https://www.pnas.org/doi/10.1073/pnas.2218380120

(機器翻譯，略譯稿)_
在之前的試驗中，用木頭製成的晶體管只能調節離子傳輸。當離子耗盡時，晶體管停止工
作。然而，來至瑞典林雪坪的研究人員開發的晶體管可以持續工作並調節電流而不會惡化。
研究人員使用輕木來製造他們的晶體管，因為所涉及的技術需要一種結構均勻的無紋木材
。他們去除了木質素，只留下長的纖維素纖維，其通道位於木質素所在的位置。
然後，這些通道被稱為 PEDOT:PSS 的導電塑料或聚合物填充，從而形成導電的木質材料
。伊薩克·恩奎斯特。
研究人員用它來製造木質晶體管，並可以證明它能夠調節電流並在選定的輸出水平上提供
連續功能。它還可以打開和關閉電源，儘管有一定的延遲——關閉電源大約需要一秒鐘；
上，大約五秒鐘。
電子設備監管
可能的應用可能包括調節電子設備，這是林雪坪大學另一個強大的研究領域。晶體管溝道
如此之大的一個優點是它可能比普通有機晶體管承受更高的電流，這對於某些未來應用可
能很重要。
Electrical current modulation in wood electrochemical transistor
Van Chinh Tran https://orcid.org/0000-0003-0122-4914, Gabriella G.
Mastantuoni, Marzieh Zabihipour, +4, and Isak Engquist
[email protected] Info & Affiliations
Edited by Peter Fratzl, Max-Planck-Institut fur Kolloid und
Grenzflachenforschung, Potsdam, Germany; received October 31, 2022; accepted
March 1, 2023 by Editorial Board Member Joanna Aizenberg
April 24, 2023
https://doi.org/10.1073/pnas.2218380120
Significance
Abstract
1. Results and Discussion
2. Conclusions
3. Experimental Section
4. Characterization
Significance
The orthotropic 3D microstructure has recently promoted wood as a template
for applications in wood-based energy and electronic devices. Different
varieties of electroconductive wood are widely reported; however, modulating
the wood’s electrical conductivity without changing its chemical composition
has not been done and remains challenging. In this work, we present an
approach to preparing conductive wood (CW), in which the electrical
conductivity can be modulated using an external potential. This has resulted
in a transistor where all three terminals are made of conductive wood and
which can be operated continuously at the selected conductivity without being
limited by, e.g., saturation effects. We expect this device and concept will
be a stepping stone for the development of wood-based electrical components.
意義
正交各向異性 3D 微結構最近將木材作為模板推廣到木基能源和電子設備中。不同品種的
導電木被廣泛報導；然而，在不改變其化學成分的情況下調節木材的導電性尚未完成，並
且仍然具有挑戰性。在這項工作中，我們提出了一種製備導電木材 (CW) 的方法，其中導
電率可以使用外部電位進行調製。這導致晶體管的所有三個端子均由導電木材製成，並且
可以在選定的導電率下連續工作，而不受例如飽和效應的限制。我們預計該設備和概念將
成為開發木質電子元件的墊腳石。
Abstract
The nature of mass transport in plants has recently inspired the development
of low-cost and sustainable wood-based electronics. Herein, we report a wood
electrochemical transistor (WECT) where all three electrodes are fully made
of conductive wood (CW). The CW is prepared using a two-step strategy of wood
delignification followed by wood amalgamation with a mixed electron-ion
conducting polymer, poly(3,4-ethylenedioxythiophene)–polystyrene sulfonate
(PEDOT:PSS). The modified wood has an electrical conductivity of up to 69 Sm1 induced by the formation of PEDOT:PSS microstructures inside the wood 3D
scaffold. CW is then used to fabricate the WECT, which is capable of
modulating an electrical current in a porous and thick transistor channel (1
mm) with an on/off ratio of 50. The device shows a good response to gate
voltage modulation and exhibits dynamic switching properties similar to those
of an organic electrochemical transistor. This wood-based device and the
proposed working principle demonstrate the possibility to incorporate active
electronic functionality into the wood, suggesting different types of
bio-based electronic devices.
Sign up for PNAS alerts.
抽象的
植物中質量運輸的性質最近激發了低成本和可持續的木基電子產品的發展。在此，我們報
告了一種木材電化學晶體管 (WECT)，其中所有三個電極都完全由導電木材 (CW) 製成。
CW 是使用木材脫木素的兩步策略製備的，然後將木材與混合電子離子導電聚合物聚（
3,4-乙烯二氧噻吩）- 聚苯乙烯磺酸鹽（PEDOT：PSS）混合。改性木材的導電率高達 69
Sm -1由 PEDOT:PSS 微結構在木材 3D 支架內的形成引起。然後使用 CW 製造 WECT，它
能夠在開/關比為 50 的多孔厚晶體管溝道 (1 mm) 中調製電流。該器件對柵極電壓調製
具有良好的響應並表現出動態類似於有機電化學晶體管的開關特性。這種基於木材的設備
和擬議的工作原理證明了將活性電子功能整合到木材中的可能性，暗示了不同類型的生物
基電子設備。
Get alerts for new articles, or get an alert when an article is cited.
As we step into the era of green technologies, there will be an increasing
distinction between complex, nanoscale electronics on one hand and simple,
large-size or large-area electronics on the other hand, the latter with
special functionalities like biosensing, biointegration, biodegradability,
etc. Bio-based materials will be the underpinning for the development of
these functionalities. During the last decades, cellulose, lignin, conducting
polymers, and other organic or bio-based materials have emerged as potential
templates or active components in various electrical devices (1, 2). Among
these materials, wood stands out when it comes to materials that have the
potential for ion transport and regulation (3). Several studies have shown
that the unique three-dimensional (3D) microstructures of wood lumina are
ideally designed for mass transport in the electrodes of electrochemical
devices (1–4). As a result, wood has been carbonized or functionalized with
conductive materials for applications in supercapacitors, batteries, and
electrochromic screens (2–6). After removing lignin, the open pathways
generated along the wood fibers have been shown to provide promising ion
transport channels in nanofluidic devices (7). Besides that, wood was also
explored in other electrical systems such as triboelectric nanogenerators and
electrical magnetic shielding (8, 9). These achievements indicate that wood
has a huge potential for energy and electronic technology. Wood is
orthotropic, and the directionality can provide advantages for organic
transistor performance. However, among many reported devices using a wood
template, there is, to the best of our knowledge, yet no report of an
electrical transistor made of wood or even of electronically induced
modulation of conductivity in wood-based conductors.
In order to transform wood to be an active component in a conventional
transistor [semiconductor transistor or organic electrochemical transistor
(OECT)], tunable electrical conduction is required. The conductivity can be
induced by either wood carbonization or wood modification with conducting
polymers such as polyaniline, polypyrrole, or poly(3,4-ethylenedioxythiophene)
–polystyrenesulfonate (PEDOT:PSS) (5, 10, 11). However, modulation of the
conductivity of carbon conductors is not possible by electronic or
electrochemical measures, which rules out carbonized wood as a transistor
channel material. This leaves wood modified with conducting polymers, which
will be the material system of choice in the present study. Earlier attempts
for “wood-based transistors” include studies focused on utilizing wood and
its derivatives as a nonconductive substrate(11) for templating a
conventional transistor. In this direction, cellulose paper is a common
choice as it has good flexibility and in some cases also high transparency
(11, 12) Cellulose fibers also show potential as a structural component in
the electrolyte of an electrolyte gate transistor (12). Although cellulose
has been used for silicon-containing transistors (13), transistors not based
on silicon technology could in the future reduce electronic waste and be
biodegradable. In a recent attempt, Li et al. reported a wood-based membrane
that could regulate ion transport through modulation of an external voltage
(7). For this purpose, the authors have coated a gate electrode layer of
silver metal on the wood surface and applied the working principle of a
gating transistor for regulating the ion movement in the membrane. This
results in a kind of ionic transistor, where the authors have proved that
ionic current can be regulated at the nanoscale of the wood scaffold. Such a
device has interesting potential but will also be limited by the need to
transfer between electrical current and ionic current at the electrodes. The
anticipated buildup of ionic charge or electrochemical reaction products at
the interfaces will inhibit the possibility to operate the device in a steady
state for longer periods of time. For the prolonged operation of a
transistor, it is required to rely upon modulation of electrical conduction
(7). Therefore, we need an approach that might utilize the wood ionic
conductivity but also includes sufficient and tunable electrical conduction
in wood for the actual transistor channel. Building on recent progress in
creating PEDOT:PSS-based conductive wood (CW) (10), the principle of
operation for OECTs should be a suitable candidate that could provide such
tunability. The OECT builds upon a transistor channel that is electrically
conductive and that can be electrochemically modulated using ionic transport
between the transistor channel and a gate electrode (14, 15). Since both
these transport mechanisms can be realized within a wood template, a fully
wood-based OECT should be possible to construct. The device is not expected
to have high performance compared to conventional transistors as, in this
work, the primary focus is to prove the hypothesis that electrical
conductivity of the CW can be modulated by using an external potential. The
result of this attempt is a wood transistor, in which all electrodes are made
of CW. We believe this device concept will be a good example for encouraging
the use of earth-abundant and sustainable resources in specific electrical
applications.
As shown in Fig. 1, the wood electrochemical transistor (WECT) was made from
three pieces of CW. The CW was prepared using a two-step strategy of wood
delignification and wood amalgamation with PEDOT:PSS (Fig. 1A and SI
Appendix, Fig. S1). Hardwood balsa was selected owing to its high strength,
low density, and high permeability, as well as its relatively homogeneous
structure with less difference between earlywood and latewood regions
compared to softwood (10, 16). In a preliminary test, we observed that balsa
performed better than birch or ash in preparing high–CW electrodes (SI
Appendix, Fig. S2). The conducting polymer was selected due to its excellent
tunable electrical conductivity shown in numerous examples to provide a
successful OECT channel (14, 17, 18) for, e.g., neuromorphic signaling and
chemical sensing applications (14). The preparation method enables the
formation of a beneficial microstructure of PEDOT:PSS in the wood scaffold
(Fig. 1A). This microstructure not only provides electrical conductivity but
also leaves room to utilize the wood 3D architecture for ionic transport
(17). The CW, after that, was used to construct a WECT with an approach that
is adopted from the fabrication of a double-gate OECT.
Fig. 1.
The schematic diagrams of (A) conductive wood preparation and (B) the wood
electrochemical transistor fabrication processes. (C) From Left to Right:
Front view photograph of a WECT, front view of the WECT configuration, and
section of a wood fiber (conceptual view of transistor’s channel zoom-in)
showing PEDOT:PSS-coated cell wall and electrolyte-transporting lumen.
The assembly of a WECT is explained in Fig. 1B, where two pieces of CW
(longitudinal × tangential × radial = 30 mm × 5 mm × 1 mm) were used as
the bottom and the top gates (denoted as WECT-Gate). Another piece of CW
(longitudinal × tangential × radial = 30 mm × 2 mm× 1 mm) was used as the
transistor channel (WECT-Channel), while a cellulose/polyester lab tissue and
a gel–electrolyte mixture were used as separator and electrolyte,
respectively. The final WECT transistor is configured as a double-gate OECT,
in which the main operation process (reduction/oxidation of the conductive
polymer) happens at the microscopic scale in the lumina that form the
transistor channel, Fig. 1C.
1. Results and Discussion
To be employed as an active component in a transistor, the wood must have
sufficient electrical conductivity. In some studies of wood coated with
conducting polymers, it has been found that wood pretreatment methods
including wood delignification can enhance the conductivity (6, 10, 19, 20).
In this work, the balsa wood was delignified before being impregnated with
the PEDOT:PSS suspension.
The effect of wood delignification on the CW conductivity was examined by
varying the wood delignification time from 0.0 (native wood) to 10.0 h, with
results shown in Fig. 2A. We see that the CW-5.0h has the highest
conductivity (69.0 ± 9.0 Sm), while the CW-Native shows the lowest value
of 3.5 ± 1.0 Sm. In native balsa wood, liquid transport occurs mainly in
the lumina of vessels. With the removal of lignin, wood is expected to gain a
higher porosity in the cell walls, and the pits in the cell walls are open
(21), which makes the lumina of fibers and parenchyma cells available to
transport PEDOT:PSS suspension (22). In addition, diffusion pathways are also
opened up in the middle lamella and cell wall corners in the delignified wood
(DW) (23). As a result, the fiber lumina, which are the dominating structures
in balsa wood, can be coated with PEDOT:PSS. The improved PEDOT:PSS diffusion
results in a higher electrical conductivity in the wood. However, when the
delignification time is longer than 5 h, we discovered that the wood fiber
cells collapse as the softened cell walls fall onto each other, causing a “
compacting of the cellular structure.” This is evident in a reduced sample
thickness in the DW-7.5 h and DW-10.0 h samples (SI Appendix, Table S1). The
collapsing hinders efficient polymer diffusion in the structure (24), thus
leading to a drop in the conductivities from 69 ± 9.0 Sm of CW-5.0h to
17.0 ± 5.0 Sm of CW-7.5 h. There is a slight increase to 24.0 ± 8.0 Sm1 for CW-10.0 h, but we judge this to be due to natural sample variations,
and we focus on the large difference compared to the CW-5.0h sample. Based on
these results and differently from other reported studies (24, 25), we found
that there is in fact an optimum delignification time. This is due to an
optimum lignin content resulting in an optimum level of sample integrity.
Accordingly, we selected 5.0h as the optimal delignification duration, and
the corresponding CW (CW-5.0h) was used for the fabrication of the WECT. The
WECT-Channel and WECT-Gate were made of CW-5.0h in sizes of 30 mm × 2 mm ×
1 mm (longitudinal × tangential × radial) and 30 mm × 5 mm × 1 mm
(longitudinal × tangential × radial), respectively. Although having a
bigger sheet area, the WECT-Gate has a lower conductivity (30.0 ± 4.0 Sm)
than the WECT-Channel (69.0 ± 9.0 Sm). This probably relates to the
poorer access of PEDOT:PSS to the interior parts of the larger wood piece,
which has a 2.5 times larger cross-section. One plausible reason for this is
that the initial PEDOT:PSS adsorption in the outer parts of the sample may
partially block PEDOT:PSS diffusion into the inner parts of the sample. There
could also be a delignification gradient contributing in the same way, but we
believe this to be less plausible since the good permeability of the balsa
wood should ensure homogeneous delignification on the 1–radial mm scale.
However, we note there is a potential for improvement in future works, where
the effects of the samples’ geometry, the delignification gradient, and the
polymer infiltration gradient should be studied (26).
Fig. 2.
(A) The conductivity of CW samples fabricated using different delignification
times (Inset: four-point probe measurement setup). (B) Cyclic voltammetry of
the WECT-Channel and WECT-Gate samples. (C) Lignin content of Native and
DW-5.0h. (D) Illustration of the ionic conductivity measurement setup. (E and
F) are the cross-sectional images of CW-Native and CW-5.0h samples,
respectively. (E, I) and (F, I) in turn are the cross-sectional SEM images
taken at the middle of supercritically point-dried CW-Native and CW-5.0h
samples. (E, II) and (E, III) are the EDX elemental mapping of oxygen and
sulfur in the corresponding SEM images of CW-Native. (F, II) and (F, III) are
the EDX elemental mapping of the oxygen and sulfur in the corresponding SEM
image of CW-5.0h. (G and H) are the SAXS patterns of DW-5.0h and CW-5.0h,
respectively. (I) The analyzing results of 1D SAXS spectra of DW-5.0h and
CW-5.0h (Inset: the calculated figures of DW-5.0h and CW-5.0h).
In addition to the electrical conductivity, the electrochemical properties
including charge storage capacitance and ionic conductivity are important for
understanding the applicability of CW as active electrodes (Gate and Channel)
in an OECT. While the device operates, the 3D structure of CW is expected to
facilitate a sufficient charge accumulation, which in turn will play a key
role in switching the current passing through the CW-based device channel
(14).
As shown in Fig. 2B, both the WECT-Gate and the WECT-Channel show good
capacitive behavior with their CV curves assuming a slightly deviated
rectangular shape. The deviation is most probably caused by the redox
activity of small amounts of native lignin, remaining in the DW (DW-5.0h),
and therefore also present in the CW-5.0h sample (10). The lignin content was
determined using the TAPPI T222 om-02 method and amounts to 7.1 ± 0.1 wt%
which is significantly lower than the 24.9 ± 0.1 wt% in the native wood
(Fig. 2C). Both the WECT-Gate and the WECT-Channel show good capacitances of
55.0 ± 5.0 mF and 31.0 ± 4.0 mF, respectively, at the scan rate of 20 mV/s
(see the specific capacitances in the SI Appendix). The higher capacitance of
WECT-Gate is a consequence of its larger size (2.5 times the volume of
WECT-Channel). For the operation of OECTs, it is advantageous if the gate
electrode has a larger capacitance than the transistor channel(26–28). As an
additional observation, we note that for other electrochemical devices
including supercapacitors, the capacitance results suggest the CW-5.0h as a
potentially useful material (10). Along with the electrodes’ capacitances,
the ionic conductivity of the WECT-Channel was also studied to understand its
capability for ion-mediated electrochemical conductivity regulation when an
external voltage is applied. By using the measurement setup shown in Fig. 2D,
the recorded ionic resistance of the WECT-Channel (or CW-5.0h sample) is
lower than that of DW-5.0h. This implies that PEDOT:PSS has played an
important role in lowering the ionic resistance, which in turn means an
increase in the ionic conductivity within the CW scaffold. More detailed
results (SI Appendix, Fig. S3) and further discussion are presented in the SI
Appendix.
Structural and morphological characterization was performed to map and
understand the wood’s morphology and PEDOT:PSS distribution and is presented
in Fig. 2 E and F. By comparing CW-5.0h with CW-Native, we observed a clear
distinction in their appearance and microstructure. As seen in Fig. 2 E and
F, where cross-sections from the middle of each sample are shown, CW-5.0h
appears dark blue throughout its thickness, while the cross-section of
CW-Native reveals its native light brown color. This indicates that PEDOT:PSS
has penetrated the entire DW structure but was not able to access the inner
parts of the native (lignified) wood. Accordingly, at a microscopic scale as
investigated by scanning electron microscopy (SEM), there is no trace of
PEDOT:PSS in the cross-section of CW-Native (Fig. 2 E, I and III). In
contrast, a PEDOT:PSS layer was seen in the fibers’ lumens (Fig. 2 F, I and
SI Appendix, Fig. S4) and vessels’ lumens (SI Appendix, Fig. S5) of the
CW-5.0h. This observation was further confirmed by Energy-dispersive X-ray
analysis (EDX) elemental mapping images, in which sulfur was mostly observed
in the wood lumen (Fig. 2 F, III). It should be noted that the PEDOT:PSS thin
film in Fig. 2F was visualized by applying supercritical point drying instead
of air-drying (SI Appendix, Fig. S4) as the liquid CO2 causes detachment of
the polymer from the wood cell wall, making it clearly visible. In the
pristine samples used in WECTs, the PEDOT:PSS film only coats the inner
surface of the lumens (10), leaving the central section open for electrolyte
transport. A similar coating phenomenon is expected to happen at the ray
cells, contributing to transverse electrical transport and helping to create
3D electrical interconnection in the wood structure. With the observed
microstructure thus formed, conceptually illustrated in Fig. 1A, PEDOT:PSS is
expected to promote both the electron and ion transport through the 3D
structure of CW-5.0h and in the WECT-Channel.
In addition to the structure observed in SEM images, the small-angle X-ray
scattering (SAXS) measurements shown in Fig. 2 present another insight into
the distribution of PEDOT:PSS in the wood cell wall of CW-5.0h. The
two-dimensional SAXS patterns of DW-5.0h and CW-5.0h are presented in Fig. 2
G and H). Anisotropic streaks were observed, but they only show small
differences between the two samples. We further analyzed the one-dimensional
(1D) data (Fig. 2I) to determine the correlation length, which means the
center-to-center distance of the cellulose fibrils. Our calculation suggests
that CW-5.0h has a slightly larger correlation length (29) than that of
DW-5.0h (.0 ± 0.10 nm compared to .55 ± 0.15 nm) which could be an
indication of penetration of PEDOT:PSS polymer chains in between the CW-5.0h
wood fibers. A control measurement indicates that DMSO only has a minor
contribution to this effect (cf. SI Appendix, Fig. S6). Although we are
unable to quantify the amount of PEDOT:PSS inside the wood cell wall, this
result suggests that some of the polymer is localized there; however, the
amount should be small compared to the amount covering the cell wall. The
presence also has some effects on the wood fibers’ arrangement, which was
studied by wide-angle X-ray scattering (WAXS) and presented in SI Appendix,
Fig. S7. The CW-5.0h sample has a smaller Herman’s orientation factor of the
200 crystal plane compared to DW-5.0h, which means the alignment of cellulose
fibrils is disturbed, probably as a result of the partial impregnation with
PEDOT:PSS. To further investigate the interaction between wood fibers and
PEDOT:PSS, ATR-FTIR measurements were carried out for both CW-5.0h and
DW-5.0h (SI Appendix, Fig. S8). The discussion on FTIR results (SI Appendix)
suggests that PEDOT:PSS has interactions with the wood fibers, which
facilitated the amalgamation of these components in the composite of CW-5.0h
(30–32).
Finally, mechanical properties (tensile strength and Young’s modulus) were
measured for native, delignified (DW-5.0h), and CW (CW-5.0h) and reveal that
the CW is similar in tensile strength and stiffness to the original balsa
wood (cf. SI Appendix, Fig. S9). Taking all the collected evidence of
electrical/electrochemical and structural properties into account, CW-5.0h
was selected as a good candidate for forming electrodes for the wood
transistor.
1.1. WECT.
OECTs can be constructed with a single gate electrode or with double gates on
either side of the transistor channel, as shown schematically in SI Appendix,
Fig. S10 A and B, respectively. A double-gate configuration is beneficial
when the transistor channel dimensions are large since it provides better and
faster access for ion transport to all parts of the transistor channel. Here,
for a WECT device with a 1-mm-thick channel, the double-gate structure would
thus be advantageous. This is experimentally proven by comparing the
switching performance of both configurations (cf. SI Appendix, Fig. S10 C and
D). In view of this result, double-gate transistors were selected as the
standard configuration for further investigations.
As illustrated in Fig. 1B, a double-gate WECT is structured with the two gate
electrodes positioned on the top and bottom sides of the transistor channel.
Both the WECT-Gate electrodes and the WECT-Channel are made from 1-mm-thick
CW-5.0h. Although 1 mm is much thicker than the ordinary thickness of a
conventional PEDOT:PSS-based OECT [less than 1 μm (26, 27, 33)], still the
device operates like an ordinary p-type OECT. The current passing through the
WECT-Channel is defined as the drain–source current (ID). At zero gate
voltage (VG), the transistor channel is open, and the transistor is ON,
whereas by applying a gate voltage of 6.0 V, the channel becomes fully
reduced, and the transistor is in the OFF state. Fig. 3B shows the transfer
curves of the device in which the ON/OFF [ID(VG = 0)/ID(VG = 6.0 V)] current
modulation reaches 1.7 orders of magnitude (50 times) for the forward sweep.
In comparison with the ON/OFF ratio (hundreds to 105) of conventional
PEDOT:PSS-based OECTs (14, 26, 33), the ratio of 50 is small but reasonable
for a transistor with a combination of high electrode thickness and limited
conductivity. In the same figure, it is observed that the transistor is
switched off when VG reaches .5 V. The switching process is repeatable,
which is shown in three consecutive switching runs presented in SI Appendix,
Fig. S11.
Fig. 3.
Wood electrochemical transistor (WECT). (A) The speculated operation
mechanism with a focus on part of a single wood fiber with the cell wall and
lumen, (B) the transfer curve, (C) the output curves at different gate
voltages, and (D) the dynamic switching characteristics at the frequency of
100 mHz. (Note: Each measurement was carried out on different devices.)
A tentative operation mechanism at a microscopic level of the WECT is
illustrated in Fig. 3A which is based on the switching performance and the
WECT-Channel’s morphology observed in Figs. 1C and 2. A section of one wood
fiber was selected to describe the working mechanism as it could represent
the current modulation principle in the entire WECT-Channel (10). Before
applying any potential to the gate electrodes (VG = 0 V), the WECT is in its
ON state. Upon applying VG > 0, the WECT is gradually switched to the OFF
state due to the electrolyte cations being driven out from the electrolyte
toward the wood cell wall surface where PEDOT:PSS is mainly localized and
held at negative potential. Here, an electrochemical reaction takes place
where the cations compensate the counter anions (PSS, and the PEDOT+ is
reduced to its nonconductive form PEDOT0 (15). As a result, the conductivity
of the WECT-Channel is decreased.
Output measurements are carried out to provide additional information about
the performance of WECTs. Here, the drain–source voltage (VD) was swept from
0 to .0 V, while VG was increased in steps from 0 to 6.0 V. In Fig. 3C,
the obtained ID–VD curves are shown. At VG = 0 V and VG = 1 V, we see that
ID increases linearly with increasing VD and reaches a plateau at around VD =
.5 V. After that, ID undergoes a slight decrease when the VD level is
increased up to .0 V. The decrease is probably related to either a slight
reduction of PEDOT:PSS at such a high voltage range or a small current
leakage between the drain and the gate. In the output curves where VG is
stepped to values higher than .0 V, the transistor is in the OFF mode, and
such behavior is not observed. This is fully in agreement with the
characteristics of ordinary OECTs.
Although the WECT is expected to be slow, we also examined the dynamic
switching capability of the device. In dynamic switching measurements, a
function generator was sourcing VG as a square wave switching between 0 and
5.0 V at 100 mHz, while a constant VD of .0 V was applied. From the graph
shown in Fig. 3D, we see that repeated dynamic switching is fully possible,
although not fast. The main part of the OFF switching happens in around 1 s,
which has to be considered as good under the circumstances (a 1-mm-thick
transistor channel). The ON switching is slower and is not fully completed in
5 s, which is probably the reason why we see a decreasing ON/OFF ratio as the
measurement in Fig. 3D progresses. For full ON/OFF dynamics, a lower
frequency than 100 mHz would have to be used. With these dynamic switching
properties, the WECT is not suitable for conventional electronic circuits but
is probably an interesting candidate for wood-integrated applications ranging
from electrochromic displays to simple logic circuits responding to sensor
input.
2. Conclusions
A transistor made of CW was successfully demonstrated. This result proves
that it is possible to modulate the electrical conductivity of the
electroactive wood by applying an external voltage. The WECT operates
according to the same principle as a double-gate OECT, where the two gates
and the transistor channel are made of delignified wood, made conductive (69
Sm) by the formation of a PEDOT:PSS layer in the lumina of the wood
structure, in particular the fiber lumina. The current modulation occurs
through electrochemical oxidation/reduction of PEDOT, with a measured ON/OFF
ratio of up to 50 times. Although the device performance is poorer than the
common PEDOT:PSS-based OECT, the WECT proves the principle and shows that
there is a possibility to transform wood into a functional transistor by
utilizing its oriented and hierarchical 3D structure, thereby introducing the
possibility to control and regulate the electronic current in CW. We also
believe there are possibilities for improvement by either optimizing the
conductivity of wood or manipulating the device configuration. Since the
mechanical stability of the transistor electrodes is as good as the original
balsa wood, strong and self-supporting devices could be readily constructed.
In view of the large interest in exploratory research concerning
bioelectronics, bio-based electronics, and plant electronics, this device and
its working principle might be a stepping stone toward different applications
in those fields.
3. Experimental Section
3.1. Materials.
Balsa (Ochroma pyramidale) veneers with an oven-dried density of ~0.22 g cm3 were purchased from Material AB (Sweden). Sodium chlorite (NaClO2, 80%),
sodium chloride (NaCl, 99%), and dimethyl sulfoxide (DMSO, 99%) were received
from Sigma Aldrich. PEDOT:PSS (Clevios PH1000, water suspension with %
solid content) was purchased from Heraeus, Germany. Carbon fibers, paraffin
wax, silver paste, and carbon paste were purchased from Sigma Aldrich and
used as received. Blue gel (250 g) was purchased from Cefar-Complex, Sweden.
3.2. Wood Delignification.
Balsa wood veneers were cut in the size of 30 mm × 10 mm × 1 mm
(longitudinal × tangential × radial). The veneers were delignified at 80 °
C in a NaClO2 (1.0 wt%) solution in acetate buffer for different reaction
times: 2.5 h, 5.0 h, 7.5 h, or 10.0 h. The obtained DW samples were
correspondingly named DW-2.5h, DW-5.0h, DW-7.5h, and DW-10.0h.
3.3. CW Preparation.
The DW samples were dried under an ambient atmosphere and cut into smaller
pieces. Samples with dimensions of 30 mm × 2 mm × 1 mm (longitudinal ×
tangential × radial) were used to prepare the CW-based transistor channel,
while samples having a dimension of 30 mm × 5 mm × 1 mm (longitudinal ×
tangential × radial) were used to prepare the CW–based gate of the WECT.
These DW pieces were thereafter impregnated in a PEDOT:PSS suspension (100 g
of PEDOT:PSS suspension mixed with 6 g of DMSO) before being oven-dried at 75
°C to achieve the CW (see SI Appendix, Fig. S1 for a visual diagram). The
final CW products were obtained after mechanically removing all the
aggregated polymer layers on the surface of the dried samples. Corresponding
with the DWs, the obtained CWs were named CW-2.5h, CW-5.0h, CW-7.5h, and
CW-10.0h, respectively. The PEDOT:PSS-coated native wood (CW-Native) was
prepared in a similar approach, in which native wood was cut in a specific
size before being impregnated in the same PEDOT:PSS suspension to achieve the
CW. All CWs used as the WECT-Channel and the WECT-Gate were dried and stored
in a controlled environment before being used for the device fabrication and
measurement.
3.4. WECT Fabrication.
The fabrication of a double-gate WECT is presented in Fig. 1B, where the two
gate electrodes are set perpendicularly to the channel electrode. One gate is
on the top surface and the other one is placed under the bottom surface of
the transistor channel. The channel and gates are separated by a
cellulose-based tissue paper before dropping the electrolyte mixture on the
crossing area of the electrodes (Fig. 1B). The electrolyte is prepared by
mixing 2 mL blue gel (Blågel, Cefar-Complex) with 1 mL NaCl (1 M) and kept
for 2 weeks before use.
4. Characterization
4.1. Morphology and Chemical Composition.
SEM/EDX: The morphology of the wood samples was analyzed by field emission
SEM (Hitachi S-4800, Japan) at a low acceleration voltage of 1 kV. The
samples were microtomed, dried either in ambient conditions or under
supercritical CO2, and coated with a platinum/palladium conductive layer
using a sputter coater (Cressington 208HR, UK). EDX was performed at an
acceleration voltage of 6 kV with an Oxford Instruments, X-MAX N 80, UK.
Lignin and Monosaccharides Content: Klason lignin content was determined by
acid hydrolysis according to the TAPPI T222 om-02 method. The samples were
analyzed in duplicates. Quantification of the neutral sugars was performed on
Dionex ICS-3000 high-performance ion-exchange chromatography (Thermo Fisher
Scientific Inc.) after acid hydrolysis. The samples were analyzed in
duplicates, and anhydrous factors were used for the monosaccharides (0.88 for
xylose and arabinose and 0.90 for glucose, mannose, and galactose). Meier’s
correlations were used to calculate the weight percentage of cellulose and
hemicellulose.
Leaching experiments were carried out for the WECT-Gate and WECT-Channel by
soaking the CW samples in deionized water for 4 d. The results indicate no
significant leaching and are presented in the Supporting Information (SI
Appendix, Fig. S12).
4.2. Conductivity Measurement.
The CW was glued on top of 4 chromium/gold electrodes by carbon paste, and
the resistance was measured using a 4-probe technique. A Keithley 2400 source
meter was used to supply the current to the two outer electrodes and to
measure the voltage between the two inner electrodes.
Based on the obtained resistance, the electrical conductivity (σ) of CW
samples is calculated using the following equation (10):
σ=1ρ=LRA,
[1]
where R is the obtained resistance, L is the distance between the two inner
electrodes, and A is the cross-sectional area of the specimen.
The sheet resistance of CW samples was calculated using the following
equation:
Rs=1σt,
[2]
where Rs is the sheet resistance, and t is the thickness of the sample.
4.3. Electrochemical Measurement.
The CW electrode for electrochemical measurement was prepared following the
same procedure as in our previous work. The CW is first connected to carbon
fibers using carbon paste before wrapping a part of the carbon fiber with
paraffin wax and Kapton tape, respectively. The electrochemical measurement
was performed in a three-electrode system configuration using a
potentiostat/galvanostat (by BioLogic, SP-200) coupled to a computer. The
capacitance of samples was calculated using the formula (34): C=1vΔV∫
V2V1idV=A2×k×ΔV
where i
is the charge/discharge current (A), A is the integral area of the CV curve,
k is the scan rate (mV/s), and ΔV is the working potential of the discharge
process.
4.4. Wood OECT Characterization.
Transfer (drain current vs. gate voltage), output (drain current vs. drain
voltage), and dynamic switching (drain current vs. time) measurements were
conducted using a semiconductor parameter analyzer (HP/Agilent 4155B) and a
function generator (Agilent 33120 A).
4.5. SAXS and WAXS Measurements.
SAXS and WAXS measurements were performed on a point-collimated Anton Paar’s
SAXS point 2.0 system equipped with a Cu Kα radiation source (wavelength
1.5418 Å and beam size of ~500 μm) and an Eiger R 1M detector with 75 × 75
μm pixel size (at RISE, Sweden). The sample-to-detector distance was set to
576 mm and 111 mm for SAXS and WAXS, respectively. The exposure time of each
measurement is 10 min, and they were performed at room temperature with a
beam path pressure of about 1 to 2 mbar. The data reduction was performed by
using SAXS analysis software (Anton Paar, Graz, Austria).
Please refer to the additional SAXS measurement, which was carried out at the
CERMAV-CNRS (France), in the SI Appendix.