Organic Electronics 14 (2013) 1323–1329

Contents lists available at SciVerse ScienceDirect

Organic Electronics

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

Origin of mechanical strain sensitivity of pentacene thin-filmtransistors

1566-1199/$ - see front matter Crown Copyright � 2013 Published by Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.orgel.2013.02.030

⇑ Corresponding authors. Tel.: +49 1786905049 (V. Scenev, J.P. Rabe).E-mail addresses: vscenev@aol.com (V. Scenev), rabe@physik.

hu-berlin.de (J.P. Rabe).

V. Scenev a,⇑, P. Cosseddu b,c, A. Bonfiglio b,c, I. Salzmann a, N. Severin a, M. Oehzelt d, N. Koch a,d,J.P. Rabe a,⇑a Humboldt-Universität zu Berlin, Institut für Physik, Newtonstr. 15D-12489 Berlin, Germanyb University of Cagliari, Dep. of Electrical and Electronic Engineering, Piazza d’Armi, 09123 Cagliari, Italyc CNR–Institute of Nanoscience, Centre S3 via Campi 213A, I-41100 Modena, Italyd Helmholtz Zentrum Berlin für Materialien und Energie–BESSY II, D-12489 Berlin, Germany

a r t i c l e i n f o

Article history:Received 20 November 2012Received in revised form 4 February 2013Accepted 22 February 2013Available online 14 March 2013

Keywords:PentaceneStrainOFETBending experimentMorphology

a b s t r a c t

We report on bending strain-induced changes of the charge carrier mobility in pentaceneorganic thin-film transistors employing a combined investigation of morphological, struc-tural, and electrical properties. The observed drain current variations are reversible if thedeformation is below 2%. The morphology and structure of the active pentacene layer isinvestigated by scanning force microscopy and specular synchrotron X-ray diffraction,which show that bending-stress causes morphological rather than structural changes,modifying essentially the lateral spacing between individual pentacene crystallites. Inaddition, for deformations >2% the rupture of source and drain gold electrodes is observed.In contrast to the metal electrodes, the modification of the organic layer remains reversiblefor deformations up to 10%, which suggests the use of soft and flexible electrodes such asgraphene or conducting polymers to be beneficial for future strain sensing devices.

Crown Copyright � 2013 Published by Elsevier B.V. All rights reserved.

1. Introduction

Thin films of conjugated organic molecules are subjectof intense research due to their applicability in novel(opto-)electronic devices, including organic thin-film tran-sistors (OTFTs). Pentacene (PEN) is the prototypical hole-conducting material with notably high charge carriermobilities of up to 5.5 cm2/V s [1–3] in p-type OTFTs.One key advantage of organic electronic devices is the pos-sibility to produce flexible all-organic OTFTs with the func-tional organic semiconductor films deposited on flexibleplastic foils like Mylar� or polyethylenetherephthalate(PET) as substrate [4]. Through their intrinsic flexibility,OTFTs can be applied as mechanical strain-sensing devices[5,6] exhibiting significant advantages over conventional

types of sensors: they can be processed under ambientconditions and are generally inexpensive to fabricate[4–11]. For sensing strain, reversible changes in the electri-cal characteristics of OTFTs were employed, includingdrain current, charge carrier mobility, threshold voltage,and contact resistance for deformations up to 1–2%[5,6,9,12,13]. In particular, the reported changes in mobil-ity were proposed to be due to morphological changes ofthe PEN layers under mechanical strain and/or the activa-tion of trap states in the PEN/electrode interface region[5,6,9,14,15]. However, up to now there has been no directexperimental evidence to support these suggestions. In or-der to enable targeted research for improving currentOTFTs for future reliable sensing applications, furtherexperimental work is needed to complete the microscopicpicture of strain-impact on OTFT performance. Here we re-port a comprehensive electrical characterization of flexibleOTFT devices under applied bending-stress, relating thedevice characteristics to morphological and structural

1324 V. Scenev et al. / Organic Electronics 14 (2013) 1323–1329

properties of the active layer, as determined by a combina-tion of scanning force microscopy (SFM) and specular X-ray diffraction (XRD). We identify the deterioration of themetal electrodes to be responsible for irreversibilities inthe device characteristics, while the observed modifica-tions of the active organic layer remain reversible fordeformations up to 10%, which suggests using soft organicelectrodes for future improved sensing applications.

2. Experimental

OTFTs were fabricated on 200 lm thin and flexiblepolyethylenetherephthalate (PET) substrates. A gold layerwas vacuum-deposited on the substrates as gate electrode.Polyvinylalcohol (PVA) with ammonium dichromate (AD)salt (cross-linking agent) was deposited from water solu-tion as gate dielectric by spin-coating and subsequentUV-curing for cross-linking. Gold source and drain elec-trodes (Fig. 1a) were patterned by thermal depositionthrough a shadow mask. Pentacene (PEN) was used as or-ganic semiconductor. PEN films were evaporated at a basepressure <10�7 mbar; the film thickness and the depositionrate (0.5 nm/min) were monitored by a quartz crystalmicrobalance placed next to the sample; the device issketched in Fig. 1a. A special apparatus for defined bendingof the OTFT (Fig. 1c) was developed to be suitable for use inboth SFM and specular XRD experiments. XRD measure-ments were carried out at the beamline W1 at DESY–HASYLAB (Hamburg, Germany) with a primary beam en-ergy of 10.5 keV, using the bending apparatus (Fig. 1c) assample holder for in situ experiments, and a standard setup(PVA samples fixed planar on a Si wafer) for reference

Fig. 1. (a) Schematic representation of the top-contact OTFT architecture. (b) Mothe device thickness. Red and dashed green lines indicate the upper and neutral pthe beam surface. (c) The bending apparatus with the OTFT device in place. (dcontacts are indicated for the drain-current sensitivity measurements in the delectrodes and active layer by SFM. (For interpretation of the references to color in

measurements. The electrical and morphological charac-terization of the PEN layers under deformation was carriedout in situ under ambient conditions. For electrical charac-terization a Keithley 2636A sourcemeter was used. Thesurface topography was imaged with SFM (Nano Wizard,JPK instruments). The PEN film strain was calculated withthe model of beam buckling (Fig. 1b), which allows to cal-culate the film strain, e, directly from the bending radius[16], as demonstrated before [9,17]. The bending radiuswas calculated taking into account the homogeneousbending of the PET foil. This setup allowed to strain PENfilms up to 10%, with the upper limit defined by the small-est accessible radius of curvature and the thickness of thePET substrate. All electrical device characterization, XRDand SFM measurements, were carried out under ambientconditions.

3. Results

Transfer curves (i.e., drain current (ID) versus gate volt-age (VG) characteristics) of the devices were recorded fordifferent degrees of strain. The bending-stress was releasedafter acquiring transfer characteristic for a particular strainin order to verify the reversibility of the process (i.e., if thecurrent recovers to the initial value). We find, as expected[9], a decrease in drain current as function of device defor-mation (Fig. 2a). For deformations of up to 1.7%, strainingof the device causes reversible variations of the transfercurves. Starting from 1.7% ID does not recover to the initialvalue of the pristine device upon bending-stress relief.Transfer characteristics exhibit a pronounced hysteresiswith the hysteresis loop-area increasing with increasing

del used to calculate PEN film strain e; R denotes the radius of curvature, tlane, respectively. Strain of PEN film is assumed to be equal to the strain of) Scheme of the device under bending stress with the SFM-tip; electricaleformed OTFT and for the in situ morphological investigations on boththis figure legend, the reader is referred to the web version of this article.)

Fig. 2. Electrical characteristics of a representative OTFT device. (a) Transfer curves recorded during successive device bending and relief of bending stressafter each deformation (indicated with e = 0%) with VD = �60 V. (b) Transfer curves representing the increase of the hysteresis loop area with increasingdeformation (with VD = �60 V). (c) Dependency of the hysteresis loop area versus bending deformation. (d) Device strain sensitivity of the drain current at�60 V gate voltage versus deformation; the red line denotes the linear fit within the range, where the device sensitivity is reversible; the green line indicatesthe transition from reversible to irreversible variation of drain current. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

V. Scenev et al. / Organic Electronics 14 (2013) 1323–1329 1325

deformation (Fig. 2b and c). This is attributed to charge car-rier trapping enhanced by the strain. The effect of bendingstrain induced activation of trap states in the active layerand its relationship to the hysteresis in transfer character-istics of OTFTs has already been studied and attributed tothe intrinsic properties of the organic semiconductor[6,15]. The sensitivity of ID to the strain is determined byplotting the difference between pristine and bended devicecurrents (recorded with �60 V drain voltage) normalizedto the pristine value versus strain (Fig. 2d).

Our electrical measurements reveal that the strain-sensitivity of PEN OTFT devices scales almost linearly withstrain for deformations up to 1.7% (Fig. 2a). Beyond this va-lue we find that the device becomes less sensitive to strain(Fig. 2b). With our experimental set up it was possible toobtain the device sensitivity up to 7.5% mechanical stress.

The finding that after a certain degree of deformationthe current does not recover to its initial value is clear evi-dence that irreversible processes occurred in the devicestructure. To shed light on this issue we performed mor-phological and structural investigations of the deformeddevice, with particular attention on the active layer andthe electrodes.

SFM height images of unstrained devices reveal uni-formly distributed grains with lateral grain sizes of fewhundreds of nm (Fig. 3a–d), typical for PEN films grownon PVA by vacuum deposition [18]. Straining the deviceup to 1.3% did not cause any noticeable variation in PEN

film morphology (Fig. 3b–d). Due to the distortions of theSFM images, which are inherent to the SFM instrumenta-tion including piezo creep, the strain of 1.3% could not bedetected directly.

The PEN film morphology on PVA substrates remainedqualitatively alike for strains beyond 2%, and the impactof bending was clearly reflected in the topography images(Fig. 4b), revealing an increased distance between remotemorphological features for higher values of induced sur-face strain (Fig. 4c). Moreover, the radius of curvaturecould be directly estimated from the cross-section analysisof the SFM micrograph (see Supporting Information,Fig. S1) to (1.1 ± 0.1) mm. Since the bending radius wasmeasured [16,19] to be 1 mm (see Supporting Information,Fig. S2), the value from cross-section analysis is in verygood agreement. However, the resolution of the SFM imag-ing was not sufficient to clearly reveal whether individualgrains became strained or whether the distance betweengrains (along grain boundaries) increased. Therefore weperformed specular XRD measurements to verify strainingof PEN crystallites as described further below.

In contrast to PEN films on PVA, we detected the forma-tion of cracks within the gold electrodes on top of PEN filmfor strains larger than 2% (Fig. 4d and e). The cracks per-sisted after release of strain.

To investigate the impact of strain on the PEN filmstructure, we performed specular XRD measurements ofPEN films in situ deformed up to 3.3% (Fig. 4f). For the un-

Fig. 3. (a) SFM topography of PEN film. (c) Zoom-in before straining; (d) an in situ 1.3% strained PEN film. (b) Provides a cross-section analysis of the SFMheight images carried out on PEN films under deformations of up to 1.3% to allow for a comparison of the lateral distance between two distinctive positions(marked with arrows in (c)). Arrows in (c) correspond to arrows in the cross-sections (b), respectively. Also for deformations of up to 1.3% no significantchanges in PEN films were found compared to the pristine device. (For interpretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)

Fig. 4. Representative SFM micrographs of (a) a pristine PEN layer, and (b) an in situ 10% strained PEN film, together with (c) the corresponding cross-sections; arrows in (a) and (b) correspond to arrows in the cross-sections. (d) SFM micrograph of the top gold electrode, demonstrating crack formationupon deformation of the gold electrode; (e) transition region (interface) of electrode (left)/pentacene (right). (f) Specular XRD data in the range of the (001)-reflection of the PEN thin-film phase (lattice spacing d (001) = 1.54 nm) recorded as function of in situ applied bending-stress e; qz denotes the verticalmomentum transfer; the peak at 0.46 ÅA

0�1 (on the black curve) labels a reflection assigned to the substrate; the peak intensity decreases with the degree of

deformation due to geometrical reasons in the experiment (i.e., less area being illuminated by the X-ray beam). Insets: (left) sketch of the (inclined)molecular orientation in PEN thin-film phase with respect to the PVA substrate (illustrated by yellow triangles); (right) specular XRD data (ex situ) for apristine sample and the same sample after releasing a stress of 4.6%. (For interpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

1326 V. Scenev et al. / Organic Electronics 14 (2013) 1323–1329

strained device (Fig. 4f, black curve) the strong peak at avalue of vertical momentum transfer (qz) of 0.41 ÅA

0�1 corre-

sponds to the (001) reflection of the PEN thin-film poly-morph [20]. This indicates that the PEN molecules adopta standing orientation within the film, however, slightly in-clined by ca. 6� with respect to the PVA substrate normal

(see yellow triangles in the left inset of Fig. 4f), and thatthe molecules are arranged in a herringbone motif typicalfor this specific polymorph [20]. Recently, bending-stress-driven phase transitions were suggested upon flatteningof inward and outward bent PEN devices [21]. To observesuch phase transitions upon deformation of the present

V. Scenev et al. / Organic Electronics 14 (2013) 1323–1329 1327

PEN/PVA films, we performed both in situ specular XRD onfilms deformed up to 3.3% (Fig. 4f), and ex situ specularXRD on a film after the release of strain as high as 4.6%(Fig. 4f, right inset). Most notably, in no case a shift ofthe (001)-peak maximum occurred, which would, how-ever, be expected if a phase transition to any other PENpolymorph [22–25] would take place (expected change invertical momentum transfer of the (001) reflection:Dqz P 0.02 ÅA

0�1), and if, therefore, the lateral intermolecu-

lar distance would accordingly change. Note that, apartfrom phase transitions, relevant stress-related changes ofthe intermolecular distance can be excluded within thisdeformation range, as they typically occur upon pressuresin the GPa region, which is reported in related high-pres-sure investigations [26]. Note, that the elastic modulus ofPEN film is 15 GPa and this large value is due to the poly-crystalline nature of PEN film [27]. In particular, already achange in lattice spacing by 2.5% would correspond toDqz � 0.01 �1 and would, thus, be well captured by ourexperiment.

This is also true for even higher deformations of up to10%, as indicated by SFM micrographs and the correspond-

Fig. 5. SFM micrographs of (a) a pristine PEN film and (b) a flattened PEN film aftebetween two positions marked with arrows in the SFM micrographs of (c) a pristof the references to color in this figure legend, the reader is referred to the web

ing cross-section analyses (Fig. 5) comparing the morphol-ogy of the same PEN area before (Fig. 5a and c) and after(Fig. 5b and d) stress release, where no significant changesin morphology are observed. Note that the identity of theinvestigated areas in the two micrographs is clearly con-firmed by characteristic morphological features presentin both the micrographs (Fig. 5a and b) and in the respec-tive cross-sections (Fig. 5c and d). Moreover, both micro-graphs exhibit the same morphological features, whichdemonstrates that the PEN film completely recovers toits initial state, even after the present severe mechanicaldeformation.

4. Discussion

The electrical measurements agree well with previousreports, where both spherical [5,6] and cylindrical [9,28]deformations showed similar dependencies of the electri-cal response on mechanical strain, i.e., a reversible varia-tion of transfer curves for small strain values andirreversible changes for large strain. Our present study re-veals two distinct deformation regimes of the flexible

r deformation to 10%. Corresponding cross-sections taken along white lineine and (d) flattened PEN film after deformation to 10%. (For interpretation

version of this article.)

1328 V. Scenev et al. / Organic Electronics 14 (2013) 1323–1329

OTFTs: reversible deformation up to 10% occurs for PEN onPVA substrates, while an irreversible deterioration of thegold electrodes is observed starting at a strain value of2%. Note that, in full agreement with this observation, thingold electrodes have been previously reported to deterio-rate for strains exceeding 2% [29]. As it has been studiedelsewhere [29,30], the electrical resistance of thin goldfilms decreases with mechanical strain. Moreover, very re-cently, A.N. Sokolov and coworkers reported about irre-versible changes observed in OTFTs upon mechanicaldeformation. In particular, it was found that for surfacestrain higher than 1.3%, source and drain electrodes buck-ling and cracking starts to appear. This issue typically ledto a decrease of the electrodes conductivity that induceda permanent decrease of the OTFT output current [31].Therefore, we propose that the observed irreversible varia-tion of transfer curves is mainly due to the deterioration ofthe gold electrodes. The reversibility of both the transfercurve variation and the film topography until gold elec-trode deterioration starts infer that the PEN film itself sus-tains large strains of up to, at least, 10%. Most importantly,our study reveals that PEN films might be used in devicesto sense substantially larger strains than previously re-ported. From these findings, hence, a design strategy for fu-ture improved organic sensors clearly arises: using soft,flexible and mechanically stable materials like grapheneor conductive polymers might enable the full reversibilityof strain-related measures in reliable future sensor devices.

Furthermore, our study allows proposing a wellfounded model for the reason of the observed reversibilityof transfer curves variations. One may speculate of tensilestress having two distinct effects on the PEN thin film: (i),the lateral size of the pentacene crystals might becomestretched, therefore modifying the PEN microstructure,which might be detrimental to charge transport, or (ii),the distance between individual neighboring PEN crystal-lites might be increased, therefore reducing the tunnelingprobability of charge carriers in the process of chargetransport through the grain boundaries. From the SFMand XRD data in our present study the related fundamentalprocess can now be clearly identified as (ii): tensile stressleads to an increase in inter-grain distance while leavingthe crystal structure unaffected, as no phase transitionswere observed. It is well established that hopping trans-port, in which a conducting path from the source to thedrain electrode is formed by coupled grain chains orientedin random directions [9] rather than band transport dom-inates conduction in polycrystalline thin organic films[9,32,32]. The increase in distance between crystallinePEN grains translates into an increase in potential barrierfor (thermally activated) tunneling between grains, which,therefore, reduces the overall charge carrier mobility inbended PEN thin film OTFTs and, thus, fully explains theobserved PEN-OTFT strain sensitivity.

5. Conclusions

The underlying mechanism of bending-strain sensitiv-ity observed for PEN OTFTs was investigated by combiningin situ electrical characterization with morphological andstructural investigations. Our results indicate that the

bending-strain sensitivity of PEN OTFTs is due to changesin the film morphology and that the correlated increasein spacing between PEN crystallites is responsible for theobserved decrease in drain current, which is found to bereversible for small deformations. From our study, the irre-versibility in electrical characteristics for deformations lar-ger than 2% is recognized to be due to crack formation ingold source- and drain electrodes. In contrast to irrevers-ible changes in the metal layers, PEN itself is found to befully reversibly modified even for extreme deformationsof up to 10%, which identifies the organic/metal interfaceas bottleneck of the investigated OTFTs and suggests theuse of soft flexible electrodes for future strain-sensingdevices.

Acknowledgements

We thank Wolfgang Caliebe (DESY–HASYLAB), TatjanaDjuric and Armin Moser (both TU-Graz) for experimentalsupport. Financing through the DFG (Germany) and theAustrian Science Fund (FWF) project P21094-N20 is grate-fully acknowledged.

Appendix A. Supplementary material

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.orgel.2013.02.030.

References

[1] T. Hasegawa, J. Takeya, Science and Technology of AdvancedMaterials 10 (2009) 024314.

[2] Y. Yamashita, Science and Technology of Advanced Materials 10(2009) 024313.

[3] N. Koch, ChemPhysChem 8 (2007) 1438.[4] P. Cosseddu, A. Bonfiglio, I. Salzmann, J.P. Rabe, N. Koch, Organic

Electronics 9 (2008) 191.[5] I. Manunza, A. Sulis, A. Bonfiglio, Applied Physics Letters 89 (2006)

143502.[6] I. Manunza, A. Bonfiglio, Biosensors & Bioelectronics 22 (2007) 2775.[7] G. Darlinski et al., Journal of Applied Physics 97 (2005) 093708.[8] H. Kawaguchi, T. Someya, T. Sekitani, T. Sakurai, Ieee Journal of Solid-

State Circuits 40 (2005) 177.[9] T. Sekitani et al., Applied Physics Letters 86 (2005) 073511.

[10] H. Sirringhaus, N. Tessler, R.H. Friend, Science 280 (1998) 1741.[11] T. Someya et al., Proceedings of the National Academy of Sciences of

the United States of America 102 (2005) 12321.[12] P. Cosseddu, A. Piras, A. Bonfiglio, Ieee Transactions on Electron

Devices 58 (2011) 3416.[13] P. Cosseddu, S. Milita, A. Bonfiglio, Ieee Electron Device Letters 33

(2012) 113.[14] F.C. Chen, T.D. Chen, B.R. Zeng, Y.W. Chung, Semiconductor Science

and Technology 26 (2011) 034005.[15] T. Lindner, G. Paasch, S. Scheinert, Journal of Applied Physics 98

(2005) 114505.[16] J.K.F. Yau, N. Savvides, C.C. Sorrell, Physica C 266 (1996) 223.[17] T. Sekitani et al., Journal of Non-Crystalline Solids 352 (2006) 1769.[18] T.B. Singh et al., Advanced Materials 17 (2005) 2315.[19] A. Jedaa, M. Halik, Applied Physics Letters 95 (2009) 103309.[20] S. Schiefer, M. Huth, A. Dobrinevski, B. Nickel, Journal of the

American Chemical Society 129 (2007) 10316.[21] C. Yang et al., Applied Physics Letters 92 (2008) 243305.[22] R.B. Campbell, J. Trotter, J.M. Robertson, Acta Crystallographica 14

(1961) 705.[23] D. Nabok et al., Physical Review B 76 (2007) 235322.[24] S.C.B. Mannsfeld, A. Virkar, C. Reese, M.F. Toney, Z.N. Bao, Advanced

Materials 21 (2009) 2294.[25] D. Holmes, S. Kumaraswamy, A.J. Matzger, K.P.C. Vollhardt,

Chemistry–A European Journal 5 (1999) 3399.

V. Scenev et al. / Organic Electronics 14 (2013) 1323–1329 1329

[26] M. Oehzelt et al., Physical Review B 74 (2006) 104103.[27] D. Tahk, H.H. Lee, D.Y. Khang, Macromolecules 42 (2009) 7079.[28] T. Sekitani, I. Shingo, Y. Kato, T. Someya, Japanese Journal of Applied

Physics Part 1-Regular Papers Brief Communications & ReviewPapers 44 2841 (2005).

[29] S.P. Lacour, S. Wagner, Z.Y. Huang, Z. Suo, Applied Physics Letters 82(2003) 2404.

[30] T. Li, Z.Y. Huang, Z. Suo, S.P. Lacour, S. Wagner, Applied PhysicsLetters 85 (2004) 3435.

[31] A.N. Sokolov, Y.D. Cao, O.B. Johnson, Z.A. Bao, Advanced FunctionalMaterials 22 (2012) 175.

[32] Y. Takamatsu, T. Sekitani, T. Someya, Applied Physics Letters 90(2007) 133516.