journal article·2013-06-20

Impact of Materials versus Geometric Parameters on the Contact Resistance in Organic Thin‐Film Transistors

Manfred Gruber, Egbert Zojer, Ferdinand Schürrer, Karin Zojer

Advanced Functional Materials· Volume 23· Issue 23
Publisher page

Abstract

Abstract The contact resistance is known to severely hamper the performance of organic thin‐film transistors, especially when dealing with large injection barriers, high mobility organic semiconductors, or short channel lengths. Here, the relative significance of how it is affected by materials‐parameters (mobility and interfacial level‐offsets) and geometric factors (bottom‐contact vs top‐contact geometries) is assessed. This is done using drift‐diffusion‐based simulations on idealized device structures aiming at a characterization of the “intrinsic” situation in the absence of traps, differences in the film morphology, or metal‐atoms diffusing into the organic semiconductor. It is found that, in contrast to common wisdom, in such a situation the top‐contact devices do not always outperform the bottom‐contact ones. In fact, the observed ratio between the contact resistances of the two device structures changes by up to two orders of magnitude depending on the assumed materials parameters. The contact resistance is also shown to be strongly dependent on the hole mobility in the organic semiconductor and influenced by the chosen point of operation of the device.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Organic Electronics and Photovoltaics · Advancements in Semiconductor Devices and Circuit Design · Thin-Film Transistor Technologies

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

References (62)

Show source-provided references
1. 10.1063/1.223472410.1063/1.2234724 ↗
2. 10.1063/1.276397310.1063/1.2763973 ↗
3. 10.1002/adfm.20070029410.1002/adfm.200700294 ↗
4. 10.1063/1.280428810.1063/1.2804288 ↗
5. 10.1002/adfm.20080146610.1002/adfm.200801466 ↗
6. 10.1038/nmat205910.1038/nmat2059 ↗
7. 10.1002/smll.20100225410.1002/smll.201002254 ↗
8. 10.1063/1.219703310.1063/1.2197033 ↗
9. 10.1109/JPROC.2005.85030110.1109/JPROC.2005.850301 ↗
10. 10.1021/ja809848y10.1021/ja809848y ↗
11. 10.1016/S1369-7021(07)70016-010.1016/S1369-7021(07)70016-0 ↗
12. 10.1002/adfm.20100203510.1002/adfm.201002035 ↗
13. 10.1063/1.278969910.1063/1.2789699 ↗
14. 10.1063/1.275998710.1063/1.2759987 ↗
15. 10.1063/1.273838210.1063/1.2738382 ↗
16. 10.1063/1.315983510.1063/1.3159835 ↗
17. 10.1063/1.311582610.1063/1.3115826 ↗
18. 10.1126/science.121882910.1126/science.1218829 ↗
19. 10.1063/1.141537410.1063/1.1415374 ↗
20. 10.1063/1.281364010.1063/1.2813640 ↗
21. 10.1016/S0038-1101(02)00204-610.1016/S0038-1101(02)00204-6 ↗
22. 10.1063/1.211218910.1063/1.2112189 ↗
23. 10.1063/1.221513210.1063/1.2215132 ↗
24. 10.1063/1.151295010.1063/1.1512950 ↗
25. 10.1063/1.347953110.1063/1.3479531 ↗
26. 10.1063/1.305864010.1063/1.3058640 ↗
27. 10.1063/1.180653310.1063/1.1806533 ↗
28. 10.1002/adma.20080273310.1002/adma.200802733 ↗
29. 10.1063/1.294240010.1063/1.2942400 ↗
30. 10.1016/j.orgel.2009.03.01210.1016/j.orgel.2009.03.012 ↗
31. 10.1016/j.orgel.2009.05.01910.1016/j.orgel.2009.05.019 ↗
32. 10.1016/j.orgel.2012.04.02910.1016/j.orgel.2012.04.029 ↗
33. 10.1002/pssa.20112759510.1002/pssa.201127595 ↗
34. 10.1016/j.orgel.2012.05.00910.1016/j.orgel.2012.05.009 ↗
35. 10.1063/1.152798310.1063/1.1527983 ↗
36. 10.1063/1.156815710.1063/1.1568157 ↗
37. 10.1016/S0038-1101(02)00210-110.1016/S0038-1101(02)00210-1 ↗
38. 10.1557/JMR.2004.026610.1557/JMR.2004.0266 ↗
39. 10.1063/1.295973210.1063/1.2959732 ↗
40. Horowitz G. · 2007
41. 10.1016/0038-1101(70)90060-210.1016/0038-1101(70)90060-2 ↗
42. 10.1016/0038-1101(76)90019-810.1016/0038-1101(76)90019-8 ↗
43. 10.1103/PhysRevB.59.751410.1103/PhysRevB.59.7514 ↗
44. 10.1063/1.36652210.1063/1.366522 ↗
45. This is at least a very good approximation as in the channel region the current flows within a few nanometers of the semiconductor/dielectric interface.
46. Note that this is not to be confused with the pinch‐off generated voltage drop at the drain that occurs when operating devices in the saturation regime. This voltage drop is observed experimentally but has nothing to do with the contact.
47. 10.1063/1.161336910.1063/1.1613369 ↗
48. Note that the specific choice of the value of the deviation between the true potential and the fit introduces some ambiguity regarding the determination of the exact position of (2); moreover the choice of the grid also has a minor impact. This prompted us not to take the potential at (2) as the reference for the determination ofVCbut rather use the value obtained in the extrapolation.
49. 10.1016/j.orgel.2005.03.00310.1016/j.orgel.2005.03.003 ↗
50. 10.1063/1.148512910.1063/1.1485129 ↗
51. 10.1063/1.242439610.1063/1.2424396 ↗
52. 10.1063/1.153210210.1063/1.1532102 ↗
53. 10.1063/1.185887410.1063/1.1858874 ↗
54. 10.1063/1.147921010.1063/1.1479210 ↗
55. Note that we could not recover theRCvalues given in ref. [20] even though we used the same device dimensions and material parameters. According to our analysis this deviation originates at least in part from a very profound difference between theRCvalues determined with our approach those deduced from the transfer‐line method employed in ref. [20].
56. 10.1002/cphc.20030094210.1002/cphc.200300942 ↗
57. In fact at very small barriers a few percent of the total current come from the top facet of the bottom contacts which is a consequence of carrier diffusion that becomes relevant due to the enhanced carrier concentration above the contact.
58. 10.1063/1.295973210.1063/1.2959732 ↗
59. Note that this effect is not to be confused with the onset‐voltage shift caused by a space‐charge or dipole layer between the semiconductor and the dielectric (see ref. [5]). Here the resulting decrease ofISDat smallVGSis much more gradual.
60. Note that due to the large aspect ratio of OTFTs with channel lengths of typically several micrometers and dielectric thicknesses on the order of 100 nm the barrier shaping field also for BC devices primarily originates fromVGSand not fromVDS.
61. This is true also when the contribution ofRCtoRtotis no longer negligible as even thenRCis directly proportional toRch.
62. Indeed we find a clearly linear relation between the channel resistance and 1/μin all investigated cases which in turn results inRCbeing directly proportional to 1/μ.