Microelectronic Engineering 121 (2014) 1–4

Contents lists available at ScienceDirect

Microelectronic Engineering

journal homepage: www.elsevier .com/locate /mee

Micro-heaters embedded in PDMS fabricated using dry peel-off process

http://dx.doi.org/10.1016/j.mee.2014.02.0290167-9317/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Address: 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505,Japan. Tel.: +81 3 5452 6224; fax: +81 3 5452 6225.

E-mail addresses: ikjbyun@iis.u-tokyo.ac.jp (I. Byun), rueno@iis.u-tokyo.ac.jp(R. Ueno), bjoonkim@iis.u-tokyo.ac.jp (B. Kim).

Ikjoo Byun, Ryohei Ueno, Beomjoon Kim ⇑CIRMM, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan

a r t i c l e i n f o a b s t r a c t

Article history:Received 14 October 2013Accepted 22 February 2014Available online 1 March 2014

Keywords:Dry peel-off processMicro-heatersPolydimethylsiloxane (PDMS)Self-assembled monolayer (SAM)Surface modification

The present report describes a reliable fabrication method of micro-heaters embedded in poly-dimethylsiloxane (PDMS), and shows the characterization of the micro-heaters. Metallization of PDMSis achieved using a dry peel-off process which involves modifying the surface properties of the substrateand metal patterns through self-assembled monolayer (SAM) and manually peeling off the PDMS withembedded metal layers. Thus, micro-heaters embedded in PDMS can be fabricated by a simpler and eas-ier way compared to a conventional method (e.g. patterning a conducting composite of PDMS using arazor blade). As a result, Au micro-heaters embedded in PDMS were successfully fabricated withoutany chemical swelling and contamination. Micro-heaters on a glass substrate were also fabricated forcomparison with those embedded in PDMS. For heating up to 90 �C, the micro-heaters embedded inPDMS needed only �90 mW compared to those fabricated on the glass substrate needed �530 mW.Moreover, we could not observe any degradation of the micro-heaters by thermal stresses that confirmedby repeatability (10 thermal cycle with a range of 25–89 �C) and stability test (20 min at 90 �C). Micro-heaters took less than 60 s to reach the target temperature (90 �C) and spent less than 60 s to drop toroom temperature. The spatial temperature distribution was not significantly varied with materials ofthe substrate (i.e. PDMS or glass).

� 2014 Elsevier B.V. All rights reserved.

1. Introduction

The development of lab-on-a-chip (LoC) systems has been oneof the dominant themes in analytical instrumentation for chemicaland biomedical analysis applications over the past decade [1–3].Among the components, heat generation and thermal control arecritical in LoC for precise control of protein synthesis and theamplification of nucleic acid molecules using polymerase chainreaction (PCR) [4–8]. Also, chemical synthesis in micro-reactors re-quires precise temperature control [9,10]. Heating blocks [4] orheating wires [10,11] can be used for the temperature control ofLoC devices while the temperature control can be achieved muchprecisely by micro-heaters [5–8].

Polydimethylsiloxane (PDMS) is one of the most popular poly-mer for LoC devices because it is optically transparent, flexible,chemically resistant, bio-compatible, inexpensive and easy to fab-ricate [7–10]. Usually, heaters for PDMS LoC devices are fabricatedon a glass substrate, and assembled with PDMS micro-fluidicchannels. On the other hand, micro-heaters embedded in PDMS

have the advantages such as flexibility, rapid prototyping andgreater compatibility with existing PDMS chips [12].

A conducting composite of PDMS has been investigated by somegroups who create a mixture of various fillers (e.g. carbon blackpowder or silver platelets) and PDMS prepolymers, thus producingan inherently conducting PDMS [12–14]. To pattern the conductingPDMS, the gel-state conducting PDMS is molded into photoresist(PR) patterns, then unnecessary portions of the gel are removedfrom the mold surface using a razor blade. However, this methodis restricted by two reasons: (1) a razor blade can damage mechan-ically the mold (e.g. PR or PDMS) and (2) large volumetric change ofconducting PDMS deteriorate the spatial resolution of heating area(coefficient of thermal expansion: 310 ppm �C�1 for PDMS [15]).Another approach, metallization of PDMS, also have been investi-gated through a self-assembled monolayer (SAM) as a molecularadhesive [16–20]. Recently, we demonstrated that (3-mercapto-propyl)trimethoxysilane (MPTMS) can drastically promote theadhesion between Au and PDMS using a liquid deposition method[18]. Also, we showed the fabrication of Au micro-patterns embed-ded in PDMS without any chemical contamination of PDMS duringthe transfer process [19].

Here, we demonstrate a simple process for fabrication of Aumicro-heaters embedded in PDMS. The key point in fabricatingthe Au micro-heaters is the embedding of a metal layer into PDMS

2 I. Byun et al. / Microelectronic Engineering 121 (2014) 1–4

carried out by a ‘‘dry peel-off’’ process, which involves modifyingthe surface properties of the substrate and metal patterns throughSAM treatment and manually peeling off the PDMS with embeddedmetal layers [19]. Moreover, characterization was conducted forevaluating resistance, time response, area for effective heating,temperature repeatability and temperature stability. Also, we char-acterize micro-heaters on a glass substrate for comparison withthose embedded in PDMS.

2. Experiment

2.1. Fabrication procedure

Chemicals and experimental details are noted at S1 in Supple-mentary material. The fabrication strategy of Au micro-heatersembedded in PDMS is schematically shown in Fig. 1. A Si waferwas treated with a piranha solution [H2SO4 (98%)/H2O2 (30%), 3:1(v/v)] for 10 min to clean the surface of the Si substrate and growa fresh thin oxide layer. After dehydration at 150 �C for 10 min, asparse MPTMS layer was formed on the Si substrate using vapordeposition for 10 min (Fig. 1a). A thin Au layer (thickness:100 nm) deposited onto the substrate by thermal evaporationwas lithographically patterned and wet-etched (Fig. 1b). The sub-strate with both Au patterns and photoresist (PR) patterns wereimmersed in perfluorodecyltrichlorosilane (FDTS) solution (5 mMin hexane) for 5 min (Fig. 1c). After PR removal, the substrate withAu patterns was treated with an ethanolic solution of 20 mM MPT-MS for 2 h (Fig. 1d). A 10:1 (by weight) mixture of PDMS base/cur-ing agent was poured on the substrate (thickness: 3 mm), thenheat-cured in an oven at 60 �C for 3 h, and then maintained at roomtemperature for 12 h (Fig. 1e). Finally, the PDMS with Au micro-patterns was manually peeled off from the Si substrate (Fig. 1f).

For the comparison, Au micro-heaters were also fabricated on1 mm thick glass substrate (Micro-slide glass, S9111, Matsunami,Japan). Au (100 nm) was deposited on the glass substrate with Cr(10 nm) as an adhesion layer. After lithography and wet-etchingof Au/Cr layers, PR was removed with acetone and ethanol, fol-lowed by rinsing with deionized water and drying with a streamof air.

2.2. Design of the micro-heaters and characterization

The micro-heaters were designed with different geometricshapes (Fig. 1g). The width of heating wire was 40, 80 and160 lm with a length of 15 mm. Pads for electrical connection topower supply were designed for 3 � 3 mm. Joule heating was ap-plied to the micro-heaters operations. To apply the voltage andmeasure the electrical resistance, DC voltage current source/mon-itor (6240A, ADCMT, Japan) was used. Copper wires and Au

Fig. 1. (a–f) Schematic illustration of the fabrication process of Au micro-heaters embed(h) before the transfer and (i) after the transfer.

micro-heaters were electrically connected with silver paste (Elec-troconductives, D-362, Fujikura Kasei Dotite, Japan).

A simple setup was utilized to characterize the micro-heaters(Fig. S1 in Supplementary material). The temperature of micro-heaters and peripheral area (i.e. PDMS or glass) were monitoredby infrared (IR) thermo-microscopy (FSV-GX7700, Apiste). Theemissivity for the IR imaging was set to 0.86 for the PDMS and0.95 for the glass substrate [21,22]. Because of a large differenceof emissivity between PDMS (0.86) and Au (0.02), it is difficult tomeasure the temperature of both PDMS and Au, simultaneously.Thus, we set the emissivity of the substrate materials (i.e. 0.86for PDMS or 0.95 for a glass) to the thermo-microscope, andmeasured the highest temperature of the target area. The pointof the highest temperature in PDMS part is adjacent to the Aumicro-heaters.

3. Results and discussions

3.1. Fabrication results

Au micro-heaters embedded in PDMS were successfully fabri-cated using a dry peel-off process (Fig. 1h and i). Because a sparseMPTMS layer between Si and Au made the moderate adhesion, theAu layer was not exfoliated during the wet-process (e.g. photoli-thography, etching, and rinsing). Buckling and wrinkles were ob-served at the surface of Au micro-heaters, but it did not criticallyaffect the performance of the micro-heaters.

3.2. Electrical resistance

At the first attempt, the electrical resistance of Au micro-heat-ers was measured. For theoretical calculation, a simple Eq. (1) forresistance was used where R is the resistance, qR is the electricalresistivity, L is the length and A is the cross-sectional area:

R ¼ qRLA

ð1Þ

We measured the resistance when the applied voltage was 1 V.The results of experiments and theoretical calculation were agreedwell (Fig. S2 in Supplementary material). Thus, we can assume thatthe micro-heaters were not broken or cracked after the transferfrom the donor substrate (i.e. Si wafer) to PDMS.

3.3. Temperature with respect to the applied voltage and powerconsumption

To have a comprehensive understanding of the relationship be-tween the temperature and applied voltage, the temperature of mi-cro-heaters was measured during increasing the applied voltage

ded in PDMS. (g) Design of the micro-heaters, an optical image of Au micro-heaters,

Fig. 2. Temperature of micro-heaters with the substrates of PDMS or glass versus (a) applied voltage and (b) power consumptions. (c) Repeatability test and (d) stability testfor micro-heaters embedded in PDMS (width of heaters: 40 lm).

Fig. 3. (a) Temperature time responses of Au micro-heaters (embedded in PDMS or fabricated on glass substrates) to a square power signal (duration = 60 s) with indicatedpower consumptions. (b) Spatial temperature distribution of PDMS and glass substrate with micro-heaters (width: 40 lm) along the line in (c and d). IR thermo-microscopicimages of micro-heaters (c) embedded in PDMS and (d) fabricated on the glass substrate.

I. Byun et al. / Microelectronic Engineering 121 (2014) 1–4 3

(Fig. 2a). Micro-heaters with smaller width needed higher voltageto be heated up compared to those with larger width. However,when the applied voltage was converted to power consumption,the temperature of micro-heaters was proportional to the powerconsumption and independent to the width of micro-heaters(Fig. 2b). For satisfy in most biological operation, the micro-heaterswere evaluated less than 90 �C [12]. For heating up to 90 �C, themicro-heaters embedded in PDMS needed only �90 mW comparedto those fabricated on the glass substrate needed �530 mW.Because of low thermal conductivity of PDMS (0.18 W mK�1)

compared to that of glass (1.05 W mK�1) [12], the micro-heatersembedded in PDMS could be heated with relatively lower voltageand power consumption.

3.4. Repeatability and stability

Square signals with power consumptions of 50, 70 and 90 mWfor an ON state and 0 mW for an OFF state were applied to the mi-cro-heaters for 10 thermal cycles. The duration of each state (i.e.ON and OFF) was 30 s. The maximum temperature at etch cycle

4 I. Byun et al. / Microelectronic Engineering 121 (2014) 1–4

was 64.4 ± 0.5 �C, 76.2 ± 0.4 �C and 89.0 ± 0.5 �C at 50 mW, 70 mWand 90 mW, respectively (Fig. 2c). In given condition, we could notobserve any degradation of micro-heaters by thermal stresses.

Also, the micro-heaters embedded in PDMS could run for morethan 20 min (Fig. 2d). The temperature of the micro-heaters was65.7 ± 0.7 �C, 76.8 ± 0.2 �C and 89.5 ± 0.9 �C at 50 mW, 70 mWand 90 mW, respectively. Overall, the repeatability and stabilitysuggest that the micro-heaters embedded in PDMS are feasiblefor both heating at a fixed temperature and running thermal cycles.

3.5. Time response

The time response to temperature rise and fall is an importantcharacteristic to most thermal devices. Fig. 3a exhibits such tem-perature responses of the micro-heaters to a square signal from apower supply. The duration of the applied square signal was 60 s.Depending on the power consumption and the materials for sub-strates, three temperatures, 35 �C, 67 �C and 90 �C were evaluated.As expected due to the low thermal conductivity of PDMS, the timeresponse of micro-heaters embedded in PDMS was slower thanthat of micro-heaters fabricated on the glass substrate (durationfor heating to 90 �C: �60 s for PDMS and �10 s for glass substrate,duration for cooling to room temperature: �60 s for PDMS and�10 s for glass substrate). However, the micro-heaters embeddedin PDMS were heated up and cooled down rapidly compared tothe conducting composite PDMS (�240 s) [12].

The spatial temperature distribution was observed when thetemperature of micro-heaters was saturated (Fig. 3b–d). As a re-sult, the spatial temperature distribution was not significantly var-ied with the materials of substrate (i.e. PDMS or glass).

Currently, we are investigating bending and stretching effectson micro-heaters embedded in PDMS as well as observing a cellbehavior during heating the single cell.

4. Conclusions

Micro-heaters fabricated on a donor substrate (i.e. Si wafer)were successfully transferred and embedded in PDMS using a drypeel-off process without any chemical swelling and contamination.This process includes only simple photolithography, metaldeposition, and surface modification by SAM treatment. Themicro-heaters embedded in PDMS were heated up to 90 �C withrelatively low power consumption (90 mW) with good repeatabil-ity and thermal stability (The temperature errors were less than1 �C). Compared to the micro-heaters fabricated on glass substrate,the micro-heaters embedded in PDMS showed slower heating and

cooling characteristics, but the spatial temperature distributionwas not significantly varied by materials of the substrate.

Compared to other techniques (e.g. patterning of conductingPDMS [12]) in terms of a fabrication method, this method is sim-pler and easier, but much reliable (c.f. large volumetric shrinkageof conducting PDMS). In terms of heating characteristics, the mi-cro-heaters fabricated using dry peel-off process shows faster timeresponse with better repeatability. Therefore, we are expectingthat flexible and wearable all-PDMS-micro-chips can be eventuallyrealized with the micro-heaters embedded in PDMS fabricatedusing the dry peel-off process.

Acknowledgement

This work has been, partially, supported by the JSPS Core-to-Core Program A (Advanced Research Networks).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.mee.2014.02.029.

References

[1] A. Manz, N. Graber, H.M. Widmer, Sens. Actuators B 1 (1990) 244–248.[2] D.J. Harrison, P.G. Glavina, A. Manz, Sens. Actuators B 10 (1993) 107–116.[3] J. El-Ali, P.K. Sorger, K.F. Jensen, Nature 442 (2006) 403–411.[4] Y.S. Shin, K. Cho, S.H. Lim, S. Chung, S.-J. Park, C. Chung, D.-C. Han, J.K. Chang, J.

Micromech. Microeng. 13 (2003) 768–774.[5] J. El-Ali, I.R. Perch-Nielsen, C.R. Poulsen, D.D. Bang, P. Telleman, A. Wolff, Sens.

Actuators A 110 (2004) 3–10.[6] D.-S. Lee, S.H. Park, H. Yang, K.-H. Chung, T.H. Yoon, S.-J. Kim, K. Kim, Y.T. Kim,

Lab Chip 4 (2004) 401–407.[7] T. Yamamoto, T. Nojima, T. Fujii, Lab Chip 2 (2002) 197–202.[8] C.-H. Wang, K.-Y. Lien, T.-Y. Wang, T.-Y. Chem, G.-B. Lee, Biosens. Bioelectron.

26 (2011) 2045–2052.[9] J.W. Ha, A. Kundu, J.H. Hang, Fuel Process. Technol. 91 (2010) 1725–1730.

[10] R. Fu, B. Xu, D. Li, Int. J. Therm. Sci. 45 (2006) 841–847.[11] N.G. Wilson, T. McCreedy, Chem. Commun. (2000) 733–734.[12] H.-S. Chuang, S. Wereley, J. Micromech. Microeng. 19 (2009) 045010.[13] L. Liu, S. Peng, X. Niu, W. Wen, Appl. Phys. Lett. 89 (2006) 223521.[14] X. Niu, S. Peng, L. Liu, W. Wen, P. Sheng, Adv. Mater. 19 (2007) 2682–2686.[15] M.V. Kunnavakkam, F.M. Houlihan, M. Schlax, J.A. Liddle, P. Kolodner, O.

Nalamasu, J.A. Rogers, Appl. Phys. Lett. 82 (2003) 1152–1154.[16] K.S. Lim, W.-J. Chang, Y.-M. Koo, R. Bashir, Lab Chip 6 (2006) 578–580.[17] K.J. Lee, K.A. Fosser, R.G. Nuzzo, Adv. Funct. Mater. 15 (2005) 557–566.[18] I. Byun, A.W. Coleman, B. Kim, J. Micromech. Microeng. 23 (2013) 085016.[19] I. Byun, A.W. Coleman, B. Kim, Microsyst. Technol. (2013), http://dx.doi.org/

10.1007/s00542-013-1923-8.[20] J.-L. Lin, S.-S. Wang, M.-H. Wu, C.-C. Oh-Yang, Sensors 11 (2011) 8395–8411.[21] I.-F. Shih, D.B. Chang, Appl. Opt. 30 (1991) 3650–3655.[22] J.L. Lin, M.H. Wu, C.Y. Kuo, K.D. Lee, Y.L. Shen, Biomed. Microdevices 12 (2010)

389–398.