Purdue UniversityPurdue e-Pubs

Birck and NCN Publications Birck Nanotechnology Center

5-2011

A batch fabricated capacitive pressure sensor withan integrated Guyton capsule for interstitial fluidpressure measurementTeimour MalekiBirck Nanotechnology Center, Purdue University

Benjamin FogleBirck Nanotechnology Center, Purdue University

Babak ZiaieBirck Nanotechnology Center, Purdue University, bziaie@purdue.edu

Follow this and additional works at: http://docs.lib.purdue.edu/nanopub

Part of the Nanoscience and Nanotechnology Commons

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu foradditional information.

Maleki, Teimour; Fogle, Benjamin; and Ziaie, Babak, "A batch fabricated capacitive pressure sensor with an integrated Guyton capsulefor interstitial fluid pressure measurement" (2011). Birck and NCN Publications. Paper 1021.http://dx.doi.org/10.1088/0960-1317/21/5/054005

A batch fabricated capacitive pressure sensor with an integrated Guyton capsule for interstitial

fluid pressure measurement

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2011 J. Micromech. Microeng. 21 054005

(http://iopscience.iop.org/0960-1317/21/5/054005)

Download details:

IP Address: 128.46.221.64

The article was downloaded on 26/07/2013 at 19:09

Please note that terms and conditions apply.

View the table of contents for this issue, or go to the journal homepage for more

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING JOURNAL OF MICROMECHANICS AND MICROENGINEERING

J. Micromech. Microeng. 21 (2011) 054005 (7pp) doi:10.1088/0960-1317/21/5/054005

A batch fabricated capacitive pressuresensor with an integrated Guyton capsulefor interstitial fluid pressure measurementTeimour Maleki1,2, Benjamin Fogle1,2 and Babak Ziaie1,2,3

1 School of Electrical and Computer Engineering, Purdue University, W Lafayette, IN, USA2 Birck Nanotechnology Center, Purdue University, W Lafayette, IN, USA3 Weldon School of Biomedical Engineering, Purdue University, W Lafayette, IN, USA

Received 1 November 2010, in final form 7 February 2011Published 28 April 2011Online at stacks.iop.org/JMM/21/054005

AbstractIn this paper, we present the design, fabrication and test of a batch fabricated capacitivepressure sensor with an integrated Guyton capsule for interstitial fluid pressure measurement.The sensor is composed of 12 μm thick single crystalline silicon membrane and a 3 μm gap,hermetically sealed through silicon–glass anodic bonding. A novel batch scale method forcreating electrical feed-throughs inside the sealed capacitor chamber is developed. The Guytoncapsule consists of an array of 10 μm diameter access holes etched onto a silicon back-plateseparated from the silicon sensing membrane by a gap of 5 μm. The presence of the Guytoncapsule (i.e. plates with access holes plus the gap separating them from the sensing membrane)allows for the ingress of interstitial fluid inside the 5 μm gap following the implantation, thus,providing an accurate measurement of interstitial fluid pressure. The fabricated sensor is3 × 2 × 0.42 mm3 in dimensions and has a maximum sensitivity of 10 fF mmHg−1.

(Some figures in this article are in colour only in the electronic version)

1. Introduction

Interstitial fluid pressure (IFP) is an important physiologicalparameter determining the exchange of fluid between capillarybeds and inter-cellular space [1]. In addition, it plays animportant role in several pathologies such as lung edema[2], renal disorders [3], hydrocephalus [4] and compartmentalsyndrome [5]. In oncology, IFP is an important clinicalparameter indicating the accessibility of many solid tumorsto therapeutic agents [6–9]. This is in light of the fact thathigher levels of IFP in tumors impede the efficient delivery ofdrugs. In addition to chemotherapy, radiation therapy can alsobenefit from lowering the IFP by increasing the oxygen partialpressure in hypoxic tumor regions.

Although several methods for lowering interstitialpressure using anti-angiogenic agents have been developed[10, 11], to date no method for continuous monitoring ofIFP has been reported. The ability to measure IFP using animplantable micro-transponder would provide valuable datafor evaluating various pharmacological agents used to lowerthe IFP. In addition, by measuring the IFP, one can tailor the

administration of IFP lowering drugs to optimize their efficacy(e.g., administration of the chemotherapeutic agent when theIFP is reduced). All IFP measurement methods reported so far,including the gold standard capillary servo-null technique, relyon inserting an externally connected fragile transducer into thetumor and measuring the pressure in acute settings [12, 13].In 1963, Guyton introduced a perforated capsule for chronicmeasurement of IFP [14]. In his pioneering experiments,Guyton implanted celluloid and methacrylate capsules (inone case the capsule was 2.5 cm in length and 0.8 cm indiameter having 100 holes of 1 mm each) in subcutaneoustissues of dogs. After 2–3 weeks the capsules were filledwith in-growing tissue and interstitial fluids (the air inside thecapsule is absorbed and replaced by interstitial fluid within 4–7 days). Subsequent insertion of a fine needle into the capsuleprovided a direct access to the interstitial fluid for chronicmeasurements of IFP. These measurements for the first timeproved the existence of negative IFP in certain tissues, thuschanging the way physiologists had envisioned fluid transportacross capillary beds.

0960-1317/11/054005+07$33.00 1 © 2011 IOP Publishing Ltd Printed in the UK & the USA

J. Micromech. Microeng. 21 (2011) 054005 T Maleki et al

Recently, MEMS technology has provided a uniqueopportunity to fabricate miniature pressure sensors that canbe implanted in tight anatomical locations (e.g., insidethe eye, brain and vascular system) [15–19]. In general,micromachined pressure sensors can be divided into twocategories, i.e. piezoresistive and capacitive. Capacitivepressure sensors have distinct advantages compared totheir piezoresistive counterparts making them an attractivealternative for many medical applications. These includehigher sensitivity, smaller temperature dependence and lowerpower consumption [20–23]. However, they require a morecomplicated fabrication process due to the requirement for ahermetic lead transfer to a sealed cavity. Several processes(both batch scale and die level) with varying complexitieshave been reported to allow electrical feed-through withoutcompromising the hermeticity [24–27].

In this paper, we present a simple fabrication process fora Guyton capsule embedded capacitive pressure sensor. Thesensor fabrication relies on a combination of (1) modified lift-off process for planarized gold feed-through interconnects,(2) anodic bonding of an SOI and a glass wafer, (3) goldelectroplating to fill the feed-through holes, and (4) dry etchingof silicon handle substrate combined with an isotropic SiO2

wet etch to create the Guyton capsule. The developed sensorcan be used in conjunction with a wireless telemetry system forchronic (long-term) measurement of IFP. The presence of theGuyton capsule will also be effective in isolating the sensingmembrane from the deleterious effects of cell and fibrous tissueencapsulation, thus reducing the long-term drift of the sensor.

2. Sensor structure and design

Figure 1(a) shows the schematic of the pressure sensor.It is composed of a 300 μm thick glass wafer (HoyaSD-2) acting as the fixed plate of the capacitor, a12 μm thick single crystalline silicon layer as the flexiblediaphragm, a sealed cavity with a 3 μm gap, and a Guytoncapsule formed by an array of 10 × 10 × 100 μm3

access holes separated from the sensing membrane by a 5 μmgap (in this paper, we designate the gap separating the accessholes from the sensing membrane as the Guyton chamberwhile reserving the term Guyton capsule for the gap and theaccess holes taken together as a unit). The membrane’s overalldimensions are 1 × 1 × 0.012 mm3. Figure 1(b) shows across-sectional view clearly demonstrating various layers andstructures.

Although many groups have reported on FEM simulationsfor capacitive pressure sensors, most of these attempts onlyconsider the membrane deflection due to the ambient pressuregradient. In our simulations, two additional elements wereadded in an attempt to improve the accuracy. The first wasthe thermal stresses generated during anodic bonding due tothe mismatches between the thermal expansion coefficients ofsilicon and glass. The second was the effects of air trappedinside the sealed cavity, since the anodic bonding used to formthe cavity is done at atmospheric pressure rather than vacuum.

A finite element simulation of the sensor was done usingthe COMSOL MultiphysicsTM software package to obtain a 3D

Hoya SD2glass

Handle layer

Devicelayer

BuriedOxide

Electricalcontacts

(a)

(b)

Guytonchamber

Hermeticallysealed cavity

Membrane

Accessholes

Figure 1. (a) Structure of the capacitive pressure sensor withintegrated Guyton capsule, (b) cross-sectional view, showing themembrane, sensor cavity, access holes, Guyton chamber and theelectrodes.

model of the electric field in the deformed geometry, defined bythe Moving Mesh (ALE) application mode. Figure 2(a) showsthe middle point deflection versus ambient pressure. As can beseen, even at atmospheric pressure, the membrane is deflectedby 1.68 μm. Capacitance versus pressure is depicted infigure 2(b), while figure 2(c) shows the pressure sensorsensitivity versus applied pressure.

3. Fabrication process

Figure 3 shows a schematic of the fabrication process. It startswith an SOI wafer (a 350 μm thick handle layer, a 5 μm thickburied oxide and a 15 μm thick device layer). To reduce thestray capacitance, the device layer was selected to be highlyresistive (>1000 � cm). A 3 μm deep rectangle was etchedin the device layer using reactive ion etching (RIE) to act asthe sensing capacitor gap, figure 3(a). The desired depth wasachieved by controlling the RIE process parameters such astime, power and gas flow rate. Two holes (50 μm wide and15 μm deep) were then created in the device layer by DRIE tobe used as access contacts for the feed-throughs, figure 3(b).Next, the buried oxide at the contact locations was removedusing wet HF (resulting in 20 μm undercuts), figure 3(c). Thisintentional undercut plays an important role in the subsequent

2

J. Micromech. Microeng. 21 (2011) 054005 T Maleki et al

1.65

1.7

1.75

1.8

1.85

1.9

1.95

750 760 770 780 790 800 810 820 830

Cen

ter

Def

lect

ion

(µm

)

Applied Pressure (mmHg)

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

5

750 760 770 780 790 800 810 820 830

Cap

acita

nce

(pF)

Pressure (mmHg)

0

2

4

6

8

10

12

760 770 780 790 800 810 820

Sens

itivi

ty(f

F/m

mH

g)

Pressure (mmHg)

(a)

(b)

(c)

Figure 2. COMSOL multiphysicsTM simulations for the pressure sensor considering the effects of the thermal mismatch and internal airpressure: (a) middle point deflection, (b) capacitance, and (c) sensitivity versus pressure.

electroplating step. Afterward, 300 nm silicon-rich LPCVDnitride layer was deposited to cover the entire wafer to act asthe passivation layer for the metal electrodes, figure 3(d). Thisstep was followed by a deposition and lift-off of 10 nm/100 nmCr/Au electrode layer on both SOI (figure 3(e)) and glasswafer (figure 3(f )). A modified lift-off process in which ashallow (100 nm) recess was etched onto the nitride prior tothe metal deposition was used to achieve a highly planarized(non-uniformity < 60 nm) surface which is crucial for thequality of the anodic bonding process.

Glass and SOI wafers were then cleaned and anodicallybonded (T = 400 ◦C, V = 1000 V), figure 3(g). After bonding,the handle layer was thinned down to ∼100 μm using TMAH(20%, 90 ◦C). Then DRIE was used to create the backsideaccess holes and open the contact area, figure 3(h). A short(3 s) dip in an ultrasonic bath was used to break the nitrideresidue on top of the feed-though access holes. Next a

10 nm/100 nm Cr/Au seed layer was sputtered, followedby 15 μm gold electroplating to fill the feed-throughs,figure 3(i). Finally, a 15 min HF etch was performed toundercut the buried oxide layer and establish the Guytoncapsule (i.e. the access holes and the gap separating themfrom the sensing membrane, interstitial fluid will accumulateinside this space following implantation), figure 3(j ).Figure 4 shows an optical micrograph of the fabricated sensorshowing the contact regions and access holes (magnifiedimage).

The above-mentioned fabrication process containedseveral critical steps designed to ensure a hermetic sealedcavity with electrical feed-throughs. These included: (1) amodified lift-off process to ensure a highly planar surfacefollowing the Cr/Au deposition, figure 5(a) shows AFMimage of the boundary between the Cr/Au interconnect andnitride layer; (2) long zigzag interconnect traces, figure 5(b);

3

J. Micromech. Microeng. 21 (2011) 054005 T Maleki et al

(a) (b)

(c) (d )

(e) ( f )

(g) (h)

(i) ( j )

Figure 3. Fabrication process for a Guyton-capsule-embedded pressure sensor.

3 mm

Electroplated contacts

Access holes

10 µm

Figure 4. Optical micrograph of a fabricated device.

(3) gold–gold contact ring around the feed-through holes,figure 5(c); and (4) electroplated gold at the feed-throughholes, figure 5(d). Leakage in this process is mostly throughthe nanochannels formed along the interconnect traces. Ithas been suggested that if the step height along the feed-through is less than 60 nm, the pressure formed during anodicbonding will close this channel, resulting in a hermetic seal[28]. The AFM in figure 5(a) shows a step height of <50 nm(except an initial edge of 140 nm formed at the sidewall ofthe photoresist which is deformed by the high pressure of theanodic bonding). Furthermore, a long and narrow zigzag linedecreases the leakage probability. The gold–gold contact ringand subsequent through-hole electroplating further blocks theremaining leakage path.

4. Results and discussion

For successful Guyton capsule access-holes fabrication, DRIEparameters (pressure, coil and bias power, and gas flow) had tobe adjusted carefully. As the access holes are deep and narrow,silicon etch may stop before reaching the buried oxide layer,figure 6(a); or be slower in the access holes compared to the

4

J. Micromech. Microeng. 21 (2011) 054005 T Maleki et al

(a)

(b) (c)

(d)

Figure 5. Four methods used to ensure a hermetic seal: (1) modified lift-off process, (a) the AFM image of the boundary between theCr/Au and nitride layer, (2) long zigzag interconnect paths (b), (3) gold–gold bond ring (c), and (4) access hole encapsulation through goldelectroplating (d).

contact areas due to the lag effect, figures 6(b) and 3(g) and (h).SEM picture of the final device with completely open accessholes is depicted in figure 6(c). The device is cleaved suchthat various layers and structures are visible (see the magnifiedimage).

Following the fabrication process, the sensors were testedfor leakage and no base-line drift was observed over a period ofseveral weeks of storage in the nitrogen box. The sensors werethen tested for sensitivity, figure 7. At low pressures (below830 mmHg) capacitance change versus pressure was linear(a sensitivity of 10 fF mmHg−1), as expected theoretically.At higher pressures (above 830 mmHg), the gap was almost

closed and a sharp increase in capacitance was observed. At870 mmHg, the gap was completely closed and two electrodeswere shorted out. No passivation layer was used in thesesensors; however, a thin PECVD nitride passivation layercan be added after electrode formation to prevent shortageat higher pressures and allow the sensor to be used also inthe touch-mode, increasing the dynamic range. Nevertheless,the maximum of the previously reported interstitial pressure is820 mmHg [7], which remains in the linear region of thepressure sensor’s operation. In the case of rare ultrahighinterstitial pressure, the nonlinearity in the output curve canbe compensated electronically using the calibration curve.

5

J. Micromech. Microeng. 21 (2011) 054005 T Maleki et al

(a) (b)

Membrane

Guyton chamber

Access holes

(c)

Figure 6. SEM of a cleaved device showing (a) etch stoppage in the access holes, (b) slower etch rate in the access holes and (c) successfulaccess holes fabrication. Magnified image shows different layers and structures.

29.0

29.5

30.0

30.5

31.0

31.5

32.0

32.5

33.0

760 770 780 790 800 810 820 830 840 850 860

Cap

acita

nce

(pF)

Pressure (mmHg)

Low IFP

High IFP

Figure 7. Experimental results showing capacitance versus pressure.

6

J. Micromech. Microeng. 21 (2011) 054005 T Maleki et al

5. Conclusions

In conclusion, we reported on the first micromachinedcapacitive pressure sensor that incorporates an integratedGuyton capsule. Such a capsule is essential in accuratemeasurement of IFP which in turn will provide criticalinformation for a variety of pathological conditions such asindicating a ‘window of opportunity’ for the administrationof chemotherapeutic agents following a reduction of the IFP.The fabrication process incorporates several steps to ensurea hermetic seal. In addition, the presence of the Guytoncapsule will isolate the sensing membrane from the deleteriouseffects of cell and fibrous tissue encapsulation, thus reducingthe long-term drift of the sensor. Experimental measurementof the sensor’s sensitivity closely matches the simulation;however, the measured base capacitance is higher than thepredicted value. This higher base capacitance is due to thelong interconnection traces and contact pads on the siliconsubstrate. Future efforts will be focused on integrationof the sensor into a wireless telemetry system for chronicmeasurement of IFP.

Acknowledgments

We would like to thank the staff of Birck NanotechnologyCenter at Purdue University for their help and assistance.Funding for this work was provided by the National Instituteof Health grant 1R21EB007256.

References

[1] Hall J E 2010 Guyton and Hall Textbook of MedicalPhysiology 12th edn (Philadelphia, PA: Saunders)

[2] Timby J, Reed C, Zeilender S and Glauser F L 1990Mechanical causes of pulmonary edema Chest 98 973–9

[3] Burnett J C, Haas J A and Larson M S 1985 Renal interstitialpressure in mineralocorticoid escape Am. J. Physiol. RenalPhysiol. 249 F396–9

[4] Matta B F, Menon D K and Turner J M 2001 Textbook ofNeuro-Anesthesia and Critical Care 1st edn (London:Greenwich Medical Media)

[5] Rorabeck C H, Bourne R B, Fowler P J, Finlay J B and Nott L1988 The role of tissue pressure measurement in diagnosingchronic anterior compartment syndrome Am. J. Sports Med.16 143–6

[6] Jain R K 1987 Transport of molecules in the tumorinterstitium: a review Cancer Res. 47 3039–51

[7] Heldin C H, Rubin K, Pietras K and Ostman A 2004 Highinterstitial fluid pressure—an obstacle in cancer therapyNat. Rev. Cancer 4 806–13

[8] Brown J M and Giaccia A J 1998 The unique physiology ofsolid tumors: opportunities (and problems) for cancertherapy Cancer Res. 58 1408–16

[9] Fukumura D and Jain R K 2007 Tumor microvasculature andmicroenvironment: targets for anti-angiogenesis andnormalization Microvasc. Res. 74 72–84

[10] Lin M I and Sessa W C 2004 Antiangiogenic therapy: creatinga unique ‘window’ of opportunity Cancer Cell6 529–31

[11] Dickson P V, Hamner J B, Sims T L, Fraga C H, Ng C Y C,Rajasekeran S, Hagedorn N L, McCarville M B,Stewart C F and Davidoff A M 2007 Bevacizumab-inducedtransient remodeling of the vasculature in neuroblastomaxenografts results in improved delivery and efficacy ofsystemically administered chemotherapy Clin. Cancer Res.13 3942–50

[12] Hargens A R and Ballard R. E. 1995 Basic principles formeasurement of intra muscular pressure Oper. Tech. SportsMed. 3 237–42

[13] Ozerdem U and Hargens A R 2005 A simple method formeasuring interstitial fluid pressure in cancer tissuesMicrovasc. Res. 70 116–20

[14] Guyton A C 1963 A concept of negative interstitial pressurebased on pressures in implanted perforated capsules Circ.Res. 12 399–414

[15] Ko W H, Hynecek J and Boettcher S F 1979 Development of aminiature pressure transducer for biomedical applicationsIEEE Trans. Electron. Devices 26 1896–905

[16] Leung A M, Ko W H, Spear T M and Bettice J M 1986Intracranial pressure telemetry using semicustom integratedcircuits IEEE Trans. Biomed. Eng. 33 386–95

[17] Chau H-L and Wise K D 1988 An ultraminiature solid-statepressure sensor for a cardiovascular catheter IEEE Trans.Electron. Devices 35 2355–62

[18] DeHennis A D and Wise K D 2006 A fully integrated multisitepressures sensor for wireless arterial flow characterizationIEEE J. Microelectromech. Syst. 15 678–85

[19] Chen P-J, Saati S, Varma R, Humayun M S and Tai Y-C 2010Wireless intraocular pressure sensing using microfabricatedminimally invasive flexible-coiled LC sensor implant IEEEJ. Microelectromech. Syst. 19 721–34

[20] Baney W, Chilcott D, Hung X, Long S, Siekkinen J, Sparks Dand Staller S 1997 A comparison between micromachinedpiezoresistive and capacitive pressure sensors SAE Trans.106 562–5

[21] Chau H-L and Wise K D 1987 Scaling limits inbatch-fabricated silicon pressure sensors IEEE Trans.Electron. Devices 34 850–8

[22] Clark S K and Wise K D 1979 Pressure sensitivity inanisotropically etched thin-diaphragm pressure sensorsIEEE Trans. Electron. Devices 26 1887–96

[23] Ko W H 1986 Solid-state capacitive pressure transducersSensors Actuators 10 303–20

[24] Sander C S, Knutti J W and Meindl J D 1980 A monolithiccapacitive pressure sensor with pulse-period output IEEETrans. Electron. Devices 27 927–30

[25] Esashi M, Sugiyama S, Ikeda K, Wang Y and Miyashita H1998 Vacuum-sealed silicon micromachined pressuresensors Proc. IEEE 86 1627–39

[26] Chavan A V and Wise K D 2001 Batch-processedvacuum-sealed capacitive pressure sensors IEEE J.Microelectromech. Syst. 10 580–8

[27] Akar O, Akin T and Najafi K 2001 A wireless batch sealedabsolute capacitive pressure sensor Sensors Actuators A95 29–38

[28] Rogers T and Kowal J 1995 Selection of glass, anodic bondingconditions and material compatibility for silicon–glasscapacitive sensors Sensors Actuators A 46–47 113–20

7