On the use of the term “ambipolar”

James G. Champlain

Citation: Applied Physics Letters 99, 123502 (2011); doi: 10.1063/1.3641898

View online: http://dx.doi.org/10.1063/1.3641898
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/99/12?ver=pdfcov
Published by the AIP Publishing

Articles you may be interested in

Direct extraction of carrier mobility in graphene field-effect transistor using current-voltage and capacitance-
voltage measurements
Appl. Phys. Lett. 101, 213103 (2012); 10.1063/1.4768690

Long-term retention in organic ferroelectric-graphene memories

Appl. Phys. Lett. 100, 023507 (2012); 10.1063/1.3676055

Characteristics of high-k Al 2 O 3 dielectric using ozone-based atomic layer deposition for dual-gated graphene
devices
Appl. Phys. Lett. 97, 043107 (2010); 10.1063/1.3467454

Tunable ambipolar Coulomb blockade characteristics in carbon nanotubes-gated carbon nanotube field-effect
transistors
Appl. Phys. Lett. 94, 022101 (2009); 10.1063/1.3065067

Ultrahigh-density nanotransistors by using selectively grown vertical carbon nanotubes

Appl. Phys. Lett. 79, 3696 (2001); 10.1063/1.1419236

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Tue, 02 Dec 2014 16:49:41
 APPLIED PHYSICS LETTERS 99, 123502 (2011)

On the use of the term “ambipolar”

James G. Champlaina)
Naval Research Laboratory, Electronics Science and Technology Division, Washington, D.C. 20375-5347,
USA
(Received 21 July 2011; accepted 31 August 2011; published online 21 September 2011)
The term ambipolar has been used extensively in association with carbon nanotube and graphene-based
field effect transistors, often in a varied manner, leading to a confused understanding of the term.
Through the use of established scientific definitions and theoretical work on device operation, this paper
attempts to clarify the understanding of the term and present a discussion of its appropriate use.
[doi:10.1063/1.3641898]

With the recent increase of interest and research in the electrons and holes, as a single entity with a single effective
area of carbon nanotubes (CNTs), graphene, and their field (ambipolar) mobility and effective (ambipolar) diffusion
effect transistors (FETs), there has been an associated constant, is ambipolar transport.
increase in the use of the term ambipolar. The term has been There are two key features of ambipolar transport to
used to describe various aspects of device and electronics note: ambipolar transport requires (local) charge neutrality,
research from physical and device phenomena (e.g., polarity the semiconductor must be charge neutral for the entire space
of the material1,2 and type of charge transport and current considered at every point (not to be confused with neutrality
conduction3–8) to more nebulous subject matter (e.g., of an overall system, such as the gate-to-channel region of a
“ambipolar property of graphene,”8 “ambipolar behavior,”9 FET), and ambipolar transport describes only the motion of
and “ambipolar nonlinear electronics”10). This varied and of- excess charge carrier concentrations, not equilibrium concen-
ten ambiguous usage of the term begs the question, “What trations. The latter is illustrated in the previous example: the
does ambipolar mean?” This paper attempts to answer this
question through the use of established scientific phenomena
and theoretical work on device operation.
Our discussion begins with examining long-accepted
definitions that exist associated with ambipolar. Reading the
literature, one will find that there are only two clearly defined
terms in solid-state physics and devices involving ambipolar.
One is the well-established phenomenon, ambipolar
transport;11–13 the other is the device concept referred to as
the ambipolar FET.14–16
Ambipolar transport is a specific solid-state physical
phenomenon that describes the motion of excess electronic
charge carriers in a (quasi-)neutral region of semiconduc-
tor.17 As an illustration, we examine a slab of n-type semi-
conductor (Fig. 1). At equilibrium, the semiconductor is
charge neutral ðq=q ¼ po þ NDþ  no ¼ 0Þ. Applying a small
electric field (shown here by applying a small potential
across the length of the slab, Ex ¼ Vx/L), such that the semi-
conductor remains quasi-neutral, a drift current is induced
proportional to the majority electron equilibrium concentra-
tion (no) moving at a velocity dictated by the strength of the
electric field and mobility of the electrons (Jno ¼ qnolnEx,
Fig. 1(a)). Next, equal concentrations of excess electrons and
holes (Dn ¼ Dp) are introduced in a defined spatial region
(via generation from light for example, Fig. 1(b)). The semi-
conductor remains charge neutral, p þ NDþ  n ¼ po FIG. 1. (Color online) Illustration of ambipolar transport. Top of figure, a
bulk slab of n-type semiconductor with a potential (Vx) applied across its
þDp þ NDþ  no  Dn ¼ po þ NDþ  no ¼ 0. As a result of length (L). (a) Steady-state conditions prior to excess charge injection. (b)
space-charge conservation, the excess electrons and holes The initial excess charge injection, illustrated by electron-hole-pair
will move as a single entity in a direction essentially dictated (Dn ¼ Dp) generation via photon absorption. (c) The subsequent motion of
by the minority carrier; in this instance, the general direction the excess charge by ambipolar transport. (i) Band diagram and applied elec-
tric field in each case; (ii) electron and hole carrier concentrations
of the applied electric field (Fig. 1(c)). This motion of excess (n ¼ no þ Dn and p ¼ po þ Dp, respectively); (iii) majority equilibrium elec-
tron and excess electron and hole fluxes (Uno, UnA, UpA ¼ UnA, respectively)
and currents (Jno ¼ (q)Uno, JnA ¼ (q)UnA, JpA ¼ (þq)UpA ¼ JnA, respec-
a)
Electronic mail: [email protected]. tively); and (iv) the total (net) current (JT ¼ Jno þ JnA þ JpA ¼ Jno).

0003-6951/2011/99(12)/123502/3/$30.00 99, 123502-1

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Tue, 02 Dec 2014 16:49:41
 123502-2 J. G. Champlain Appl. Phys. Lett. 99, 123502 (2011)

equilibrium electron (majority) concentration (no) drifts in a nowhere in Pfleiderer’s articles is it invoked in establishing
direction opposite to the applied electric field, as expected; this concept.
whereas the excess electron (Dn) and hole (Dp) concentra- There are two key points in this definition for the ambipo-
tions drift in a direction with the applied electric field, as dic- lar FET. First, the device can be switched between n- and p-
tated by ambipolar transport (Figs. 1(b) and 1(c)). channels. This mode of operation is not unique to so-called
A direct effect of ambipolar transport is that in any two “ambipolar materials” or devices. Si-MOSFETs that comprise
terminal device or three terminal device with an insulated the CPU in a common desktop computer switch between n-
third terminal (e.g., an insulated gate FET), the net current and p-channels (i.e., accumulation-inversion and vice-versa)
associated with excess charge is zero. This is also illustrated billions of times a second. Additionally, all semiconductor
in the presented example with the fluxes of the excess electron FETs, from graphene to GaN, can theoretically be modulated
and hole concentrations being equal in proportion and direc- between n- and p-channel operation, with practical demonstra-
tion (UnA ¼ UpA), leading to equal but opposing electrical cur- tions shown in semiconductors

with bandgaps up to that of
rents (JnA ¼ JpA; Figs. 1(b) and 1(c)). Summing all currents GaAs E g ¼ 1:4 eV .18 Clearly, this is not a material-centric
in the system, the ambipolar electron and hole currents cancel, quality. Second, the requirement that the contact is ohmic to
leaving only the drift current of the equilibrium electron con- both “majority and minority carriers,” as Pfleiderer writes.
centration (JT ¼ Jno þ JnA þ JpA ¼ Jno; Figs. 1(b) and 1(c)). What he perhaps means here is that the contact is ohmic to
This leads one to ask, “Can ambipolar transport be the both electrons and holes, whose roles as majority and minority
mechanism for charge transport leading to current conduc- carrier switches depending on the polarity of the channel. This
tion in the channel of a FET?” As just stated, in any three is much more a practical consideration of the device than a
terminal device with an insulated third terminal (e.g., CNT- physical property of the semiconductor.
and graphene-based FETs), if ambipolar transport were to For example, if one were to take a Si-MOSFET (n- or p-
occur, it would lead to zero net current. Therefore, ambipolar channel) and place pþ wells next to nþ wells, shorted to-
transport cannot be considered the mechanism for charge gether via a metal contact, to form both the source and drain
transport responsible for current in these devices. The answer contacts, one would arguably produce an ambipolar FET, as
to this question can be alternatively viewed by considering defined by Pfleiderer. Alternatively, in the antimonide (Sb)-
the modulation of mobile charge in the channel. For ambipo- based material system, simply placing metal on a narrow
lar transport to occur, the total charge in the channel alone bandgap semiconductor results in an ohmic contact to both
must sum to zero to maintain (local) space-charge neutrality n- and p-type material, albeit with different contact resistan-
(i.e., qchannel ¼ 0). Enforcing this condition results in the ces,19 and these materials have shown inversion in MOS-like
channel being “invisible” to any applied gate field structures;20 making an ambipolar FET also possible in the
ðr  D ¼ qchannel ¼ 0Þ. In other words, the gate and channel Sb-based material system.
would not communicate, there would be no modulation of Ultimately, the ambipolar FET, as defined by Pfleiderer,
charge (no field effect), and, therefore, no FET. is not a description of the physical phenomena within the
The conclusion is that ambipolar transport cannot be device; it simply describes a device concept, based on
considered the mechanism for charge transport and, subse- well-established physics readily observed in nearly any insu-
quently, current conduction in the channel of a FET, includ- lated gate FET (i.e., the accumulation of carriers and inver-
ing CNT- and graphene-based FETs. Therefore, statements sion of channel type). With appropriately fabricated source-
that ambipolar transport (or conduction) has been observed drain contacts, any insulated gate FET could be termed an
in these devices or responsible for the observed current are ambipolar FET.
incorrect.3–5,7,8 Having established that the ambipolar FET is solely the
We now turn to the second definition, the ambipolar FET. description of a device concept (e.g., modes of operation,
The ambipolar FET was defined by H. Pfleiderer in the mid- contact characteristics) and that there is no connection
1980s.14–16 An excerpt from one of his articles, between the operation of these devices and ambipolar trans-
port, our discussion turns toward understanding the mecha-
The usual field-effect transistor (FET) with insulat- nism for charge transport in these devices. It should be
ing gate is known as a “unipolar” device with either stated that the purpose of this paper is not to introduce a
an n-channel or a p-channel. The purpose of this new naming convention for these devices. The author
paper is to present another FET concept. A thin- accepts that the term ambipolar FET has been used to such
film transistor (TFT) with source and drain contacts an extent that it has been established as describing a cate-
which are ohmic for majority and minority carriers gory of FET operation. However, the term ambipolar FET
[and] can be switched between n-channel and lends no insight into the physical phenomena or processes
p-channel modes. This kind of device is termed an in these device.
“ambipolar” FET.16 What we know of these devices is that the physical proc-
esses within them involve both electrons and holes, which
In these articles, Pfleiderer was not considering CNTs or upon examination leads to the associated transport of these
graphene as the channel material, but instead amorphous sili- charge carriers necessarily being bipolar transport. From the-
con (a-Si). One should notice that a physical phenomenon is oretical work on graphene-based FETs, the transfer charac-
not being defined here, per se; it is solely the description of a teristic of a graphene-based FET consists of two unipolar
device concept based on modes of operation. Additionally, modes of operation (n- and p-channels) with a mixed or
this definition does not mention ambipolar transport, and bipolar channel mode separating the two (Fig. 2).21,22 Here,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Tue, 02 Dec 2014 16:49:41
 123502-3 J. G. Champlain Appl. Phys. Lett. 99, 123502 (2011)

cinct and clear communication of ideas and concepts. Sec-

ond, the phenomena within these devices can be described
with already established physics. By adding an overlapping
and redundant definition to understood phenomena simply
confuses the issue, which appears to have become the case
with CNT- and graphene-based FETs.
In conclusion, as scientist and engineers we should
strive to understand and bring clarity to the behavior and
function of new materials and devices, and in doing so avoid
confusion and obscurity. This clearly necessitates an under-
standing of established phenomena. Occasionally, adopting
new terminology and interpretations of physical phenomena
is required. However, more often than not, this is not neces-
sary and “conventional” (i.e., well-established) physics are
more than sufficient for describing the observed phenomena.
This work was supported by the Office of Naval
Research.
1
I. Meric, M. Y. Han, A. F. Young, B. Ozyilmaz, P. Kim, and K. L. Shepard,
Nature Nanotechnol. 3, 654 (2008).
2
M. Freitag, C. Hsin-Ying, M. Steiner, V. Perebeinos, and P. Avouris,
Nature Nanotechnol. 5, 497 (2010).
3
R. Martel, V. Derycke, C. Lavoie, J. Appenzeller, K. K. Chan, J. Tersoff,
FIG. 2. (Color online) Plot of a transfer characteristic of a graphene-based and P. Avouris, Phys. Rev. Lett. 87, 256805 (2001).
4
FET with the total drain current (JD, solid line) and the (mean value along Y.-F. Chen and M. S. Fuhrer, Phys. Rev. Lett. 95, 236803 (2005).
5
the channel) electron and hole components of the total drain current (JD;n , S. A. Thiele, J. A. Schaefer, and F. Schwierz, J. Appl. Phys. 107, 094505
solid circles, and JD;p , open circles, respectively). Above the transfer charac- (2010).
6
teristic are plotted the local electron (n, solid line) and hole (p, dashed line) J. S. Moon, D. Curtis, S. Bui, M. Hu, D. K. Gaskill, J. L. Tedesco, P.
concentrations and local electron (JD,n, solid line) and hole (JD,p, dashed Asbeck, G. G. Jernigan, B. L. VanMil, R. L. Myers-Ward, C. R. Eddy, P.
line) currents along the channel for p-channel ðVGS  VDirac Þ, bipolar channel M. Campbell, and X. Weng, IEEE Electron Device Lett. 31, 260 (2010).
7
(here, at VGS ¼ VDirac), and n-channel ðVGS  VDirac Þ operation (Ref. 22). Y. M. Lin, H. Y. Chiu, K. A. Jenkins, D. B. Farmer, P. Avouris, and A.
Valdes-Garcia, IEEE Electron Device Letters 31, 68 (2010).
8
H. Wang, A. Hsu, J. Wu, J. Kong, and T. Palacios, IEEE Electron Device
Lett. 31, 906 (2010).
9
ambipolar transport has not been applied, instead the total A. Javey, M. Shim, and D. Hongjie, Appl. Phys. Lett. 80, 1064 (2002).
10
T. Palacios, A. Hsu, and H. Wang, IEEE Commun. Mag. 48, 122 (2010).
current in the channel is simply the sum of electron and hole 11
C. Herring, Bell Syst. Tech. J. 28, 401 (1949).
currents, JD(x) ¼ JD,n(x) þ JD,p(x). Of note is the striking sim- 12
J. R. Haynes and W. Shockley, Phys. Rev. 81, 835 (1951).
13
ilarity of the charge and current profiles under bipolar chan- W. van Roosbroeck, Phys. Rev. 91, 282 (1953).
14
nel operation to that of forward biased pn junctions.23 H. Pfleiderer, W. Kusian, and B. Bullemer, Siemens Forschungs- und
Entwicklungsberichte 14, 69 (1985).
We now turn to the impetus for the writing of this paper 15
H. Pfleiderer, W. Kusin, and B. Bullemer, Siemens Forschungs- und
and what seems to have occurred in the scientific and engi- Entwicklungsberichte 14, 114 (1985).
16
neering communities: the ambiguous naming of any phe- 17
H. Pfleiderer, IEEE Trans. Electron Devices ED-33, 145 (1986).
nomena that is associated with electrons and holes within Ambipolar transport was first examined in plasma physics, however we
are only concerning ourselves with the solid-state physics definition.
these materials and devices as ambipolar, without any scien- 18
M. Passlack, M. Hong, and J. P. Mannaerts, Solid-State Electron. 39, 1133
tific reasoning or foundation, other than perhaps the term (1996).
19
ambipolar FET. This practice, if not incorrect, is very mis- J. G. Champlain, R. Magno, and J. B. Boos, Electron. Lett. 43, 1315
(2007).
leading, as we have shown that the phenomena within these 20
A. Ali, H. S. Madan, A. P. Kirk, D. A. Zhao, D. A. Mourey, M. K. Hudait,
devices are not related to ambipolar transport. R. M. Wallace, T. N. Jackson, B. R. Bennett, J. B. Boos, and S. Datta,
Should we then expand the definition of ambipolar? The Appl. Phys. Lett. 97, 143502 (2010).
21
argument against this comes in two parts. First, if we were to M. G. Ancona, IEEE Trans. Electron Devices 57, 681 (2010).
22
J. G. Champlain, J. Appl. Phys. 109, 084515 (2011).
expand the definition, then the already defined phenomenon 23
The informed reader will realize that ambipolar transport occurs in for-
of ambipolar transport becomes obscured. This runs counter ward biased pn junctions; however, it results in zero net current and is not
to the purpose of using unambiguous terminology: the suc- the major charge transport mechanism within the device.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Tue, 02 Dec 2014 16:49:41