Professional Documents
Culture Documents
2005 Mechanical Force Sensors Using Organic Thin - LM Transistors
2005 Mechanical Force Sensors Using Organic Thin - LM Transistors
Grzegorz Darlinski, Ulrich Böttger, Rainer Waser, Hagen Klauk, Marcus Halik, Ute Zschieschang, Günter Schmid
, and Christine Dehm
Low-temperature processable inherently photosensitive polyimide as a gate insulator for organic thin-film
transistors
Appl. Phys. Lett. 86, 133508 (2005); 10.1063/1.1894587
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
195.194.14.2 On: Thu, 27 Mar 2014 15:06:22
JOURNAL OF APPLIED PHYSICS 97, 093708 共2005兲
impacts only the pentacene channel, but not the contacts. Furthermore, the absolute and relative pressure-induced
These measurement did not show any change in drain current changes in the drain current shown in Fig. 4共b兲 indicate that
关see Fig. 1共c兲兴. the observed effect is not a leakage current through the tung-
The transfer characteristics shown in Figs. 1共c兲 and 2共c兲 sten needle, which would be expected to result in a constant
were obtained with a pressure of about 106 Pa, calculated by current, i.e., a current independent of the gate-source voltage,
dividing the force 共measured with the balance兲 by the esti- which is obviously not the case. The data in Fig. 4共b兲 also
mated tip area. Raising the pressure above 106 Pa did not show that the observed increase in drain current is not merely
lead to any further changes in current. due to a pressure-induced reduction in gate dielectric thick-
One feasible mechanism for the observed effect could be ness 共which would cause an increase in the gate field E for
that the increase in drain current is caused by leakage cur- the same gate-source voltage VGS, since E = VGS / t, where t is
rents, either due to a short circuit through the tungsten needle the dielectric thickness兲. If this was the case we would ex-
or by crushing the dielectric 共creating shorts to the gate elec- pect the relative change in drain current to be independent of
trode兲. To analyze this possibility we have measured all three the gate-source voltage, which is obviously not the case here.
TFT currents 共drain current ID, source current IS, and gate Therefore, we can rule out that the observed effect is due to
current IG兲 for both pressure states. The results 关see Fig. 4共a兲兴 elastic material effects or erroneous measurement setup.
indicate that, although leakage paths between the gate elec- While we see two stable states for slow measurements
trode and the source/drain contacts exist, the pressure- 共i.e., measurements where the gate-source voltage is swept
induced changes are entirely in the drain and source currents slowly between positive and negative values兲, fast measure-
while the gate current is not affected by the mechanical pres- ments with total sweep times of about 200 ms show a
sure. We therefore conclude that the gate dielectric is not smooth transition from the “low-pressure” state to the “high-
damaged by the needle and that the observed effect is not pressure” state 关see Figs. 5共a兲 and 5共b兲兴. To examine the time
due to an increase in gate leakage. dependence of the pressure effect we measure the drain cur-
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
195.194.14.2 On: Thu, 27 Mar 2014 15:06:22
093708-3 Darlinski et al. J. Appl. Phys. 97, 093708 共2005兲
rent at fixed gate and drain voltages selecting the voltages for
maximum pressure sensitivity 共VGS = + 10 V , VDS = −20 V兲,
while moving the needle up and down 关see Fig. 5共c兲兴. There
was no measurable delay in response, but it takes about 20 s
for the drain current to reach steady states. The rise and fall
times 共i.e., the time it takes for the current to change from
0% to 70% or from 100% to 30%兲 are in the range of a few
hundred milliseconds. The response appears to be much
shorter when the current is not measured at fixed gate-source
voltages but is swept between +10 and −20 V. Extending the
waiting time at positive VGS accelerates the transition. From
this we conclude that the transition between the two steady
states is mediated by trapping and detrapping of charges.
Trapped charges will influence a thin-film transistor in
many ways, depending on the exact location of the traps in
the device. From simply looking at the transfer characteris-
tics it remains unclear which physical parameters of the TFT
are influenced by the mechanical pressure. To further exam-
ine this we have used a SPICE program 共simulation program
with integrated circuit emphasis兲 to simulate the device.
As discussed elsewhere12,13 the SPICE metal-oxide-
semiconductor field-effect transistor 共MOSFET兲 model is in-
sufficient to simulate organic thin-film transistors. Therefore,
we have modified the standard level-2 p-MOSFET model by
introducing branches which form paths for leakage currents,
as shown in Fig. 6. By adjusting the SPICE model parameters
and improving the result by fitting the leakage currents we
found a system which describes our device reasonably well.
By changing MOSFET parameters we are able to simulate
the transition from “pressure” state to “no-pressure” state
with good accuracy. Figure 4共c兲 shows the currents calcu- FIG. 4. 共a兲 Measured transfer characteristics of a TFT with interdigitated
lated by SPICE. The agreement between experimental and source/drain contacts and a channel length of 10 m, showing drain current,
simulation data is very good for positive gate-source volt- source current, and gate current; 共b兲 measured relative 共䊊兲 and absolute 共䊐兲
changes in drain current between pressure and no-pressure states; 共c兲 simu-
ages. For negative gate-source voltages the simulation is lated transfer characteristic of the same device; and 共d兲 simulated relative
dominated by 共gray兲 and absolute 共black兲 changes.
ID = − K P 冉 冊冉
W
Leff
VGS − VTH −
VDS
2
冊
VDS共1 − VDS兲, 共1兲
− VTH兲 and VGS ⬍ VTH. The theoretical ID vs VGS relationship
which describes the behavior of inorganic transistors in the is nearly linear in this range of gate-source voltages, but for
linear regime, but which is inaccurate for organic devices.14 organic semiconductors it is rather exponential. This inaccu-
In Eq. 共1兲 ID denotes the drain current for 0 ⬎ VDS ⬎ 共VGS racy of the model is insignificant for the discussion of the
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
195.194.14.2 On: Thu, 27 Mar 2014 15:06:22
093708-4 Darlinski et al. J. Appl. Phys. 97, 093708 共2005兲
IV. CONCLUSIONS
In summary, we have shown that organic thin-film tran-
sistors are influenced by direct mechanical pressure. By
simulating the results we have shown that mechanical pres-
sure applied directly to the TFT 共including the contact re-
gions兲 results in changes in carrier mobility, threshold volt-
age, and contact resistances. We have shown a dependence
on materials and interfaces as the effect was only measured
when contact regions were affected. The origin of the ob-
served effect is not completely clear yet; we believe that
trapped charges are responsible. However, the trapping
mechanism and the nature of the traps are still under
discussion.
1
C. D. Sheraw et al., Appl. Phys. Lett. 80, 1088 共2002兲.
FIG. 5. 共a兲 “Slow” measurement showing drain current in two steady states 2
B. K. Crone, A. Dodabalapur, R. Sarpeshkar, A. Gelperin, H. E. Katz, and
at VDS = −20 V, 共b兲 “fast” measurement showing shifting drain current be- Z. Bao, J. Appl. Phys. 91, 10140 共2002兲.
tween the two steady states at VDS = −20 V, and 共c兲 time dependence of drain 3
T.-T. Zhu, J. T. Mason, R. Diekmann, and G. G. Malliaras, Appl. Phys.
current at VGS = 10 V and VDS = −20 V when applying and removing Lett. 81, 4643 共2002兲.
pressure. 4
J. A. Rogers et al., Proc. Natl. Acad. Sci. U.S.A. 98, 9 共2001兲.
5
T. W. Kelley, L. D. Boardman, T. D. Dunbar, D. V. Muyres, M. J. Peller-
observed pressure effect since it can be corrected by fitting ite, and T. P. Smith, J. Phys. Chem. B 107, 24 共2003兲.
6
G. B. Blanchet, Y.-L. Loo, J. A. Rogers, F. Gao, and C. R. Fincher, Appl.
the model equation only. Phys. Lett. 82, 463 共2003兲.
The good agreement between our model and the experi- 7
U. Zschieschang, H. Klauk, M. Halik, G. Schmid, and C. Dehm, Adv.
mental results is evident from Figs. 4共b兲 and 4共d兲, which Mater. 共Weinheim, Ger.兲 15, 1147 共2003兲.
show the pressure-induced relative and absolute changes in
8
M. Halik et al., Nature 共London兲 431, 963 共2004兲.
9
T. Someya and T. Sakurai, Tech. Dig. - Int. Electron Devices Meet. 2003,
drain current, suggesting that our simulations provide a use- 203.
ful analysis of the mechanism causing the observed pressure- 10
H. Klauk, M. Halik, U. Zschieschang, F. Eder, and G. Schmid, Appl. Phys.
induced change in drain current. To achieve this agreement Lett. 82, 4175 共2003兲.
11
M. Halik, H. Klauk, U. Zschieschang, T. Kriem, G. Schmid, and W. Ra-
between simulation and experiment we modified the thresh-
dlik, Appl. Phys. Lett. 81, 289 共2002兲.
old voltage 共VTO兲, mobility 共UO兲, source resistance 共RS兲, 12
R. Brederlow et al., Dig. Tech. Pap.-IEEE Int. Solid-State Circuits Conf.
and drain resistance 共RD兲 SPICE-model parameters as a re- 2003, 378.
13
sponse to applied mechanical pressure. We therefore con- P. V. Necliudov, M. S. Shur, D. J. Gundlach, and T. N. Jackson, J. Appl.
Phys. 88, 6594 共2000兲.
clude that the force-induced change in drain and source cur- 14
Explanation can be found in every SPICE documentation on MOSFETs.
rents is due to the distribution and activity of trap states at or 15
J. A. Nichols, D. J. Gundlach, and T. N. Jackson, Appl. Phys. Lett. 83,
near the semiconductor/dielectric interface. This assumption 2366 共2003兲.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
195.194.14.2 On: Thu, 27 Mar 2014 15:06:22