www.advmat.de

2104075 (1 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

Review

A Critical Outlook for the Pursuit of Lower Contact 
Resistance in Organic Transistors

James W. Borchert,* R. Thomas Weitz, Sabine Ludwigs, and Hagen Klauk*

J. W. Borchert, R. T. Weitz
1st Institute of Physics
Georg August University of Göttingen
Friedrich-Hund-Platz 1, 37077 Göttingen, Germany
E-mail: James.Borchert@uni-goettingen.de
S. Ludwigs
IPOC - Functional Polymers
Institute of Polymer Chemistry
University of Stuttgart
Pfaffenwaldring 55, 70569 Stuttgart, Germany
H. Klauk
Max Planck Institute for Solid State Research
Heisenbergstraße 1, 70569 Stuttgart, Germany
E-mail: H.Klauk@fkf.mpg.de

The ORCID identification number(s) for the author(s) of this article 
can be found under https://doi.org/10.1002/adma.202104075.

DOI: 10.1002/adma.202104075

1. Introduction

Organic semiconductors have been imple-
mented in a variety of electronic devices, 
including organic light-emitting diodes 
(OLEDs),[1] organic solar cells,[2] organic 
photodetectors,[3] and organic transis-
tors of various forms.[4–7] An essential 
requirement for all these devices is the 
efficient injection and/or extraction of 
charges across the interfaces between the 
organic semiconductor(s) and the elec-
trical contacts.[8] Extensive research and 
development efforts into the active mate-
rials and the device engineering required 
for achieving efficient charge injection/
extraction have therefore been instru-
mental in enabling, for instance, the 
commercialization of OLEDs. Progress in 
this area has now proceeded to the extent 
that charge injection and extraction are 
not the most critical limiting issues for 
the state-of-the-art of OLEDs compared 
to other aspects, such as effective carrier 
and exciton confinement, energy transfer, 
outcoupling, and lifetime.[9–12] Much the 
same is true for organic solar cells, in that 
the primary focus and source of improve-

ments recently have been more-closely linked to the develop-
ment of nonfullerene acceptors.[13] Organic transistors of all 
types, on the other hand, which have been touted as a leading 
alternative to transistors based on inorganic semiconductors 
for novel large-area integrated-circuit applications for many 
years,[14,15] have yet to attain widespread adoption in consumer-
electronics applications.

Several of the disadvantages of organic transistors compared 
to inorganic transistors, such as generally lower charge-carrier 
mobilities, poorer device uniformity, and reduced reliability,[16] 
have seen significant improvements over time, such that a few 
commercially available devices utilizing organic thin-film tran-
sistors (TFTs) are now becoming available.[17] However, the con-
tact resistance (RC) persists as a major impediment toward fur-
ther development of circuits based on organic transistors.[18–21] 
This is especially true for the development of organic TFTs for 
low-power, high-frequency applications, such as mobile active-
matrix displays, since a high RC limits the maximum unity-
current-gain-cutoff (transit) frequency that would otherwise 
be expected through device miniaturization.[22] High contact 
resistance in organic TFTs continues to be a major problem 
despite significant strides in expanding both the breadth of 

To take full advantage of recent and anticipated improvements in the 
performance of organic semiconductors employed in organic transis-
tors, the high contact resistance arising at the interfaces between the 
organic semiconductor and the source and drain contacts must be 
reduced significantly. To date, only a small portion of the accumulated 
research on organic thin-film transistors (TFTs) has reported channel-
width-normalized contact resistances below 100 Ωcm, well above what is 
regularly demonstrated in transistors based on inorganic semiconductors. 
A closer look at these cases and the relevant literature strongly suggests 
that the most significant factor leading to the lowest contact resistances 
in organic TFTs so far has been the control of the thin-film morphology of 
the organic semiconductor. By contrast, approaches aimed at increasing 
the charge-carrier density and/or reducing the intrinsic Schottky barrier 
height have so far played a relatively minor role in achieving the lowest 
contact resistances. Herein, the possible explanations for these observa-
tions are explored, including the prevalence of Fermi-level pinning and the 
difficulties in forming optimized interfaces with organic semiconductors. 
An overview of the research on these topics is provided, and potential 
device-engineering solutions are discussed based on recent advance-
ments in the theoretical and experimental work on both organic and 
inorganic semiconductors.

© 2021 The Authors. Advanced Materials published by Wiley-VCH 
GmbH. This is an open access article under the terms of the Creative 
Commons Attribution License, which permits use, distribution and re-
production in any medium, provided the original work is properly cited.

Adv. Mater. 2022, 34, 2104075



www.advmat.dewww.advancedsciencenews.com

2104075 (2 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

knowledge of the physicochemical properties of metal–organic 
semiconductor interfaces[23,24] and the development of a diverse 
selection of material-engineering methods for tailoring the 
contact interface to enhance charge injection/extraction in 
organic semiconductors.[20,25,26] Indeed, the vast majority of the 
reported methods consider a “successful” method to be one 
that can reduce exceptionally high channel-width-normalized 
contact resistance (RCW) often exceeding 100 kΩcm down 
to 100 Ωcm at best,[20] which is still generally unacceptable 
for mobile applications that rely on lithium ion batteries.[22] 
To date, only a handful of reports showing RCW of less than 
100 Ωcm in organic field-effect transistors, including TFTs and 
electrolyte-gated organic field-effect transistors (EGOFETs),[6] 
have been published, most of which having come only in the 
last half decade (Table 1).[27–37] Put simply, there are many ways 
to reduce very high contact resistance in organic field-effect 
transistors, but it is evidently much harder to improve upon 
already “low” contact resistances of around 100 Ωcm, especially 
in TFTs.

It is important to appreciate that even the best metrics 
of 10 Ωcm in organic TFTs[34] and 1 Ωcm in EGOFETs[28] are 
orders of magnitude higher than what is achievable in nearly 
all transistors based on inorganic semiconductors. The closest 
comparison that can be made is to state-of-the-art indium 
gallium zinc oxide (IGZO) TFTs, which often show RCW of 
around 10 Ωcm[38] and can reach as low as 0.8 Ωcm.[39] The 

difference is even more pronounced for transistors based on 
other classes of semiconductors. In fact, in transistors based 
on highly crystalline inorganic semiconductors, the contact 
resistance can be reduced nearly to the quantum limit deter-
mined by the 2D charge-carrier density accumulated in the 
charge-carrier channel of the transistor.[40–42] This corresponds 
to around 0.01 Ωcm in both silicon metal–oxide–semiconductor 
field-effect transistors (MOSFETs)[43] and gallium nitride high-
electron-mobility transistors (HEMTs).[44] 2D semiconduc-
tors, of which transition-metal dichalcogenides (TMDs) are of 
particular interest for some of the same application spaces as 
organic semiconductors and IGZO, typically achieve RCW  <   
1 Ωcm these days,[40–42] including a recent demonstration of 
0.012 Ωcm in MoS2 transistors.[42]

It is thus no surprise that, due to the inexorable link between 
contact resistance and dynamic performance, even state-of-
the-art low-voltage organic transistors still struggle to achieve 
a transit frequency on the order of 10  MHz,[22,31,34,36,37,45–47] 
with only a single recent result of over 100  MHz that notably 
required relatively high operating voltages, in part to combat the 
high contact resistance.[48] In stark contrast, TMD and IGZO 
TFTs are able to already achieve maximum transit frequencies 
exceeding 1 GHz.[49,50] While the charge-carrier mobility is also 
typically higher in TMDs and IGZO, it is worth noting that a 
transit frequency of 1  GHz was demonstrated in IGZO TFTs 
with an effective carrier mobility of only 1.2 cm2 V−1 s−1 and an 

Adv. Mater. 2022, 34, 2104075

Table 1. Reports of organic field-effect transistors with channel-width-normalized contact resistance (RCW) of less than 100 Ωcm.

Refs. (Year) Semiconductor  
(morphology)

Transistor 
architecture

Concept stated to be critical for low RCW RCW  
[Ωcm]

2D charge-carrier density 
[1013 cm−2]a)

[27]
(2007)

Pentacene (polycrystalline) Coplanar UV/ozone-treated Au 80 0.5

[28]
(2010)

P3HT (polycrystalline) Staggered High charge-carrier density from electrolyte gate 
insulator

1 20

F8T2 (polycrystalline) Staggered 14 20

[29]
(2013)

DNTT (polycrystalline) Staggered Contact doping and minimized semiconductor film 
thickness

80 0.5

[30]
(2017)

C10-DNTT (polycrystalline) Coplanar Chemisorbed monolayer treatment of Au contacts 
and annealing in N2

75 0.1

[31]
(2018)

C8-DNBDT-NW (bilayer crystalline) Staggered Contact doping and low-dimensional semiconductor 
film

47 0.6

[32]
(2019)

PDPP (polycrystalline) Staggered High charge-carrier density from electrolyte gate 
insulator

2.7 3

[33]
(2019)

DPh-DNTT (polycrystalline) Staggered Au contacts only 56 1.0

DPh-DNTT (polycrystalline) Coplanar Chemisorbed monolayer treatment of Au contacts 
and thin gate insulator

29 1.0

[34]
(2020)

DPh-DNTT (polycrystalline) Coplanar Chemisorbed monolayer treatment of Au contacts 
and thin gate insulator

10 0.8

C10-DNTT (polycrystalline) Coplanar 31 0.8

[35]
(2020)

C10-DNTT (monolayer crystalline) Staggered Transferred Au contacts and low-dimensional semi-
conductor film

40 0.9

[36,37]
(2020)

C9-DNBDT-NW (bilayer crystalline) Staggered Contact doping and low-dimensional semiconductor 
film

60, 50 0.4, 0.3

a)If not explicitly reported, estimated based on the maximum drain current, channel width, and a channel “thickness” of 5 nm or the gate overdrive voltage and gate-
dielectric capacitance.



www.advmat.dewww.advancedsciencenews.com

2104075 (3 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

intrinsic channel mobility of 18.2 cm2 V−1 s−1,[45] similar to what 
is achievable in some premier organic semiconductors.[18,51] 
While a transit frequency of 10 MHz is sufficient for some tar-
geted applications, such as low-end flexible mobile displays,[52] 
broadening the range of possible applications for organic tran-
sistors clearly requires significant further reduction of the con-
tact resistance. However, this goal is likely only to be achieved 
through more-focused efforts toward understanding the lim-
iting factors in organic transistors that already show “low” con-
tact resistances.

Here, a focused critical analysis of the various approaches 
that have yielded RCW  <  100 Ωcm in organic transistors is 
presented, with further discussion of the prospects that any of 
these given methods will enable further improvements beyond 
the state-of-the-art, which as of writing stands at 10 Ωcm for 
high-frequency organic TFTs.[34] A necessary brief summary 
of the physical origins of the contact resistance is provided to 
facilitate a meaningful discussion of the merits of the discussed 
methods for reducing the contact resistance and about poten-
tial further routes toward improvement. The primary goal of 
this report is to clarify the important factors and mechanisms 
behind the methods that have led to the lowest contact resist-
ances in organic transistors. For a more in-depth review of the 
physics of charge injection in organic transistors and modeling 
of the contact resistance, the interested reader is referred to 
other recent and comprehensive reviews.[20,25,26,53,54] After this 
more general overview, focused discussions are provided on the 
potentially augmented role of Fermi-level pinning (FLP)[55,56] at 
the metal–organic semiconductor interfaces in organic transis-
tors with low contact resistance, since this imposes potentially 
severe limitations on the capabilities of any method aimed 
at reducing the injection-barrier height. Further discussion 
is provided on methods to combat FLP, including the role of 
the organic semiconductor itself[56–58] and the difficulties in 
forming more idealized interfaces between the contacts and the 
organic semiconductor,[35,59–61] as these are potentially the key 
parameters for enabling further reduction of the contact resist-
ance in state-of-the-art high-frequency organic TFTs. Consid-
eration will also be given to the extent to which methods for 
improving the contact resistance are applicable to the miniatur-
ization of organic transistors to nanoscale dimensions, which 
remains the essential route toward enhancing the dynamic 
performance of integrated circuits based on organic transis-
tors.[22,62,63] The discussion will include key results and ideas 
from the literature on these topics, including potential insights 
from outside the organic-electronics community, especially 
with regards to contacts to 2D semiconductors like TMDs.[41]

2. Contact Resistance in Organic Transistors

In this section, we will give a brief overview of the most impor-
tant aspects of the contact resistance in organic transistors.

2.1. Critical Impacts of the Device Architecture

There are many intrinsic and extrinsic factors determining 
the origins of the contact resistance in organic transistors. 

Potentially, the most impactful and widely studied extrinsic 
factor in organic TFTs is the choice of the device architec-
ture.[64–69] The different device architectures for organic TFTs 
can generally be grouped into coplanar and staggered configu-
rations (Figure 1a). This distinction refers simply to the position 
of the source and drain contacts with respect to the interface 
between the organic semiconductor and the gate dielectric. 
Both architectures have shown advantages and disadvantages 
in terms of the impacts on various performance characteristics 
of the TFTs, such as gate coupling and charge injection.[66,68–70] 
The conventional wisdom is that staggered TFTs should have 
smaller contact resistance than coplanar TFTs, based on experi-
mental data from organic TFTs comprising the same materials 
and film thicknesses,[64,67,71,72–74] and from theoretical device 
simulations and physical models.[66,67,73,74] The standard argu-
ment is that the effective area for charge injection is larger and 
hence the contact resistance is smaller in staggered TFTs than 
in coplanar TFTs.[67,75,76] However, there are other important fac-
tors for consideration that make a firm statement about which 
architecture is more advantageous difficult to completely justify.

The charge-injection efficiency in staggered TFTs is highly 
sensitive to the thickness of the organic-semiconductor layer due 
to the direct effects that this has on the access resistance.[29,75,77] 
In addition, a gate-voltage dependence of the contact resistance 
arises due to screening of the gate field by charges accumu-
lated in the semiconductor between the source contact and the 
gate.[70] In the case of top-contact (TC) staggered TFTs, the sem-
iconductor morphology and the method by which the contacts 
are deposited can also have significant impacts on the metal–
organic semiconductor interface and on the efficiency of charge 
transport away from the contact region.[35,59,78,79] Likewise, in 
bottom-contact (BC) coplanar TFTs, any nonidealities in the 
semiconductor morphology on the contact surface and along 
the edges of the contacts may drastically degrade the charge-
injection efficiency.[60,80,81] Coplanar contacts have been shown 
to provide low contact resistance, so long as the semiconductor 
morphology is sufficiently controlled at the interfaces between 
the contacts, the gate dielectric and the organic semicon-
ductor.[27,81,82] Moreover, it was recently demonstrated that when 
a very thin gate dielectric is implemented in coplanar TFTs, the 
contact resistance can be even lower than in comparable stag-
gered TFTs.[33,83] Due to these and other complex competing 
factors that can arise in the fabrication of organic transistors, 
the historical record on which device architecture is capable of 
showing the lowest contact resistance in organic transistors is 
somewhat ambiguous.[34] Suffice to say, at this point, the dif-
ferences in the contact resistances between sufficiently compa-
rable state-of-the-art coplanar and staggered organic TFTs are 
relatively small, as will become clear in this report.

2.2. Discerning “High” and “Low” Contact Resistance

Regardless of the specific architecture or other potentially 
contributing extrinsic issues from fabrication, the contact 
resistance is generally identified by the observation that a 
significant portion of the drain–source voltage (VDS) drops 
across the contacts, rather than along the channel region of 
the transistor.[84–86] This parasitic voltage drop (VC) tends to 

Adv. Mater. 2022, 34, 2104075



www.advmat.dewww.advancedsciencenews.com

2104075 (4 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

be non-Ohmic in organic transistors with significant depend-
ence on the applied electric fields[26,76,86–88] but can attain 
Ohmic-like character under certain conditions.[75] Injection 
can additionally be influenced by the mobility and the den-
sity of the charge-carriers in the organic semiconductor near 
to the contact interface, especially in the case of highly disor-
dered semiconductors dominated by hopping (diffusion-lim-
ited) transport.[89–92] Furthermore, there is evidence to suggest 
that the carrier mobility in the channel region also can have 
a direct impact on the contact resistance, though the precise 
nature of this influence and how it should be modeled is still 
under active investigation.[54,73,87,92,93]

Generally speaking, in transistors with sufficiently long 
channel lengths such that the greatest portion of the total 
resistance (R) is determined by the sheet resistance of the 
semiconductor layer (Rsh), the current–voltage characteristics 
of the transistors can give the impression of “Ohmic” injec-
tion, in which case the contact resistance is usually considered 

“low.” Ohmic contact resistance is achieved when the voltage 
drop across the contacts is negligible compared to the total 
voltage drop in the device.[26,84] Truly “Ohmic” contact resist-
ance is only accomplished in organic transistors when VDS is 
sufficiently low that fewer charge carriers per unit volume are 
injected than are present in the bulk via thermal activation 
or application of a gate–source voltage (VGS). Owing to the 
low permittivity of organic semiconductors, as VDS and VGS 
are increased, the Schottky barrier is reshaped and reduced 
in height by the Coulombic interactions between the injected 
charges in the semiconductor and the corresponding image 
charges on the metal contact.[53,70,89,92,94] The excess injected 
charges are diffusively transported away from the contacts 
according to space-charge limitations in the organic semi-
conductor, and the current–voltage characteristics become 
nonlinear.[95,96] A non-Ohmic (“high”) contact resistance can 
often be clearly observed in the output curves of transistors 
with sufficiently small channel length (L), where a distinctive 

Adv. Mater. 2022, 34, 2104075

Figure 1. a) Schematic cross-sections of the two main device architectures utilized for organic transistors. The source and drain contacts are separated 
by a channel of length L. b) Output curves of two organic transistors with a relatively short channel of L = 2 µm. The blue curve shows evidence of a 
high, non-Ohmic contact resistance compared to the red curve. c) Gated transmission line method (TLM) analysis corresponding to the transistors in 
panel (b) showing that they are identical other than in terms of the width-normalized contact resistance (RCW). d) Schematic diagram of the current 
density (J) in the semiconductor near the source contact in staggered and coplanar transistors, separating the contributions to the total resistance (R) 
into the interface (RI) and access (RA) resistance components of the contact resistance (RC) and the channel resistance (Rch). e) Simplified schematic 
energy diagram of the interface between a metal and an organic-semiconductor film in thermal equilibrium. The diagram depicts the situation in a 
p-channel transistor without applied voltages. ϕm: work function of the metal. Δ: interface dipole. ΦB,n: energy barrier for electrons. ΦB,n: energy barrier 
for holes. Nint: interface states. dvdW: width of the van der Waals gap. Wd: width of the depletion region. EA: electron affinity. IE: ionization energy. e) 
Adapted with permission.[58] Copyright 2017, IOP Publishing.



www.advmat.dewww.advancedsciencenews.com

2104075 (5 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

“S” shape appears when the injection barriers are large 
(Figure 1b).[70]

For the rest of this report, the terms “low” and “high” for the 
contact resistance are used with a cutoff set at RCW = 100 Ωcm, 
rather than as synonyms for Ohmic or non-Ohmic contacts, 
respectively. Indeed, some state-of-the-art organic TFTs with 
the lowest overall contact resistances can still show non-Ohmic 
behaviors for sufficiently small channel lengths, such as field 
dependence and barrier-lowering effects.[87,97]

2.3. Measurement of the Contact Resistance

The assumption of an Ohmic contact resistance along with 
adherence to the gradual channel approximation (GCA) in 
the channel of the TFT[98] allows the contact resistance to be 
measured using the gated transmission line method (TLM), 
wherein the transfer characteristics of transistors with dif-
ferent channel lengths are measured at a low drain–source 
voltage (VDS), i.e., in the linear regime of operation.[99] In a 
TLM analysis, the source resistance and the drain resistance 
are lumped into a single term denoted as RC, which gener-
ally includes contributions from the interface resistance (RI) 
and the access resistance arising from the resistivity of the 
organic semiconductor (RA) (Figure 1d). Approaches to reduce 
the contact resistance are generally aimed at addressing one 
or more of these contributions, which are rooted in the phys-
icochemical characteristics of the interface between the (typi-
cally) metal contacts and the organic semiconductor.[100] Since 
the contact resistance is known to depend on the charge-car-
rier density and field dependence of the mobility, the most 
common practice for reporting a single value for the contact 
resistance from a TLM analysis is to take the value extracted at 
the highest gate overdrive voltage (VGS–Vth), where Vth is the 
threshold voltage (Figure  1c). In staggered transistors, a crit-
ical further quantity that can be estimated using TLM is the 
transfer length (LT) that, along with the channel width (W), 
determines the area over which the majority of charge-carrier 
injection occurs.[76,101] In light of the discussion in the pre-
vious section, it should be noted that the key premise of the 
TLM, namely, that the contact resistance is Ohmic and inde-
pendent of the channel length, does not necessarily hold, due 
to the fact that the contact resistance is inherently nonlinear 
and therefore does show a non-negligible dependence on the 
channel length, even for relatively ideal TFTs.[76,87,97,102] Nev-
ertheless, while single-device characterization methods, such 
as the gated four-probe (gFP) method[65,79] and Kelvin probe 
force microscopy (KPFM),[85,86] can provide more detailed and 
potentially more accurate descriptions of the potential profiles 
at the contacts, the TLM is by far the most common approach 
for determining the contact resistance of organic as well as 
inorganic TFTs, due to its simplicity of implementation and 
its advantages in terms of providing minimal statistics for sev-
eral devices.[99]

There are a limited number of reliable approaches for the 
measurement of the contact resistance while the TFTs are oper-
ated in the saturation regime.[103] It was recently shown that 
measurements of the transit frequency (fT) of TFTs with mul-
tiple channel lengths may also be used for this purpose.[34]

2.4. Physical Origins of the Contact Resistance in Organic 
Transistors

When two dissimilar materials are brought into physical con-
tact, charge rearrangement spontaneously occurs across the 
interface to bring the combined system to thermal equilibrium, 
i.e., a common Fermi level (EF) is established.[84] Figure  1e 
shows a simplified energy diagram for a high-work func-
tion metal making contact to an organic semiconductor in a 
p-channel transistor.[58] The resulting equilibrium (re)arrange-
ment of the energy levels and charge distributions across the 
interface between the metal contact and the semiconductor will 
then to a first degree determine the charge-transport processes 
that dominate during device operation, which can be broadly 
grouped into thermally activated and quantum-tunneling-
mediated transport of carriers across the interface.[104] The most 
crucial aspects to consider for these charge injection/extrac-
tion processes between the metal and the organic semicon-
ductor are the energy (Schottky) barriers (ΦB) that form at the 
interface, the width of the interfacial van der Waals gap (dvdW) 
between the metal contact and the first layer of the organic-
semiconductor film, the presence of an interface dipole (Δ), 
and the width of the depletion region (Wd) determined by the 
energetic landscape and charge-carrier density in the semicon-
ductor.[61] For the organic transistors with the lowest contact 
resistances, charge injection can generally be described well by 
thermionic-emission theory for Schottky diodes in which the 
current density (J)follows[25,41,84]

= −Φ





 − − ∆Φ +
















α∗J A T

k T

qV

k T
DSexp 1 expB

B

B

B

 (1)

where A* is the effective Richardson constant, α is the dimen-
sionality constant (equal to 2 for bulk semiconductors and 3/2 
for 2D semiconductors), q is the elementary charge, and ΔΦB is 
the reduction in the Schottky barrier height due to image forces. 
For VDS = 0 and α = 2, the contact resistance then follows

=










Φ





∗R

k

qA T k T
expc

B B

B

 (2)

The exponential dependence of the contact resistance on ΦB 
means that this is in principle the most critical parameter for 
improving the contact resistance.[25,63] That being the case, the 
material parameter that has often been given the most atten-
tion is the work function of the contacts (ϕC), defined as the dif-
ference between the vacuum energy level (Evac) and the Fermi 
energy (EF) of the isolated contact material. For a pure metal 
contact, this is taken to be the work function of the metal (ϕm), 
and thus this is what is used in the simplified energy diagram 
in Figure 1e. Assuming vacuum-level alignment, comparison of 
ϕm to the energies of the transport levels of the isolated semi-
conductor (ionization energy (IE) and electron affinity (EA)) 
then gives an initial guideline for the expected magnitude of the 
injection barrier for holes (ΦB,p) and electrons (ΦB,n). While this 
can serve as a starting point for contact engineering, this sim-
plified view is essentially never entirely accurate, since multiple 
other extrinsic factors contribute to the final equilibrium state 
that is realized after contact formation (Figure  1e).

[23,100,105,106] 

Adv. Mater. 2022, 34, 2104075



www.advmat.dewww.advancedsciencenews.com

2104075 (6 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

While this is generally true of all metal–semiconductor con-
tacts, there are notable unique aspects that dominate the 
detailed energetic characteristics of metal–organic semicon-
ductor interfaces that ultimately determine the charge-injection 
physics.[100]

Perhaps the most well-known and studied phenomenon in 
the research into the complex physics of metal–organic semi-
conductor interfaces is the common presence of a pronounced 
interface dipole (Δ),[107] which can be quantified, e.g., by ana-
lyzing the distribution of the occupied states in the organic 
semiconductor across the semiconductor thickness by ultra-
violet photoelectron spectroscopy (UPS).[108,109] The interface 
dipole has many noted possible origins, relating primarily to 
the localized energy levels in organic semiconductors.[110] This 
includes charge transfer across the interface leading to the for-
mation of a surface sheet charge on the metal with an associ-
ated image charge-carrier density in the organic semiconductor, 
strong chemical reactions between the organic-semiconductor 
molecules and ultraclean metal surfaces, variations in the ori-
entation of the molecules at the surface and inside the semi-
conductor film, intrinsic dipole moments within the adsorbate 
molecules themselves, and the reduction of the metal work 
function by the interaction of the π-electron clouds of a phys-
isorbed organic layer with the metal electron tail states (referred 
to as the “pillow” or “push-back” effect).[26,107–111] Furthermore, 
the redistribution and accumulation of charges in the organic 
semiconductor leads to band bending.[58] Band bending is typi-
cally quantified by the extent over which the transport levels 
vary, proceeding from the metal surface into the organic semi-
conductor away from the contact interface (Wd).[92,107] While 
it is common practice to represent the bending of the energy 
levels as continuous, this is not strictly speaking appropriate for 
organic semiconductors, due to the more localized nature of the 
molecular orbitals. This is especially true for small-molecule 
semiconductors that tend to grow in a layer-by-layer mode,[112] 
as shown schematically in Figure  1e.[58] Finally, depending on 
the shape and distribution of the density of states (DOS) of the 
highest occupied molecular orbital (HOMO) and lowest unoc-
cupied molecular orbital (LUMO) (represented by Gaussian 
distributions in Figure 1e) and the potential presence of inter-
facial mid-gap states (Nit) arising from disorder or impurities, 
the equilibrium position of the Fermi level with respect to the 
transport levels of the organic semiconductor may become 
“pinned,” such that changes in the contact work function do 
not affect the final barrier heights (ΦB,p, ΦB,n).[23,56,55,77,110,113] 
This effect is known as FLP and is discussed in greater detail 
in a latter section.

The further adherence to thermionic-emission theory for 
charge-carrier transport across the interface between the contact 
and the organic semiconductor depends on the charge-carrier 
density and the efficiency of the charge-carrier transport in the 
semiconductor layer.[84,91,92,114] A key aspect that determines the 
efficiency of charge-carrier transport in organic semiconduc-
tors, as well as the quality of the contact–semiconductor inter-
face, is that the molecules interact with each other and crystal-
lize through weak van der Waals forces, rather than covalent 
bonds.[112] The DOS of the HOMO and LUMO are thus strongly 
affected not only by the chemical structure, but also by both the 
arrangement of the molecules in the organic semiconductor 

and the concentration of any dopants that could potentially be 
introduced to change the charge-carrier-transport characteris-
tics.[57,115] Efficient transport of charge in organic semiconduc-
tors relies in large part on the degree of overlap of the π-orbitals 
between adjacent molecules, which can be drastically affected 
by any differences in the semiconductor morphology.[51,116] 
The mean free path of mobile charge carriers in organic semi-
conductors is thus typically very small compared to inorganic 
semiconductors (on the order of the molecular separation dis-
tances), and is highly sensitive to, e.g., thermal fluctuations and 
extrinsic factors, such as the surface roughness of the substrate 
and surface contaminants, leading to small transfer integrals 
and diffusion-limited transport of charges.[51,91,92,115]

In most cases, the real physical interfaces formed between 
the contacts and the organic-semiconductor layer in a device 
are not as atomically sharp as can be achieved in inorganic-
semiconductor devices, but are rather ill-defined and prone 
to increased disorder depending on the device-fabrication 
approach that is implemented. Diffusion of metal nanoclus-
ters and thermal damage to the organic semiconductor often 
occur in the case of metals deposited directly onto the organic-
semiconductor layer, such as in top-contact TFTs.[59,79] On the 
other hand, amorphous or polycrystalline thin films tend to 
form when the organic semiconductor is deposited directly 
onto a bare metal contact due to large surface roughness and/
or suboptimal control of the surface energy, which is typically 
different on the contacts than on the surface of the gate dielec-
tric, such as in bottom-contact TFTs.[60,81,117] A short mean free 
path leads to diffusion-limited hopping transport of the injected 
charge, resulting in a wide depletion region and subsequently 
increased dependence of the charge-injection efficiency on the 
charge-carrier mobility in the organic semiconductor close to 
the contacts due to space-charge limitations.[26,92,93,118] This 
was shown to be particularly true of many amorphous semi-
conducting polymers where the mobility of the charge carriers 
tends to be low and recombination rates with image charges at 
the metal contacts are high.[92] By contrast, inorganic semicon-
ductors characterized by highly delocalized transport of charges 
in well-defined energy bands can be described by thermionic 
emission rather well, since the mean free path of the carriers 
is large and contributions from diffusion-limited transport are 
negligible.[41,84] In more-crystalline organic-semiconductor thin-
films, such as low-dimensional liquid crystals,[119] the charge-
carrier injection from the metal contacts can also follow thermi-
onic emission quite closely,[35,120,121] and the contact resistance 
in transistors based on such materials can be quite small.

In summary, the physical origins of the contact resistance 
in organic transistors are a complex combination of competing 
intrinsic and extrinsic factors, all of which have received some 
level of attention in the community in order to optimize charge-
carrier injection. However, as will be seen in the following 
critical overview of the lowest contact resistances so far dem-
onstrated, because of the above-outlined substantial effects that 
the thin-film morphology of the organic semiconductor has 
on the local energetics and on charge-carrier transport, con-
trolling the morphology of the organic semiconductor close to 
the metal contacts is the most prevalent requirement. Indeed, 
while minimization of the height of the Schottky barrier, e.g., 
by tuning the contact work function, has received a large 

Adv. Mater. 2022, 34, 2104075



www.advmat.dewww.advancedsciencenews.com

2104075 (7 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

amount of the attention, it is evidently of secondary importance 
when attempting to achieve the most efficient carrier injection 
possible regardless of the organic TFT architecture.[77,80]

3. Survey of Organic Transistors Showing 
RCW < 100 Ωcm
In this section, the handful of reports of organic TFTs and 
EGOFETs that have shown channel-width normalized contact 
resistances of less than 100 Ωcm (Table  1) are discussed in a 
mostly chronological fashion. It is worth highlighting that 
these contact resistances have been accomplished using various 
methods rather than one ubiquitous approach, each in some 
way affecting one or more aspects of the contact–organic semi-
conductor interface. With only one exception,[34] the contact 
resistances in these reports were measured using the gated 
TLM.[99]

3.1. First Realization of the Critical Role of Semiconductor-
Morphology Control

The first report of organic transistors with an RCW of less than 
100 Ωcm was published by Stadlober et al. in 2007 (Figure 2).[27] 
Here, a coplanar bottom-gate, bottom-contact architecture was 
implemented, using a 100 nm thick SiO2 gate dielectric, vac-
uum-deposited pentacene as the active layer and gold source 
and drain contacts that were exposed to UV/ozone to produce 
a thin layer of gold oxide on the contact surface. As the title 
of the article states, the best result (RCW  = 80 Ωcm) was at 
the time (and remains still) orders of magnitude lower than 
in most organic transistors.[34] This was a significant work for 
several other reasons. The channel lengths (L) that were imple-
mented in this study were as small as 370  nm, by virtue of 

using nanoimprint lithography.[122] This showed that the UV/
ozone treatment for reducing the contact resistance is readily 
suitable for device scaling, which is essential for increasing 
the dynamic performance of organic transistors.[22] To address 
the mechanisms behind the contact-resistance improvement,  
Stadlober et  al. noted that the gate-independent contribution 
to the contact resistance (RC,0), determined by the convergence 
point of the linear fits in the TLM analysis of the width-normal-
ized total on-state resistance (RonW) at different VGS as a func-
tion of L (Figure  2c),[78,99] was still relatively large (≈50 Ωcm) 
compared to the total contact resistance. Using the common 
interpretation for RC,0 to be representative of the interface 
resistance (RI) arising from the height of the Schottky barrier 
at the interface,[77,99] this implies that the deciding factor for 
the lower RC compared to the transistors with untreated gold 
contacts was the improvement in the gate-voltage-dependent 
access resistance (RA) to the channel region through the pen-
tacene layer by virtue of an improvement in the thin-film mor-
phology both on the treated contacts and in the channel region 
directly next to the contacts (Figure 2d,e). This provided one of 
the first pieces of experimental evidence that the strong influ-
ence of the semiconductor morphology at the interface with the 
contacts on charge exchange is of vital importance for reducing 
RCW below 100 Ωcm, perhaps even exceeding the importance 
of reducing the nominal Schottky barrier height.

3.2. Ultrahigh Charge-Carrier Density in Electrolyte-Gated 
Organic Transistors

The results from Stadlober et  al.[27] unfortunately did not 
lead to such low contact resistances becoming immediately 
common-place in organic TFTs. Indeed, despite a variety of 
new methods developed in the following decade,[20,25] only three 
additional reports of RCW  <  100 Ωcm were published.[28–30] It 

Adv. Mater. 2022, 34, 2104075

Figure 2. Output curves of bottom-gate, bottom-contact pentacene TFTs with a) untreated gold contacts and b) gold contacts exposed to UV/ozone 
prior to the pentacene deposition. c) TLM analysis of TFTs with UV/ozone-treated contacts. d) AFM topographical scan of the channel region of a TFT 
with untreated contacts. e) AFM topographical scan of the channel region of a TFT with UV/ozone-treated contacts. f) Schematic cross-section of the 
TFTs with the UV/ozone-treated contacts. Adapted with permission.[27] Copyright 2007, John Wiley & Sons.



www.advmat.dewww.advancedsciencenews.com

2104075 (8 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

is also intriguing that in these three cases, three entirely dif-
ferent methods were implemented, to the notable exclusion 
of the already proven benefit of UV/ozone treatment of the 
contacts. In 2010, Braga et al. reported on EGOFETs based on 
poly(3-hexylthiophene) (P3HT) and poly(9,9-dioctylfluorene-co-
bithiophene) (F8T2) that had exceptionally low contact resist-
ances (as low as RCW = 1 Ωcm for P3HT; shown in Figure 3).[28] 
These impressively small contact resistances were achieved 
by virtue of extremely high charge-carrier densities induced 
in the semiconductor film, with an estimated 2D carrier den-
sity of 2 × 1014 cm−2. This is afforded by the ion-gel electrolyte 
employed as the gate insulator (Figure 3b), which forms an elec-
tric double layer with high electric field strength[123,124] and from 
which ions can diffuse into the organic semiconductor under 
the influence of the applied electric field.[124] This importantly 
includes the semiconductor film in the region near to the con-
tact interface, and is thought to reduce the width of the deple-
tion region, significantly reducing the access resistance. More 
recently, Lenz et al. demonstrated that this operating principle 
can be applied to ultrashort channel, vertical organic field-effect 
transistors (VOFETs), with RCW = 2.7 Ωcm measured in other-
wise similar lateral EGOFETs.[32]

Ion diffusion into the semiconductor is also a key oper-
ating principle behind organic electrochemical transistors 
(OECTs)[125] that show in some cases even lower contact resist-
ances than EGOFETs, owing to mixed ion and electron con-
duction.[126,127] As of the writing of this report, EGOFETs and 
OECTs stand as the overall benchmark to beat for organic tran-
sistors in terms of the contact resistance. It is worth noting that 
the current record standing at 0.5 Ωcm for OECTs was based 
on crystallized poly(3,4-ethylenedioxythiophene) polystyrene 
sulfonate (PEDOT:PSS), highlighting again the importance of 
the organic semiconductor morphology on the contact resist-
ance.[127] EGOFETS and OECTs also include the only exam-
ples of RCW  <  100 Ωcm yet reported for transistors based on 
polymer semiconductors. Unfortunately, EGOFETs and OECTs 
are limited in their suitability for high-frequency applications, 
since switching between the “on” and “off” states of the tran-
sistor is achieved in the case of EGOFETs through the move-
ment of ions in the electrolyte and in the case of OECTs through 
electrochemical doping of the semiconductor, both of which are 

diffusion-limited processes that are inherently slow compared 
to promoting channel formation in a field-effect transistor.[6,125] 
It is additionally unclear whether the high charge-carrier den-
sities achievable in EGOFETs and OECTs can ever be realized 
in organic TFTs through, e.g., molecular doping (see following 
section). Nonetheless, these results provide some motivational 
proof that organic TFTs based on conventional (i.e., nonelectro-
lyte) gate dielectrics could possibly attain contact resistances on 
the order of those based on some inorganic semiconductors, 
like IGZO,[39] if the charge-carrier density can be sufficiently 
enhanced in semiconductor region close to the contacts.

3.3. Contact Doping and Reduction of the Semiconductor 
Thickness

Rather than ionic doping from an electrolyte, contact doping 
with ionic or organic molecules in the organic semiconductor 
host presents a potential alternative approach to achieve 
higher charge-carrier densities in organic semiconduc-
tors[128–131] and has also seen a measure of success in reducing 
the contact resistance in nanoscale organic TFTs[109] and in 
organic permeable-base transistors (OPBTs).[47] In 2013, Mat-
sumoto et  al. demonstrated RCW  <  100 Ωcm using a small-
molecule dopant to modify the contact properties of the TFTs 
(Figure  4).[29] There, the acceptor 1,3,4,5,7,8-hexafluorotetracy-
anonaphthaquinodimethane (F6-TNAP),[133] which has a deep 
LUMO level of 5.37  eV, was used to dope the contact regions 
of bottom-gate, top-contact TFTs based on vacuum-deposited 
dinaphtho[2,3-b:2′,3′-f]thieno[3,2-b]thiophene (DNTT) and 
2,9-didecyldinaphtho[2,3-b:2′,3′-f]thieno[3,2-b]thiophene (C10-
DNTT).[134] The potentially most crucial finding in this work 
was the demonstration of the critical importance of minimizing 
the thickness of the organic semiconductor to reduce the access 
resistance in staggered devices (Figure 4b). Indeed, it is evident 
that reducing the thickness of the DNTT layers to 15 nm was 
instrumental in achieving RCW  <  100 Ωcm (Figure  3b). This 
report is also notable in that it was the first demonstration of 
the potential for low contact resistance in organic TFTs based 
on DNTT and its functional derivatives, which unsurprisingly 
also show among the highest hole mobilities out of all organic 

Adv. Mater. 2022, 34, 2104075

Figure 3. The “ultralow” contact resistance of electrolyte-gated organic transistors is achieved by penetration of the ions from the ion-gel electrolyte 
into the organic-semiconductor layer near the contacts. a) TLM analysis. b) Results from TLM for different thicknesses of the P3HT layer. Adapted with 
permission.[28] Copyright 2010, American Institute of Physics.



www.advmat.dewww.advancedsciencenews.com

2104075 (9 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

semiconductors.[51,135–137] Combining these factors resulted in a 
lowest RCW of ≈80 Ωcm in their report (Figure 4a).

3.4. Chemical Modifications to the Contact Surfaces and 
Annealing

Rather than increasing the charge-carrier density through 
contact doping, the contact resistance can also be reduced in 
principle by the insertion of various interlayers between the 
contacts and the organic-semiconductor layer, with the inten-
tion of enabling more efficient injection from the metal to 
the transport levels by changing the contact work function or 
by introducing intermediate states to assist in charge trans-
port.[138–141] In the case of bottom-contact device architectures, 
this can take the form of either physisorbed or chemisorbed 
species. A particularly common and versatile method that is 
primarily used to improve the organic-semiconductor mor-
phology and reduce the trap-state density on different surfaces 
is to modify the surface energy and reactivity by forming a self-
assembled monolayer (SAM) on the surface.[142] SAMs have 
been an especially effective tool for tuning the surfaces of both 
the gate dielectric and the contacts in organic TFTs to improve 
the charge-transport characteristics of the organic-semicon-
ductor layer.[143–145] On noble-metal contacts, such as gold or 
silver, molecules with a thiol (SH) anchoring group are used 
to selectively bind to the surface of the contacts through oxida-
tive addition of the SH bond to the metal surface, followed 
by a reductive elimination of the hydrogen.[60,74,81,117,141,146–158] 

These thiol-based SAMs can be used to modify the work func-
tion of the metal contacts, with the magnitude of the energy 
shift depending on the collective dipole moment of the SAM 
that is controlled by the structural chemistry of the molecules 
to be assembled on the surface, such as by simply changing 
the position of a polar group relative to the thiol anchor.[151,159] 
The dipolar character and the influence of a SAM can be ben-
eficial for lowering the contact resistance in organic TFTs by 
tuning the work function by up to a few hundred millielec-
tronvolts, bringing it closer to the desired transport level of 
the semiconductor, and simultaneously improving the organic-
semiconductor morphology both on the contact surface and 
along the edges of the contacts.[74]

Since its first demonstration in organic transistors in 
2005,[160] pentafluorobenzenethiol (PFBT) has become by 
far the most widely utilized and effective thiol molecule to 
date for improving the hole injection in p-channel organic 
TFTs.[30,33,34,46,82,155–157] In many ways, PFBT provides an 
ideal SAM for contact modification to improve the injection 
of holes in p-channel TFTs, because it forms stable, well-
ordered monolayers[158] with a large collective downward 
dipole moment[161] and promotes the formation of an organic-
semiconductor thin-film morphology with excellent in-plane 
π-stacking and a high charge-carrier mobility.[33,82,162] The 
aptitude of PFBT SAMs on gold contacts for enabling RCW 
below 100 Ωcm was first demonstrated in 2017 by Kitamura 
(Figure  5).[30] In this case, PFBT-treated gold contacts in 
bottom-gate, bottom-contact TFTs were combined with a vac-
uum-deposited C10-DNTT layer with enhanced charge-carrier 

Adv. Mater. 2022, 34, 2104075

Figure 4. Results of TLM analysis of bottom-gate, top-contact organic TFTs utilizing contact doping. The small-molecule acceptor F6-TNAP (also 
referred to as F6-TCNNQ) was used to dope the contact regions to enhance the hole transport in thin films of the small-molecule semiconductors DNTT 
and C10-DNTT. a) RCW as a function of the gate overdrive voltage. b) Demonstration of the influence of the thickness of the organic-semiconductor 
layer on the contact resistance. Adapted with permission.[29] Copyright 2013, Elsevier. c) Schematic cross-section of the TFTs.

Figure 5. Low contact resistance achieved in bottom-gate, bottom-contact organic TFTs through implementation of a self-assembled monolayer (SAM) 
of pentafluorobenzenethiol (PFBT) on the gold contacts and annealing of the transistors at a temperature of 100 °C in nitrogen to enhance the charge-
carrier mobility. a) Effective carrier mobility measured in the linear regime (µlin) plotted as a function of the channel length. b) Width-normalized contact 
resistance (RCW) determined by TLM as a function of the gate overdrive voltage. Adapted with permission.[30] Copyright 2017, The Japan Society of 
Applied Physics. c) Schematic cross-section of the TFTs.



www.advmat.dewww.advancedsciencenews.com

2104075 (10 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

mobility achieved through postdeposition annealing at a tem-
perature of 100 °C in nitrogen to remove defects in the crystal 
packing (Figure  5a).[163] This led to a lowest RCW of 72 Ωcm 
at VGS–Vth  =  −10  V from the TLM analysis performed at the 
lowest VDS (−0.2 V; Figure 5b).

3.5. Low-Dimensional Liquid Crystals of Small-Molecule 
Semiconductors

In the years following 2017, at least one report of RCW  < 
100 Ωcm has been published by a few different research 
groups each year. These results have been driven primarily 
by the implementation of high-performing functional deriva-
tives of DNTT and other highly π-extended small-molecule 
semiconductors[164] in combination with optimized fabrica-
tion methods. In particular, the liquid-crystalline properties 
of these and other organic semiconductors have seen rapid 
development in recent years.[165–167] Owing to enhanced con-
trol of the crystallinity and morphology through various 
means, such as annealing and directional solution-shearing, 
liquid-crystalline organic semiconductors have shown the 

potential to realize charge-carrier mobilities in excess of 
10 cm2 V−1 s−1 in extremely thin films that can be fabricated 
into large-area single crystals by solution processing.[167] The 
potential for realizing additional improvements to the con-
tact resistance by using liquid crystals was unequivocally 
demonstrated in a seminal paper by the Takeya group in 
2018.[31] The team compared bottom-gate, top-contact TFTs 
based on monolayer (1L), bilayer (2L), and trilayer (3L) crys-
tals of 3,11-dioctyldinaphtho[2,3-d:2′,3′-d′]benzo[1,2-b:4,5-b′]
dithiophene (C8-DNBDT-NW) and a thin contact-doping layer 
of 2,3,5,6-tetrafluoro-7,7,8,8,-tetracyanoquinodimethane (F4-
TCNQ) (Figure 6). While it might have been expected that the 
lowest contact resistance would be obtained for the monolayer 
crystals (since in staggered TFTs this would reduce the access 
resistance to a practical minimum by positioning the con-
tacts as close as possible to the channel[168]), the lowest con-
tact resistance (RCW = 47 Ωcm) was actually demonstrated in 
the bilayer TFTs (Figure 6b). This was attributed to a smaller 
density of defects in the crystal packing compared to the 
monolayer crystals, as evidenced by atomic force microscopy 
(AFM) and small-angle electron diffraction measurements 
(Figure  6c,d). Two follow-up publications by the same group 

Adv. Mater. 2022, 34, 2104075

Figure 6. a) Two-terminal conductivity (σ2T) as a function of the gate–source voltage measured on bottom-gate, top-contact organic TFTs based on 
monolayer (1L), bilayer (2L), and trilayer (3L) single crystals of the small-molecule semiconductor C8-DNBDT-NW. b) Width-normalized contact resist-
ance as a function of the gate–source voltage for the bilayer and trilayer TFTs. c) TEM images and d) small-angle electron diffraction patterns of the 
1L and 2L crystals of C8-DNBDT-NW. Adapted with permission.[31] Copyright 2018, American Association of the Advancement of Science. e) Schematic 
cross-section of the TFTs that showed the lowest contact resistance. A thin layer of thermally evaporated F4-TCNQ was implemented as a contact 
dopant.



www.advmat.dewww.advancedsciencenews.com

2104075 (11 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

using similar devices have reported comparably low contact 
resistances,[36,37] leading to a concomitant enhancement in the 
on/off current ratio (up to 1010) and the transit frequency (up 
to 45 MHz).[36]

While analyses of the crystal structure suggested that the 
monolayer C8-DNBDT-NW crystals contained a higher den-
sity of packing defects that limited the TFT performance 
compared to the multilayer crystals, an additional factor that 
was not discussed is that the uppermost layer of the crys-
tals may have degraded due to thermal damage during the 
deposition of the gold contacts (by thermal evaporation in 
vacuum).[61] To combat this potentially critical effect, the 
Chan group implemented mechanically transferred gold 
contacts gently placed onto large-area monolayer crystals of 
C10-DNTT (Figure  7a).[35] Indeed, comparison of transmis-
sion electron microscopy (TEM) cross-sections showed evi-
dence of a sharper interface to the monolayer crystal with 
the transferred contacts compared to contacts deposited by 
thermal evaporation (Figure  7b), and TLM showed a low 
minimum RCW = 40 Ωcm (Figure 7d). Of all organic transis-
tors covered in this review, these devices share perhaps the 
greatest degree of similarity with FETs based on 2D inorganic 
semiconductors, especially TMDs.[40,41] This includes the use 
of mechanically transferred contacts, which is an approach 
that has also been implemented in 2D TMD transistors to 
form so-called “van der Waals contacts” characterized by 
atomically sharp interfaces without chemically or kinetically 
damaging the 2D semiconductor.[106,169] The necessity of this 
and some other elaborate approaches for forming contacts 

to TMDs arises because substitutional doping is difficult to 
effectively implement or in many cases detrimental to the 
transistor characteristics including the contact resistance.[41] 
This parallels the limitations of doping the contact regions in 
organic crystals and polycrystalline films, where the dopants 
may disrupt the molecular order and severely degrade the 
in-plane charge-carrier transport[129] or poorly intercalate the 
crystal domains.[139]

3.6. Low-Voltage Organic TFTs Using a Thin Gate Dielectric

The final examples of low contact resistance (Figure  8), 
including the standing record for organic TFTs of 10 Ωcm 
reported in 2020,[34] were a result of several combined advan-
tageous factors, but notably did not rely on the use of low-
dimensional crystals as in the previous section. The first key 
factor to achieving low contact resistance was a robust approach 
for making a very thin gate dielectric that enables low-voltage 
operation and high-mobility thin films of small-molecule semi-
conductors.[170] The importance of using a thin gate dielectric 
as it relates to contact resistance became apparent through a 
re-evaluation of the effects of the device architecture (coplanar 
vs staggered) on the charge injection in organic TFTs.[33] This 
was motivated by previous theoretical studies[68,83] in which it 
was shown through drift-diffusion-based simulations of organic 
TFTs that if the Schottky barrier is sufficiently minimized and 
the gate-dielectric thickness is reduced, charge injection can 
become more efficient in coplanar TFTs, to the extent that a 

Adv. Mater. 2022, 34, 2104075

Figure 7. a) Photograph of bottom-gate, top-contact organic TFTs formed by placing gold contacts by mechanical transfer onto the surface of a C10-
DNTT monolayer crystal. b) Cross-sectional TEM images of the interfaces between the SiO2 gate dielectric, the C10-DNTT monolayer, and gold contacts 
deposited by thermal evaporation or mechanical transfer. c) Width-normalized contact resistance and intrinsic channel mobility as a function of the 
gate overdrive voltage determined using TLM. Adapted with permission.[35] Copyright 2020, Wiley-VCH. d) Schematic cross-section of the TFTs.



www.advmat.dewww.advancedsciencenews.com

2104075 (12 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

smaller contact resistance is achievable than in otherwise com-
parable staggered devices (Figure 8a). The second most impor-
tant factor was the implementation of high-quality thin films 
of small-molecule semiconductors that can show high intrinsic 
hole mobilities.[171] In particular, C10-DNTT and 2,9-diphenyl-
dinaphtho[2,3-b:2′,3′-f]thieno[3,2-b]thiophene (DPh-DNTT) 
were chosen based on previous demonstrations of low con-
tact resistance in similarly fabricated TFTs based on these 
molecules.[137] Finally, in the BC TFTs that ultimately showed 
the lowest contact resistances (29[33] and 10 Ωcm[34] for DPh-
DNTT and 31 Ωcm[34] for C10-DNTT), a SAM of PFBT was 
implemented on the gold source and drain contacts. The effects 
of PFBT on the DPh-DNTT morphology are readily apparent 
from comparing grazing-incidence X-ray diffraction (GIXRD) 
measurements of DPh-DNTT films deposited onto bare Au, 

PFBT-treated Au, and the gate-dielectric surface (Figure 8b). On 
PFBT-treated Au, the (110), (020), and (120) diffraction peaks are 
clearly distinguished, indicating in-plane π–π stacking in the 
DPh-DNTT film,[172] similar to the measurement of DPh-DNTT 
on the optimized gate dielectric (Al2O3 passivated with n-tetra-
decylphosphonic acid, TDPA). By contrast, a clear degradation 
of the in-plane π-stacking is observed for the bare Au surface.

It is also intriguing to note that in this study, even the TC 
TFTs showed exceptionally low RCW = 56 Ωcm (Figure 8c). Per-
haps most significant is that this was demonstrated without the 
use of contact dopants or other interfacial layers of any kind, 
using gold contacts deposited by thermal evaporation directly 
onto a relatively thick (20  nm) polycrystalline film of DPh-
DNTT. This potentially indicates that the vertical access resist-
ance component from the vacuum-deposited polycrystalline  

Adv. Mater. 2022, 34, 2104075

Figure 8. a) Simulated normalized output curves of bottom-contact (BC) organic TFTs as a function of the gate-dielectric thickness (tdiel). Adapted 
with permission.[83] Copyright 2015, American Physical Society. b) Grazing-incidence X-ray diffraction measurements on 30 nm thick DPh-DNTT films 
on surfaces consisting of Au (black), PFBT-treated Au (red), and atomic-layer-deposited Al2O3 passivated with an n-tetradecylphosphonic acid (TDPA) 
SAM (blue). The (110), (020), and (120) diffraction peaks are clearly distinguished in both of the two latter cases, signifying strong in-plane π–π stacking, 
while on bare Au, only the (110) peak is present. c) Width-normalized contact resistance (RCW) as a function of the gate overdrive voltage of top-contact 
(TC) and BC TFTs. Adapted with permission.[33] Copyright 2019, Springer Nature. d) Transit frequency (fT) as a function of the channel length (L) of BC 
TFTs. Adapted with permission.[34] Copyright 2020, American Association for the Advancement of Science. e) Schematic cross-section of the TC TFTs. 
f) Schematic cross-section of the BC TFTs.



www.advmat.dewww.advancedsciencenews.com

2104075 (13 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

DPh-DNTT thin-films is exceptionally low in comparison to 
thin-films of other organic semiconductors, making DPh-
DNTT potentially relevant also for vertical transistors, such as 
OPBTs.[7,47]

4. Present Challenges for Realizing a Contact 
Resistance of 1 Ωcm or Lower in Organic TFTs
From the empirical evidence reviewed in the previous section, 
we will now try to elucidate a general approach for obtaining 
state-of-the-art contact resistances in organic TFTs in either the 
coplanar or the staggered device architecture. The most critical 
factor regardless of the device architecture is undoubtedly to 
minimize the access-resistance component of the contact resist-
ance. This is primarily achieved through control of the mor-
phology of the organic semiconductor to minimize structural 
defects affecting charge transport both near the interfaces with 
the contacts and within the channel region close to the contacts. 
This is essential for promoting the most efficient charge injec-
tion and transport and must be considered as a prerequisite for 
future studies seeking to develop new methods to reduce the 
contact resistance. For staggered TFTs, this often includes an 
additional requirement that the thickness of the organic semi-
conductor be as small as possible. Furthermore, comparing 
the evidence from the above reports suggests that the myriad 
methods for fine-tuning other aspects of the interface, such 
as the contact work function and charge-carrier density, have 
played a relatively secondary role in achieving the state-of-the-
art thus far.

To make an assessment about what the primary limiting 
factors are once the organic-semiconductor morphology is 
sufficiently controlled, it should first be acknowledged that 
even for some of the above-discussed organic transistors, 
which have nearly “ideal” characteristics (including low-
dimensional highly crystalline semiconductor films, high car-
rier mobilities, sharp contact interfaces, thin gate dielectrics), 
the contact resistance is still no better than about 10 Ωcm and 
thus still substantially higher than in state-of-the-art inorganic 
transistors.[40–42] This hints that there is a common limita-
tion that has so far not been fully addressed in these devices 
which prevents the contact resistance from being reduced 
much further below 10 Ωcm. A compelling explanation for 
this observation is that the Schottky-barrier height has not 
been sufficiently reduced by the various methods that have 
so far been implemented to improve the interface energetics, 
such as using strongly dipolar SAMs.[77,149,161] The inability 
to tune the Schottky-barrier height can likely be attributed  
to the evidently ubiquitous occurrence of FLP in the contacts to  
organic semiconductors.[55,56,173] The likeliness of this explana-
tion is supported by experimental and theoretical work both 
from within the organic-semiconductor community and the 
broader transistor community, especially with respect to 2D 
semiconductors.[40–42]

In the next section, the origins of FLP in metal–semicon-
ductor contacts in general are reviewed, and further unique 
aspects related to organic semiconductors are expounded upon. 
Finally, the possible solutions to overcoming the effects of FLP 
to enable lower contact resistance are explored.

4.1. Fermi-Level Pinning at Metal–Organic Semiconductor 
Interfaces

Since the equilibrium position of the Fermi level (EF) in a 
metal–semiconductor contact depends most strongly on the 
work function of the contact material (ϕC), this is often a major 
focal point of efforts to tune the interface energetics to enable 
efficient injection of holes or electrons into the semiconductor.[8] 
The heights of the hole-injection barrier (ΦB,p) and electron-
injection barrier (ΦB,n) are taken to be the energy difference 
between the respective transport levels (HOMO and LUMO) 
and EF (Figure 1b). Many reports have demonstrated that using 
a different contact metal[64] or otherwise modifying ϕC through 
chemical means like dipolar SAMs[117,151] can improve the con-
tact resistance in organic transistors, ostensibly due to a reduc-
tion in the relevant injection-barrier height. This is at least part 
of the explanation that is given for the often-excellent contact 
resistance improvement of p-channel bottom-contact organic 
TFTs, which utilized PFBT and other fluorinated SAMs,[154,174] 
including the experimental instances reviewed in the previous 
section.[30,33,34] This has thus motivated investigations into other 
thiol molecules capable of forming dipolar SAMs on the metal 
contacts with some particular emphasis on those with a larger 
number of fluorine atoms to induce a work-function shift to 
higher energies beyond that obtained with PFBT.[153,154] How-
ever, the efficacy of these approaches relies on adherence to the 
so-called Schottky–Mott limit, wherein ΦB,p and ΦB,n scale uni-
formly with respect to ϕC according to

ϕ ϕ= Φ = − Φ =S cd /d d /d 1B,n B,p c  (3)

While S = 1 can be observed in many cases for values of ϕC 
that fall within the bandgap of the semiconductor,[55,173] S can 
also be significantly smaller than unity and may gradually or 
abruptly decrease to zero at low and/or high ϕC. The deviation 
of S from the Schottky–Mott limit is known generally as FLP 
and has many potential physical origins.[55,57,107,109,175–177] There 
is growing evidence that the benefits of dipolar SAMs and sim-
ilar approaches are inherently limited by FLP.[77,178]

FLP in metal–organic semiconductor interfaces has been 
most prominently observed and studied through investigations 
using UPS.[109] UPS is a powerful method that can be used to 
evaluate the interface energetics by measuring the semicon-
ductor work function (ΦORG) and its dependence on the sub-
strate work function (ΦSUB) (Figure 9).[23] In this way, deviations 
from the Schottky–Mott limit at high or low substrate work 
functions have been observed for many organic semiconduc-
tors, including small-molecule semiconductors and polymers 
(Figure 9a,b).[55,173]

FLP has also been known for many years from studies of 
metal contacts to inorganic semiconductors, of which GaAs is 
a notorious example. In this case, the Fermi level can become 
pinned due to the formation of interface states with ener-
gies within the bandgap that occur due to adatom-induced 
defects, metal-induced gap states, chemical impurities, and/
or crystallographic dislocations at the surface of the GaAs.[179] 
While some of these explanations are applicable to metal–
organic semiconductor contacts as well, the precise mecha-
nisms for the occurrence of FLP in organic semiconductors 

Adv. Mater. 2022, 34, 2104075



www.advmat.dewww.advancedsciencenews.com

2104075 (14 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

are somewhat different and still under active investigation, 
given that the experimental evidence can support multiple 
nuanced interpretations.[55–58] One of the most prominent 
explanations for the occurrence of FLP at metal–organic 
semiconductor interfaces is that interfacial polarons, which 
are created through the spontaneous equilibration process 
across the interface lead also to integer charge-transfer (ICT) 
states in the organic semiconductor with fixed energy levels 
for electrons (EICT−) or holes (EICT+) that are lower (higher) 
in energy than the bulk EA (IE) (Figure  9c).[23,110] Charge 
transfer between these states then results in an additional 
contribution to the interface dipole that leads to FLP. How-
ever, recent electrostatic modeling efforts (Figure  10a,b)[57] 
have shown that the presence of an ICT state is insufficient 
to account for the large minimum injection-barrier heights 
that are often observed at interfaces that show FLP, such as 
Au–pentacene (ΦB,p = 0.4 eV in reference [180]). Experimental 
FLP results for several systems, including the dependence on 
film thickness, could potentially more generally be explained 
by the intrinsically broad DOS of the HOMO/LUMO levels 
that is a ubiquitous characteristic of organic semiconductors 

(Figure 10c).[56,176] Furthermore, the universality of this expla-
nation extends beyond potential effects related to disorder at 
the interface and was expounded upon by Yang et  al.[56,58] It 
was found that while the common assertion that the pres-
ence of a separate density of gap states in the bandgap of the 
semiconductor due to disorder can indeed contribute to FLP, 
it is not a prerequisite. Instead, the occurrence of FLP and 
by extension the magnitude of the minimum heights of the 
injection barriers that are possible for a given contact–organic 
semiconductor system could be predicted based entirely on 
the standard deviations of the assumed Gaussian HOMO and 
LUMO distributions in the organic semiconductor in weak 
contact with a metal (Figure 10d).[56]

The dependence of FLP on the broadness of the DOS of 
the LUMO or HOMO provides an implication that the route 
toward lower injection barriers and thus lower contact resist-
ances in organic transistors coincides with narrowing the 
DOS. Indeed, this agrees well with the relative prominence 
of low-dimensional liquid-crystalline films of organic semi-
conductors in the above survey of the lowest contact resist-
ances[31,35–37] and aligns with the general motivation for 

Adv. Mater. 2022, 34, 2104075

Figure 9. a) Survey of results from ultraviolet photoelectron spectroscopy (UPS) studies of various small-molecule and polymeric semiconductors 
showing the dependence of the work function of the organic semiconductors (ΦORG) on the substrate work function (ΦSUB). Data points for a few semi-
conductors relevant for organic transistors are highlighted. Adapted with permission.[55] Copyright 2009, Wiley-VCH. b) Survey of additional organic 
semiconductors of particular interest for the development of organic solar cells. In both panels (a) and (b), strong Fermi-level pinning is evident from 
the sharply decreased scaling factors at low and high ΦSUB. c) Schematic diagrams outlining the integer charge-transfer (ICT) explanation for FLP. 
Adapted with permission.[173] Copyright 2019, John Wiley & Sons.



www.advmat.dewww.advancedsciencenews.com

2104075 (15 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

fabricating 2D organic crystals for transistors.[167,181] Fortu-
nately, this may then put the elimination of FLP and thus 
realizing lower contact resistances in close alignment with the 
wider research interests in developing organic-semiconductor 
thin films with higher intrinsic charge-carrier mobilities, since 
this generally coincides with a narrower DOS.[182,183] While 
promising, more efforts are clearly needed to enable tran-
sistors with lower contact resistance using other promising 
highly crystalline organic semiconductors, such as 2-decyl-
7-phenyl[1]-benzothieno[3;2-b][1]benzothiophene (Ph-BTBT-
C10)[168,184] and rubrene.[185] One such limitation that remains 
even in the case of a narrower DOS is that these transistors 
are predominantly implemented in bottom-gate, top-contact 
architectures where the access resistance can remain a major 
detriment due to generally lower out-of-plane charge-carrier 
mobilities.[51,168] Therefore, the development of an effective 
method for fabricating transistors based on low-dimensional 
single crystals in a coplanar architecture may help to alleviate 
this disadvantage and take full advantage of the enhanced 
charge-carrier mobility to reduce the contact resistance.[186] 
The primary difficulty for such an approach is to “match” the 

surface-energy characteristics of the contacts and the gate 
insulator using, e.g., SAMs[187] to avoid disruption of the crys-
tallization of the organic semiconductor across the edges of 
the contacts.

4.2. Depinning the Fermi Level at Metal–Organic  
Semiconductor Interfaces

Besides the potential to eliminate FLP through narrowing the 
DOS, forming a more ideal metal–organic semiconductor inter-
face and decoupling other sources of pinning, such as interface 
traps, might be achieved by inserting a thin interlayer. Such an 
approach has been shown to be effective for “depinning” the 
Fermi level in inorganic devices, e.g., using graphene or hex-
agonal boron nitride (hBN) interlayers to make contacts in 2D 
transistors based on TMDs,[41,106] and has to some extent been 
applied to organic transistors (Figure  11).[188,189] Quasi-Ohmic 
and tunneling contacts were also demonstrated between penta-
cene and gold by sputtering a thin interlayer (≈6 nm) of silicon 
nitride onto the pentacene film prior to depositing the gold 

Adv. Mater. 2022, 34, 2104075

Figure 10. a) Simulations of the energy-level offsets of pentacene showing the effects of a broader density of states via the standard deviation (σL) for 
the LUMO on the Fermi-level-pinning (FLP) behavior at low substrate work functions. b) Simulations showing the effects of different pentacene film 
thicknesses on the FLP behavior indicated by reductions in S at the extreme ends of the effective work function. Adapted with permission.[57] Copyright 
2014, Springer Nature. c) High-resolution ultraviolet photoelectron spectroscopy (UPS) measurements of the HOMO of pentacene in various states. 
In general, horizontally oriented monolayers (ML) show a much sharper density of states (DOS) comparable to the gas phase than “upright” oriented 
monolayer crystals or disordered layers, which are more relevant for thin-film transistors. Adapted with permission.[58] Copyright 2017, IOP Publishing. 
d) Analysis of the barrier height for holes (Δh) of pentacene thin films as a function of the work function of the substrate. Using a film with smaller 
thickness (down to 1.6 nm) shows a significant reduction in the pinning level. Adapted with permission.[56] Copyright 2017, Elsevier.



www.advmat.dewww.advancedsciencenews.com

2104075 (16 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

contacts (Figure 11a).[190] The accompanying theoretical simula-
tions predicted an ideal thickness for the interlayer based on 
the intrinsic pinning properties of the contact–organic semi-
conductor system to balance depinning and tunneling of car-
riers across the interlayer (Figure  11b). However, the results 
showed that even for the ideal silicon-nitride thickness of about 
2 nm, the contact resistance would still be significantly higher 
than in the absence of an energy barrier at the interface.

Alternatively, SAMs may provide a more versatile avenue 
for introducing an interlayer to depin the Fermi level. Along 
with the added advantage of being inherently self-patterning, 
the chain length of the SAM interlayer can be modified to tune 
its thickness, thus minimizing the tunneling barrier between 
the metal and organic semiconductor.[150] The potential for 
SAMs to optimize the interface energetics extends beyond 
tuning the interlayer thickness and the work function.[117] For 
instance, Lindell et  al.[113] showed that the pinning level in 
F4-TCNQ could be adjusted using alkanethiol SAMs with dif-
ferent tail groups (CH3, COH, COOH) on gold contacts. Thus, 
while work-function tuning with SAMs is limited by FLP, 
choosing an appropriate tail group for the SAM to change 
the charge-screening properties and induce a more-preferable 
pinning-energy level within the organic semiconductor may 
be possible. However, the precise mechanism behind the 
changes in pinning level with the different tail groups needs 
more detailed investigation.

Finally, akin to the development of contact heterostruc-
tures in OLEDs, the interlayer need not be composed of 
a single component material but rather a stack of mate-
rials selected based on their energetic properties. This was 

demonstrated in recent work by Blom and co-workers[138,139] 
using a strategy with a combined interlayer of a high-work-
function metal oxide (MoO3) and an organic layer with a deep 
IE to improve hole injection in organic devices (Figure  11c). 
While this strategy proved to be a promising route for dras-
tically improving the injection into relatively amorphous 
organic semiconductors by orders of magnitude, the results 
of further application to high-mobility polycrystalline layers 
of 2,7-dioctyl[1]benzothieno[3,2-b][1]benzothiophene (C8-
BTBT) showed less-compelling results with a lowest RCW  ≈ 
15 kΩcm(Figure 11d), significantly higher than what has been 
demonstrated using, e.g., contact doping.[191] The explanation 
given by the authors was that the tendency of the interlayer 
to agglomerate at the edges of the molecular terraces in the 
C8-BTBT film prevents sufficient tuning of the interface ener-
getics (Figure  11e). This problem could potentially be solved 
by adopting a coplanar device architecture, since controlling 
the homogeneity of the interlayer will in principle not depend 
strongly on the organic-semiconductor morphology.

4.3. Doping the Organic-Semiconductor Host

Doping with strong acceptors (donors) has proven to be a pow-
erful tool for enhancing hole (electron) transport and injec-
tion into organic semiconductors, primarily by increasing the 
mobile charge-carrier density and helping to alleviate some of 
the effects of molecular disorder, such as by filling trap states in 
the host organic semiconductor.[128–132] Contact doping has been 
shown to be especially effective for narrowing the space-charge 

Adv. Mater. 2022, 34, 2104075

Figure 11. a) Measurements of the total and normalized relative contact resistivity in gold contacts to pentacene including a Si3N4 interlayer. b) Simu-
lations evaluating the relative improvement in the contact resistance with the insulating layer thickness. Adapted with permission.[190] Copyright 2010, 
American Physical Society. c) Schematic energy diagrams showing the effect of adding an interlayer (IL) with deep-lying LUMO to the metal–organic 
semiconductor contact. d) Contact resistance as a function of the gate–source voltage in bottom-gate, top-contact organic transistors based on poly-
crystalline C8-BTBT. Improvement in the contact resistance can clearly be seen by using MoO3 and an IL. e) Atomic force microscopy (AFM) height 
scans of the C8-BTBT thin films without (left) and with (right) the IL deposited on top. Adapted with permission.[139] Copyright 2020, John Wiley & Sons.



www.advmat.dewww.advancedsciencenews.com

2104075 (17 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbH

region as the primary mechanism leading to lower contact 
resistance in organic transistors.[192] However, while contact 
dopants were also implemented in some of the above cases 
with RCW < 100 Ωcm,[29,31,36,37] it is not actually clear if the pres-
ence of contact dopants was instrumental in achieving these 
results. A much larger portion of the contact resistance in 
these devices seems to have instead been eliminated by mini-
mizing the access resistance through reduction of the organic-
semiconductor thickness (see, e.g., Figures 4b and 6a).[29,31,35,168] 
This is further supported when considering the other cases 
surveyed above where dopants are completely absent and use 
either the same organic semiconductors or similar fabrication 
approaches.[30,34,35]

There are multiple possible explanations for this perceived 
lack of effectiveness of contact doping in the cases where the 
contact resistance is already below 100 Ωcm. To start, doping 
in organic semiconductors, especially in low-dimensional crys-
tals, has so far shown limited efficiency compared to doping 
inorganic semiconductors[128–130,193] or through ion penetration 
from an electrolyte as in EGOFETs and OECTs.[28,123–127] Often 
very high dopant concentrations are required to meaningfully 
enhance the conductivity of intrinsic organic-semiconductor 
hosts primarily through trap filling, as shown, e.g., by Méndez 
et al. using F4-TCNQ to dope C8-BTBT (Figure 12a–c).[129] Fur-
thermore, there is typically an upper limit to the doping con-
centration that can be added, since adding too many dopant 

molecules eventually proves to be detrimental to charge 
transport, due to disruption of the molecular packing of 
the host material and degradation of the transfer integrals 
(Figure  12c).[128–130,193] Some additional lack of efficiency may 
also be a result of the method by which contact doping is 
usually implemented in the fabrication process: In the above 
organic transistors that used contact doping showing low 
contact resistance[29,31,36,37] and in most other instances, it is 
common practice to add pure dopant molecules via thermal 
sublimation as an additional interlayer deposited after the 
organic-semiconductor host and prior to the deposition of 
the contacts. This can potentially limit the efficient integra-
tion of dopants into the host-semiconductor films where they 
could provide the greatest benefit for enabling efficient charge 
transfer. Furthermore, while dopants are clearly effective for 
reducing the width of the space-charge region next to the metal 
contacts (Figure  12d), dopants may not be sufficient to coun-
teract the limitations imposed by FLP. This was demonstrated 
by Olthof et  al.[194] using UPS measurements of various host 
films and dopants that showed that the Fermi level remains 
pinned to a fixed level determined by the host, regardless of 
the dopant strength and concentration (Figure  12c–e). None-
theless, combining an efficient dopant to minimize the space-
charge region with an effective depinning strategy is a poten-
tially promising route toward minimizing the effective injec-
tion barrier and thereby realizing lower contact resistances.

Adv. Mater. 2022, 34, 2104075

Figure 12. a) Prototypical example of a host–dopant system of C8-BTBT and TCNQ with various degrees of fluorination. b) Conductivity as a function of 
doping concentration of C8-BTBT doped with F4-TCNQ. c) Specular X-ray diffraction of C8-BTBT doped with F4-TCNQ. Higher concentrations result in 
a degradation of the crystal packing of the host (yellow highlighted peaks). Adapted with permission.[129] Copyright 2013, Wiley-VCH. d) Hole-injection 
barrier (φ) as a function of layer thickness of MeO-TPD doped with F4-TCNQ. e) UPS spectra of the HOMO region in MeO-TPD with three different 
dopants. f) UPS spectra of three different hosts doped with F4-TCNQ. The hole-injection barrier is dominated by the pinning level determined by the 
host. Adapted with permission.[194] Copyright 2009, American Institute of Physics.



www.advmat.dewww.advancedsciencenews.com

2104075 (18 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbHAdv. Mater. 2022, 34, 2104075

4.4. What Does an Ideal Contact Look Like?

During the writing of this report, an annealing-free approach 
was introduced wherein the contact resistance in transistors 
based on TMDs could be almost completely eliminated by 
including a semimetal interlayer to form contacts free of metal-
induced gap states and with a Schottky barrier height reduced 
to zero.[42] Using a semimetallic bismuth interlayer with MoS2 
yielded an ultralow RCW of 0.012 Ωcm, within an order of mag-
nitude of the quantum limit over the measured charge-carrier 
density range.[40]

Mastery over the contacts to TMDs did not happen over-
night. Several challenges that share some analogue with 
the most-advanced contacts to organic semiconductors had 
to be overcome through clever contact engineering.[41] In 
particular,

– there typically is a van der Waals gap between the metal con-
tact and the top-most semiconductor layer (Figure 1e);

– substitutional doping is limited or impossible;
– out-of-plane mobility is much smaller than in-plane mobility;
– FLP limits the degree to which the Schottky barrier can be 

reduced by work-function tuning.

The general unavailability of substitutional doping is the 
most critical limitation, as this constrains the degree to which 
the semiconductor itself can be modified to enhance the mobile 
charge-carrier density. These issues have been partially circum-
vented in 2D transistors with a variety of methods, such as by 
forming so-called edge contacts where the metal contact is fab-
ricated such that charges are injected at the high-mobility in-
plane edge of the 2D crystal.[195] Since organic molecular crys-
tals are formed by van der Waals interactions between mole-
cules, edge contacts may not be feasible due to disturbances to 
the molecular-packing structure upon contact formation.[59,61] 
Interlayers may therefore continue to play a role as they have in 
some TMD devices, such as MoTe2 transistors with scandium 
contacts, where a single-atom-thick layer of hexagonal boron 
nitride can be inserted to prevent FLP and the (in this case) 
detrimental reaction of the first few layers of MoTe2 with the 
scandium.[196] In any case, for practical contacts where an out-
of-plane contribution to charge injection is inevitable, a semi-
hierarchical series of contact structures can be envisioned, fol-
lowing inspiration from the contacts to TMDs (Figure  13). In 
Figure  13, a comparison is made between demonstrated low-
RC structures in organic transistors and their potentially more 
“ideal” counterparts.

Figure 13. Contact idealization in organic transistors may take many forms. The multilayer organic crystal with a staggered architecture is limited 
by a larger access resistance (RA). The transfer length (LT) is indicative primarily of the sheet resistance under the contacts but can also be affected 
by the interface resistance (RI), with a larger LT arising for a constant contact resistivity (ρC). In addition, the contact overlap length (Lov) needs to 
be sufficiently larger than LT. In the coplanar architecture, an interlayer is necessary to control the semiconductor morphology; RA is reduced and RI 
dominates the contact resistance. Reduction to a monolayer crystal minimizes RA and in combination with a depinning layer may lead to minimization 
of RI. Recessing the contacts into the gate insulator may yield additional improvement to the morphology transitioning from the contact to insulator 
surface. The potentially “ideal” case is that of a metallized organic layer in the contact region, which minimizes LT, RI, and RA. The “contact” will be 
formed by two phases of the same organic semiconductor (a highly conducting phase and the semiconducting phase).



www.advmat.dewww.advancedsciencenews.com

2104075 (19 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbHAdv. Mater. 2022, 34, 2104075

To achieve an ideal situation, the van der Waals gap would 
likely need to be minimized or eliminated entirely by hybridiza-
tion of the metal with the first layer of molecules in the semi-
conductor to minimize this contribution to the injection bar-
rier.[41,169] In the case of a single-monolayer crystal, ideally the 
entire layer of metallized semiconductor underneath/above the 
metal contact would then act as the contact to the nonmetal-
lized channel region. This can be achieved in TMD transistors 
by careful selection of the contact metal and the semiconductor 
with additional steps to remove surface contaminants.[40] Poten-
tial drawbacks to be aware of are that hybridization of the metal 
with the TMD is not always beneficial,[196] and this approach 
often requires subsequent high-temperature annealing pro-
cesses to complete the reaction at the interface (this of course 
excepts the recent development discussed at the start of this 
section[42]). For organic transistors, a low-temperature process 
would need to be implemented to form similarly intimate con-
tact interfaces, so as not to destroy the semiconductor and to 
maintain applicability to flexible substrates.

The development of low-temperature solutions for forming 
more intimate interfaces to organic semiconductors might 
directly benefit from the collective research efforts on molecular 
electronic junctions[197] and hybridized-molecular monolayers 
on metals.[24,100] In particular, efforts to understand the role 
of anchoring groups on the efficiency of charge transfer from 
a metal contact to a chemisorbed molecule may prompt the 
design of more effective SAMs or organic semiconductors with 
prospects for direct hybridization with the metal contacts.[198] 
It may also be possible to implement a metal–organic hybridi-
zation strategy with a donor (acceptor) molecule that forms 
a single-atom-thick chemisorbed layer on metal contacts to 
improve electron (hole) injection. Various studies have revealed 
methods for achieving such selective formation of strongly 
bonded molecular monolayers to metals, including F4-TCNQ 
on gold[199] and hexaazatriphenylene-hexanitrile (HATCN)[200] 
and phthalocyanines on silver,[201] showing that hybridized 
states occur upon chemisorption to the metal surface. Reduc-
tion of the bandgap upon interfacing with a metal surface has 
also been commonly observed in some other aromatic mole-
cules,[105,202] indicating the possibility to metalize the first layer 
of the semiconductor or the interlayer, but these approaches 
have to our knowledge not yet been applied to contacts in state-
of-the-art organic TFTs showing low contact resistance.

Finally, it should be reemphasized that the design of an 
“ideal” contact will likely remain largely holistic, with a “uni-
versal” strategy unlikely to ever be fully realized. In all of these 
strategies, care would have to be taken to mitigate detrimental 
effects on a case-by-case basis, such as the interface dipole intro-
duced by, e.g., the push-back effect arising from the van der 
Waals and Pauli exclusion interactions between the metal con-
tact and the organic molecules.[203] A persistent major challenge 
will also be attempting to implement a well-ordered organic-
semiconductor layer across the contact-to-channel interface, 
which would presumably have different surface chemistries 
and reactivities (metals, oxides, or other organics). Further, a 
hybridized interface does not necessarily do anything to help 
with FLP, and the so-far developed depinning approaches, such 
as the use of an insulating interlayer,[190] are mutually exclu-
sive in principle to the hybridization approaches for removing 

the van der Waals gap. In general, development of all of these 
approaches will likely be best guided by corresponding density-
functional-theory calculations of interface layers on metals,[204] 
in combination with experimental investigations focusing on 
the organic semiconductors that have already shown excellent 
contact resistance, such as DNTT and its derivatives, wherever 
possible.

5. Summary and Outlook

A variety of methods have been developed to improve the con-
tact properties in organic transistors over the past 20+ years. 
However, large contact resistance continues to be a major 
source of nonideality in organic transistors and presents a 
severe limitation for applying organic transistors in widespread 
commercial electronics applications. For the few reported cases 
covered in this report that have shown width-normalized con-
tact resistances below 100 Ωcm, the most impactful parameter 
nearly to the exclusion of all others so far has been control 
of the intrinsic organic-semiconductor morphology near the 
contacts. Other methods developed for improving the contact 
resistance, such as contact doping and modifying the energy-
level alignment of metal contacts to the charge-transport levels, 
by contrast, have evidently played a comparatively limited role 
in achieving the state-of-the-art so far. This hints that a limiting 
factor persists once the organic-semiconductor morphology is 
sufficiently controlled and has yet to be addressed adequately by 
these other methods. A compelling explanation may be found 
in the evidently ubiquitous and obtrusive occurrence of Fermi-
level pinning in metal–organic semiconductor interfaces. Given 
that there is apparently a strong link between the appearance 
of Fermi-level pinning and the generally broad density of states 
of the transport levels in condensed thin films of organic semi-
conductors, it can be predicted that molecular engineering 
of the organic-semiconductor layer specifically to narrow the 
density of states is paramount for reducing the contact resist-
ance beyond what is currently achievable in the state-of-the-art 
systems discussed in this report. The significant recent strides 
in developing ultrathin monolayer crystals for organic transis-
tors has shown some promise and will more than likely play 
a key role going forward. The essentiality of contact doping 
remains to be fully proven, but the development of optimized 
dopant–host systems that sufficiently reduce the width of the 
space-charge region may enable further improvements by 
minimizing the need to reduce the height of the Schottky bar-
rier, thus potentially circumventing the detrimental effects of 
Fermi-level pinning. Chemical modification of the metal con-
tacts using SAMs remains relevant, but the working principles 
must necessarily be extended beyond their effects on the work 
function of the contacts to include the detailed effects that the 
SAM chemistry has on the interfacing organic-semiconductor 
layer. Beyond these more well-known or proven methods, 
some potential avenues to improve the contact–organic semi-
conductor interface remain to be thoroughly explored, such as 
the use of interlayers to depin the Fermi level or methods to 
metallize the first interfacing layer(s) of the organic semicon-
ductor. In any case, we emphasize that it would be beneficial 
going forward to require any new methods or claims of further 



www.advmat.dewww.advancedsciencenews.com

2104075 (20 of 24) © 2021 The Authors. Advanced Materials published by Wiley-VCH GmbHAdv. Mater. 2022, 34, 2104075

progress in reducing the contact resistance to first unequivo-
cally account for any potential effects related to changes in the 
organic-semiconductor morphology, given its clear dominance 
on the charge-injection characteristics.

Acknowledgements
J.W.B. and H.K. gratefully acknowledge stimulating discussions with 
Florian von Wrochem and Paddy K. L. Chan. R.T.W. acknowledges 
funding by the Deutsche Forschungsgemeinschaft (DFG, German 
Research Foundation) under Germany’s Excellence Strategy – EXC 
2089/1−390776260 (e-conversion). S.L. acknowledges support from the 
Center for Integrated Quantum Science and Technology (IQST) and the 
Carl Zeiss Foundation.

Open access funding enabled and organized by Projekt DEAL.

Conflict of Interest
The authors declare no conflict of interest.

Keywords
charge injection, contact resistance, Fermi-level pinning, interfaces, 
organic transistors

Received: May 28, 2021
Revised: August 20, 2021

Published online: October 7, 2021

[1] Organic Light Emitting Devices (Eds: K. Müllen, U. Scherf), Wiley-
VCH, Weinheim, Germany 2006.

[2] A. W. Hains, Z. Liang, M. A. Woodhouse, B. A. Gregg, Chem. Rev. 
2010, 110, 6689.

[3] P. Peumans, A. Yakimov, S. R. Forrest, J. Appl. Phys. 2003, 93, 3693.
[4] H. Sirringhaus, Adv. Mater. 2014, 26, 1319.
[5] C. D. Dimitrakopoulos, P. R. L. Malenfant, Adv. Mater. 2002, 14, 99.
[6] M. J. Panzer, C. D. Frisbie, Adv. Mater. 2008, 20, 3177.
[7] E. Guo, F. Dollinger, B. Amaya, A. Fischer, H. Kleemann, Adv. Opt. 

Mater. 2021, 9, 2002058.
[8] N. Koch, ChemPhysChem 2007, 8, 1438.
[9] S.  Reineke, F.  Lindner, G.  Schwartz, N.  Seidler, K.  Walzer, 

B. Lüssem, K. Leo, Nature 2009, 459, 234.
[10] J.-H. Jou, S. Kumar, A. Agrawal, T.-H. Li, S. Sahoo, J. Mater. Chem. 

C 2015, 3, 2974.
[11] S. Scholz, D. Kondakov, B. Lüssem, K. Leo, Chem. Rev. 2015, 115, 

8449.
[12] G.  Sini, M.  Schubert, C.  Risko, S.  Roland, O. P.  Lee, Z.  Chen, 

T. V.  Richter, D.  Dolfen, V.  Coropceanu, S.  Ludwigs, U.  Scherf, 
A. Facchetti, J. M. J. Fréchet, D. Neher, Adv. Energy Mater. 2018, 8, 
1702232.

[13] A.  Wadsworth, M.  Moser, A.  Marks, M. S.  Little, N.  Gasparini, 
C. J.  Brabec, D.  Baran, I.  McCulloch, Chem. Soc. Rev. 2019, 48, 
1596.

[14] G.  Schweicher, G.  Garbay, R.  Jouclas, F.  Vibert, F.  Devaux, 
Y. H. Geerts, Adv. Mater. 2020, 32, 1905909.

[15] G.  Gelinck, P.  Heremans, K.  Nomoto, T. D.  Anthopoulos, Adv. 
Mater. 2010, 22, 3778.

[16] T. Yang, Q. Wu, F. Dai, K. Huang, H. Xu, C. Liu, C. Chen, S. Hu, 
X.  Liang, X.  Liu, Y.-Y.  Noh, C.  Liu, Adv. Funct. Mater. 2020, 30, 
1903889.

[17] J. Harding, Inf. Disp. 2019, 35, 9.
[18] A. F. Paterson, T. D. Anthopoulos, Nat. Commun. 2018, 9, 5264.
[19] P. K. L. Chan, Adv. Electron. Mater. 2019, 5, 1900029.
[20] M. Waldrip, O. D. Jurchescu, D. J. Gundlach, E. Bittle, Adv. Funct. 

Mater. 2020, 30, 1904576.
[21] H. Klauk, Adv. Electron. Mater. 2018, 4, 1700474.
[22] U.  Zschieschang, J. W.  Borchert, M.  Giorgio, M.  Caironi, 

F.  Letzkus, J. N.  Burghartz, U.  Waizmann, J.  Weis, S.  Ludwigs, 
H. Klauk, Adv. Funct. Mater. 2020, 30, 1903812.

[23] M.  Fahlman, S.  Fabiano, V.  Gueskine, D.  Simon, M.  Berggren, 
X. Crispin, Nat. Rev. Mater. 2019, 4, 627.

[24] N. Koch, J. Phys.: Condens. Matter 2008, 20, 184008.
[25] C. Liu, Y. Xu, Y.-Y. Noh, Mater. Today 2015, 18, 79.
[26] D. Natali, M. Caironi, Adv. Mater. 2012, 24, 1357.
[27] B. Stadlober, U. Haas, H. Gold, A. Haase, G.  Jakopic, G. Leising, 

N.  Koch, S.  Rentenberger, E.  Zojer, Adv. Funct. Mater. 2007, 17, 
2687.

[28] D. Braga, M. Ha, W. Xie, C. D. Frisbie, Appl. Phys. Lett. 2010, 97, 
193311.

[29] T. Matsumoto, W. Ou-Yang, K. Miyake, T. Uemura, J. Takeya, Org. 
Electron. 2013, 14, 2590.

[30] M. Kitamura, Oyo Buturi 2017, 86, 122.
[31] A.  Yamamura, S.  Watanabe, M.  Uno, M.  Mitani, C.  Mitsui, 

J. Tsurumi, N. Isahaya, Y. Kanaoka, T. Okamoto, J. Takeya, Sci. Adv. 
2018, 4, eaao5758.

[32] J.  Lenz, F.  del  Giudice, F. R.  Geisenhof, F.  Winterer, R. T.  Weitz, 
Nat. Nanotechnol. 2019, 14, 579.

[33] J. W. Borchert, B. Peng, F. Letzkus, J. N. Burghartz, P. K. L. Chan, 
K. Zojer, S. Ludwigs, H. Klauk, Nat. Commun. 2019, 10, 1119.

[34] J. W. Borchert, U. Zschieschang, F. Letzkus, M. Giorgio, R. T. Weitz, 
M. Caironi, J. N. Burghartz, S. Ludwigs, H. Klauk, Sci. Adv. 2020, 6, 
eaaz5156.

[35] B. Peng, K. Cao, A. H. Y. Lau, M. Chen, Y. Lu, P. K. L. Chan, Adv. 
Mater. 2020, 32, 2002281.

[36] T.  Sawada, A.  Yamamura, M.  Sasaki, K.  Takahira, T.  Okamoto, 
S. Watanabe, J. Takeya, Nat. Commun. 2020, 11, 4839.

[37] A.  Yamamura, T.  Sakon, K.  Takahira, T.  Wakimoto, M.  Sasaki, 
T.  Okamoto, S.  Watanabe, J.  Takeya, Adv. Funct. Mater. 2020, 30, 
1909501.

[38] N.  Münzenrieder, G. A.  Salvatore, L.  Petti, C.  Zysset, L.  Büthe, 
C.  Vogt, G.  Cantarella, G.  Tröster, Appl. Phys. Lett. 2014, 105, 
263504.

[39] S.  Knobelspies, A.  Takabayashi, A.  Daus, G.  Cantarella, 
N. Münzenrieder, G. Tröster, Solid-State Electron. 2018, 150, 23.

[40] D. Jena, K. Banerjee, G. H. Xing, Nat. Mater. 2014, 13, 1076.
[41] A. Allain, J. Kang, K. Banerjee, A. Kis, Nat. Mater. 2015, 14, 1195.
[42] P.-C.  Shen, C.  Su, Y.  Lin, A.-S.  Chou, C.-C.  Cheng, J.-H.  Park, 

M.-H.  Chiu, A.-Y.  Lu, H.-L.  Tang, M. M.  Tavakoli, G.  Pitner, X.  Ji, 
Z. Cai, N. Mao, J. Wang, V. Tung, J. Li, J. Bokor, A. Zettl, C.-I. Wu, 
T. Palacios, L.-J. Li, J