A04 16.51-33.02 C04 12.97-25.93 A05 24.79-49.58 C05 19.47-38.94

PAGE 79

66 Square needle-width 40u-Fracture Force vs Length0 25 50 050100150200250300350 Length(um)Fracture Force (gF ) Eulers relation Compressive strength Experimental Figure 5.3 Force-Length Relation for 40 m Wide Square Microneedle Circular needlewidth 40um-Fracture force vs Lengt h0 25 50 050100150200250300350 Length(um)Fracture Force(gF) Eulersforce Compressivestress Experimental Figure 5.4 Force-Length Relation for 40 m Wide Circular Microneedle

PAGE 80

67 Square needlewidth 60um-Fracture Force vs Length0 25 50 75050100150200250300 Length(um)Fracture(gF) Eulers relation Compressivestress Experimental Figure 5.5 Force-Length Relation for 60 m Wide Square Microneedle Circular-width 60um-Fracture Force vs Length0 25 50 050100150200250300 Length umForce(gF) Eulers force Compressive_stress Experimental Figure 5.6 Force-Length Relation for 60 m Wide Circular Microneedle

PAGE 81

68 Fracture Force vs width (Length=75)0 0.5 1 1.5 2 2.5 3 010203040506070 Width(um)Force(gF) Figure 5.7 Fracture Force vs. Width (Length=constant) for Circular Geometry Figure 5.8 Wall Buckling after TMAH Etch (10x-Optical Microscope) 5.2 Insertion Tests on Skin like Polymer Characterized soft polymer (size: 1cmx1cm) was used for prel iminary microneedle insertion testing. The polymer was mounted on the skin holding fixt ure using double sided sticky tape. The block holding the microneedle chip was forced into the polymer speci men using manual micrometer translation (approximate insertion rate of 10 m/sec). Figure 5.9 illustrates the intact condition of one of the tested 5x5 microneedle array (width 40 m, pitch 150 m and length 75 m) indicating sturdiness after the insertion test. The needle shaf t showed traces of polymer confirming penetration and not surface indentation. This was further validated by Energy Dispersive X-ray Spectrum (EDAX) measurement which indicate d presence of polymer constituents. Buckled edges for 40 m square needles

PAGE 82

69 Penetration force plot for this array has been illustrated i n Figure 5.10 displaying a discontinuity peak at approximately 0.2gF, however due to lack of automation software for data acquisition system and motorized micrometer it was difficult to mark this value as penetration force. However it was difficult to observe any cut marks on the polymer surface. Further testing with needle arrays reproduced similar results making quantifica tion difficult. Later the automation software was purchased and data points could then sav ed and exported into Microsoft Excel. Figure 5.11 illustrates the penetration plot for a 25x25 array (width=60 m, pitch=150 m, length=75 m) with 0.8Gf force. However this value could be a rough estimate since the polymer is isotropic in mechanical properties. Figure 5.9 Illustrating the Condition of 5x5 Microneedle Array (Width 40 m, Pitch 150 m and Length 75 m) Before and After Needle Insertion. Needles show Polymer Trace in the N eedle Lumen after insertion

PAGE 83

70 Figure 5.10 Penetration Plot for 5x5 Array with Penetration Discontinuity Obser ved at 0.2gF Penetration force vs time for 25x25 array(width=60u pitch=150u)-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 135791113151719212325272931 time(sec) Penetration force (Gf) (gF) Figure 5.11 Penetration Plots for 25x25 Array (Width=60 m, Pitch=150 m, Length=75 m) with First Force Peak at 0.8 gF

PAGE 84

71 5.3 Penetration Testing on Split Thickness Skin 5.3.1 Skin Tests Using Manual Translation For this study, penetration tests were performed on 1cm x 1cm excis ed split-thickness skin specimen obtained from U.S. Cell and Tissue Bank, Ohio. Room tempera ture (70-73 0 F) and humidity (40%) were kept constant throughout the experimentation. The c ryo-preserved samples were thawed to room temperature and were observed to dehydrate rapidly. The hydration issue related to the skin was resolved by physiological saline appli cation after every 5 minutes. Excess water from Stratum corneum surface was wiped using Kim wipe, en suring close to 100% humidity for the skin samples. Each chip was observed under optical m icroscope before every experiment for verification of overall needle integrity SEM images of the needles were obtained to observe the condition of needles after penetration. Due to nonava ilability of motorized system, the needles were displaced 500 m into the skin specimen utilizing manual micrometer movement at the rate of 10-50 m/sec(approximate). Black India ink was used as a stain to fac ilitate easy observation of penetration marks. Preliminary tests conducted using skin specimen suggested that needle lengths need to be increased from 50 m to 100 m for effective penetration due to skin viscoelasticity [33] since black ink stain marks was not v isible on the skin specimen even after 500 m insertion into skin. The typical penetration force–time plot is shown in Figure 5.12. This plot shows presence of a number of peaks owing to non uniform insertion speeds during experim entation. However the first sharp peak observed has been characterized as penetrat ion force value since it represents cutting of stratum corneum. Figure 5.13 illustrates the black ink marks on skin specimen after insertion tests for two different needle configurations. SEM im age in Figure 5.14 demonstrates the microneedle penetration while being retracted from skin w hile Figures 5.15 5.17 highlight the sturdiness of microneedles with skin clogging in the needle lumen. The penetration results

PAGE 85

72 have been summarized in the Table 5.4 with force of the order of 1 -4 gF for 25x25 array and 0.30.6gF for 5x5 array. As observed in this table, penetration force/ needle could be approximated as ~0.01gF while fracture force/needle was 1~3gF, explaining needle robu stness during skin insertion Insertion force-time plot-25X25 array (width=40u,pi tch=200u)-92 -91 -90 -89 -88 -87 -86 -85 14710131619222528313437404346Time(sec)Force (gF) Figure 5.12 Typical Force-Time Plot for 25x25 Array Marked by a Number of Peaks Due to N on Uniform Insertion Rate (a) (b) Figure 5.13 Optical Microscope Image Illustrating Black India Ink Stains on C adaver Skin after Insertion for (a) 25x25 Array –Circular Needles (Width=20 m, Pitch=100 m, Height=125 m) (b) 25x25 Array –Square Needles (Width=40 m, Pitch=200 m, Height=125 m)

PAGE 86

73 Figure 5.14 Microneedle While Being Removed from the Skin Sample Suggesting Penetrati on Figure 5.15 SEM Image of 25x25 Array of Circular Needles (Width=20 m, Pitch=100 m, Height=125 m) after Skin Insertion

PAGE 87

74 Figure 5.16 SEM Image Illustrating Clogging of Needle with Skin Figure 5.17 SEM Image Showing Effective 4.5 m Clogging of Needle Lumen

PAGE 88

75 Table 5.4 Penetration Force Summary with Needle Specifications Microneedle specifications Needle width ( m) Needle pitch ( m) Needle length ( m) Observed Penetration force (gF) Penetration force/needle (gF) 25x25array-square 40 200 125 1.57 0.003 25x25array-circular 40 100 140 1.75 0.003 25x25array-square 40 100 125 No peak 25x25array-circular 60 200 130 No peak 25x25array-square 60 200 75 1.05 0.003 25x25array-square 60 200 105 4.40 0.007 5x5array-squaretest1 40 100 90 0.58 0.023 5x5array-squaretest2 40 100 90 0.35 0.014 5x5array-square 40 200 95 0.68 0.027 5x5array-square 20 100 90 0.30 0.012 5x5array-square-2 repeat tests 20 100 100 No peaks 5.3.2 Skin Tests Using Motorized Micrometer Since manual translation did not assure constant skin insertion r ate and appropriate force quantization, few enhancements were made to the existing setup. The sampling rate for the data acquisition system for the load cell was set at maximum (10 Hz) Penetration experiments were repeated using motorized micrometer (Courtesy: Star Center, La rgo) with insertion rate set at 50 m/sec. These results were found to be encouraging since some res ults exhibited a sharp peak suggesting possible stratum corneum penetration with similar India ink marks. A sample penetration plot has been shown in Figure 5.18 for a 25x25 needle array (needle width 60 m, pitch150 m, length 105 m) with force peak of 4.03 gF. The needle chip was forced a dist ance of 1150 m into the skin with insertion force plotted over the entire distance as shown in Figure 5.19. The results from the penetration tests have been summarized in T able 5.5 with penetration force per needle approximated to ~0.01gF. The table also shows that s ome tests did not display a sharp penetration peak, thereby leading to non force quantification. This i nability could be attributed to

PAGE 89

76 following factors: wavy nature of skin, non uniform microneedle edg es, low rate sampling rate of data acquisition system and insufficient load cell sensitivi ty. Since the cadaver specimen were wavy and exhibited high surface roughness, all needles must not have penetrated at the same instant; hence the point loading may be significantly higher than the rea d out values Another approach to analyze the microneedle contribution during c hip insertion into the skin was to test using a plain chip without needles on skin under similar displacement conditions. Work done by the needles was then related by computing the area be tween the force– displacement curves with and without needles. Figure 5.20 illustrat es this approach for above used 25x25 array. However this approach doesn’t confirm the needle pene tration or indentation since penetration peak observed was at low displacement value s where no significant difference between the curves is observed. Also the motorized micrometer e mployed lacked a feedback system which could be interfaced with the PC to co-ordinate the needle displacement and penetration force time intervals. In current measurement setu p, this co-ordination was achieved manually using a stop clock. Penetration force vs time for 25x25 array_60x150um55 56 57 58 59 60 61 62 63 01234567891011121314151617181920Time(sec)Force(gF) Figure 5.18 Magnified Peak in Skin Penetration Force-Time Plot for 25X25 Array (Circular Needle-Width 60 m, Pitch 150 m, Length 105 m) Indicating 4.03gF as Penetration Force

PAGE 90

77 Figure 5.19 Typical Skin Penetration Force-Time Plot for 25X25 Array (Circular Needle-Width 60 m, Pitch 150 m, Length 105 m) with Motorized Micrometer. Penetration force vs. time for 25x25 array_60x150um 55 65 75 85 95 105 115 125 135 145 155 165 175 185 195 205 215 225 235 245 255 265 275 285 295 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Time(sec ) Force(gF)

PAGE 91

78 Table 5.5 Split Thickness Penetration Results Using Motorized Meter Microneedle Specifications Needle width ( m) Needle pitch ( m) Needle length ( m) Observed Penetration force (gF) Penetration force/needle (gF) 5x5array-circular 40 100 100 8.10 0.324 5x5array-circular-repeat 40 100 100 0.23 0.009 5x5 array-square 20 150 90 No peak 5x5array-circular 20 200 110 0.13 0.005 5x5 array-square 40 200 100 0.15 0.006 5x5 array-circular 40 200 95 No peak 5x5 array-circular 60 200 100 No peak 25x25 array circular 20 100 95 No peak 25x25array-square 20 100 95 0.13 0.0002 25x25array-circular 40 200 110 2.18 0.003 25x25array-square 40 200 150 2 peaks observed0.28 & 1.98 0.0004/0.003 25x25 array-circular 60 150 105 4.03 0.006 25x25array-circular-repeat 60 150 105 2 peaks observed1.63 & 3.15 0.003/0.005 25x25array-squaretests on different locations of the same skin sample 40 200 110 No peak Force-displacement plot with and without needles50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 0 400 800 1200Displacement(um)Force(gF) Chip with needles Chip without needles Figure 5.20 Plot Illustrating Work Done by 25 X 25 Needle Array (Width 60 m, Pitch150 m, Length 105 m )-Area Between Force Displacement Curves for Chips with and without Needl es.

PAGE 92

79 5.4 Confirmation of microneedle penetration as against indentation on split thickness skin During insertion testing, it was crucial to confirm that th e needles actually penetrate through the corneocytes as against surface indentation. Though the black India ink marks and cutting peaks observed in force plots for some needle configurati ons were useful evidences favoring skin penetration, it was necessary to determine the pene tration depth through sectioning. Several attempts were made in this direction; however the nature of cadaver skin impeded some of these endeavors. These procedures have been reviewed below: The skin specimen on which penetration tests were performed using 60 m diameter 25x25 microneedle arrays of length 60m and 100 m were delivered to School of Medicine and University of Cincinnati respectively for 10-20 m sectioning followed by Eosin and Hematoxylin staining. However these sections did not show any cuts or penetration marks. This was attributed to sealing back of skin after puncture since the microneedles were removed after penetration impeding puncture inspecti on. Another split thickness skin specimen tested with 100m was sub jected to frozen sectioning using cryotome at Shriners Hospital. The specimen w as vertically sectioned (10-12m) after freezing them in small plastic holders using c ryotome. However this approach was again in vain owing to above stated reasons. In this procedure, skin-needle fixation was performed prior to sec tioning to overcome the skin folding issue. The 25x25 microneedle array (width=20m, intraneedle pitch =100 m, height=100 m) was gently pushed against the skin specimen using f inger pressure. The specimen (with the microneedle array) held between two plast ic slides was fixed in formalin solution for 48 hrs. This sample was then stained, repea tedly dehydrated in alcohol baths before vertical sectioning in cryotome. On optical microscopy, some sections demonstrated 60-90 m deep and 20 m wide penetration mar ks into the epidermis layer as illustrated in Figure 5.21 and 5.22. However the intra-array pitch was

PAGE 93

80 observed to be approximately 50m as against 100 m. Multiple in sertions during skin fixation along with non uniform sections due to skin wrinkling could ac count for this irregularity. Figure 5.21 Optical Microscope Image (4X Magnification) Illustrating Microneedl e Penetration Marks (a) (b) Figure 5.22 Optical Microscope Image (20X Magnification) Illustrating (A) Spli t Thickness Skin Section Without Penetration (B) Split Thickness Skin Section with 90 m Deep,20 m Wide Needle Marks into Epidermis

PAGE 94

81 5.5 Penetration Tests on Isolated Stratum Corneum The ability of the microneedles to penetrate enzymatically is olated Stratum Corneum (obtained from U.S. Cell and Tissue Bank, Ohio) was tested. The s pecimen received on a filter paper was initially placed in DI water to determine the str atum corneum side (floating on top). The SC was positioned on the soft polymer for manual testing. Befor e testing, a 1.5 l pipette drop of 0.9% physiological saline was placed on stratum corneum to avoid excessive dryness during testing. 25x25 microneedle array (Circular needles-width 40 m, pitch 200 m and length 100 m) was manually pushed onto the specimen uniformly applying tweez er pressure. The needles made contact with the specimen in certain region and rippe d SC layers, leaving behind a thin layer with holes. Another SC specimen placed on soft polymer was mounted on the skin holdi ng block and subjected to manual insertion of India ink stained microneedle array (25x25 square needle arraywidth 40 m, pitch150 m, length 100 m). The insertion marks are visible in Figure 5.23. Also the plots exhibited several non uniformities due to uneven microneedl e motion as shown in Figure 5.24. The test results which exhibited sharp penetration peaks have bee n summarized in Table 5.6. The graphs as well as the images indicate possible penetrat ion into SC. The large force values are due to dehydration of SC, transforming into a hard material. Also the stratum corneum specimen was of smaller dimension as compared to the area occupied by the needle arr ay.

PAGE 95

82 Figure 5.23 Penetration Marks of 25X25 Microneedle Array (Square Needle-Width 40 m, Pitch 150 m, Length 100 m) on SC after Insertion Test with Magnified Single Needle Mark SC-Penetration force vs time for A29-70 -60 -50 -40 -30 -20 -10 0 10 20 30024681012141618Time(sec)PenetrationForce(gF) Figure 5.24 SC Penetration Plot for 25X25 Microneedle Array ((Square Needle-Wi dth 40 m, Pitch 150 m, Length 100 m) Table 5.6 Stratum Corneum Penetration Results Microneedle specifications Needle width ( m) Needle pitch ( m) Needle length ( m) Observed SC Penetration force (gF) 25x25 array-square 40 150 100 84.68 25x25 array-square 40 100 100 77.05

PAGE 96

83 Penetration experiments provide an approximate quantitative anal ysis for the insertion force into split thickness and stratum corneum (t he topmost layer of the skin) despite specimen surface morphology and unavailability of resources. The needl es were found to be mechanically sturdy to insert skin without fracture. As compar ed to force requirement for metal and polymer microneedles (8-300gF/needle) [9,11], SiO 2 array required very small force of the order of tens of gF for effective penetration.

PAGE 97

84 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK 6.1 Conclusions Hollow biocompatible microneedle arrays were designed, fabrica ted and mechanically evaluated for reliability in transdermal applications. Two d ifferent bulk micromachining techniques Coherent Porous Silicon and DRIE technology were succ essfully implemented to fabricate sharp SiO 2 needles with inherent length controllability feature that pot entially enhance different target applications. The effect of geometry (shap e, needle width, intra array pitch and needle length) on mechanical characterization was studied experime ntally on multidimensional DRIE based needle array chips owing to speedy and reliable fabri cation capability of designed needle patterns from single wafer using the technique. An e xperimental setup was developed in house for fracture and skin penetration force measurement. Microneedle r eliability was quantified experimentally by fracture strength. Results suggested that t he needle failure did not conform to Euler’s force-length inverse parabolic relation (design assumpt ion) and fractured at lower values near material compressive strength due to brittle nature. It was observed that the fracture force decreased with increasing length at constant widths. The force also increased by 2.25 gF as width was increased from 10-40 m at constant 75 m micro needle length (circular geometry). Circular geometry fractured at lower compressive force as compared to square counterparts due to smaller interfacial micro needle area. The fracture force was obser ved to be lower than that obtained theoretically from compressive strength for SiO 2 This could be attributed to defects in brittle

PAGE 98

85 microstructure leading to crack initiation and propagation and proc essing issues like non uniform needle edge and buckling of oxide walls during needle release due t o compressive stresses at SiSiO 2 interface generated during oxidation process. Since the existing setup did not facilitate shear force measurements, these values were determined analytically using Finite Element analysis. Compressive and indentation tests on skin substitute polymer conf irmed that the polymer behaved very different unlike skin due to its isotropic, elast ic and non viscous nature. Hence this polymer was used for preliminary characterization and as a so ft cushion for stratum Corneum penetration testing. Real time microneedle insertion tests into cadaver skin and isolated Stratum Corneum test experimentally highlight the robustness of the needle s to penetrate without fracture. SEM images and sharp peaks observed in the insertion force -time plots suggesting that the needles go through the corneocytes and don't merely indent them. Howeve r confocal microscopy of fixed skin would give a reasonable estimate of penetration dep th. The penetration force could be quantified only for a few microneedle chips whose force plots displayed a sharp peak indicating cutting of stratum corneum. The penetration force require d for 25x25 array was of the order of 1-5 gF as compared to 0.25-0.7 gF for 5x5 array indicating nearl y no force for skin insertion. However measured penetration force/needle (~0.01gF) was several magnitudes smaller compared to fracture force (~1-12gF) supporting microneedle robustne ss during testing. The point loading may be significantly higher than the read out va lues because of skin roughness and not all needles contacting the same time. The results from other tests could not be analyzed due to absence of sharp peaks owing to following limitation: low sam pling rate of data acquisition system for load cell, greater sensitivity of load cell, high r esolution viewing and motorized micrometer. However after each penetration test, distinct needle st ain marks were observed on the skin samples hinting needle penetration. This characterization research confirmed mechanical robustness of SiO 2 microneedles for transdermal applications.

PAGE 99

86 6.2 Recommendations for Future W ork The ultimate goal for out of plane hollow SiO 2 needle fabrication was to investigate their capability for drug delivery and bio-fluid extraction. For this pur pose, initial needle penetration depth into cadaver split thickness skin needs to be investig ated. This can be achieved by injecting fluorescent dye through hollow needles into cadaver skin and then det ermine the penetration depth through confocal microscopy. Investigation of fluid deliv ery capability of the needles needs to be explored by performing fluidic studies at different flow ra tes. Also the needle mechanical strength needs to be enhanced by biocompatible polymer coating to cont ain the residue, in case of in vivo fracture.

PAGE 100

87 REFERENCES 1. Cross S. E. and Roberts M.S., “Physical Enhancement of Transdermal Drug Application: Is Delivery Technology Keeping up with Pharmaceutical Development” Current Drug Delivery 2004, 1, pp 81-92. 2. S. P. Davis, B. J. Landis, Z. H. Adams, M. G. Allen, M. R. Prausnitz, “ Journal of Biomechanics” 2004,Vol. 37, No. 8, pp.1155-1163. 3. Zahn J. ,” Microfabricated Microneedles for minimally Invasive Drug D elivery, Sampling and Analysis”, PhD Dissertation, 1995. 4. McAllister D., Allen M. and Prausnitz M. ,” Microfabricated Microneedles for Gene and Drug Delivery ”, Annual Review of Biomedical Engineering, 2000,2, pp:289-313. 5. Chen J., Wise K., Hetke J. and Bledsoe S., “ A multichannel neural Probe for selective Chemical Delivery at the Cellular level ”, IEEE Transaction on Biomedical Engineering”, 1997, vol. 44, 8, pp 760-769. 6. Henry S., McAllister D., Allen M. and Prausnitz M., “ Micromachined needles for transdermal delivery of drugs ”, IEEE MEMS, Germany, 1998, pp 494-498. 7. Chandrasekharan S., Brazzle J. and Frazier A., “ Surface Micromachined Metallic Microneedles ”, Journal of Microelectromechanical systems, June 2003, vol. 12, 3, pp 281-288. 8. Brazzle J., Papautsky I. and Frazier A., “ Hollow Metallic Micromacined Needle Arrays ”, Biomedical Microdevices, 2000, 2:3, pp 197-205. 9. Chandrasekharan S. and Frazier A., “ Mechanical Characterisation of Surface Micromachined Microneedle Array ”, Journal of Microelectromechanical systems, June 2003, vol 12, 3, pp 289-295. 10. Gardeniers H., Luttge R., Berenschot E., Boer M., Yeshurun S., Hefetz M., Oever R. a nd Berg A. “ Silicon Micromechanical Hollow Microneedles for Transdermal Liquid Transport ”, Journal of Microelectromechanical systems, Dec 2003, vol 12, 6, pp 855862. 11. Stoeber B. and Liepmann D., “ Fluid injection through out-of-plane microneedles ”, Proc. 10th International Conference on Solid State Sensors and Actuators,1999, Senda i,Japan.

PAGE 101

88 12. Zahn J., Talbot N., Leipmann D. and Pisano A., “ Microfabricated Polysilicon Microneedles for Minimally Invasive Biomedical Devices ”, Biomedical Microdevices 2000, 2:4, pp 295-303. 13. Lin L. and Pisano A., “ SiliconProcessed Microneedles ”, Journal of Microelectromechanical systems, March 1999, vol 8, 1, pp 78-84. 14. http://www.essentialdayspa.com/Skin_Anathomy_and_Physiology.htm 15. Forslind, Bo; Lindberg, Magnus; Albano, Robert E., “ Skin, Hair, and Nails: Structure and Function, Basic and Clinical Dermatology” New York Marcel Dekker, Inc., 2004, ISBN 082474313X. 16. McElhany et al, “ Handbook of Human Tolerance ” 17. Nordin M. and Frankel V., “ Basic Biomechanics of the Musculoskeletal System ” 18. IEEE Transactions on electron devices, Vol.ED25, No.10, Oct1978, p.1249 19. http://www.sfu.ca/adm/materials.html 20. Kovacs G., “ Micromachined Transducers Sourcebook ”, WCB McGraw-Hill, ISBN 0-07290722-3, pp 69. 21. Chun K., Haschigushi G.,Toshiyoshi H., Fujita H., Pioufile B., Kikuche Y. Ishikawa J., Murakami Y., Tamiya E., “ DNA injection into cell conglomerates by micromachined hollow microcapillary arrays ”, Transducers, 1999, pp 44-47. 22. Canham L., “ Porous Silicon as a Theurapeutic biomaterial ”, IEEE-EMBS Special talk Conference on Microtechnologies in Medicine and Biology”, 2000, pp 109-112. 23. Lehmann V., “Porous SiliconA new material for MEMS ”, IEEE Proceedings, 1996, pp 1-6. 24. Lehmann V., “ The Physics of Macropore formation in Low Doped n-type Silicon ”, Journal of Electrochemical Society, October 1993, Vol. 140, No. 10, pp 2836-2843. 25. Brent VanDyke, “Development of Coherent Porous Silicon for use in Biological and Optical Applications”, M.S Thesis, May 2000, University of Cincinnati. 26. Potluri K., “Multi material micro needle fabrication using porou s silicon technology”, M.S Thesis,April 2003, University of South Florida. 27. Lehmann V. and Gruning U., “ The limits of macropore array fabrication”, Thin Solid Films, 1997, vol. 297, pp 13-17. 28. http://www.sigma-koki.com/english/D/ActuatorSystems/Optmic /OMDC-B!BJ/OMDCB!BJ.html

PAGE 102

89 29. Pringle D., Koob T. and Kim H., “ Indentation properties of growing femoral head following ischemic necrosis ”, Journal of Orthopedic Research, Jan 2004, Vol 22(1), pp 122-130. 30. http://web.mit.edu/6.777/www/matprops/sio2.htm 31. Lawn B., “Fracture of brittle solids”, Cambridge University Press, 1993. 32. Wilmsen C., Thompson E. and Meissner G., “ Buckling of Thermally-Grown SiO 2 Thin Films ”, IEEE Transaction on Electron Devices, January 1972, pp 122. 33. Teo M., Shearwood C., Ng K., Lu J. and Mocchala Shabbir, ‘ In vitro and In vivo characterization of MEMS Microneedles ”, Biomedical Microdevices, 2005, vol. 7:1, pp 47-52.