Organic Electronics 11 (2010) 2037–2044

Contents lists available at ScienceDirect

Organic Electronics

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

Fully gravure and ink-jet printed high speed pBTTT organic thinfilm transistors

Alejandro de la Fuente Vornbrock ⇑, Donovan Sung, Hongki Kang, Rungrot Kitsomboonloha,Vivek SubramanianDepartment of Electrical Engineering and Computer Sciences, University of California, Berkeley, CA 94720, USA

a r t i c l e i n f o

Article history:Received 30 April 2010Received in revised form 31 August 2010Accepted 1 September 2010Available online 28 September 2010

Keywords:GravureThin film transistorspBTTTInk-jetSwitching speedPrinted electronics

1566-1199/$ - see front matter � 2010 Elsevier B.Vdoi:10.1016/j.orgel.2010.09.003

⇑ Corresponding author. Present address: Hewleries, Long Down Avenue, Stoke Gifford, BristolKingdom. Tel.: +44 07788 852 123.

E-mail address: delafuentea@gmail.com (A. de la

a b s t r a c t

Organic thin film transistors with channel lengths below 20 lm have been fabricated onplastic substrates using a combination of rotogravure and ink-jet printing exclusively. Gra-vure is utilized to deposit thin, smooth, and narrow metal lines ideal for gate electrodes; apoly(4-vinylphenol) dielectric; and a poly(2,5-bis(3-tetradecylthiophene-2-yl)thieno[3,2-b]thiophene) (pBTTT) semiconductor using a heated roll. A novel fluid guiding techniqueis used to maintain closely spaced ink-jet printed source and drain (S/D) contacts. Togetherthese printing processes yield aggressively scaled yet easily manufacturable TFTs withoperating frequencies of 18 kHz.

� 2010 Elsevier B.V. All rights reserved.

1. Introduction

Printed electronics has received attention as a technol-ogy for the realization of low-cost circuits such as flexibledisplays, RFID tags, and sensors. While substantial work to-wards these goals has been performed using lithographyand ink-jet printing, relatively little progress has beenmade in transferring these advances to higher speed pro-cesses such as gravure printing, flexography, etc. A fewearly demonstrations of OTFTs as well as simple circuitshave been made [1–4] proving the feasibility of using roto-gravure in combination with other printing techniquessuch as flexo and inkjet. However these devices sufferedfrom large feature sizes, hampering their usability in anyrealistic applications.

We have developed a custom high-resolution rotogra-vure printer suitable for the development of printed elec-

. All rights reserved.

tt Packard Laborato-, BS34 8QZ, United

Fuente Vornbrock).

tronics processes on a laboratory scale. Using thismachine, an organometallic silver precursor ink and plana-rized polyethylene napthalate (PEN) films, we have previ-ously demonstrated printed lines with feature sizes assmall as 17.5 lm and a line edge roughness of <2 lm [5].Scaling transistors with printed contacts, particularly gates,has been challenging to date because the processes usedhave provided relatively thick and rough films causing earlybreakdown and decreased yield [6]. In this work, by exploit-ing the small, well-controlled volume of our gravure wellsand the use of an organometallic precursor ink, we demon-strate patterns of sufficient quality for use in fully-inte-grated devices with films as thin as 100 nm, allowing forthe realization of high-performance, well-scaled devices.

In this paper we fabricate thin film transistors usingthese features, thus realizing the smallest rotogravureprinted devices to date. In particular, we introduce ahybrid process using gravure for patterning high-resolu-tion lines, thin dielectrics, and pBTTT semiconductor whichuntil now, has been difficult to print; and using inkjet witha novel fluid guiding technique for precisely aligned upperelectrodes to achieve an overall optimized process flow for

2038 A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044

highly scaled printed devices offering high performance ina scalable, manufacturing-friendly process.

2. Experiment

Both a top gate and bottom gate transistor process wasdeveloped to explore the possibility of using gravure as aprinting technique for TFTs. In the bottom gate process(Fig. 1 (left)), the gate and dielectric were printed usinggravure followed by ink-jet printed source and drain con-tacts and gravure printed semiconductor. In the top gateprocess (Fig. 1 (right)) gravure was used to print the sourceand drain contacts, followed by the semiconductor anddielectric, and finally the gate was printed using ink-jet.A direct comparison of representative devices can be madein Figs. 5 and 6.

Details of the process to print and scale the feature sizesof the first gravure printed metal layer and dielectric aredescribed elsewhere [5]. In brief, the metal lines were gra-vure printed using Inktec PR-010 silver precursor ink onplanarized Q65FA PEN substrates by DuPont Teijin, whichhad been cleaned using an N2 gun, and treated with UVozone for 5 min. The line patterns where composed of30 lm wide gravure cells with �32 lm spacing such thatthe wells almost touched. For all gravure steps reportedhere prints were made at a speed of 1 m/s, which despitebeing slow compared with the speed used in the graphicarts (>10 m/s), is suitable for electronics. A doctor bladepressure of 12 PSI per inch, and a 3 mm nip contact whereused. After printing, the lines were annealed in an oven at150 �C for 15 min. Line widths were 20–21 lm with athickness of 70 nm and a dome shaped profile giving alow roughness of 5.8 nm RMS and 20 nm peak to peak.Poly(4-vinylphenol) with 20% poly(melamine co-formalde-hyde)methylated cross-linker was used as a dielectric bydissolving 1 g of the dielectric and cross-linker into

Fig. 1. Hybrid gravure/ink-jet process flow for bottom gate transistors(left) and top gate transistors (right).

15 mL of propylene glycol monomethyl ether acetate andprinting with electromechanically engraved gravure wellson a 40 l/cm pitch yielding an areal volume of 6.6 mL/m2.After printing, the dielectric was cross-linked at 220 �Cfor 5 min.

2.1. Ink-jet printed contacts

A unique printing process was developed to allow thesource and drain contact lines to be ink-jet printed close en-ough to each other to ensure overlap with the gate (<20 lmline spacing), yet avoid shorting the source and drain due tobulging, which is a common problem found when ink-jetprinting two straight rows of drops (Cabot CCI-300 nano-particle silver ink) in close proximity as seen in Fig. 2.

The growth mechanism for the bulges seen on the leftside of the micrograph is also presented in Fig. 2. Dropsmerging at the beginning of an ink-jet printed line willtry to adopt the least energetic shape possible: a larger, cir-cular drop. As the line begins to get drawn out, ink depos-ited at the end of the line will flow towards the initial dropdue to a pressure differential between the small depositeddrop, the line, and the larger drop or bulge that was ini-tially formed and is now growing. Eventually the pressurebetween the bulge and the line is equilibrated in part dueto the increased size of the bulge, and in part due to theflow resistance of the long line, and the bulge ceases togrow [7,8].

The bulge is thus a natural result of steady state line for-mation that could be minimized using a few approachesincluding the use of larger drop spacing at the beginningof a line and higher print speeds [8]. Such a strategy is

Fig. 2. Two printed thin film transistors (top) with channel lengths�15 lm showing the risk of a source/drain short failure due to the bulgeon the left side of the ink-jet printed lines. The bulge in these lines is dueto the nature of line formation, as shown in the diagram (bottom).

Fig. 3. Optical micrograph of a pBTTT TFT with ink-jet printed contactsemploying fluid guiding to allow for a 15 lm channel length (top).Diagram of the fluid guiding process flow (bottom). Guide drops printedabove and below contact lines, direct fluid flow away from the channelduring the initial process of line formation.

A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044 2039

not attractive for high-speed manufacturing, since raster-ized constant-space printing is much more attractive froma throughput perspective. Instead, a novel fluid guiding ap-proach was implemented where guiding drops are placedat the beginning of the source and drain lines, off-axisand away from the channel, such that the bulge that formsearly in the print process contacts the guiding drops and isguided away from the channel (Fig. 3). Lines printed in thismanner have a much reduced line edge waviness in thechannel region allowing for much better control of the

Table 1Gravure cell parameters–cell depth and areal volume for wells electromechanicallysolutions of pBTTT as shown.

pBTTT concentration % Vol. Cell depth (lm) 45

(mg/mL) (mg/mg) Area volume (mL/m2) 6.1

5 3.85 0.5 Thickness (nm): 30.52 2.31 0.3 18.33 1.54 0.2 12.21 0.77 0.1 6.1

channel length and overlap to the gate, these features arenow limited by the deposition accuracy of the drop whichis ±5 lm. Transistor structures with channel lengths assmall as 15 lm and leakage below 10�6 A/cm2 at 1 MV/cm fields as tested before semiconductor deposition canbe reliably made. These S/D contacts compare favorablywith gravure printed ones where channel lengths below20 lm are currently difficult to achieve.

2.2. Gravure printed semiconductor

pBTTT has been demonstrated to provide some of thehighest field effect mobility and ambient stability obtainedfrom a polymeric semiconductor [9]. However the polymeris difficult to dissolve into a solution and requires the useof hot solvents, which make it difficult to print this mate-rial with techniques such as ink-jet (due to the rapid evap-oration of the heated solvents, causing clogging) orflexography (due to the swelling effects organic aromaticsolvents can have on polymer masters). Gravure rolls anddoctor blades are metallic and thus impervious to organicsolvents regardless of their temperatures. Further theycan be heated [10] to maintain pBTTT solubility and alsoto maintain consistent ink viscosity, thus making gravurea good candidate for depositing this material.

To test the thickness and uniformity of pBTTT filmsformed with gravure, an electromechanically engraved rollwith eight different engravings as detailed in Table 1 wasused. The gravure roll was heated prior to printing in anoven. Because of its high thermal mass, the roll could easilymaintain its temperature through the print process whichlasted approximately 5 min. Ink was delivered to the rollvia a heated syringe just before the print. Substrates weretreated with UV ozone for 4 min prior to printing to im-prove wetting of the pBTTT ink. After printing, the sub-strates were dried in an oven at 100 �C.

The optimal film thickness for pBTTT is known to be be-tween 30 and 60 nm to provide high-quality film forma-tion while minimization bulk leakage effects. Typical inktransfer volumes were found to be approximately halfthe cell volume; therefore, optimal ink dilutions could beestimated and inks with five different concentrations wereprepared. Table 1 provides the expected film thicknessesprinted from these solutions.

The prints were visually inspected to determine whichpatterns produced the most uniform films, and which con-centration provided the most adequate thickness. Directmeasurements of the film thickness where difficult toobtain because the flexibility of the PEN substrate prevents

engraved on a 40 l/cm pitch- and estimated film thicknesses provided from

50 55 60 65 70 75 80

8 10 13.1 15.6 19.3 20.7 23.3

40 50 65.5 78 96.5 103.5 116.524 30 39.3 46.8 57.9 62.1 69.916 20 26.2 31.2 38.6 41.4 46.6

8 10 13.1 15.6 19.3 20.7 23.3

Fig. 4. Schematic of film thickness variation (top) and equivalent gravure cell volume (bottom) for gravure printed pBTTT films. Cells are spaced by 250 lmon the gravure. Large cells (a) merge in the vertical print direction causing band-like thickness variation. Small cells (c) are too far apart to allow ink touniformly spread. An ideal cell spacing (b) allows for fluid flow to form uniform films.

1.E-04 -1.E-05VG=0V

1.E-05 -1.E-05VG 0V

VG=-20V1.E-06

VD=-42.5V VG=-40V

1.E-07

D(A

) VD 42.5V

VD=-80V

G@

A)

VG=-60V

1.E-08ID IG@VD=-42.5

IG@VD=-80V

-8.E-06

-6.E-06ID(

VG=-80V

1.E-09 -4.E-06

1.E-10

1.E-11

1.E-1280604020020

VG (V

0.E+00

-2.E-06

0 00-80-60-40-20020

( )

-80.00-60.00-40.00-20.00.

VD (V)

1.E-04 4.5E-3

1.E-05

1.E 04

4.0E-3

4.5E 3

1.E-06 3.5E-3 ]

1.E-073.0E-3

A^1

/2]

1.E-082.5E-3

D| [

A]

|I D| [

A

1.E-091 5E 3

2.0E-3|I

Sq

rt

Sqrt |ID|

1.E-10 1 0E-3

1.5E-3Sqrt |ID|

Vt

1.E-11 5.0E-4

1.0E 3|ID|

St

0+E0.021-E.1-100-50050

V [V]VG [V]

(a) (b)

(d)(c)

Fig. 5. (a) Transfer characteristics, (b) output characteristics, (c) transfer characteristic andffiffiffiffiffiIDp

showing square law behavior with a �4 V VT, andsubthreshold slope of 9.5 V/dec. The transfer characteristic shows a low gate current throughout the operation range, and provides a lSAT = 1.5e–2 cm2/Vs,on/off of 3.7e–5, lLIN = 2.6e–2 cm2/Vs. d) Shows an optical micrograph of the device (center), W = 754 lm and L = 18 lm.

2040 A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044

a flat reference to be obtained for step-height measure-ments and is too wavy for optical methods such as ellips-ometry. The optimal pattern/concentration was thus

determined by looking for the best uniformity and compar-ing the color of the printed film to a test sample which wasspin coated with a 50 nm thick pBTTT film.

A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044 2041

Variation of film thickness was observable when view-ing the substrate at an oblique angle despite appearinguniform when viewed head on. The variation was too faintto be captured by camera or microscope. Instead, a sche-matic of these results is shown in Fig. 4. Patterns printedwith a high print volume (21 or 23 mL/m2) presented anoscillatory thickness perpendicular to the print directionbut good uniformity in the print direction (Fig. 4(a)). Pat-terns printed with a low print volume (6 and 8 mL/m2)had an incomplete film as wells were spaced too far apartfor the deposited ink to spread and merge. Patterns with aslightly higher print volume (10 mL/m2) were continuousfilms, but exhibited a similar color oscillation as for thedithered patterns (Fig. 4(c)), while a few print volumes(13 and 15.6 mL/m2) yielded optically uniform films(Fig. 4(b)). Best results were found using a concentrationof 3 mg/mL of pBTTT in dichlorobenzene and a 13.1 mL/m2 pattern.

3. Results and discussion

3.1. DC results

DC characteristics where obtained in air using an Agi-lent 4156C semiconductor parameter analyzer. Mobilitieswhere extracted using the square-law equations for linearand saturation mobility as appropriate and the peak trans-conductance extracted from the first derivative of thetransfer characteristics in order to avoid including errorsin extracting the threshold voltage.

As printed, the bottom gate devices exhibited reason-able saturation mobilities of 0.06 cm2/Vs and a modeston/off ratio of 103. It was determined that the poor on/offratios were due to leakage through the gate to the probetip during testing, as the semiconductor had not been pat-terned over the gate contact. After eliminating the semi-conductor around the gate contact, the off currentdecreased dramatically yielding devices with on/off ratiosof 106. pBTTT is quite air-stable, and mobility stabilizedat 0.01–0.02 cm2/Vs after storing in air for 3 weeks. Trans-fer and output characteristics for a prototypical device arepresented in Fig. 5. The onset voltage and threshold voltagefor these devices have had a positive shift from their initialstate, which could be an indication of the incorporation of

ID (A

)

1E-8

1E-7

1E-6

1E-5

-40-30-20-100

ID (A

)

VG (V)

VD=-5VVD=-22.5VVD=-40VIG

Fig. 6. Transfer and output characteristics for top gate, gravure prin

moisture into the film [9]. These are among the highestperforming gravure printed organic TFTs to date.

A top gate device structure was also investigated. In theprocess flow for this device, shown on the right of Fig. 1,gravure is utilized to define the source and drain. Thesecontacts are printed first, using PR-010 and are printed inthe same manner as the gate contacts for bottom gate de-vices. A key advantage to printing S/D contacts with gra-vure is that the spacing between the lines is determinedon the roll, and can be accurately controlled during theprint process because of the contact nature of the printtechnique. However a key disadvantage is that the non-im-age areas of a gravure roll remain coated with residual inkas ink removal by the doctor blade is never perfect [5]. Thisresidual ink can cause chemical contamination of the semi-conductor, and increase the roughness of the surface ontowhich the semiconductor is to be deposited decreasingits crystallinity and performance [11,12].

The electrical characteristics of these devices, shown inFig. 6, exhibit very high off currents. The off-state leakagewas determined not to be caused by gate leakage, but byS/D leakage. This could be caused by a poor semiconductorlimited by the roughness of the surface onto which it wasdeposited, or by chemical doping and contamination fromthe residual ink in the channel. Further, test devices fabri-cated on silicon samples rule out that the performance ofthe material was compromised by subsequent processingof the dielectric and gate. Considering that the channel forthese devices is formed on the top surface of the semicon-ductor which because of its roughness should cause in-creased carrier scattering, the saturation mobility,0.015 cm2/Vs, is comparable to that observed on bottomgate devices.

Removal of the scum deposited during gravure printingof the contacts is thus imperative to improving the perfor-mance of these devices, and will be necessary to realizetop-gated devices by gravure printing.

3.2. AC performance

The DC characterization of printed pBTTT TFTs is essen-tial to understanding the materials properties and elec-tronic behavior of the devices. Performance parametersextracted from DC characterization are used to model how

-4E-6

-4E-6

-3E-6

-3E-6

-2E-6

-2E-6

-1E-6

-5E-7

0E+0-40.00-30.00-20.00-10.000.00

VD (V)

VG=0V

VG=-10V

VG=-20V

VG=-30V

VG=-40V

ted TFTs. W = 854 lm, L = 21 lm, lFE = 6.8e–3, on/off = 17.5.

vin

rin

Cgs gmvgs ro Cop

Cgd

rop rtest

Fig. 7. Schematic of experimental setup (top) for fT extraction and equivalent small signal circuit (bottom).

2042 A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044

a device will operate in a real circuit. However, for actualapplications, AC analysis and characterization is crucial,since DC characterization alone does not provide sufficientinformation about how a transistor will operate under theseconditions. At high frequencies, capacitances dominate atransistors performance, shorting the signal from the gateto channel due to overlap capacitance, decreasing the gainof a transistor in an analog circuit; and ever increasing thedrive current required to charge capacitive loads, limitingthe rise and fall times of digital signals.

For organic thin film transistors, the operating speed ofthe device can also be limited by charge injection from thesource/drain contacts into the channel [13–15]. In this caseinjection is limited by trapping and de-trapping rates atthe metal/semiconductor interface, which can be difficultto measure directly.

A direct form of assessing the AC performance of a tran-sistor is through the measurement of its transition fre-quency, fT. The transition frequency is defined as thefrequency at which an unloaded transistor exhibits unitycurrent gain. This measurement is particularly useful foranalog circuitry where fT gives a clear bound on the maxi-mum frequency an analog circuit can be expected to pro-vide amplification. The relation of fT to device parametersand its importance in understanding the scalability of such

devices has been recently reviewed in this journal by Hop-pe and coworkers [16].

In high performance transistors, the transition fre-quency is measured using high frequency test structureswith load matched transmission lines connecting the con-tacts of the device and measurement results are fitted to asmall signal model to obtain fT [17,18]. For OTFTs, operat-ing frequencies are much lower and microwave techniquescan be avoided. Instead, a direct measurement of fT can bemade by the circuit in Fig. 7. The transistor is operated atits standard DC bias point (saturation), a small signalcurrent is applied to the gate, and the corresponding out-put current is measured at the drain of the device.

In an ideal case, the ammeter at the drain behaves asa short in small signal and thus the output resistance ofthe transistor and any output capacitance are shorted.Ignoring rin, which should be negligible for a good voltagesource, v in ¼ vgs and the input current can then beexpressed as

iin ¼ jxðCgs þ CgdÞvgs ð1Þ

the output current is

io ¼ gmvgs ð2Þ

Fig. 8. Input and ouput current vs. frequency for an all printed pBTTTtransistor. fT is measured at the intersection of the two curves, indicatingunity current gain. Data is for transistor presented in Fig. 5.

A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044 2043

by substituting vgs from Eq. (1), the current gain is given as

io

iin¼ gm

jxðCgs þ CgdÞð3Þ

Which, when ioiin¼ 1 gives fT

fT ¼gm

2pðCgs þ CgdÞð4Þ

A low current small load ammeter was emulated byplacing a test resistor in series with the drain such thatrtest << ro and connecting a high impedance probe to accu-rately measure a voltage drop across the load such thatthe output current could be measured

io ¼vo

rtest: ð5Þ

To further improve measurement accuracy, the imped-ance of the high impedance probe was taken into accountand the circuit model in Fig. 7 was used to accurately ex-tract io.

The input capacitance of the transistor was determinedby using the measured dielectric capacitance per unit area,the size of the transistor, and the overlap area of the sourceand drain contacts (determined using optical microscopy)such that

Cgs þ Cgd ¼ CinsW23

Lþ LSOL

� �þ CinsWLDOL ð6Þ

where Cins is the dielectric capacitance per unit area in F/cm2 W and L are the transistor width and length, respec-tively, and LSOL and LDOL are the overlap lengths for thesource and drain contacts. The channel capacitance was as-sumed to be 2/3 of the channel length in saturation as iscommon practice. Assuming that the overlap lengths arethe same for both source and drain, and expressing theselengths as a percentage / of the channel length, the inputcapacitance can be written as

Cgs þ Cgd ¼ CinsW23

Lþ 2/L� �

ð7Þ

With these equations and the measured output voltage,the small signal output current (iout) and the input current(iin) can be plotted to determine the fT of the device. Biasingan all printed pBTTT TFT with VDS = �60 V and VGS = �56 V,

and applying a 2 V peak-to-peak sinusoidal signal to thegate while sweeping its frequency measurements of iinand io where made. The data in Fig. 8 show a transition fre-quency of 18 kHz which is an excellent result that comparesquite favorably to other gravure printed transistors [1,2].

To better understand the source of this performance, Eq.(4) can be expanded to reflect the influence of deviceparameters by substituting gm ¼ lCins

WL ðVGS � VTÞ and Eq.

(7) to yield

fT ¼lðVGS � VTÞ

2p 23 L2 þ 2/L2� � ð8Þ

which is similar to Eq. (11) from [16], where the contactresistance term has been ignored.

For the device presented here, a transition frequency of27 kHz was calculated with Eq. (8), which is in good agree-ment with the measured value.

4. Conclusion and outlook

In this work, we have presented highly scaled gravureprinted organic transistors offering excellent performance,including some of the highest switching speeds reportedto date for fully-printed transistors. It is evident thatthough clear improvements to fT can be obtained throughimprovements in field effect mobility, the operation of thedevice is extremely sensitive to its channel length, and sub-stantial improvement can be achieving through proper de-vice scaling. Herein lays the success of the presenteddevices where gravure printed features were scaled downto 20 lm providing devices with channel lengths of18 lm. By comparison, the work from Huebler [1] and Jung[2], where channel lengths of 100 and 200 lm were printed,respectively, yielded devices with operation frequencieswell below 1 kHz. With a 10� improvement in mobilitythat can be expected from the pBTTT semiconductor de-vices operating near 200 kHz could be fabricated using thisprocess. Further, if the channel length can be reduced to10 lm, a transition frequency of 900 kHz would be ex-pected making these printed devices suitable for standardscompliant RFID and high-resolution displays.

References

[1] A.C. Huebler, F. Doetz, H. Kempa, H.E. Katz, M. Bartzsch, N. Brandt, I.Hennig, U. Fuegmann, S. Vaidyanathan, J. Granstrom, et al., Ringoscillator fabricated completely by means of mass-printingtechnologies, Organic Electronics 8 (2007) 480–486.

[2] M. Jung, J. Kim, J. Noh, N. Lim, C. Lim, G. Lee, J. Kim, H. Kang, K. Jung,A.D. Leonard, J.M. Tour, G. Cho, All-printed and roll-to-roll-printable13.56-MHz-operated 1-bit RF tag on plastic foils, IEEE Transactionson Electron Devices 57 (2010) 571–580.

[3] G. Cho, R2R Printed TFTs and RFID Tags: Demonstration of All Printed13.56 MHz Operated 1 Bit RFID Tags, Hsinchu, Taiwan: ISFED, 2008.

[4] J. Jo, J. Yu, T. Lee, D. Kim, Fabrication of printed organic thin-filmtransistors using roll printing, Japanese Journal of Applied Physics 48(2009) 04C181.

[5] A. de la Fuente Vornbrock, J.M. Ding, D. Sung, Huai-Yuan Tseng, V.Subramanian, Printing and Scaling of Metallic Traces and CapacitorsUsing a Laboratory-Scale Rotogravure Press, 2009 FlexibleElectronics and Displays Conference and Exhibition, Phoenix, AZ,USA, 2009, pp. 1–7.

[6] D. Sung, A. de la Fuente Vornbrock, V. Subramanian, Scaling andoptimization of gravure-printed silver nanoparticle lines for printed

2044 A. de la Fuente Vornbrock et al. / Organic Electronics 11 (2010) 2037–2044

electronics, IEEE Transactions on Components and PackagingTechnologies 33 (2010) 105–114.

[7] P.C. Duineveld, The stability of ink-jet printed lines of liquid withzero receding contact angle on a homogeneous substrate, Journal ofFluid Mechanics 477 (2003).

[8] D. Soltman, V. Subramanian, Inkjet-printed line morphologies andtemperature control of the coffee ring effect, Langmuir 24 (2008)2224–2231.

[9] I. McCulloch, M. Heeney, C. Bailey, K. Genevicius, I. MacDonald, M.Shkunov, D. Sparrowe, S. Tierney, R. Wagner, W. Zhang, M.L.Chabinyc, R.J. Kline, M.D. McGehee, M.F. Toney, Liquid-crystallinesemiconducting polymers with high charge-carrier mobility, NatureMaterials 5 (2006) 328–333.

[10] H. Kipphan et al., Handbook of Print Media, Springer, Heidelberg,2001.

[11] M.L. Chabinyc, R. Lujan, F. Endicott, M.F. Toney, I. McCulloch, M.Heeney, Effects of the surface roughness of plastic-compatibleinorganic dielectrics on polymeric thin film transistors, AppliedPhysics Letters 90 (2007) 233508.

[12] Y. Jung, R.J. Kline, D.A. Fischer, E.K. Lin, M. Heeney, I. McCulloch, D.M.DeLongchamp, The effect of interfacial roughness on the thin filmmorphology and charge transport of high-performancepolythiophenes, Advanced Functional Materials 18 (2008) 742–750.

[13] Y. Majima, D. Kawakami, S. Suzuki, Y. Yasutake, Simultaneousmeasurements of drain-to-source current and carrier injectionproperties of top-contact pentacene thin-film transistors, JapaneseJournal of Applied Physics Part 1 Regular Papers Short Notes andReview Papers 46 (2007) 390.

[14] T. Miyadera, T. Minari, K. Tsukagoshi, H. Ito, Y. Aoyagi, Frequencyresponse analysis of pentacene thin-film transistors with lowimpedance contact by interface molecular doping, Applied PhysicsLetters 91 (2007) 013512.

[15] K. Seshadri, C.D. Frisbie, Potentiometry of an operating organicsemiconductor field-effect transistor, Applied Physics Letters 78(2001) 993.

[16] A. Hoppe, D. Knipp, B. Gburek, A. Benor, M. Marinkovic, V. Wagner,Scaling limits of organic thin film transistors, Organic Electronics,2010.

[17] S. Lee, Direct extraction technique for a small-signal MOSFETequivalent circuit with substrate parameters, Microwave andOptical Technology Letters 39 (2003) 344–347.

[18] J. Raskin, R. Gilon, G. Dambrine, J. Chen, D. Vanhoenacker, J. Colinge,Accurate characterization of silicon-on-insulator mosfets for thedesign of low-voltage, Low-Power RF Integrated Circuits, AnalogIntegrated Circuits and Signal Processing 25 (2000) 133–155.