Contact effects in organic thin-film transistor sensors

Organic Electronics 10 (2009) 233–239

Contents lists available at ScienceDirect

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

Contact effects in organic thin-film transistor sensors Luisa Torsi *, Francesco Marinelli, M. Daniela Angione, Antonio Dell’Aquila, Nicola Cioffi, Elvira De Giglio, Luigia Sabbatini Dipartimento di Chimica, Università degli Studi di Bari, via Orabona 4, 70126 Bari, Italy

a r t i c l e

i n f o

Article history: Received 20 October 2008 Received in revised form 12 November 2008 Accepted 13 November 2008 Available online 30 November 2008

PACS: 72.80.Le 73.40.Cg 85.30.Tv 87.85.fk

a b s t r a c t Contact effects in organic thin-film transistors (OTFTs) sensors are here investigated specifically respect to the gate field-induced sensitivity enhancement of more than three orders of magnitude seen in a DHa6T OTFT sensor exposed to 1-butanol vapors. This study shows that such a sensitivity enhancement effect is largely ascribable to changes occurring to the transistor channel resistance. Effects, such as the changes in contact resistance, are seen to influence the low gate voltage regime where the sensitivity is much lower. Ó 2008 Elsevier B.V. All rights reserved.

Keywords: Thin films Transistors Chemo-bio sensors Conducting polymers Molecular electronics

1. Introduction The scientific and technological driving force towards the development of performing conducting polymer (CP) based solid state sensors is still very strong despite this field has been initiated more than 30 years ago [1]. Recently, organic thin-film transistor (OTFT) sensors have risen the interest of the scientific community for their enhanced level of performance [2–6]. In this configuration, highly repeatable responses were measured by properly gate biasing the device [2,3] whereas broad chemical selectivity was conferred either by covalently bound side groups [7] or by means of CP blends [8]. Besides, an important assessment of their selectivity capabilities was achieved with a chiral OTFT, that exhibited field-effect * Corresponding author. E-mail address: [email protected] (L. Torsi). 1566-1199/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.orgel.2008.11.009

amplified sensitivity allowing detection of optical isomers in the tens part-per-million (ppm) concentration range [9], i.e. with a three order of magnitude sensitivity improvement [10,11]. This report aims to demonstrate that the sensing process, in particular the field-induced sensitivity enhancement, can be largely ascribed to changes of the channel resistance Rch, eventually ruling out dominating contributions from of the contact resistance (Rc) or leakage current variations. The results reported can also shed light on the origin of the contact resistance in OTFTs. 2. Experimental methods 2.1. Organic thin-film transistors fabrication The transistors used for this study have a bottom-gate structure and consist of a highly n-doped silicon wafer

234

L. Torsi et al. / Organic Electronics 10 (2009) 233–239

(resistivity: 0.02–1 X/cm) coated by a 300 nm thick SiO2 thermal oxide (dielectric capacitance per unit area: Ci = 10 nF cm2). Each sample was fabricated on a cleaved wafer piece of ca. 2 cm2. The silicon substrate with a gold pad acts as the gate contact (G) while the silicon dioxide is the gate dielectric. The SiO2 surface was alkyl functionalized by treatment in 1,1,1,3,3,3-hexamethyldisilazane saturated vapors for 24 h. The dielectric surface was covered by a a,x-dihexyl-hexathiophene (DHa6T) thin-film deposited by thermal evaporation at a base pressure of 8  107 with the substrate kept at room temperature. A series of 30 gold source (S) and drain (D) contacts were defined, by thermal evaporation through a shadow mask, directly on the organic films. Before the measurements, the OTFTs were annealed under a vacuum of 105 Torr at 100 °C for 30 min. Transistors were produced with different channel lengths (distance between the two probed pads), namely: L = 0.2 mm, 0.6 mm and 1 mm, while the channel width (the pads’ longer dimension, W), was kept constant at 4 mm. The sensors were fabricated and measured in a standard laboratory environment and operated at room temperature. The synthetic approach for preparing the a,xdihexylsexithiophene comprised a Stille coupling between the 5,50 -bis-trimethylstannyl-2,20 -bithiophene and the 5-bromo-50 -hexyl-2,20 -bithiophene. The product precipitated from the reaction medium as a red powder and was purified from soluble byproducts by subsequent Soxhlet washing with methanol, chloroform, acetone and n-hexane. 2.2. Sensing measurements The analyte, in a nitrogen stream and at a controlled concentration, was delivered through a nozzle (positioned at a fixed distance of few mm from the device channel surface) directly onto the active layer surface, as depicted in Scheme 1. The Ids–Vg transfer characteristics were measured by biasing the device in a common source configuration by sweeping Vg (from positive to negative potentials) at a fixed Vds, namely 50 V. The current, Ids, flowing in the channel region was measured in nitrogen and in a flux of controlled butanol concentration in N2, obtained through a system of computer controlled flow meters. The response measurements were performed at relatively high analyte concentrations (3750–11,250 ppm) as this condition allowed to better discriminate the effects of the contact resistance variation. Both the transfer characteristics (measured in N2 and in the analyte atmosphere), lasted for 45 s with a total flux of 200 ml/min. Before starting

each run the unbiased device was conditioned by exposure to the analyte atmosphere for 45 s. 3. Results and discussion The system investigated is a a,x-dihexylsexithiophene (DHa6T) OTFT with a bottom-gate top contact configuration ( Scheme 1). Such a device has been widely investigated for 15 years now [12–17] and can, in many respects, be taken as a model system. The DHa6T thin-film, exhibiting a morphology formed of a three dimensional percolation-type network, with nano-sized crystalline domains separated by voids of comparable size [18], acts both as transistor channel material and as sensing layer. In agreement to what already reported for OTFTs exposed to volatile organic compounds, changes in the drifting source-drain current were seen upon exposure of the DHa6T OTFT to 1-butanol and the response and recovery times were generally quite short, falling in the 10–100 s of seconds range at most [2,3]. Weak chemical interactions are likely to occur between the DHa6T thin-film and the alcohol molecules, the latter being eventually partitioned between the solid (the DHa6T thin-film) and the gaseous phase. Since diffusion of the molecules into the crystalline grains is unlikely, such molecules largely percolate through the voids around the grains till the interface with the dielectric is reached. The analyte molecules are physisorbed at the grains surface and this can enhance the potential barriers at the boundaries [19], generating also charge trapping effects. The degree of physisorption is a function of the chemical affinity between the active layer and the analyte [7], and it occurs in the DHa6T bulk, as well as at the interface with the dielectric, were the twodimensional transport takes place. Typical DHa6T OTFT current–voltage characteristics in the inert N2 atmosphere are reported in Fig. 1a. Here, the source-drain current (Ids) is reported as a function of the source-drain bias (Vds) for different gate biases (Vg). As DHa6T is a p-type semiconductor, negative values of Vds and Vg (|Vg| > |Vt|) drive the device in the on-state (accumulation mode), while Vg values below the threshold voltage, Vt, generate a regime of charge depletion (off-state or depletion mode) [2]. The field-effect mobility (l) and the threshold voltage, Vt, are graphically extracted from the relevant square root of Ids vs. Vg plot (Fig. 1b) resulting in l = 0.1 cm2/Vs, Vt below 2 V while a current amplification of 104 can be achieved. These figures of merit are state of the art values for DHa6T [12–15], although best mobility

Scheme 1. DHa6T thin-film transistor sensor structure upon exposure to 1-butanol vapors along with biasing details.

235

L. Torsi et al. / Organic Electronics 10 (2009) 233–239

a

0.012

b

-80

Vt = -1.72 V

0.010

μ = 0.1 cm2/ V s

square root - I ds(A)

Ids (μA)

-60

-40

-20

0

0.008 0.006 0.004 0.002 0.000

0

-20

-40

-60

-80 -100

20

0

-20 -40 -60 -80 -100

Vg (V)

Vds(V)

Fig. 1. (a) DHa6T current–voltage characteristics in a nitrogen flux. Gate voltages go from 20 V to 100 V in steps of 10 V. (b) The relevant square root of source-drain current as a function of the gate bias is reported (L = 0.2 mm).

values can reach 0.5–1 cm2/Vs [16,17]. Data reported in the supporting information show Ids values at zero Vds biases at different Vg. It is apparent that the highest leakage current (Ids values at zero Vds) is at most of ca. 60 nA. Since the Ids currents measured reach 100 lA, the sum of the leakage currents is below 0.1% of Ids, this being a figure generally considered as adequate. The field-effect mobility in a OTFT is generally gate bias dependent and can be better estimated by considering the channel conductance

g ds ¼

@Ids @V g

ð1Þ g ds 

measured at low drain voltages (Vds < 5 V) [20]. In this case being Vds  Vg, gds can be expressed as follows:

g ds

@Ids W ¼  lC i ðV g  V t Þ L @V ds

ð2Þ

-1.4

Vt= -7.2 V

-1.2

2

µ= 0.07 cm / V s

-1.0

gds(1/M ohm)

This equation is, at first approximation, valid also for non-constant mobility. In Fig. 2 the data for the channel conductance, gds, extracted from the I–V characteristics of Fig. 1 at low source-drain voltage (0 < Vds < |5 V|), are reported as a function of Vg. The extracted mobility is of 0.07 cm2/Vs in this case. The data modeling within this framework allows also to give a first estimate of the contact resistance effects. In the case of zero contact resistance Eq. (2) holds, at low Vds (linear region). Introducing a series resistance as contact resistance, Rc, results in changing Vds by a VdsIdsRc and Eq. (2) becomes

-0.8 -0.6 -0.4 -0.2 0.0 0

-20

-40

-60

-80

-100

Vg (V) Fig. 2. DHa6T OTFT channel conductance, gds, plotted as a function of the gate bias.

W=L  C i lðV g  V t Þ 1 þ W=LC i lRc ðV g  V t Þ

ð3Þ

The transport in the channel region appears as not limited by the contact resistance, since gds shows a quite good linear dependence from Vg (correlation factor of 0.9997). A slight deviation from linearity at higher Vg biases is indeed observed and this calls for further investigations. To quantitatively extract the contribution of the contact resistance the transfer line method, TLM [21,22] was used. Accordingly, the I–V characteristics of a set of three DHa6T OTFTs, with channel lengths of L = 0.2, 0.6 and 1 mm were measured in N2. The total device resistance, R, was extracted from the linear region (0 < Vds < |10 V|) for each of the three OTFTs. In Fig. 3 the data are reported as (R  W) vs. L at different Vg voltages. According to the transfer line method, the total resistance, R, is the sum of the channel resistance, Rch, and the contact resistance Rc, the latter being graphically extrapolated at L = 0. In Fig. 4 the contact (Rc) and channel (Rch) resistances of DHa6T OTFTs with different channel lengths are plotted vs. Vg. As expected, in longer channel devices the contact resistance weights much less, while in the shorter channel devices they are of comparable size, at least at higher gate biases. This provides preliminary evidences that contact related effects are not dominating the on-state OTFT transport properties. Besides, the curves in Fig. 4 exhibit contact resistance at high Vg biases falling in the 105 X cm range. Although lower Rc have been reported in some cases for

236

L. Torsi et al. / Organic Electronics 10 (2009) 233–239

Vg=-20 V

8

N2

Vg= -40 V

RxW (MΩ cm)

Vg= -60 V Vg=-80 V

6

Vg= -100 V

4

2

0 0.0

0.2

0.4

0.6

0.8

1.0

Channel Length - (mm) Fig. 3. The total device resistances (R) extracted from the Ids–Vds curves (0 < Vds < |10 V| and Vg = 20 V, 40 V, 60 V, 80 V, and 100 V) of DHa6T OTFTs of different channel lengths exposed to nitrogen. W is the transistor channel width.

N2

R (MΩ)

10

1 R contacts Rch (L= 0.2 mm) Rch (L= 0.6 mm) Rch= (L= 1 mm)

0.1 20

40

60

80

100

Vg(V) Fig. 4. DHa6T OTFT contact and channel resistances as extracted from the plots in Fig. 3.

pentacene organic semiconductors [22,23], values in the 105 X cm are typical for thiophene based materials contacted with gold electrodes [24,25]. The same DHa6T OTFT devices were then exposed to saturated vapors of 1-butanol. A comparison of the device I–V and transfer characteristics in nitrogen and butanol are reported in Fig. 5. The DHa6T OTFT Ids–Vg, transfer characteristics, are reported in Fig. 5a, along with the relevant log-plot (Fig. 5b), while typical device Ids–Vds characteristics are reported in panel (c) and (d). The black solid curves in Fig. 5a and b show the current drifting in the channel region between the source and the drain contacts at a fixed drain-source bias (Vds = 50 V) while sweeping Vg between 20 V and 100 V. Generally, as it is apparent in Fig. 5 and as already widely reported, a on-state Ids decrease occurs upon exposure of a thiophene based OTFT to alcohol vapors [3,25]. This can be explained by the elicited poten-

tial barrier increase at grain boundaries upon exposure to the analyte or, equivalently, by a charge trapping effect. However, a closer inspection of Fig. 5b, reveals that for gate biases falling below a cross-over point, a Ids current increase occurs upon exposure to the analyte. This effect has been seen consistently in most sensing OTFT, using differently substituted thiophene as well as phenylenethiophene based organic semiconductor, when exposed to organic vapors as well as to inorganic species [9]. It can be also seen on a recently published report, by a different laboratory [6]. In this case a doping effect can be postulated as an increase of the off-current is seen. Sample dependent features were for instance the gate bias at which the cross-over took place and in the intensity of the current differential change. The response of the OTFT sensor DI was defined as follows:

DI ¼ Ids ðN2 Þ  Ids ð1  butanolÞ

ð4Þ

In Fig. 5b it can be seen that different DI values can be measured at different Vg biases and as Ids are negative values, DI is negative above the cross-over while is positive below. The DHa6T I–V characteristics, exposed to butanol at different concentrations, were measured for devices with different channel lengths and plots, similar to those reported in Fig. 3, were constructed. The relevant data are reported in Fig. 6a. Also in this case the Rch and Rc data are extracted by the TLM and a comparison of the contact and channel resistances in N2 (data as in Fig. 4) and in butanol are given in Fig. 6b and c, respectively. In Fig. 7 the relevant Rch and Rc variation are reported and the following features can be outlined. The DRc values can be several times lower then DRch, especially for longer channel devices. This holds true also for the relative percentage variations as DRc/R is ca. 8% while DRch/R can be as high as 20%. Moreover, the variations of the contact and channel resistances, upon exposure to 1-butanol vapors, go in opposite directions as DRch are negative values while DRc are positive ones. Such a result allows establishing some correlations. As already mentioned, a cross-over point, discriminating negative and positive DI current changes, was identified in the transfer characteristics (curves in Fig. 5b being a typical example). It is straightforward to associate negative resistance changes to negative current variations and vice versa. This means that, when a negative DI is seen, it is the channel resistance that is changing, while when DI is positive, contact resistance variations are dominating. Therefore, the plot of the OTFT sensor transfer characteristics, measured in an inert and in the analyte atmosphere, gives an indication of the voltage range in which the contribution of contact effects dominates. To better understand the interplay occurring between the channel and the contact resistance changes in OTFT sensors a modeling of the data is in progress. As a further step, the DHa6T OTFT transfer characteristics have been measured also in atmospheres of butanol vapors at different concentrations. The DI values were evaluated at different gate biases and at constant concentration as shown in Fig. 5b. The DI values at a fixed Vg bias

237

L. Torsi et al. / Organic Electronics 10 (2009) 233–239

a

-60

-4

10

b

N2

N2

-50

-5

Butanol -40

ΔI

-30 -20

ΔI<0

-6

10

|Ids(A)|

Ids(μA)

Butanol

10

-7

10

-10

ΔI>0

-8

10 0 20

0

-20

-40

-60

-80

-100

20

0

-20

Vg(V) -50

d

N2

-40

-30

-30

-10

0

0 -20

-40 -60 Vds (V)

-80

-100

-100

-20

-10

0

-80

Butanol

-40

-20

-60

-50

Ids (μA)

Ids (μA)

c

-40 Vg(V)

0

-20

-40 -60 Vds (V)

-80

-100

Fig. 5. DHa6T OTFT transfer characteristics in N2 and 1-butanol are reported in panel (a). Panel (b) shows the relevant log-plot. I–V characteristics curves in N2 and 1-butanol are reported in panels (c) and (d), respectively. The OTFT channel length is L = 0.2 mm and the analyte concentration is 11,250 ppm.

and at different concentration constitute the data points for one calibration curve, the slope (m) of each curve being the device sensitivity. The relevant m data are given in Fig. 8 for L = 0.2 mm and L = 1.0 mm DHa6T OTFTs. The occurrence of a gate field dependent sensitivity is apparent and a sensitivity enhancement of more than three orders of magnitude is seen when the device is driven from the off to the on-state. Similar data were gathered for a phenylene– thiophene [9], as well as for differently substituted thiophene oligomers exposed to organic and inorganic species, showing that this can be a general property of OTFT sensors. Fig. 8 further shows that the slope values too can be positive (hollow red points) and negative (solid black points). Also in this case it is possible to associate the negative slopes to the channel resistance variation, while the positive ones to the contact resistance changes. This means that contact resistance changes dominate in the low gate voltage regime while channel resistance variations are responsible for the much higher sensitivity at gate biases that drive the transistor in the on-state. Since red points values are consistently orders of magnitude lower than negative ones, the contact resistance variations affect the field-enhanced sensitivity only negligibly. Besides, it is also

important to outline that the associated standard error on the response repeatability is much lower in the on regime than in the off-one. All the evidences provided so far show that the enhanced sensitivity observed is in fact not dominated by contact related effects. Such effects are present, but influence the TFT sensor low gate voltage regime. Interestingly, in this operating regime an OTFT is much more likely to behave as a chemiresistor. As to the leakage is concerned, it has been shown in the Supplementary information file that the leakage is limited to 0.1% of the maximum Ids current, while the total relative current variation can be as high as 20%. Therefore, leakage current contribution to the total current variation is negligible too. A final remark is due on the sensing mechanism. Upon exposure of the OTFT active layer to a volatile organic vapor a partition of the analyte molecules between the solid and the gaseous phase occurs. The chemical affinity between the analyte and the active layer modulates the degree of physisorption of the analyte molecules at the grains’ surface. This in turn can enhance the potential barriers at the grain boundaries, eventually lowering the intensity of the drifting source-drain current, as seen in preliminary evidences [19,26]. This effect involves the

238

L. Torsi et al. / Organic Electronics 10 (2009) 233–239

a

Vg= -20 V

8

ΔRch @L=0.2 mm

1-butanol

ΔRch @L=0.6 mm

Vg= -40 V

ΔRch @L=1 mm

Vg= -60 V Vg= -80 V Vg=-100V

Δ R(MΩ)

RxW (M cm)

6

ΔRc

0

10

4

2

-1

10

0 0.0

0.2

0.4

0.6

0.8

-20

1.0

-40

-100

-80

Fig. 7. Differential changes of Rc (hollow red squares) and Rch (black solid symbols) parameters upon exposure to 1-butanol at different gate voltages and for different channel lengths (triangles: L = 0.2 mm, circles: L = 0.4 mm, stars: L = 1 mm). Hollow red squares are positive values while solid black symbols are negative ones. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

b 10

Rc (MΩ)

-60

Vg(V)

Channel Length - (mm)

1 1E-9 Rc 1-Butanol Rc N2

1E-10

-20

-40

-60

-80

-100

Vg(V)

c

1-Butanol

10

Rch (L= Rch

1E-11

0.2 mm)

(L= 0.6 mm)

Rch (L=

m(A/ppm)

0.1

1E-12

1.0 mm)

Rch (MΩ)

1E-13 40

20

0

-20

-40 Vg(V)

-60

-80

-100

1 Fig. 8. Slopes (m) of the calibration lines (DI vs. [c]) measured at different Vg biases. Hollow red points stand for positive values, while solid black are negative ones. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

N2 Rch (L=

0.2 mm)

Rch (L=

0.6 mm)

Rch (L=

1.0 mm)

0.1 -20

-40

-60 Vg(V)

-80

-100

Fig. 6. (a) The total device resistances (R) extracted from the Ids–Vds curves (0 < Vds < |10 V| and Vg = 20 V, 40 V, 60 V, 80 V, 100 V) of DHa6T OTFTs of different channel lengths exposed to 1-butanol, W is the transistor channel width. (b) Comparison of the contact resistances in 1-butanol and in N2. (c) Comparison of the channel resistances in 1-butanol and in N2 for devices of different channel lengths.

whole bulk of the DHa6T film down to the interface with the gate dielectric where the two-dimensional transport occurs [27]. The two-dimensional nature of the sensing

mechanism was demonstrated too, as DI responses were seen to be active layer thickness independent [9]. The field-induced sensitivity enhancement, can be seen as correlated to the comparably larger number of charges drifting in the device’s channel in the on-state accumulation regime with the respect to the charges, drifting in the bulk, in the device off state. The data reported can also contribute to shed light on the origin of contact effects in OTFTs. A simple thermionic emission is generally postulated although this hypothesis is not supported by evidences such as the contact activation energy being much smaller than the estimated potential barriers determined by pho-

L. Torsi et al. / Organic Electronics 10 (2009) 233–239

toemission spectroscopy. Also the channel potential measurements by Kelvin probe force microscopy and the four-probe methods, indicate that the contact resistances and temperature dependences associated with the individual source and drain electrodes are nearly identical [21]. Alternatively, the transport through the contacts can be limited by a disordered depletion region near the contacts, this being even more plausible in top contact devices. It is received that the charged carrier transport in OTFTs proceeds through thermionic emission through grains [28]. The data in Fig. 8 shows that, upon exposure to an alcohol, the contact resistance and the channel one change in opposite fashion, implying that they can be generated by two different mechanisms. This can be of further support to the hypothesis that the carrier diffusion through a disordered depletion region is the dominant contributing mechanism to OTFT contact resistance. A doping near the contacts could also be favored by the amorphous nature of the organic semiconductor in this region. 4. Conclusions A systematic study is reported on the role of contact resistance effects on OTFT sensors, in particular as it concerns the gate field-induced sensitivity enhancement. The on-state sensitivity enhancement is largely ascribable to changes occurring in the transistor channel transport and effects, such as contacts resistance or leakage current variations do not dominate the on-state sensor behavior. Their effect is seen in the low gate voltage regime where sensitivities are much lower. Besides, the evidences reported bring further support to the hypothesis of the OTFT contact resistances being mainly due to carried diffusion though a disordered depletion region. Acknowledgments Prof. P. Mastrorilli, Prof. G.P. Suranna and Dr. G. Romanazzi are acknowledged for the synthesis of the high quality DHa6T oligomers and Prof. F. Palmisano is acknowledged for useful discussions. This work was partially supported by the ‘‘PRIN-06 Project – 2006037708 – Plastic bio-FET sensors”.

239

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.orgel.2008. 11.009. References [1] K.C. Persaud, Mater. Today 8 (2005) 38. [2] L. Torsi, A. Dodabalapur, Anal. Chem. 77 (2005) 380A. [3] B. Crone, A. Dodabalapur, A. Gelperin, L. Torsi, H.E. Katz, A.J. Lovinger, Z. Bao, Appl. Phys. Lett. 78 (2001) 2229. [4] D.A. Bernards, D.J. Macaya, M. Nikolou, J.A. DeFranco, S. Takamatsu, G.G. Malliaras, J. Mater. Chem. 18 (2008) 116. [5] L. Torsi, Anal. Bioanal. Chem. 384 (2006) 309 (and papers therein). [6] K.C. See, A.Becknell, J. Miragliotta, H.E.Katz,Adv. Mater. 19 (2008)3322. [7] L. Torsi, M.C. Tanese, N. Cioffi, M.C. Gallazzi, L. Sabbatini, P.G. Zambonin, G. Raos, S.V. Meille, M.M. Giangregorio, J. Phys. Chem. B 107 (2003) 7589. [8] J. Huang, J. Miragliotta, A. Becknell, H.E. Katz, J. Am. Chem. Soc. 129 (2007) 9366. [9] L. Torsi, G.M. Farinola, F. Marinelli, M.C. Tanese, O. Hassan Omar, L. Valli, F. Babudri, F. Palmisano, P.G. Zambonin, F. Naso, Nature Mater. 7 (2008) 412. [10] E.J. Severin, R.D. Scanner, B.J. Doleman, N.S. Lewis, Anal. Chem. 70 (1998) 1440. [11] B.P.J. de Lacy Costello, N.M. Ratcliffe, P.S. Sivanand, Synth. Met. 139 (2003) 43. [12] F. Garnier, A. Yassar, R. Hajlaoui, G. Horowitz, F. Deloffre, B. Servet, S. Ries, P.J. Alnott, Am. Chem. Soc. 115 (1993) 8716. [13] H.E. Katz, L. Torsi, A. Dodabalapur, Chem. Mater. 7 (1995) 2235. [14] H.E. Katz, A. Dodabalapur, L. Torsi, D. Elder, Chem. Mater. 7 (1995) 2238. [15] H.E. Katz, J.G. Laquindanum, A. Lovinger, J. Chem. Mater. 10 (1998) 633. [16] C.D. Dimitrakopoulos, B.K. Furman, T. Graham, S. Hegde, S. Purushothaman, Synth. Met. 92 (1998) 47. [17] M. Halik, H. Klauk, U. Zschieschang, G. Schmid, S. Ponomarenko, S. Kirchmeyer, W. Weber, Adv. Mater. 15 (2003) 917. [18] A.J. Lovinger, H.E. Katz, A. Dodabalapur, Chem. Mater. 10 (1998) 3275. [19] A. Dodabalapur, SPIE Optics and Photonics 2007 – 26–30 August, San Diego, CA. [20] G. Horowitz, M.E. Hajlaoui, R. Hajlaoui, J. Appl. Phys. 87 (2000) 4456. [21] M.J. Panzer, C.D. Frisbie, in: Z. Bao, J. Locklin (Eds.), Organic FieldEffect Transistors, CRC Press Taylor and Francis Group 2007 (Chapter. 2.4). [22] J. Zaumseil, K.W. Baldwin, J.A. Rogers, J. Appl. Phys. 93 (2003) 6117. [23] P.V. Pesavento, K.P. Puntambekar, C.D. Frisbie, J.C. McKeen, P.P. Ruden, J. Appl. Phys. 99 (2006) 094504. [24] E.J. Meijer, G.H. Gelinck, E. van Veenendaal, B.-H. Huisman, D.M. de Leeuw, T.M. Klapwijk, Appl. Phys. Lett. 82 (2003) 4576. [25] B.H. Hamadami, D. Natelson, J. Appl. Phys. 97 (2005) 064508. [26] T. Someya, H.E. Katz, A. Gelperin, A.J. Lovinger, A. Dodabalapur, Appl. Phys. Lett. 81 (2002) 3079. [27] A. Dodabalapur, L. Torsi, H.E. Katz, Science 268 (1995) 270. [28] G. Horowitz, M.E. Haifaoui, Adv. Mater. 12 (2000) 1046.