Enhancement of carrier mobility in pentacene thin-film transistor on SiO2 by controlling the initial film growth modes

Enhancement of carrier mobility in pentacene thin-film transistor on SiO2 by controlling the initial film growth modes

Applied Surface Science 255 (2009) 5096–5099 Contents lists available at ScienceDirect Applied Surface Science journal homepage: www.elsevier.com/lo...

480KB Sizes 0 Downloads 17 Views

Applied Surface Science 255 (2009) 5096–5099

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Short communication

Enhancement of carrier mobility in pentacene thin-film transistor on SiO2 by controlling the initial film growth modes Qiong Qi, Aifang Yu, Peng Jiang, Chao Jiang * National Center for Nanoscience and Technology, No.11, Beiyitiao Zhongguancun, Beijing 100190, China

A R T I C L E I N F O

A B S T R A C T

Article history: Received 4 September 2008 Received in revised form 18 November 2008 Accepted 8 December 2008 Available online 13 December 2008

Pentacene thin-film transistors (TFTs) were fabricated on thermally grown SiO2 gate insulator under the conditions of various pre-cleaning treatments. Initial nucleation and growth of the material films on treated substrates were observed by atomic force microscope. The performance of fabricated TFT devices with different surface cleaning approaches was found to be highly related to the initial film morphologies. In contrast to the three-dimensional island-like growth mode on SiO2 under an organic cleaning process, a layer-by-layer initial growth occurred on the SiO2 insulator cleaned with ammonia solution, which was believed to be the origination of the excellent electrical properties of the TFT device. Field effect mobility of the TFT device could achieve as high as 1.0 cm2/Vs on the bared SiO2/Si substrate and the on/off ratio was over 106. ß 2009 Elsevier B.V. All rights reserved.

PACS: 85.30.Tv 72.20.Jv 72.80.Le Keywords: Organic thin-film transistor Pentacene Surface treatment Initial growth mode

1. Introduction To realize the reliable device performance in the realistic applications, organic thin-film transistor (OTFT) needs high carrier mobility and large saturation current to be applied to the fast switch and feasible driving circuits. For this purpose, lots of approaches have been utilized to improve pentacene (C22H14)-based transistor’s mobility, such as improving the surface roughness of the dielectric materials [1], modification the dielectric surface by self-assembled monolayer [2–7] and by optimizing the film deposition conditions [2,8,9]. Recently, several research groups report that single monolayer of pentacene with less boundary defects is virtually essential to achieve high mobility, because grain boundaries strongly affect the charge transport in pentacene films [10–12]. We also noted that, up to date, less attention was paid on the correlation between the pentacene initial growth modes and precleaning treatments for the thermally oxidized SiO2 which has been widely utilized as a dielectric layer in OTFT fabrication. In this letter, we focus on the establishment of correlation between the surface cleaning treatment and the initial pentacene film growth modes. We have found a layer-by-layer growth mode of pentacene was

* Corresponding author. Tel.: +86 10 82545563; fax: +86 10 62656756. E-mail address: [email protected] (C. Jiang). 0169-4332/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2008.12.022

established on SiO2 under a suitable surface pre-cleaning process utilizing ammonia solution. The mobility in saturation region can be high to reach 1 cm2/Vs and the saturated current to be 300 mA under the VDS = 40 V, indicative of a promising novel pentacene thin-film transistor (TFT) fabrication approach for realistic application. 2. Experimental The pentacene TFTs were fabricated on heavily doped n-type (0 0 1) silicon wafer having a 250 nm SiO2 layer grown by a wet thermal oxidization process. The substrates were ultrasonically cleaned by three various approaches. Method I: cleaning with the ammonia solution NH4OH:H2O = 1:6, then deionized water, each for 10 min. The ammonia solution was used to etch and clean SiO2 dielectric. Method II: cleaning with the NH4OH:H2O2:H2O = 1:1:5 (APM) and HCl:H2O2:H2O = 1:1:6 (HPM), and then deionized water for 10 min, respectively. This is a traditional cleaning method used in silicon industry. The additive H2O2 is thought to oxidize and remove organic contaminators from the substrate, and also can protect the substrate Si from being etched by ammonia. Method III: cleaning with only organic solvents using acetone, ethanol and deionized water each for 10 min, respectively. All cleaning methods have not further introduced substrate roughness which is 0.3 nm via an AFM investigation.

Q. Qi et al. / Applied Surface Science 255 (2009) 5096–5099

5097

Fig. 1. 3 mm  3 mm AFM images of 0.5 ML pentacene film grown on SiO2 insulator surface cleaned by (a) ammonia, (b) APM–HPM and (c) acetone and ethanol.

50 nm pentacene (Aldrich Co.) films calibrated by a quartz oscillator analyzer were deposited on the SiO2 substrates by a commercial thermal evaporator Auto-306 (BOC-Edwards Co.) deposition system. During deposition, the back pressure reached 7  10 5 Pa with deposition rate 0.02 nm/s under the room temperature. Finally, 50 nm thick Au source-drain electrodes were deposited through a shadow mask onto the pentacene film at a rate of 0.4 nm/s to finish the top-contact TFTs fabrication. The conductance channel length and width were 50 and 2000 mm, respectively. The morphology of the pentacene film was characterized by a Nanoscope III (Veeco Co.) AFM using a tapping mode. The contact angle was measured using distilled water and the transfer curve of TFT was characterized using a Kethley-4200 semiconductor analyzer.

contains less defects and nucleation sites, and hence pentacene can possess longer migration length before the adsorbed molecules join into the boundaries and form larger islands on SiO2 insulator, as indicated in Fig. 1a. The rocking curves by X-ray diffraction for the three 50 nm thick pentacene films show a pure thin-film phase characterized (Fig. 2). The intensity of peak is significantly stronger on gate dielectrics cleaned with the ammonia and the APM–HPM approaches than that cleaned with the organic solvent, indicative

3. Results and discussion Fig. 1 shows the AFM images of pentacene film grown on various treated SiO2 surfaces with a nominal thickness of 0.5 monolayer (ML). On the surface treated with ammonia (Fig. 1a), pentacene molecules homogeneously nucleate and form flat islands with round-disk shapes and smooth borders. These islands connect with each other to construct a sub-monolayer with a typical height of approximately 1.5 nm, being very similar to what was ever observed under hyper thermal molecular beam deposition [9]. On the contrary, the pentacene islands are located disorderly on the substrate with different sizes, heights, and shapes on the surface treated with the acetone and the ethanol, respectively (Fig. 1c). The contact angle of the surface treated with ammonia is much smaller than that of the surface treated with acetone and ethanol (Table 1). It indicates that most of the organic contaminators on the surface of SiO2 have been washed away by the ammonia, and a more hydrophilic surface was formed. We infer that this surface had higher surface energy, which has been demonstrated to be essential for forming large pentacene islands in the initial stages of growth [13]. When SiO2 surface are treated with the HPM and APM (Fig. 1b), the morphology shows an intermediate situation. We inferred that the addition of H2O2 may reduce the cleaning efficiency of ammonia solution in two aspects. First, addition of H2O2 is thought to lessen the dissolving rate of SiO2. Second, H2O2 can oxidize the contaminator Si which induced during the incision process of SiO2/Si substrate into contaminator SiO2, which is more difficult to be removed. A cleaner oxide surface

Fig. 2. XRD patterns of 50 nm thick pentacene deposited on SiO2 insulator cleaned by (a) ammonia, (b) APM–HPM and (c) acetone and ethanol.

Table 1 Summarized properties resulting from three different surface cleaning treatments. m is the charge carrier mobility, VT is the threshold voltage and Q is the water contact angle of different cleaned surfaces. Surface treatment Ammonia APM–HPM Acetone–ethanol

m (cm2/Vs) 1.0 0.72 0.21

VT (V) 1.1 1.0 0.90

Q (degree)

Ion/Ioff 6

1  10 5  104 3  103

51 50 62

Fig. 3. Electrical-transfer characteristics of TFTs fabricated on SiO2 insulator cleaned by (a) ammonia, (b) APM–HPM and (c) acetone and ethanol.

5098

Q. Qi et al. / Applied Surface Science 255 (2009) 5096–5099

Fig. 4. (a–c) 3 mm  3 mm AFM images of pentacene film with increasing thickness of 1, 2 and 4 ML thick pentacene deposited on SiO2 surface cleaned by ammonia. The bottom inserts reveal the line scan derived from the AFM measurement along the white lines.

of much better crystal quality and more arrangement ordering of pentacene molecules in the former than the later. The electrical transfer characteristics under the drain voltage of 40 V for all three TFTs with three different cleaning approaches are shown in Fig. 3. Saturation behavior of drain current at high drain-source voltage (VDS) was observed for all devices and its values are markedly dependent on the surface treatment procedures. The maximum saturation currents above 300 mA was obtained for the OTFTs based on the ammonia treatment when the drain voltage was swept under a constant gate-source voltage 40 V. The carrier mobility of the TFTs was calculated in saturation regime from Fig. 3 by fitting the data to the equation used in literature [14]. We also summarize the results for threshold voltage VT, on/off ratio and the calculated mobility for three cases in Table 1. The mobility of 1 cm2/Vs is one of the best results obtained on the bared SiO2 insulator under feasible growth conditions and by simple evaporation equipment. The electrical-characteristics of pentacene TFTs as well as the ordering of pentacene molecules (not shown) are significantly improved by utilizing the ammonia-cleaning method. This may infer to some correlation between the electric performance and the growth mode of the deposited pentacene. Especially the several initial deposited monolayers were ever proved to be virtually important because of the observation for the mobility saturation with the deposition layers up to few monolayers [12]. In order to clarify the high mobility mechanism in TFTs grown on the surface cleaned by the ammonia solution, we carried out the film structure analysis by consecutively introducing 1, 2 and 4 ML pentacene thicknesses as shown in Fig. 4a–c, respectively. AFM images prove a layer-plus-island growth mode, e.g. Stranski– Krastanov mode (S–K mode) occurred during the whole film deposition process. From the line profile analysis as shown in bottom inset of Fig. 4a–c, we can find that the steps have a typical height of 1.5 nm, approximately corresponding to one monolayer height. Comparing Fig. 1a with Fig. 4a, we find the nucleation density is almost identical. This means that there is no nucleation of new islands and the islands grow large up to coalesce each other and the domain gaps decrease with the increasing of pentacene molecule thickness from 0.5 to 1 ML, indicating that the growth during this period is under a diffusion-mediated growth mode [15]. Therefore, it is reasonable to conceive from a favorable energy consideration that the incoming pentacene molecule aggregates into the existed islands only, even though it needs to migrate across half of the domain radius (around 250 nm) in average and finally arrive at the domain boundary to find a suitable position to join into the crystal.

When pentacene is further deposited up to a nominal thickness of 2 ML, as indicated in Fig. 4b, the first pentacene layer is almost completed although there are still some voids, while the second and third, even fourth, pentacene monolayer begin to nucleate and grow up (see height profile). From the first layer region of Fig. 4b, we can find that the first layer consists of several large islands connected together where the grain boundary voids disappear and the grains become a unity. This indicates that most large islands have coalesced completely in the first monolayer. When the thickness of deposited pentacene increases to 4 ML as shown in Fig. 4c, each grain turns out a layer-like structure and the dendritic morphology appears. However, the above deposited layers other than the first monolayer cannot merge because of this twodimensional plus three-dimensional growth mode. We believe the high carrier conductance observed in Fig. 3 is probably attributed to the perfect healed grain boundaries of the first monolayer, as indicated in Fig. 4b by an arrow. 4. Conclusions In summary, we have investigated the impacts of cleaning method on structure, morphology, as well as TFT device performance of pentacene film. Pentacene grown on SiO2 with surface cleaned by acetone–ethanol was dominated by threedimensional growth mode. Whereas, when the surface was cleaned with the ammonia solution, the S–K growth mode occurred. The electrical characteristics of pentacene-based TFT devices, especially mobility, grain size and the ordering of pentacene molecules are significantly improved by cleaning SiO2/Si substrate with ammonia. With this surface cleaning treatment, the mobility of pentacene OTFT can reach 1.0 cm2/Vs and the on/off ratio increases to 106. Therefore, pentacene-based OFET performance is determined by the quality of first few pentacene monolayers, especially the first monolayer. Acknowledgement This research was supported by the ‘‘Hundred Talent Program’’, China. References [1] S.E. Fritz, T.W. Kelley, C.D. Frisbie, J. Phys. Chem. B 109 (2005) 10574. [2] M. Shtein, J. Mapel, J.B. Benziger, S.R. Forrest, Appl. Phys. Lett. 81 (2002) 268. [3] Y. Jang, J.H. Cho, D.H. Kim, Y.D. Park, M. Hwang, K. Cho, Appl. Phys. Lett. 90 (2007) 132104. [4] D. Kumaki, M. Yahiro, Y. Inoue, S. Tokito, Appl. Phys. Lett. 90 (2007), 1p. 33511.

Q. Qi et al. / Applied Surface Science 255 (2009) 5096–5099 [5] M. McDowell, I.G. Hill, J.E. McDermott, S.L. Bernasek, J. Schwartz, Appl. Phys. Lett. 88 (2006) 073505. [6] D.H. Kim, H.S. Lee, H. Yang, L. Yang, K. Cho, Adv. Funct. Mater. 18 (2008) 1363. [7] H. Ma, O. Acton, G. Ting, J.W. Ka, H. Yip, N. Tucker, R. Schofield, A.K.-Y. Jen, Appl. Phys. Lett. 92 (2008) 113303. [8] S. Pratontep, M. Brinkmann, F. Nu¨esch, L. Zuppiroli, Synth. Met. 146 (2004) 387. [9] T. Toccoli, A. Pallaoro, N. Coppede`, S. Iannotta, F. De Angelis, L. Mariucci, G. Fortunato, Appl. Phys. Lett. 88 (2006) 132106.

[10] [11] [12] [13]

5099

S. Verlaak, C. Rolin, P. Heremans, J. Phys. Chem. B 111 (2007) 139. S. Verlaak, P. Heremans, Phys. Rev. B 75 (2007) 115127. R. Ruiz, A. Papadimitratos, C. Alex, G.G. Malliaras, Adv. Mater. 17 (2005) 1795. S.J. Jo, C.S. Kim, J.B. Kim, J. Kim, M.J. Lee, H.S. Hwang, H.K. Baik, Y.S. Kim, Appl. Phys. Lett. 93 (2008) 083504. [14] H.E. Katz, Chem. Mater. 16 (2004) 4748. [15] R. Ruiz, D. Choudhary, B. Nickel, T. Toccoli, K.C. Chang, A.C. Mayer, P. Clancy, J.M. Blakely, R.L. Headrick, S. Iannotta, G.G. Malliaras, Chem. Mater. 16 (2004) 4497.