Carbon Energy - 2020 - Zhang - Carbon and Carbon Composites For Thermoelectric Applications

Received: 7 June 2020 | Revised: 24 June 2020 | Accepted: 27 June 2020
DOI: 10.1002/cey2.68
REVIEW
Carbon and carbon composites for thermoelectric

applications
Yichuan Zhang1 | Qichun Zhang2,3 | Guangming Chen1
1
College of Materials Science and
Engineering, Shenzhen University, Abstract
Shenzhen, P. R. China The urgent need for consistent, reliable, ecofriendly, and stable power sources
2
School of Materials Science and drives the development of new green energy materials. Thermoelectric (TE)
Engineering, Nanyang Technological
University, Singapore materials receive increasing attention due to their unique capability of rea-
3
Department of Materials Science and lizing the direct energy conversion between heat and electricity, showing di-
Engineering, City University of Hong verse applications in harvesting waste heat and low‐grade heat. Carbon
Kong, Kowloon, Hong Kong SAR, China
materials such as carbon nanotubes (CNTs) and graphene have experienced a
Correspondence rapid development as TE materials because of their intrinsic ultrahigh elec-
Qichun Zhang, School of Materials Science trical conductivity and light weight. Besides, polymer‐based carbon composites
and Engineering, Nanyang Technological
University, 50 Nanyang Ave., Singapore
are particularly fascinating as the combination of the merits of polymers and
639798, Republic of Singapore. filler materials leads to high TE performance and superior flexibility. Herein,
Email: qczhang@ntu.edu.sg the recent TE advances are systematically summarized in the studied popu-
Guangming Chen, College of Materials larity of carbon materials (ie, CNTs and graphene) and the category of poly-
Science and Engineering, Shenzhen mers. The conducting polymer‐based carbon materials are particularly
University, Shenzhen, 518055 Guangdong,
highlighted. Finally, the remaining challenges and some tentative suggestions
P. R. China.
Email: chengm@szu.edu.cn possibly guiding future developments are proposed, which may pave a way for
a bright future of carbon and carbon composites in the energy market.
KEYWORDS
carbon nanotubes, conducting polymers, graphene, thermoelectric materials, thermoplastic
polymers
1 | INTRODUCTION very regrettable, and therefore, it is highly and urgently

required to seek a sustainable solution to address this
Nowadays, increasing attention is being paid for search- issue. Fortunately, thermoelectric (TE) materials offer us
ing alternative energy sources (such as solar energy, hy- an appealing option as they can directly convert waste
drogen energy, biomass energy, and secondary batteries, heat or low‐grade heat into usable electrical energy via
and so forth) to address the energy crisis and global cli- the mobility of internal charge carriers when exposed on
mate warming.1‐4 At present, almost 90% of power supply temperature gradients. Besides, the realization process of
comes from conventional fossil fuels.5,6 Among the en- this direct energy conversion does not produce any
ergy consumption of fossil fuels, around 70% has been moving parts, annoying noise, and exhausted gas.
wasted in the form of industrial waste heat, car exhausted The performance of TE materials is generally eval-
gas, and so forth.5,7 Such tremendous energy waste is uated by a dimensionless TE figure of merit (ZT), defined
-------------------------------------------------------------------------------------------------------------------------------------------
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
© 2020 The Authors. Carbon Energy published by Wenzhou University and John Wiley & Sons Australia, Ltd.
408 | wileyonlinelibrary.com/journal/cey2 Carbon Energy. 2020;2:408–436.

26379368, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cey2.68 by Venezuela Regional Provision, Wiley Online Library on [16/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHANG ET AL. | 409
as ZT = S2σT/κ, where S, σ, T, and κ are the Seebeck survey, several review articles have been reported to
coefficient (or called the thermopower), the electrical cover organic polymer TE materials filled by carbonac-
conductivity, the absolute temperature, and the thermal eous fillers and/or inorganic TE nanoparticles such as
conductivity, respectively. In cases of systems with low or tellurium, and TE devices.20‐24 Nevertheless, an overview
similar κ, the simplified criterion power factor (PF = S2σ) exclusively highlighting carbon and carbon composites
is also adopted to access TE performance. Accordingly, remains absent, motivating us to summarize the recent
large S, high σ, and low κ are desirable for high TE advances of this topic in the present review. After the
performance. However, in practical cases, it is always introduction, we will individually present CNTs and
difficult to decouple the three TE parameters due to their graphenes as TE materials and the major factors affecting
strongly interdependent relation, resulting in tough tasks TE performance. Afterward, the conducting polymer‐
in achieving the maximization of TE performance. based carbon materials, carbon composites with ternary
In the past few decades, conventional inorganic TE or more components, and thermoplastic‐based carbon
materials including Bi2Te3, PbTe, Sb2Te3, and their alloys materials will be successively described. The final section
were studied intensively.8 However, their applications are draws a brief conclusion, presents future challenges, and
impeded by their intrinsic rigidity and poor processability. proposes some tentative suggestions for this blooming
Thus, searching new materials to address these issues is field. We hope that the summarized works in this review
urgent and highly important. Among various novel can- are most lately reported or sufficiently representative in
didates (ie, conducting polymers, organic small molecules, this review, thereby providing a timely and valuable re-
and inorganic‐organic hybrid materials),9‐11 carbon mate- ference for researchers working on TE materials.
rials, especially for carbon nanotube (CNT) and graphene,
have shown great potential in TE materials because of
light weight, nontoxicity, and ultrahigh σ.12‐14 However, 2 | CARBON MATERIALS AS
their intrinsic high κ (>3000 W m−1 K−1)15 prevents them THERMOELECTRIC MATERIALS
from achieving high TE performance, and hence various
approaches of rational nanostructuring (such as Carbon is found in the form of a variety of allotropes with
nanomeshes or nanoribbons)16,17 or physicochemical unique properties and applications.25,26 Figure 1 illus-
treatment18 are delicately utilized to decrease the κ. trates the nanostructure of carbon allotropes and their
Among them, composing composites with polymers is derivative relationship. It indicates that graphene is the
a very appealing option, because the low κ of polymers basic building unit for graphitic materials of all other
(typically in the range of 0.1‐0.5 W m−1 K−1)19 can lead to dimensionalities (Figure 1). The zero‐dimensional full-
a low κ of the carbon composites. Besides, the combina- erenes, one‐dimensional (1D) CNTs, or three‐
tion of carbon materials and polymers is not a simple dimensional (3D) graphite can be obtained if two‐
sum of each component, as energy‐filtering effect may be dimensional (2D) graphene is wrapped, rolled, or
realized at the polymer/carbon interfaces, achieving the stacked. The use of fullerenes and graphite is incon-
decoupling of the relation and optimization of TE per- spicuous as TE materials possibly because of their rela-
formance. It should be noted that composing composites tively low σ compared with CNTs and graphene.26
does not mean the absolute enhancement in TE proper- However, the ultrahigh κ values of CNTs and graphene
ties unless the band structures of carbon fillers and are still a great challenge for TE applications.28
polymers are rationally designed to induce energy‐
filtering effect. Chen et al20 pointed out that an appro-
priate energy barrier (ΔEg) in the range of 0.04 to 0.1 eV 2.1 | Carbon nanotubes as TE materials
at the interface can induce an effective energy‐filtering
effect. Therefore, judiciously chosen carbon fillers and 2.1.1 | Effect of CNT species and
polymers with appropriate band structures are significant alignment
to achieve advantageous carbon composites.
Indeed, a remarkable progress in TE studies has been It is reported that 1D CNTs have better TE properties
achieved in carbon and carbon composites. To date, the than 2D or 3D counterparts, and single‐walled CNTs
carbon materials focus on CNTs and graphenes, and the (SWCNTs) have better TE performance than multiwalled
organic components of carbon composites are in- CNTs (MWCNTs).29,30 Therefore, SWCNTs are con-
trinsically conducting polymers that mainly include sidered as one of the most promising candidates for TE
polyaniline (PANI), poly(3,4‐ethylenedioxythiophene) applications. The electronic properties of CNTs depend
(PEDOT), and polypyrrole (PPy) as well as some in- on the curvatures determined by the chiral vector, which
sulating thermoplastics. According to our literature can be used to calculate the circumference in terms of the
410 | ZHANG ET AL.
F I G U R E 1 Structures of carbon
nanomaterials. Graphene is a two‐
dimensional building unit for zero‐
dimensional fullerenes, one‐dimensional
nanotubes, and three‐dimensional
graphite. Reproduced with permission:
Copyright 2019, Springer Nature27
graphene lattice vectors.31 SWCNTs, depending on their TE performance than semiconducting SWCNTs. As
chiral vectors, can be classified as metallic SWCNTs with shown in Figure 2B, TE enhancement can be observed
a zero energy gap (the valence and conduction bands when EF is moved far away from CNP and tuned to co-
touching each other) or semiconducting SWCNTs with a incide with a vHs. By varying metallic SWCNT contents
finite energy gap.31 (Figure 2C‐E), they32 found that S monotonically in-
As shown in Figure 2A, their maximum S can be creased with σ for purely metallic films (sample #4),
achieved for the semiconducting SWCNTs by tuning the implying that the trade‐off between S and σ for semi-
Fermi energy (EF) or the charge neutrality point (CNP) conducting SWCNTs can be solved by the use of purely
with an amount around the thermal energy kBT (kB is metallic SWCNTs.
Boltzmann constant). It implies that there is a trade‐off Besides, Ichinose et al32 suggested that the alignment
between S and σ for semiconducting SWCNTs. However, of SWCNTs led to an enhanced σ but unchanged S for the
despite the relatively small σ, compared to metallic aligned metallic film (sample #5), and a PF value of
SWCNTs, semiconducting SWCNTs still exhibit higher around 300 μW m−1 K−2 was achieved. Similar conclu-
TE performance, which is mainly due to a large S in the sion was also drawn in the study by Fukuhara et al33 on
range of several hundred to over a thousand μV/K.31,35‐38 SWCNT film (a mixture of metallic and semiconducting
Therefore, the promising studies of TE properties on SWCNTs). The testing direction was illustrated in
semiconducting SWCNTs have laid a solid groundwork Figure 2F and SWCNT alignment degree was evaluated
and the encouraging energy harvesting results.39,40 Fur- by polarization‐dependent Raman spectra (Figure 2G).
ther, many TE studies have largely disregarded metallic The σ values along the parallel direction were larger than
SWCNTs due to their typical small S under normal cir- that in the perpendicular direction (Figure 2H), whereas
cumstances,39 and even many efforts have been devoted the S values were the same between both directions
to remove small amounts of metallic SWCNTs from covering the p‐type and n‐type regimes at different gate
semiconducting‐enriched SWCNT specimens.40 Recently, voltages (Figure 2I). This led to a strong anisotropy of PF
a different approach was reported by Ichinose et al32 who for the aligned SWCNT film (Figure 2J). They explained
used metallic SWCNTs possessing van Hove singularity that the cancellation in the anisotropy of mean distance
(vHs)41 in density of states (DOS) to demonstrate a higher between the hopping sites was responsible for the
ZHANG ET AL. | 411
F I G U R E 2 A, B, The behaviors of S and σ of semiconducting and metallic carbon nanotubes as a function of Fermi energy (EF)
measured from the charge neutrality point (ECNP). C, Schematic for the varying ratios of five samples with (6,5) being
semiconducting and #4 being metallic. D, E, S and power factor (PF) of single‐walled carbon nanotube (SWCNT) films with different
semiconductor‐metal ratios. Reproduced with permission: Copyright 2019, American Chemical Society.32 G, Schematic of the
parallel and perpendicular directions in an aligned SWCNT film. H, Raman spectra taken with excitation light polarized parallel and
perpendicular to the aligned axis. H‐J, σ, S, and PF as a function of gate voltage in the parallel and perpendicular directions in an
aligned SWCNT thin film. Reproduced with permission: Copyright 2018, American Institute of Physics.33 K, S of the aligned SWNT
sample before and after annealing. Reproduced with permission: Copyright 2000, American Institute of Physics34
412 | ZHANG ET AL.
isotropic S according to Mott's theory.47 Besides, S also authors ascribed the variation of TE parameters of
showed identical temperature dependencies in the par- SWCNTs to the conversion of carbon hybridization from
allel and perpendicular directions for aligned SWCNT sp2 to sp3 when SWCNT was functionalized, as demon-
samples (Figure 2K).34 strated by Raman spectra (Figure 3H) showing an in-
creased peak intensity of D band and indicating the
introduction of defects in the SWCNTs. In addition to the
2.1.2 | Effect of CNT functionalization chemical method, physical treatments (such as plasma
treatment and irradiation treatment) were also employed
The functionalization of CNTs is a very common method to functionalize SWCNTs.18,46,57 For example, UV/O3 can
to moderate TE performance of CNT materials because introduce epoxide and carboxylic groups on the surface of
both σ and S can be tuned by the introduction of func- CNTs. Nonoguchi et al46 found, with increasing irradia-
tional groups on the SWCNT sidewall, both being at- tion time, S monotonously decreased, whereas σ in-
tributed to the position of the Fermi energy and their creased and subsequently reduced (Figure 3I). The
diameter.39,48,49 However, there is a certain level of con- authors claimed that short UV/O3 irradiation may induce
troversy regarding the effects of CNT functionalization on hole doping (Figure 3J) to improve the σ while long
TE parameters, especially when CNT is modified by acid. UV/O3 irradiation could cause the high‐density defects
On one hand, acid treatment (ie, HCl, H2SO4, HNO3, (Figure 3J) to suppress the TE properties. As a result, an
and mixed acids) not only removes the external layers optimized PF of 140 μW m−1 K−2 was obtained after ir-
with defects on the CNT surface (Figure 3A),42 the acids radiation for about 120 seconds (Figure 3K), which was
also play a role as p‐type dopants.50‐52 Both factors con- approximately three times larger than the value of the
tribute to the improvement of σ.53 On the other hand, reported chemically doped thin films.
excessive acid treatment may destroy the π‐electron
conjugated systems of CNTs that can suppress the carrier
transport due to carrier scattering. This can be verified by 2.1.3 | Effect of annealing
Raman spectra (Figure 3B),50 in which the ratio of the G
and D band intensities (IG/ID) is related to the structural Annealing is another method to tune the TE performance
integrity of the sp2‐hybridized carbon atoms and dis- of CNT materials, which can effectively modify the in-
ordered sp3‐hybridized carbon atoms as well as the de- terfacial structure and hence affect the transport prop-
fects of nanotubes. The decrease of IG/ID ratio indicates erties, allowing the enhancement of TE performance.58,59
that the acid treatment may have destroyed the π‐π However, the mechanism underling this enhancement is
conjugation structure in CNTs, leading to a decrease not clear till now. Recently, Hada et al56 studied the
of σ.43,54 potential application of CNT yarns as TE materials, in
Our group44 studied the effect of chemical doping of which the amorphous carbons at the interface of the
SWCNTs by the mixed acids of HNO3 and H2SO4 on TE CNTs can impede the carriers and phonons in transport.
performance of SWCNT film (Figure 3C‐E). The acid‐ They described how interfacial amorphous carbon was
doped SWCNTs with the flat and smooth morphology converted to graphene fragments at CNT interface by
(Figure 3D) favored the charge transfer among adjacent annealing. Upon annealing, the number of sp2‐
SWCNTs relative to pristine SWCNTs with the wrinkled hybridized bonds gradually increased, whereas that of
aggregation (Figure 3C), leading to a much higher σ of sp3‐hybridized bonds gradually decreased at the same
acid‐doped SWCNTs than that of pristine SWCNTs rate (Figure 4A‐D). After 1 minute of Joule annealing, the
(Figure 3E). However, the S decreased after the rich CNT yarns produced high S values of ±100 μV/K
oxygen‐containing functional groups (such as carbonyl, (Figure 4E) and enhanced σ (Figure 4F), the former of
carboxyl, and hydroxyl groups) were introduced on the which were close to those of inorganic TE materials.60,61
SWCNTs during the acid treatment process. Similar re- It resulted in a high PF of 1000 μW m−1 K−2 for n‐type
sults have been reported by Fischer et al.55 As a con- CNT yarns (Figure 4G). Besides, the TE parameters dis-
sequence, the PF of the acid‐doped SWNCT was lower played a positive correlation with the G/D ratio of the
than that of the pristine SWCNT. In sharp contrast with Raman spectra (Figure 4H‐J). Indeed, the authors adop-
the results of Figure 3E, Jiang et al45 found the opposite ted the realistic time‐scale kinetic simulations to exhibit
results. The σ of functional SWCNT films was, respec- the conversion between sp2‐hybridized and sp3‐
tively, much lower but S is much higher than those of hybridized carbons at the CNT interface and deepened
pristine SWCNT films (Figure 3F). Specifically, the the understanding of the enhanced mechanism on TE
SWCNT‐OH film showed a higher PF compared to other performance by annealing. However, the experimental
functionalized and pure SWCNTs (Figure 3G). The characterization on this conversion is intrinsically
ZHANG ET AL. | 413
F I G U R E 3 A, Schematic of structural evolution during HCl‐treated multiwalled carbon nanotube (MWCNT) process.
Reproduced with permission: Copyright 2010, Taylor & Francis Group.42 B, Raman spectra of pristine CNT fiber, CNT fiber acidized
for 15 and 30 minutes, respectively, with mixed acids of HNO3 and H2SO4 (1:3). Reproduced with permission: Copyright 2014,
Elsevier.43 C, D, Scanning electron microscopy (SEM) images of the pristine SWCNT and the acid‐doped SWCNT. E, Thermoelectric
(TE) performances of original SWCNT and the acid‐doped SWCNT. Reproduced with permission: Copyright 2019, Royal Society of
Chemistry.44 F, G, TE performances of SWCNT, SWCNT‐OH, SWCNT‐COOH, and SWCNT‐NH2 films. H, Raman spectra of
SWCNT, SWCNT‐OH, SWCNT‐COOH, and SWCNT‐NH2. Reproduced with permission: Copyright 2019, Springer Nature.45 I, J, σ of
SWCNT films upon UV/O3 treatment and D/G ratio at 785‐nm excitation. K, PF of SWCNT films upon UV/O3 treatment.
Reproduced with permission: Copyright 2019, American Institute of Physics.46 PF, power factor; SWCNT, single‐walled carbon
nanotube
414 | ZHANG ET AL.
F I G U R E 4 A, Structural transition‐route map of carbon atoms at the interface of the CNTs. B‐D, The initial carbon,
intermediate, and completely annealed carbon structure, respectively. The sp3‐hybridized bonds are shown in gray, while the
sp2‐hybidized bonds are shown in light blue. E‐G, S, σ, and PF of untreated, Joule‐annealed, polyethyleneimine (PEI)‐doped, as well
as PEI‐doped and Joule‐annealed CNT yarns. H‐J, S, σ, and PF of the CNT yarns as a function of the G/D ratio. Reproduced with
permission: Copyright 2019, American Chemical Society.56 CNT, carbon nanotube; PF, power factor
difficult and still a challenge due to the limitations of the the use of graphene as a TE material is debatable to
current characterization techniques.62 some extent.1 Indeed, graphene shows an ultrahigh
σ (~106 S/cm) at room temperature.15 However, its
maximum S was reported to be about 80 μV/K,64,65
2.2 | Graphenes as TE materials which is lower than the corresponding maximum value
of CNTs. The κ of graphene is in the range of 4840 to
Compared to CNTs, research interest in graphene as TE 5300 W/mK, originating from the lattice vibration of
materials seems to grow relatively slowly as pristine strong covalent sp2 bonding and long mean free path of
graphene often exhibits a disappointingly poor TE per- phonon at room temperature,66 and such high κ severely
formance with ZT values as low as 10−4 to 10−3.63 Even degrades the TE performance. Overall, the origins of
ZHANG ET AL. | 415
relatively low TE performance mainly lies in the fol- to concurrent phonon Boltzmann transport simulations,
lowing two major factors: (a) its gapless band structure the reduction of κph was mainly attributed to quasi‐
(zero‐bandgap nature) leading to a relatively small S and ballistic phonon transport. S of the GNR changed from
(b) a very small σ/κ ratio stemming from both superior σ −87.7 to −125.7 μV/K in the temperature range shown in
together with very high κ.67 Figure 6C, which was several times larger than that of
Many efforts have been devoted to theoretically bulk graphene at the corresponding temperatures. The
and experimentally enhance TE performance of gra- enhanced S was mainly attributed to the bandgap open-
phene.64,65,68‐70 Some theoretical studies67,71‐73 demon- ing. This finally yielded a high ZT values of approxi-
strated, to some extent, that rational nanostructuring of mately 0.1 (Figure 6D), which was 2 to 4 orders of
graphene can enhance S via bandgap opening, reduce κ, magnitude higher than all the reported ZT values of
and keep σ, yielding a high TE performance. Typically, an graphene‐based materials including bulk graphene,63
obvious bandgap can be induced due to quantum nanomeshes,16 and doped graphene.79 This indicated the
confinement effect when graphene is modified to be defect‐free graphene with appropriate nanostructures
graphene nanomeshes (GNMs) or nanoribbons may be very promising TE materials.
(GNRs).17,74‐76 Oh et al16 demonstrated that the nano-
mesh approach is effective to improve TE performance.
They fabricated single‐ and bilayer‐graphene nano- 3 | P O L Y M E R‐ B A S E D CA R B O N
meshes (SGNMs and BGNMs) (the fabrication schematic COMPOSITES
is shown in Figure 5A) with different neck widths
(Figure 5B‐E). The SGNMs or BGNMs showed largely Polymer‐based carbon composites can synergistically
reduced κ (Figure 5F) and enhanced S (Figure 5H) combine the advantages of both carbon particles and
compared to single‐layer graphene and bilayer graphene, polymer materials.80,81 The composites are endowed with
but their σ values also were significantly reduced (Fig- the high σ of carbon materials and the unique char-
ure 5G). The BGNM8 sample having 8‐nm neck width acteristics of polymers such as low κ, flexibility and
showed a very low κ of 78 W m−1 K−1 and a high S of solution‐ or melt‐based processability.19,82‐84 Besides, the
−190 μV/K at 300K and −520 μV/K at 520K (Figure 5H). interface between polymers and carbon particles is deli-
A high PF value of around 8200 μW m−1 K−2 was cately utilized to realize energy‐filtering effect, optimiz-
achieved at 520K. Using classical and quantum me- ing the TE performance.30,85 Among the carbon particles,
chanical calculations, the authors claimed that the CNTs and graphene are the most widely used to compose
quantum confinement and phonon‐edge scattering with intrinsically conducting polymers and/or isolating
within sub‐10‐nm neck width scale were responsible for polymers.24,30,85
better TE efficiency.
It is well known that κ comprises phononic (lattice)
part (κph) and electronic part (κe).23 σ is a function of κe 3.1 | Conducting polymer‐based carbon
according to the Wiedemann‐Franz law, that is, κe = composites
σTL0, where L0 (2.44 × 10−8 WΩ/K2) is the standard
Lorenz number for graphene.77 Based on this law, the 3.1.1 | Conducting polymers/CNT
largely reduced κ always leads to a great reduction of σ composites
(e.g., in Figure 5F,G). Ouyang et al67 and Haskins et al72
theoretically explained that the defects and impurities Polyaniline/CNT composites
were responsible for the poor σ/κ ratio because they PANI is one of main TE materials because of its high σ,
suppressed electron transport more significantly than facile workability, and environmental stability. It has
phonon transport. Therefore, the decoupling of electrical metallic domains and microscopic disordering domains
and thermal transports for graphene is highly challen- (localization), the former of which shows a higher σ
ging. Recently, Li et al78 decoupled the electron and than that in the later one.86 Consequently, the TE
phonon transports by size and defect control using the performance of PANI strongly hinges on the ordering
concept in Figure 6A to simultaneously enhance σ/κ ratio degree of PANI chains that have great effects on the
and S of the as‐grown suspended GNR. This σ/κ ratio of carrier mobility by adjusting the hopping free path and
the GNR is 1 to 2 orders of magnitude higher than hopping active energy.87 It was pointed out that the π‐
micrometer‐sized graphene and GNMs (Figure 6B). It bonded surface of CNTs can interact strongly with the
was enabled by the disparate electron and phonon mean conjugated structure of PANI through π‐π interactions,
free paths as well as the clean samples (as‐grown GNR which facilitates the growing of an ordered chain along
does not almost have defects and impurities). According the CNT surface.88,89
416 | ZHANG ET AL.
F I G U R E 5 A, The schematic of fabrication and thermal measurements of a sub‐10‐nm graphene nanomeshes. B‐E, Typical
transmission electron microscopy (TEM) images of the single‐layer graphene nanomesh (SGNM) films and bilayer graphene
nanomesh (BGNM) films. The numbers after SGNM or BGNM denotes the average neck width in unit nm. F‐H, κ, σ, and S of pristine
single‐layer graphene (SLG), bilayer graphene, SGNM16, and SGNM12 as a function of temperatures. Reproduced with permission:
Copyright 2017, Elsevier16
The first study concerning PANI/MWCNT TE nano- 15.8 wt% (Figure 7C), which should be attributed to the
composites was reported by Meng et al30 by a two‐step size‐dependent energy‐filtering effect caused by the na-
method. The typical morphology (Figure 7A) demon- nostructured PANI‐coating layer enwrapped around the
strated the nanostructure of MWCNT bundles coated MWCNTs. The authors30 supposed that, with increasing
with PANI. Clear interfaces (indicated by white arrows) PANI content, thicker PANI‐coating layer weakened the
was seen around the outer walls of the MWCNTs, evi- energy‐filtering effect due to the decreasing quantum
dencing that the PANI chains grew on the outer walls of states in the quantum region. Later, Chen et al90,91 per-
the MWCNTs. The largest S of the composite sheet was formed a series of TE investigations on PANI/SWCNT
28.6 μV/K, and σ values were in the range of 6147 to composites. The PANI/SWCNT composite90 was synthe-
4441 S/m for the nanocomposites with 15.8 to 57.5 wt% sized by in situ polymerization of anilines using SWCNT
PANI (Figure 7B). The maximum PF value of as a template, and its typical transmission electron mi-
5.0433 μW m−1 K−2 was achieved at the PANI content of croscopy (TEM) is shown in Figure 7D. It showed a
ZHANG ET AL. | 417
F I G U R E 6 A, Concept of how to enhance TE performance of as‐grown suspended graphene nanoribbons. B, The room‐
temperature σ/κ ratio and κe/κ ratio compared with the reference data as a function of the characteristic width. C, S of this graphene
nanoribbon and bulk SLG below 300K. D, the high ZT value compared with the reference data. Reproduced with permission:
Copyright 2019, American Chemical Society.78 SLG, single‐layer graphene; TE, thermoelectric
simultaneous enhancement of σ and S (Figure 7E) due to than the calculated values (the inset of Figure 7I). It led
the increasing carrier mobility stemming from the for- to a very high TE performance (PF = 176 μW m−1 K−2,
mation of the ordered chain structure. Furthermore, it ZT = 0.12) in organic‐inorganic composites. The authors
kept a low κ (Figure 7F), which was attributed to the ascribed the significantly enhanced TE performance to
phonon‐scattering effect of nanointerfaces. It was also the highly ordered PANI interface layer on the SWNT
found that the significantly increased σ and reduced κ surface, which was formed by the synergetic effect of
were obtained compared to the calculated values based chain expansion by the chemical interactions between
on the particle mixture rule (dashed lines in Figure 7E,F). PANI and the m‐cresol solvent as well as the orderly
The authors used X‐ray diffraction patterns and Raman arrangement of PANI molecules due to the π‐π con-
spectra to confirm the growth of PANI along the CNT jugation between PANI and SWCNTs (Figure 7J).
surface and the formation of an ordered chain structure Therefore, the interface design between conducting
during the SWNT‐directed polymerization process. It polymers and CNTs including energy‐filtering effect, re-
indicated that the decoupling of TE parameters and the inforcement of the ordering degree of molecular con-
TE improvement of PANI/SWCNT composite (PF = 2 × formation, and π‐π conjugation effect, can realize the
10−5 W m−1 K−1, ZT = 0.004) were achieved by con- decoupling of the three TE parameters and effectively
structing highly ordered PANI chain structure. After- enhance the TE performance of composites.
ward, Chen et al91 further improved the TE performance In addition, some efforts were also contributed to
of the composite by using camphor sulfonic acid (CSA)‐ tailor the characteristics of CNTs, such as oxidation and
doped PANI. The TEM image (Figure 7G) confirmed that alignment. King et al92 tuned the TE performance of
the SWNT was coated by PANI with a thickness of about PANI/CNT films by keeping CNT concentration constant
10 to 20 nm. The CSA‐doped PANI/SWCNT showed that and changing the nature of CNTs, for example, SWCNT,
σ first increased to the maximum value of 769 S/cm MWCNT, oxidized and unoxidized ones. The PANI/un-
and then decreased with the increasing SWNT content oxidized SWCNT showed best TE performance with a PF
(Figure 7H), whereas S increased monotonically with a value of 0.6 μW m−1 K−2 compared to other counterparts.
maximum value of 65 μV/K (Figure 7I). σ was, respec- Besides, the effect of CNT alignment on TE performance
tively, much higher than the calculated values (dash for PANI/CNT composite was exploited. The PANI/
lines) based on the mixture model while κ is much lower MWCNT composite nanofibers93 were prepared with
418 | ZHANG ET AL.
F I G U R E 7 A, TEM image of polyaniline (PANI)/MWCNT composite. B, C, S and PF of PANI/MWCNT composite. Reproduced

with permission: Copyright 2010, Wiley.30 D, TEM image for SWNT/PANI composites. E, F, σ, S and κ of SWNT/PANI composites
with different SWNT contents. Reproduced with permission: Copyright 2010, American Chemical Society.90 G, TEM image of
camphor sulfonic acid (CSA)‐doped PANI/SWCNT composites. H, I, σ, S and κ of CSA‐doped PANI/SWCNT composites.
J, Schematic representations of the formation of an ordered PANI interface layer induced by the synergistic effects of the solvent process
and the π‐π conjugation between PANI and SWNT. Reproduced with permission: Copyright 2014, Royal Society of Chemistry.91
MWCNT, multiwalled carbon nanotube; SWCNT, single‐walled carbon nanotube; TEM, transmission electron microscopy
both PANI chain and MWCNT orienting along the fiber “DM‐PANI/MWCNT”) showed random orientation
axis. The PANI/MWCNT fiber had PANI molecule or- (Figure 8A). The oriented structure triggered an en-
ientation by in situ polymerization and electrospinning, hanced σ and an unchanged S in the parallel direction
whereas the arrangement of PANI chains inside PANI/ (Figure 8B,C). Therefore, the anisotropic TE properties
MWCNT fibers by directly mixed method (denoted as gained significant improvements compared with the
ZHANG ET AL. | 419
isotropic DM‐PANI/MWCNT. Besides, Chen et al94 used excellent thermal stability, where CNT is usually used for
the cold pressing method to cause preferential distribu- the enhancement of TE properties of PEDOT. With in-
tion of MWCNTs perpendicular to the pressure creasing CNT contents, σ of the composite increases
(illustrated in Figure 8D), leading to an anisotropy of TE drastically, whereas the κ and S remain nearly un-
performance. The σ showed an obvious anisotropy and S changed. Yu et al96,97 reported some interesting in-
showed an isotropy (Figure 8E,F). Further experiments vestigations on PEDOT:PSS/CNT TE composites. As
were performed on powder‐like PNAI/SWCNT through shown in Figure 9A and scanning electron microscopy
the cold pressing method (Figure 8G) by the same (SEM) (Figure 9B), the decoration of PEDOT:PSS parti-
group.95 Similar results of PNAI/SWCNT were also ob- cles on the CNT surface caused electrical junctions, and
served for the directions along and vertical to the cold hence σ was enhanced significantly (Figure 9C). The
pressing (Figure 8H,I). junction also played an important role in filtering low‐
energy electrons, keeping relatively constant S. As a re-
PEDOT/CNT composites sult, the weak correlation between σ and S led to a high
PEDOT is also one of the most promising TE candidates PF value (PF = 160 μW m−1 K−2) at 60 wt% CNTs. Be-
due to the environmental stability, good flexibility, and sides, κ remained low because the mismatches of
F I G U R E 8 A, Schematic representations of the formation mechanism of DM‐PANI/MWCNT and PANI/CNT fibers. B, C, σ and
S of PANI/MWCNT and DM‐PANI/MWCNT fibers with different MWCNT contents in the directions parallel and perpendicular to
the fiber axis. Reproduced with permission: Copyright 2012, Royal Society of Chemistry.93 D, Schematic illustration of the fabrication
procedure for powder‐like MWCNT/PANI composites. E, F, σ and S of the cold‐pressed PANI/MWCNT composites with different
CNT contents for different directions. Reproduced with permission: Copyright 2017, Elsevier.94 G, Schematic illustration of the
fabrication procedure and measured directions parallel and perpendicular to the cold pressing pressure axis. H, I, Temperature
dependence of σ and S of PANI/SWCNT composites with SWCNT contents. Reproduced with permission: Copyright 2018, Elsevier.95
MWCNT, multiwalled carbon nanotube; PANI, polyaniline; SWCNT, single‐walled carbon nanotube
420 | ZHANG ET AL.
vibrational spectra between CNTs and PEDOT:PSS ef- TE performance of PEDOT:PF6/SWCNT film can even be
fectively block the phonon transport. further improved to be 350.0 µW m−1 K−2 (Figure 10L) by
Our group44,98‐100 carried out a series of TE studies of chemical doping of SWCNTs.44
PEDOT/CNT composites. We reported the in situ poly- The PEDOT/CNT composites were also tuned by
merization of 3,4‐ethylenedioxythiophene (EDOT) the posttreatment. For instance, σ of PEDOT:PSS/
monomers on CNT surfaces to construct PEDOT:PSS/ MWCNT composites was significantly enhanced after
CNT TE nanocomposite.98 As illustrated in Figure 10A, the treatment of ethylene glycol (EG) because the re-
the PSS and MWCNT were, respectively, used as the moving of noncomplexed PSS chains results in the
dispersant and template, EDOT was in situ polymerized decrease in interbundle distances.101 As a result,
and enwrapped the MWCNT surfaces. The PEDOT EG‐treated nanocomposite films with 20 wt% CNTs
coating was directly confirmed by TEM image exhibited a high PF of 151 μW m−1 K−2, which was
(Figure 10B). With increasing MWCNT:EDOT mass ra- much higher than the nontreated sample with
tios (Figure 10C), σ increased and S remained nearly 21.6 μW m−1 K−2.101 Unlike typical adversely corre-
unchanged. PEDOT:PSS/MWCNT nanocomposites ex- lated S and σ, Hsu et al102 reported that the remarkable
hibited greatly enhanced TE performance (Figure 10D) increases in σ by the dimethyl sulfoxide (DMSO) or
compared with those of neat PEDOT:PSS. Later, we99 formic acid (FA) treatment did not decrease S. At 6.7
considered the effect of the reaction medium (hexane or wt% CNT, DMSO‐ and FA‐treated PEDOT:PSS/CNT
xylene) on the PEFOT‐coating morphology (Figure 10E). composites exhibited PF values of 464 and
The cable‐like (in hexane) or coral‐like (in xylene) mor- 407 µW m−1 K−2, respectively. The DMSO or FA treatment
phology was proven by TEM. It was found that the mainly removed insulating PSS, improving the carrier mo-
PEDOT/SWCNT with unusual coral‐like morphology bility and thereby σ. The energy‐filtering effect played a role
displayed significantly enhanced TE performance relative in maximizing S at an intermediate level of CNT con-
to the composites with a common cable‐like morphology centration. Accordingly, the highest PF was obtained at an
(Figure 10F‐H). The higher content of the ordered intermediate CNT concentration of DMSO‐ and FA‐treated
structures for PEDOT macromolecular alignments and PEDOT:PSS/CNT composites.
levels of doping and oxidation were the main reasons
contributing to the greatly enhanced TE performances. Polypyrrole/CNT composites
Recently, we reported the high‐performance PED- Despite possessing the striking stability and easy‐to‐
OT:PF6/SWCNT composite film via dynamic three‐phase prepare processing procedure, PPy/CNT composites re-
interfacial electropolymerization and subsequent physi- ceived much less attention compared to PANI/CNT and
cal mixing.100 This composite film exhibited Janus PEDOT/CNT composites. Cai et al103 prepared PPy/
morphologies with the dense and smooth water‐side and MWCNT composites by in situ oxidative polymerization.
the porous honeycomb‐like oil‐side (Figure 10I,J). This With the increase of MWCNT contents, σ showed a
film showed a maximum PF of 253.7 µW m−1 K−2 gradual increasing and subsequently a rapid decreasing
(Figure 10K). The PEDOT molecular conformation tend, whereas S monotonously increased. The maximum
transition from the coiled benzoid type to the extended value was 2.079 μW m−1 K−2 for the 20 wt% PPy/
quinoid type as well as the formation of the ordered MWCNT nanocomposite, which was almost 26 times as
structure contributed to the high TE performance. The high as that of pure PPy.
F I G U R E 9 A, CNT is coated by poly(3,4‐ethylenedioxythiophene) (PEDOT):PSS particles, making CNT‐PEDOT:PSS‐CNT

junctions in the composites. B, SEM image of the composite. C, σ and S of the composites at different CNT concentrations with inset
showing the PF value. Reproduced with permission: Copyright 2011, American Chemical Society.96 CNT, carbon nanotube;
PF, power factor; SEM, scanning electron microscopy
|421
ET AL.
ZHANG
422 | ZHANG ET AL.
Our group104‐106 performed a series of TE studies of PPy/ solution casting methods to obtain PANI/GNs pellet and
CNT composites. A layered morphology containing na- film. Both pellets and films showed simultaneous in-
nosheets of SWCNTs sandwiched by PPy nanowires was crease of σ and S with the increasing content of GNs. This
observed via convenient physical mixing, followed by va- was attributed to a substantial increase in carrier mobi-
cuum filtration.104 The effect of PPy surfactants using cetyl- lity, whereas carrier concentration was not significantly
trimethylammonium bromide and sodium dodecyl benzene changed. The highest PF of 5.6 μW m−1 K−2 was obtained
sulfonate was studied in the fabrication procedures (Figure with a PANI/GNs composite pellets at 50 wt% GNs.
11A). The layered morphologies of both PPy/SWCNT com- Recently, a very high ZT value of 0.74 for PANI/gra-
posites are presented in Figure 11B,C. Both SWCNT content phene system was reported by Lin et al.108 As illustrated
and surfactant type have obvious effects on the TE perfor- in Figure 12A, they synthesized p‐phenediamino‐
mance of composites, showing a maximum PF of 21.7 μW modified graphene (PDG) by chemical decoration with
m−1 K−2 (Figure 11D). Similar to PEDOT/CNT composite amino‐aniline groups, and then fabricated a series of
(Figure 10E‐H), the reaction medium also played important PANI/PDG composites comprising unique semi‐
roles in the coating morphologies and TE performance of the interpenetrating networks (SIPNs). The structure of
PPy/CNT composite.105 The dependence of the coating SIPNs was evidenced by Raman spectra (Figure 12B‐D).
morphology on reaction mediums, water (H2O) or aqueous In particular, the brighter short lines and darker regions
ethanol (H2O + EtOH) is presented in Figure 11E. The outer (Figure 12C,D) were responsible for PANI and the gra-
coating layers for those prepared in aqueous ethanol were phene sheets, respectively. The optimized ZT value of
distinctly uniform in morphology relative to those prepared 0.74 (PANI/PDG30‐3) was obtained (Figure 12E), which
in water. The TE performance showed a distinct dependence originated from the simultaneously enhanced S and σ. In
of both the SWCNT:PPy mass ratio and reaction medium, addition to high TE performance, Hsieh et al109 devel-
exhibiting a maximum PF value of 19.7 μW m−1 K−2 for PPy/ oped the PANI/three‐dimensional graphene (3DG) com-
SWCNT composite using aqueous ethanol (Figure 10F). The posites with reasonable mechanical strength. The 3DG
composite did not show obvious deterioration of TE perfor- matrix was synthesized by chemical vapor deposition
mance even after mechanical bending for 1000 times or with nickel catalysts and used as a scaffold for the com-
stretching (Figure 11G). Besides, the PPy/acid‐doped posites (Figure 12F). The typical morphology was shown
SWCNT composite was prepared in our recent study106 by in Figure 12G. The composites were endowed with TE
combining dynamic three‐phase interfacial electro- and mechanical integrity (Figure 12H,I). It showed a
polymerization and SWCNT chemical doping. The typical maximum PF of 81.9 μW m−1 K−2 (Figure 12I), and
morphology is shown in Figure 11H. By this methodology, highly reproducible TE performance after repeated me-
the PF value reached as high as 240.3 μW m−1 K−2 chanical bending tests (Figure 12J).
(Figure 11I), which may be the highest value for PPy and its
composites reported so far. PEDOT/graphene composites
The PEDOT/graphene composites are mainly prepared
by physical mixing, in situ polymerization, spin‐coating,
3.1.2 | Conducting polymers/graphene and so forth. Kim et al110 reported the TE performance of
composites PEDOT:PSS/graphene composites containing 1 to 3 wt%
graphene fabricated by solution spin‐coating method.
PANI/graphene composites The PF and ZT values of the composites of 2 wt% gra-
In 2012, Cai et al107 reported PANI/graphene nanosheets phene were 11.09 μW m−1 K−2 and 2.1 × 10−2, respec-
(GNs) TE composite. They adopted cold‐pressing and tively. This enhancement arose from the facilitated
F I G U R E 1 0 A, Schematic illustration of preparation for PEDOT:PSS/MWCNT via a template‐directed in situ polymerization

approach. B, TEM of PEDOT:PSS/MWCNT. C, D, σ, S and PF of pure PEDOT:PSS and PEDOT:PSS/MWCNTs composites.
Reproduced with permission: Copyright 2015, Wiley.98 E, Schematic illustration of the preparation process and the TEM images of
PEDOT/SWCNT composites using hexane and xylene as reaction media. F‐H, σ, S and PF for the neat PEDOT and the PEDOT/
SWCNT composites prepared in hexane or xylene. Reproduced with permission: Copyright 2015, Elsevier.99 I, J, SEM images of the
surface of the water‐phase side and the oil‐phase side for the pure PEDOT:PF6. K, Effects of SWCNT contents on TE performance of
PEDOT:PF6/SWCNT composites. Reproduced with permission: Copyright 2018, Royal Society of Chemistry.100 L, Effect of
3,4‐ethylenedioxythiophene (EDOT):PF6 molar ratio on TE performance PEDOT:PF6/acid‐doped SWCNT composites. Reproduced
with permission: Copyright 2019, Royal Society of Chemistry.44 MWCNT, multiwalled carbon nanotube; PEDOT,
poly(3,4‐ethylenedioxythiophene); PF, power factor; SEM, scanning electron microscopy; SWCNT, single‐walled carbon nanotube;
TE, thermoelectric; TEM, transmission electron microscopy
|423
ET AL.
ZHANG
424 | ZHANG ET AL.
carrier transfer between PEDOT:PSS and graphene as enhancement of PPy/rGO composites in terms of three
well as the high electron mobility of graphene. Yoo aspects: improving the dispersion of rGO nanosheets and
et al111 adopted the in situ polymerization method to dissolution of pyrrole monomers in water, accelerating
synthesize PEDOT:PSS/graphene composite. The as‐ the polymerization reaction of pyrrole, and increasing the
obtained composite containing 3 wt% of graphene ex- σ of PPy. The PF value of the PPy/rGO composites with a
hibited a maximum PF value of 45.68 μW·m−1 K−2. wrapping core‐shell structure reached 3.01 μW m−1 K−2
In 2013, our group first reported a template‐directed that was 84 times greater than that of the pure PPy.
in situ polymerization strategy to prepare a pie‐like Afterward, the TE performances of PPy nanowire/rGO,
structure of PEDOT/reduced graphene oxide (rGO) na- PPy nanotube/rGO, PPy particles/graphene, and PPy
nocomposite with greatly enhanced TE performance.112 nanowire/graphene composites were reported succes-
As shown in Figure 13A, in situ polymerization reaction sively in our group115 and other groups.116‐118 It is a little
of EDOT monomers occurred on the rGO surfaces, bit strange that no new studies on PPy/graphene com-
leading to thick and uniform coatings of PEDOT layers posites were updated till the second half of 2018118 ac-
on both sides of the rGO nanosheet surfaces cording to Web of Science on 1 June 2020. The possible
(Figure 13B,C). It exhibited the highest PF of reason might be the as‐obtained TE performance is re-
5.2 μW m−1 K−2 at room temperature, which was 13.3 latively very low compared to other conducting polymer‐
times greater than that of neat PEDOT (Figure 13D). based composites.
Then, we reported a further work on PEDOT/rGO com-
posites with three different in situ polymerization pre-
paration routes, that is, spin‐coating and subsequent 3.2 | Carbon composites with ternary or
liquid layer polymerization (Method A); spin‐coating, more components
followed by vapor phase polymerization (Method B); and
in situ polymerization and then posttreatment by EG Carbon composites with ternary or more components
immersion (Method C).113 The schematic illustrations of were considered to make better use of synergetic effects
the three preparation procedures and the corresponding of carbon materials and conducting polymers. Zhang
SEM images can be found in Figure 13E. It was con- et al119 incorporated fullerene into PEDOT:PSS/graphene
firmed that the surface of rGO nanosheets was wrapped to improve the TE performance. In this system, fullerene
by PEDOT layers for these three methods. All the com- interacted with noncovalently functionalized graphene at
posites prepared by three methods displayed greatly en- a liquid‐liquid interface. Such architecture increased σ
hanced TE performances compared to the corresponding from 10 000 to 70 000 S/m, enhanced S by around four-
neat PEDOT (Figure 13F‐H). fold, and improved κ from 0.2 to 2 W K−1 m−1. As a re-
sult, this ternary TE composite demonstrated a ZT value
PPy/graphene composites of 0.067, indicating an enhancement of more than 1 order
Compared to PANI/graphene and PEDOT/graphene of magnitude in comparison to binary TE composites.
composites, much less attention has been paid to the Kim et al120 synthesized PEDOT:PSS in the presence of
PPy/graphene composites in terms of TE studies. We graphene sheets and MWCNTs through an in situ poly-
prepared PPy/rGO composites by a template‐directed in merization method. The PEDOT/graphene/MWCNTs
situ oxidative polymerization of pyrrole monomers on the exhibited a ZT value of 0.031 with a 80% enhancement
surface of rGO in sodium dodecyl sulfate (SDS) aqueous relative to PEDOT:PSS. This enhancement came from the
solution.114 SDS played an important role in the synergic effects of excellent electrical bridging and
F I G U R E 1 1 A, Schematic illustration of the preparation process for the polypyrrole (PPy)/SWCNT nanocomposite. B, C, SEM
images of the self‐assembled layered morphology for C/S‐ and S/S‐ PPy/SWCNT nanocomposites; The abbreviation of C/S stands for
the dispersing agents of cetyltrimethylammonium bromide and sodiumdodecyl benzene sulfonate (SDBS) to disperse PPy nanowires
and SWCNTs, respectively, while the S/S represents the employment of SDBS to disperse both PPy nanowires and SWCNTs. D, PF
for the neat PPy nanowires and C/S‐ and S/S‐PPy nanowire/SWCNT composite films. Reproduced with permission: Copyright 2016,
Elsevier.104 E, Schematic illustration showing the preparation procedure for the PPy/SWCNT nanocomposites via a template‐directed
in situ polymerization approach with various reaction mediums and the corresponding SEM images. F, PF for PPy/SWCNT
nanocomposites. G, S, σ, and κ for the composite before and after bending Reproduced with permission: Copyright 2016, Royal
Society of Chemistry.105 H, The SEM image of PPy/acid‐doped SWCNT composite films. I, The PF of PPy/acid‐doped SWCNT
composite film. Reproduced with permission: Copyright 2019, Elsevier.106 PF, power factor; SEM, scanning electron microscopy;
SWCNT, single‐walled carbon nanotube
|425
ET AL.
ZHANG
426 | ZHANG ET AL.
electronic coupling between PEDOT and the two carbon ordered molecular structure was prepared via sequential
nanomaterials. In addition to binary carbonaceous na- layering of PANI, graphene, and DWNT with prepara-
nofillers, binary polymers were adopted. For example, tion procedure in Figure 15A. The quadlayered archi-
Wang et al121 prepared PPy/GNs/PANI composite tecture resulted in synergy within the 3D conjugated
through in situ polymerization and solution process. The network, where PANi‐covered DWNT bridged the gaps
combination of PANI and PPy with similar molecular between graphene sheets, thus increasing σ and S due to
chain structures can effectively promote the carrier improved carrier mobility (Figure 15B). The resulting
transmission. Therefore, the PPy/GNs/PANI composite PANI/graphene/PANI/DWNT composites showed the
with 32 wt% graphene demonstrates a PF value of maximum PF of 1825 µW m−1 K−2 (Figure 15C) that can
52.5 μW m−1 K−2, much higher than pure PANI, pure rival inorganic tellurides.125 Then, they used LBL
PPy, PPy/GNs composite, and PANI/GNs composite. Our deposition to fabricate PANI/graphene‐PEDOT:PSS/
group122 developed a unique layered morphology of PANI/DWNT/PEDOT:PSS composites (Figure 15D), in
PPy/PEDOT:PSS/SWCNT ternary TE composites. The as‐ which graphene and DWNT were stabilized with PED-
obtained ternary hybrids possessed the highest PF of OT:PSS (Figure 15E).126 Such composite film with 80
45.3 μW m−1 K−2 at room temperature, approximately quadlayers yielded an unprecedented PF value of
14.4 times of that of the binary composites of PPy/ 2710 μW m−1 K−2 (Figure 15F). This outstanding TE
PEDOT:PSS nanowires. properties were attributed to the improved carrier mo-
A fully aligned polymer backbone can lead to a strong bility in the interconnected and layered architecture of
enhancement in the conjugation and facilitates the PEDOT:PSS‐covered DWNT that bridged between gra-
transport of carriers. Wang et al123 reported a striking and phene sheets. This LBL technique was also used to
unique approach to improve the alignment of the PANI prepare n‐type composites with good air stability.127,128
main chain by the introduction of pyrrole unit. In the The composite film was assembled through alternately
simulation (Figure 14), PPy had a fully aligned backbone, depositing DWNT stabilized by polyethyleneimine (PEI)
and main PANI chain displayed unaligned character- and graphene stabilized by polyvinylpyrrolidone
istics. When pyrrole was introduced onto the PANI (Figure 15G).127 This carbon composite with 80 bilayers
backbone, the conformation was greatly improved by exhibited a relatively stable PF of 190 μW m−1 K−2
forming a more aligned structure (Figure 14C), that is, (Figure 15H), and TE performance can remain relatively
the random copolymer (PANIPy). Such PANIPy was air‐stable over 60 days without any protection. Likewise,
realized by a sustainable oxidative polymerization n‐type multilayered DWNT‐PEI/rGO composites were
method. The PANIPy/SWCNT composite exhibited the prepared by LBL technique and the thermal reduction of
highest PF of 70.4 μW m−1 K−2 at 300K, which was far GO.128 They found that this n‐type composite could
higher than the common PANIPPy/SWCNT composites exhibit an extended period of air‐stability time when the
(PF = 50.7 μW m−1 K−2). top was the deposited layers containing montmorillonite
Among the most important methods to prepare car- clay. The n‐type characteristics of the DWNT‐PEI/rGO
bon nanocomposites (such as in situ polymerization, were relatively unstable with S becoming positive within
template synthesis, and solution mixing), the layer‐by‐ 30 days, whereas the DWNT‐PEI/rGO decreased by 45%
layer (LBL) technique shows the unique and attractive over 90 days but reached a stable plateau after
advantages of high‐performance TE composites. Grunlan depositing clay‐based thin films (Figure 15I). The au-
et al124‐128 employed this technique to conduct a series of thors attributed the improved air‐stability to the very
TE studies. First, the multiple‐layered PANI/graphene/ low oxygen permeability of clay nanoplatelet‐based
PANI/double‐walled CNT (DWNT) composites124 with assemblies.
F I G U R E 1 2 A, Schematic diagram of semi‐interpenetrating networks (SIPNs) constructed by chemically bonded

p‐phenediamino‐modified graphene (PDG) and linear PANI. B, Raman spectrum of PANI/PDG30‐3 washed by N‐methyl‐2‐
pyrrolidone. PANI/PDGx‐y, where x is the modification temperature in °C and y is the percentage of PDG in the sample. C, D,
Raman mapping at a specific Raman wavenumber of 1200 cm−1 for C‐H bending of the quinoid ring. E, ZT values of the PANI/
graphene and PANI/PDGx‐y composites with various fraction wt% of graphene or PDG. Reproduced with permission: Copyright
2018, American Chemical Society.108 F, Schematic illustration of the synthesis procedure used in fabrication of PANI/three‐
dimensional graphene (3DG) composite. G, Typical SEM image of PANI/3DG composite. H, Stress‐strain curves for the PANI/3DG
composites. I, PF of PANI/3DG composites. J, TE performance for the 3DG‐PANI containing 80 wt% PANI before and after bending.
Reproduced with permission: Copyright 2019, Royal Society of Chemistry.109 PANI, polyaniline; PF, power factor; SEM, scanning
electron microscopy; TE, thermoelectric
ZHANG ET AL. | 427
F I G U R E 1 3 A, Schematic illustration showing the preparation process of the PEDOT/reduced graphene oxide (rGO) composite.
B,C, SEM and TEM images of the PEDOT/rGO composite. D, PF of PEDOT and the PEDOT/rGO composite; the grid‐filled column
stands are the calculated value for the nanocomposite based on mixing rule. Reproduced with permission: Copyright 2013, Royal
Society of Chemistry.112 E, Three preparation routes and SEM images of PEDOT/rGO composites; Route A: Spin‐coating and
subsequent liquid layer polymerization, Route B: Spin‐coating followed by vapor phase polymerization, Route C: In situ
polymerization and then posttreatment by immersion in ethylene glycol. F‐H, TE performances of PEDOT/rGO nanocomposites
prepared by Route A (F), Route B (G), and Route C (H). Reproduced with permission: Copyright 2015, Wiley.113 PEDOT,
poly(3,4‐ethylenedioxythiophene); PF, power factor; SEM, scanning electron microscopy; TE, thermoelectric; TEM, transmission
electron microscopy
3.3 | Thermoplastics‐based carbon processed and hence are limited to the solution‐
composites processing method due to their strong interchain inter-
actions, usually making them difficult into mass pro-
Undoubtedly, conducting polymers cover most of the duction. Most thermoplastic polymers, displaying distinct
organic components in TE carbon composites. However, advantages of very low costs and easy processability
these conducting polymers usually cannot be melt‐ containing both solution‐ and melt‐processability,129‐131
428 | ZHANG ET AL.
F I G U R E 1 4 Top view (left) and side view (right) of the partial chain segment conformation of PANI (A), PPy (B), and PANIPy
(C) under the lowest energy state. Reproduced with permission: Copyright 2019, Elsevier.123 PANI, polyaniline; PPy, polypyrrole
represent an interesting alternative to conducting (u‐MWCNT) resulted in the lowest value of 7.5 μV/K
polymers as organic components for TE composites. (Figure 16A). The authors explained that the oxygen
Similar to the fabrication of conducting polymers, the groups attached on the MWCNTs, both carboxyl as well
solution‐processed method was also applied to prepare as hydroxyl, acted as p‐dopants. The processing additive,
thermoplastics‐based TE composites via dissolving poly- that is, cyclic butylene terephthalate (CBT) oligomer,
mers into specific solvent83,129,132 or adopting polymer was added to improve TE properties of PC/MWCNTs.141
emulsion.97,133 The most representative studies are from This improvement was assumed to origin from a
Grunlan et al97,133‐135 on TE composites based on the transesterification reaction between ‐COOH groups of
aqueous polymer emulsion. They fabricated a poly(vinyl MWCNTs with the CBT oligomer, increasing the electro-
acetate)/CNT composite with a segregated structure and nic DOS of MWCNTs. An increase of ZT from 10−17
a relatively high ZT value over 0.006 was obtained with (PC/MWCNT‐COOH) to 10−8 (PC/MWCNT‐COOH/CBT)
20 wt% CNTs.133 A maximum ZT of approximately 0.02 could be achieved by varying CBT and processing
for 35 wt% was achieved in CNT‐filled segregated poly- temperature (Figure 16B,C). They142 also reported that the
mer composite97 by adding PEDOT:PSS, yielding ther- p‐type polypropylene (PP)/CNT/CuO TE composite could
mally disconnected and electrically connected contact be switched into stable n‐type TE composites by adding
junctions between CNTs. Besides, the role of semi- polyethylene glycol (Figure 16D), and S vales of these
conducting stabilizers in increasing TE performances of n‐type composites remain negative after being exposed to
the segregated polymer composites was studied.134,135 air for 8 months. Concerning melt‐processing n‐type TE
From the perspective of mass production, the melt‐ composites based on thermoplastics, a recent study by
processed thermoplastics‐based TE composites are easily from Paleo et al139 also reported PP/carbon nanofibers
scaled up using current industrial techniques relative to (CNF) n‐type composites without adding any additives,
solution‐processed counterparts. Antar et al136 demon- and they ascribed the negative S value to the CNF double
strated the possibility of conductive polymer composites layer structure surrounding the tubular core, in which
as TE material. They reported melt‐processed poly(lactic n‐type contribution (electron conduction) from the inner
acid) (PLA)/MWCNT/expanded graphite materials with layer would exceed the p‐type contribution (hole conduc-
the highest ZT of 7 × 10−5 at room temperature. Then the tion) from the outer shell. Besides, Pötschke et al143
melt‐processed thermoplastics‐based TE composites studied the effect of SWCNT‐induced crystallization of
emerged from different research groups.137‐141 polyetherimide (PEI) on TE performance of PEI/SWCNT
Pötschke et al140‐146 performed a series of TE research composite based on a nonlinear PEI (aBPDA‐P3) and a
on melt‐processed thermoplastics/CNT composites in linear PEI (ODPA‐P3). The molecular structures of
terms of various aspects. They compared the effect of aBPDA‐P3 and ODPA‐P3 are shown in Figure 16E. The
surface functionalization of MWCNT on the TE perfor- crystallinity of ODPA‐P3/SWCNT was confirmed by dif-
mance of melt‐mixing polycarbonate (PC)/MWCNT at ferential scanning calorimetry test (Figure 16F). The
fixed MWCNT loading.140 The highest S was 11.3 μV/K crystalline ODPA‐P3/SWCNT nanocomposites exhibited
for PC/carboxyl‐functionalized MWCNTs (MWCNT‐ remarkably higher values for S and PF (Figure 16G,H),
COOH), whereas the PC/unfunctionalized MWCNT compared to the amorphous aBPDA‐P3/SWCNT
ZHANG ET AL. | 429
F I G U R E 1 5 A, Schematic of the layer‐by‐layer (LBL) process for PANI/graphene/PANI/DWNT. B, Carrier transport in PANI/
graphene/PANI/DWNT composite. C, PF of PANI/graphene (triangle), PANI/DWNT (square), and PANI/graphene/PANI/DWNT
(circle) as a function of deposition circles. Reproduced with permission: Copyright 2015, Wiley.124 D, Schematic of the LBL
deposition process for PANI/graphene‐PEDOT:PSS/PANI/DWNTPEDOT:PSS. E, Atomic force microscopy images of the
corresponding suspensions. F, PF of PANI/graphene‐PEODT:PSS (triangle), PANI/DWNT‐PEDOT:PSS (square), and PANI/
graphene‐PEDOT:PSS/PANI/DWNT‐PEDOT:PSS (circle) as a function of deposition circles. Reproduced with permission: Copyright
2016, Wiley.126 G, Schematic of LBL deposition process for DWNT‐PEI/graphene‐polyvinylpyrrolidone (PVP). H, S of DWNT‐PEI/
graphene‐PVP as a function of bilayer numbers. Reproduced with permission: Copyright 2016, Elsevier.127 I, Air stability of S under
ambient conditions of DWNT‐PEI/rGO with (bottom) and without (top) clay‐based thin films. Reproduced with permission:
Copyright 2018, Wiley.128 PANI, polyaniline; PEDOT, poly(3,4‐ethylenedioxythiophene); PF, power factor
composites, due to the formation of polymer crystals in are the most commonly studied carbon materials. We
ODPA‐P3/SWCNT contributing to the enhanced charge presented the factors affecting TE performances of CNTs
carrier transport through the polymer‐CNT interface or in terms of species, alignment, functionalization, and
between CNT‐CNT junctions. annealing, and then approaches to enhance TE perfor-
mance of graphenes. Subsequently, the carbon compo-
sites with polymer components that contain conducting
4 | SUMMARY AND polymers and isolating thermoplastics were discussed,
P E R S P E C T I VE S where the conducting polymer‐based carbon composites
have been highlighted. The multicomponent carbon
This review focused on the recent advances of carbon and composites consisting of polymer/binary carbon fillers,
carbon composites as TE materials. CNTs and graphenes binary polymers/carbon filler, and binary polymers/
430 | ZHANG ET AL.
F I G U R E 1 6 A, S of polycarbonate (PC) composites containing MWCNTs with different functional groups. Reproduced with
permission: Copyright 2014, Elsevier.140 B, C, PF of PC/MWCNT‐COOH and PC/MWCNT‐COOH/CBT composites processed at 240
and 270°C. Reproduced with permission: Copyright 2014, Elsevier.141 D, The effect of polyethylene glycol (PEG) content on S of PP/
CNT/CuO composites. PP‐xCNT‐yCuO‐zPEG represents a composite having x wt% CNTs, y wt% CuO, and z wt% PEG. Reproduced
with permission: Copyright 2017, Elsevier.142 E, Molecular structures of aBPDA‐P3 and ODPA‐P3. F, Differential scanning
calorimetry curves for semi‐crystalline ODPA‐P3/SWCNT composites. The neat ODPA‐P3 is amorphous. G, H, S and PF of aBPDA‐
P3 and ODPA‐P3 composites. Reproduced with permission: Copyright 2018, Elsevier.143 MWCNT, multiwalled carbon nanotube;
PF, power factor
binary carbon fillers are a new TE trend, having attracted carbon or carbon composites. However, a unified TE
extensive attention. Although great efforts have been theory is still missing. One possible root lies in the
dedicated to the development of carbon and carbon difficulty of experimental characterization at the
composites for TE purposes, the applications of these carbon‐carbon or carbon‐polymer interfaces at na-
materials in real commercial products are still in their noscale based on the current characterization techni-
infancy due to the currently obtained low TE perfor- ques. Alternatively, the simulations may be a powerful
mance (usually ZT < 1). Furthermore, there are several tool to offer more details on behaviors of carries and
challenges from theory to application that must be ad- phonons at these interfaces.56,77,78
dressed before this new set of carbon materials can be put 2. For carbon TE materials, it is of vital importance to
into applications and commercialization. Therefore, in adopt rational methodology to enhance TE perfor-
our humble opinion, some challenges and tentative sug- mance in the future, especially for the existence of
gestions are provided in the following. some disputed results. For example, the role of CNT
chemical functionalization of CNTs in the enhance-
1. The TE mechanism at the nano level remains far from ment of TE performance is a matter of debate to some
clear, though some concepts like quantum confine- extent.44,45 Precise control of CNT modification para-
ment, carrier or phonon scattering, and energy filter- meters, including functional groups, density of groups
ing have been proposed to account for some results of on the surface, and location of groups on the surface,
ZHANG ET AL. | 431
deserves an intense investigation. It may help us to reverse is also true for conducting polymer‐based
optimize the TE performance and further unravel the composites. This subject on improving the trade‐off
role of CNT functionalization. between TE and mechanical performances may be
3. For graphene TE materials, it still a challenge to sig- another research direction in the TE community,
nificantly reduce κ and thereby increase TE perfor- since it can undoubtedly widen versatile applica-
mance. Novel approaches are strongly expected to tions of the waste‐heat harvesting electronics.
reach a further significantly enhanced TE perfor-
mance to a great extent. Other than the reported ACKN OWLEDGMENTS
modification of GNMs or GNRs, the chemical This study was financially supported by the National
decoration may be also considered.108 However, the Natural Science Foundation of China (No. 51973122) and
related reports are still rare for graphene TE materials. Guangdong Basic and Applied Basic Research Founda-
4. For conducting polymer‐based carbon composites, the tion (No. 2019A1515111196). QZ acknowledges financial
intrinsic advantages of each component, such as the support from AcRF Tier 1 (RG 111/17, RG 2/17, RG 114/
flexibility and easy processability of conducting poly- 16, RG 8/16) and Tier 2 (MOE 2017‐T2‐1‐021 and MOE
mers, have not been made full use. The flexible TE 2018‐T2‐1‐070), Singapore. QZ also thanks the support
composites are greatly anticipated for curved or irre- from State Key Laboratory of Supramolecular Structure
gularly shaped heat sources such as human body,23 and Materials, Jilin University (sklssm2020041).
exhibiting their great potential as power source for
flexible electronics. ORCID
5. In addition, to further enhance TE performance of Yichuan Zhang https://orcid.org/0000-0001-9260-161X
carbon composites, decoupling the TE parameters is Qichun Zhang http://orcid.org/0000-0003-1854-8659
necessary and feasible by constructing an electrically Guangming Chen https://orcid.org/0000-0002-
connected and thermally disconnected interfaces. The 9848-9101
increase in the ordered structure of conducting poly-
mers and the molecular design of nanostructures RE FER E NCES
(such as nanotube, nanowire, and so forth) are worth 1. Gayner C, Kar KK. Recent advances in thermoelectric mate-
being considered for the future breakthrough of TE rials. Prog Mater Sci. 2016;83:330‐382.
performance.91,93 2. Sun T, Xie J, Guo W, Li DS, Zhang Q. Covalent–organic fra-
6. Obviously, the research of thermoplastics‐based TE meworks: advanced organic electrode materials for recharge-
composites lags far behind the conducting polymer‐ able batteries. Adv Energy Mater. 2020;10:1904199.
based counterparts, mainly due to their low σ and 3. Wang L, Si W, Tong Y, et al. Graphitic carbon nitride
TE performance. Considering their merits such as (g‐C3N4)‐based nanosized heteroarrays: promising materials
for photoelectrochemical water splitting. Carbon Energy. 2020;
convenience to scale up for mass production and
2:223‐250.
high mechanical properties, the utilization of in- 4. Lin X, Sun Q, Doyle Davis K, Li R, Sun X. The application of
sulating thermoplastics may be also desirable for TE carbon materials in nonaqueous Na‐O2 batteries. Carbon
materials. The viscosity, surface tension, or flowing Energy. 2019;1:141‐164.
behavior of single‐ or two‐component thermo- 5. Yang L, Chen Z‐G, Dargusch MS, Zou J. High performance
plastics130,147‐149 can be tailored to help achieve a thermoelectric materials: progress and their applications. Adv
particular distribution (e.g., alignment or selective Energy Mater. 2018;8:1701797.
6. Yao C‐J, Zhang H‐L, Zhang Q. Recent progress in thermo-
distribution) of carbon particles and enhance σ and
electric materials based on conjugated polymers. Polymers.
thereby TE performance.
2019;11:107.
7. Compared with the boosting TE performance, much 7. Koumoto K, Terasaki I, Funahashi R. Complex oxide mate-
less attention has been paid to the trade‐off between rials for potential thermoelectric applications. MRS Bull. 2006;
TE and mechanical performances, the latter of 31:206‐210.
which (such as stretchability, mechanical bending, 8. Tan G, Zhao L‐D, Kanatzidis MG. Rationally designing high‐
Young's modulus, and so forth)105,109,150 are performance bulk thermoelectric materials. Chem Rev. 2016;
definitely vital for practical applications. 116:12123‐12149.
9. Wu J, Sun Y, Pei W‐B, Huang L, Xu W, Zhang Q. Polypyrrole
Thermoplastics‐based composites have obvious ad-
nanotube film for flexible thermoelectric application. Synth
vantages of mechanical performances over con-
Met. 2014;196:173‐177.
ducting polymers based‐composites, but their very 10. Xie J, Zhao Ce, Lin Zq, Gu Py, Zhang Q. Nanostructured
low TE performances trigger a poor trade‐off conjugated polymers for energy‐related applications beyond
between TE and mechanical performances. The solar cells. Chem Asian J. 2016;11:1489‐1511.
432 | ZHANG ET AL.
11. Gao C, Chen G. In situ oxidation synthesis of p‐type compo- 32. Ichinose Y, Yoshida A, Horiuchi K, et al. Solving the ther-
site with narrow‐bandgap small organic molecule coating on moelectric trade‐off problem with metallic carbon nanotubes.
single‐walled carbon nanotube: flexible film and thermo- Nano Lett. 2019;19:7370‐7376.
electric performance. Small. 2018;14:1703453. 33. Fukuhara K, Ichinose Y, Nishidome H, et al. Isotropic See-
12. Kawamoto M, He P, Ito Y. Green processing of carbon beck coefficient of aligned single‐wall carbon nanotube films.
nanomaterials. Adv Mater. 2017;29:29. Appl Phys Lett. 2018;113:243105.
13. Yuan Y, Lu J. Demanding energy from carbon. Carbon 34. Hone J, Llaguno M, Nemes N, et al. Electrical and thermal
Energy. 2019;1:8‐12. transport properties of magnetically aligned single wall car-
14. Wang H, Cui Y. Nanodiamonds for energy. Carbon Energy. bon nanotube films. Appl Phys Lett. 2000;77:666‐668.
2019;1:13‐18. 35. Yanagi K, Kanda S, Oshima Y, et al. Tuning of the thermo-
15. Dey A, Bajpai OP, Sikder AK, Chattopadhyay S, Khan MAS. electric properties of one‐dimensional material networks by
Recent advances in CNT/graphene based thermoelectric polymer electric double layer techniques using ionic liquids. Nano Lett.
nanocomposite: a proficient move towards waste energy 2014;14:6437‐6442.
harvesting. Renew Sustain Energy Rev. 2016;53:653‐671. 36. Jiang JW, Wang JS, Li B. A nonequilibrium Green's function
16. Oh J, Yoo H, Choi J, et al. Significantly reduced thermal study of thermoelectric properties in single‐walled carbon
conductivity and enhanced thermoelectric properties of nanotubes. J Appl Phys. 2011;109:014326.
single‐ and bi‐layer graphene nanomeshes with sub‐10 nm 37. Norton‐Baker B, Ihly R, Gould IE, et al. Polymer‐free carbon
neck‐width. Nano Energy. 2017;35:26‐35. nanotube thermoelectrics with improved charge carrier
17. Nakada K, Fujita M, Dresselhaus G, Dresselhaus MS. Edge transport and power factor. ACS Energy Lett. 2016;1:
state in graphene ribbons: nanometer size effect and edge 1212‐1220.
shape dependence. Phys Rev B. 1996;54:17954‐17961. 38. Avery AD, Zhou BH, Lee J, et al. Tailored semiconducting
18. Zhao W, Fan S, Xiao N, et al. Flexible carbon nanotube papers carbon nanotube networks with enhanced thermoelectric
with improved thermoelectric properties. Energy Environ Sci. properties. Nat Energy. 2016;1:16033.
2012;5:5364‐5369. 39. Nakai Y, Honda K, Yanagi K, et al. Giant Seebeck coefficient
19. Chen G, Xu W, Zhu D. Recent advances in organic polymer in semiconducting single‐wall carbon nanotube film. Appl
thermoelectric composites. J Mater Chem C. 2017;5:4350‐4360. Phys Express. 2014;7:025103.
20. Wang Y, Yang L, Shi XL, et al. Flexible thermoelectric ma- 40. Hayashi D, Ueda T, Nakai Y, et al. Thermoelectric properties
terials and generators: challenges and innovations. Adv Mater. of single‐wall carbon nanotube films: effects of diameter and
2019;31:1807916. wet environment. Appl Phys Express. 2016;9:025102.
21. He R, Schierning G, Nielsch K. Thermoelectric devices: a re- 41. Hicks L, Dresselhaus MS. Thermoelectric figure of merit of a
view of devices, architectures, and contact optimization. Adv one‐dimensional conductor. Phys Rev B. 1993;47:16631‐16634.
Mater Technol. 2018;3:1700256. 42. Fan Q, Qin Z, Liang X, Li L, Wu W, Zhu M. Reducing defects
22. Yao H, Fan Z, Cheng H, et al. Recent development of ther- on multi‐walled carbon nanotube surfaces induced by low‐
moelectric polymers and composites. Macromol Rapid power ultrasonic‐assisted hydrochloric acid treatment. J Exp
Commun. 2018;39:1700727. Nanosci. 2010;5:337‐347.
23. Liu Z, Chen G. Advancing flexible thermoelectric devices with 43. Wang K, Li M, Liu Y, Gu Y, Li Q, Zhang Z. Effect of acid-
polymer composites. Adv Mater Technol. 2020;5:2000049. ification conditions on the properties of carbon nanotube fi-
24. Soleimani Z, Zoras S, Ceranic B, Shahzad S, Cui Y. A review bers. Appl Surf Sci. 2014;292:469‐474.
on recent developments of thermoelectric materials for room‐ 44. Fan W, Liang L, Zhang B, Guo C, Chen G. PEDOT thermo-
temperature applications. Sustain Energy Technol Assess. electric composites with excellent power factors prepared by
2020;37:100604. 3‐phase interfacial electropolymerization and carbon nano-
25. He R, Huang X, Chee M, Hao F, Dong P. Carbon‐based per- tube chemical doping. J Mater Chem A. 2019;7:13687‐13694.
ovskite solar cells: from single‐junction to modules. Carbon 45. Lan X, Liu C, Wang T, et al. Effect of functional groups on the
Energy. 2019;1:109‐123. thermoelectric performance of carbon nanotubes. J Electron
26. Toshima N. Recent progress of organic and hybrid thermo- Mater. 2019;48:6978‐6984.
electric materials. Synth Met. 2017;225:3‐21. 46. Nonoguchi Y, Tani A, Kitano T, Kawai T. Enhanced ther-
27. Baby R, Saifullah B, Hussein MZ. Carbon nanomaterials moelectric properties of semiconducting carbon nanotube
for the treatment of heavy metal‐contaminated water and en- films by UV/ozone treatment. J Appl Phys. 2019;126:135108.
vironmental remediation. Nanoscale Res Lett. 2019;14:1‐17. 47. Mott NF, Davis EA. Electronic Processes in Non‐Crystalline
28. Balandin AA. Thermal properties of graphene and nanos- Materials. New York: Oxford University Press; 1979.
tructured carbon materials. Nat Mater. 2011;10:569‐581. 48. Hung NT, Nugraha AR, Hasdeo EH, Dresselhaus MS, Saito R.
29. Hicks L, Dresselhaus MS. Effect of quantum‐well structures on Diameter dependence of thermoelectric power of semi-
the thermoelectric figure of merit. Phys Rev B. 1993;47: conducting carbon nanotubes. Phys Rev B. 2015;92:165426.
12727‐12731. 49. Green AA, Hersam MC. Colored semitransparent conductive
30. Meng C, Liu C, Fan S. A promising approach to enhanced coatings consisting of monodisperse metallic single‐walled
thermoelectric properties using carbon nanotube networks. carbon nanotubes. Nano Lett. 2008;8:1417‐1422.
Adv Mater. 2010;22:535‐539. 50. Sankapal BR, Setyowati K, Chen J, Liu H. Electrical properties
31. Hung NT, Nugraha AR, Saito R. Thermoelectric properties of of air‐stable, iodine‐doped carbon‐nanotube–polymer com-
carbon nanotubes. Energies. 2019;12. posites. Appl Phys Lett. 2007;91:173103.
ZHANG ET AL. | 433
51. Villalpandopaez F, Son H, Nezich D, et al. Raman spectro- 71. Sevinçli H, Cuniberti G. Enhanced thermoelectric figure of
scopy study of isolated double‐walled carbon nanotubes with merit in edge‐disordered zigzag graphene nanoribbons. Phys
different metallic and semiconducting configurations. Nano Rev B. 2010;81:113401.
Lett. 2008;8:3879‐3886. 72. Haskins J, Kınacı A, Sevik C, Sevinçli H, Cuniberti G,
52. Graupner R, Abraham J, Vencelová A, et al. Doping of single‐ Çağın T. Control of thermal and electronic transport in
walled carbon nanotube bundles by Brønsted acids. Phys defect‐engineered graphene nanoribbons. ACS Nano. 2011;
Chem Chem Phys. 2003;5:5472‐5476. 5:3779‐3787.
53. Liu P, Hu DC, Tran TQ, Jewell D, Duong HM. Electrical 73. Yamawaki M, Ohnishi M, Ju S, Shiomi J. Multifunctional
property enhancement of carbon nanotube fibers from post structural design of graphene thermoelectrics by Bayesian
treatments. Colloids Surf A Physicochem Eng Asp. 2016;509: optimization. Sci Adv. 2018;4:eaar4192.
384‐389. 74. Han MY, Özyilmaz B, Zhang Y, Kim P. Energy band‐gap
54. Zhang X, Sreekumar TV, Liu T, Kumar S. Properties and engineering of graphene nanoribbons. Phys Rev Lett. 2007;98:
structure of nitric acid oxidized single wall carbon nanotube 206805.
films. J Phys Chem B. 2004;108:16435‐16440. 75. Li X, Wang X, Zhang L, Lee S, Dai H. Chemically derived,
55. Zhou W, Vavro J, Nemes NM, et al. Charge transfer and Fermi ultrasmooth graphene nanoribbon semiconductors. Science.
level shift in p‐doped single‐walled carbon nanotubes. Phys 2008;319:1229‐1232.
Rev B. 2005;71:205423. 76. Dai H, Jiao L. Narrow graphene nanoribbons from carbon
56. Hada M, Hasegawa T, Inoue H, et al. One‐minute Joule an- nanotubes. Nature. 2009;458:877‐880.
nealing enhances the thermoelectric properties of carbon 77. Vallabhaneni AK, Singh D, Bao H, Murthy J, Ruan X. Re-
nanotube yarns via the formation of graphene at the interface. liability of Raman measurements of thermal conductivity of
ACS Appl Energy Mater. 2019;2:7700‐7708. single‐layer graphene due to selective electron‐phonon
57. Sham ML, Kim JK. Surface functionalities of multi‐wall car- coupling: a first‐principles study. Phys Rev B. 2016;93:
bon nanotubes after UV/Ozone and TETA treatments. 125432.
Carbon. 2006;44:768‐777. 78. Li Q‐Y, Feng T, Okita W, et al. Enhanced thermoelectric
58. Niven JF, Johnson MB, Juckes SM, White MA, Alvarez NT, performance of as‐grown suspended graphene nanoribbons.
Shanov V. Influence of annealing on thermal and electrical ACS Nano. 2019;13:9182‐9189.
properties of carbon nanotube yarns. Carbon. 2016;99:485‐490. 79. Ma W, Liu Y, Yan S, et al. Chemically doped macroscopic
59. Yamamoto G, Shirasu K, Nozaka Y, Sato Y, Takagi T, graphene fibers with significantly enhanced thermoelectric
Hashida T. Structure–property relationships in thermally‐ properties. Nano Res. 2018;11:741‐750.
annealed multi‐walled carbon nanotubes. Carbon. 2014;66: 80. Zhang Y, Pang H, Dai K, et al. Conductive network formation
219‐226. during annealing of an oriented polyethylene‐based compo-
60. Snyder GJ, Toberer ES. Complex thermoelectric materials. site. J Mater Sci. 2012;47:3713‐3719.
Nat Mater. 2008;7:105‐114. 81. Wang H, Yu C. Organic thermoelectrics: materials prepara-
61. Goldsmid HJ. Bismuth telluride and its alloys as materials for tion, performance optimization, and device integration. Joule.
thermoelectric generation. Materials. 2014;7:2577‐2592. 2019;3:53‐80.
62. Crispin X, Geskin V, Crispin A, et al. Characterization of the 82. Zhang Y, Dai K, Tang J, Ji X, Li Z. Anisotropically conductive
interface dipole at organic/metal interfaces. J Am Chem Soc. polymer composites with a selective distribution of carbon
2002;124:8131‐8141. black in an in situ microfibrillar reinforced blend. Mater Lett.
63. Seol JH, Jo I, Moore AL, et al. Two‐dimensional phonon 2010;64:1430‐1432.
transport in supported graphene. Science. 2010;328:213‐216. 83. Piao M, Kim G, Kennedy GP, Roth S, Dettlaff‐Weglikowska U.
64. Bolotin K, Sikes KJ, Jiang Z, et al. Ultrahigh electron mobility in Thermoelectric properties of single walled carbon nanotube
suspended graphene. Solid State Commun. 2008;146:351‐355. networks in polycarbonate matrix. Phys Status Solidi B. 2013;
65. Wei P, Bao W, Pu Y, Lau CN, Shi J. Anomalous thermoelectric 250:1468‐1473.
transport of Dirac particles in graphene. Phys Rev Lett. 2009; 84. Zhang YC, Dai K, Pang H, Luo QJ, Li ZM, Zhang WQ.
102. 166808‐166808. Temperature and time dependence of electrical resistivity in
66. Balandin AA, Ghosh S, Bao W, et al. Superior thermal con- an anisotropically conductive polymer composite with in situ
ductivity of single‐layer graphene. Nano Lett. 2008;8:902‐907. conductive microfibrils. J Appl Polym Sci. 2012;124:1808‐1814.
67. Ouyang Y, Guo J. A theoretical study on thermoelectric 85. Gao C, Chen G. Conducting polymer/carbon particle ther-
properties of graphene nanoribbons. Appl Phys Lett. 2009;94: moelectric composites: emerging green energy materials.
263107. Compos Sci Technol. 2016;124:52‐70.
68. Dragoman D, Dragoman M. Giant thermoelectric effect in 86. Du Y, Shen SZ, Cai K, Casey PS. Research progress on
graphene. Appl Phys Lett. 2007;91:203116. polymer–inorganic thermoelectric nanocomposite materials.
69. Gunst T, Markussen T, Jauho A‐P, Brandbyge M. Thermo- Prog Polym Sci. 2012;37:820‐841.
electric properties of finite graphene antidot lattices. Phys Rev 87. Wu J, Sun Y, Xu W, Zhang Q. Investigating thermoelectric
B. 2011;84:155449. properties of doped polyaniline nanowires. Synth Met. 2014;
70. Mazzamuto F, Nguyen VH, Apertet Y, et al. Enhanced ther- 189:177‐182.
moelectric properties in graphene nanoribbons by resonant 88. Zengin H, Zhou W, Jin J, et al. Carbon nanotube doped
tunneling of electrons. Phys Rev B. 2011;83:235426. polyaniline. Adv Mater. 2002;14:1480‐1483.
434 | ZHANG ET AL.
89. Sainz R, Benito AM, Martinez MT, et al. Soluble self‐aligned 105. Liang L, Gao C, Chen G, Guo CY. Large‐area, stretchable,
carbon nanotube/polyaniline composites. Adv Mater. 2005;17: super flexible and mechanically stable thermoelectric films of
278‐281. polymer/carbon nanotube composites. J Mater Chem C. 2016;
90. Yao Q, Chen L, Zhang W, Liufu S, Chen X. Enhanced thermo- 4:526‐532.
electric performance of single‐walled carbon nanotubes/polyani- 106. Fan W, Zhang Y, Guo C‐Y, Chen G. Toward high thermoelectric
line hybrid nanocomposites. ACS Nano. 2010;4:2445‐2451. performance for polypyrrole composites by dynamic 3‐phase in-
91. Yao Q, Wang Q, Wang L, Chen L. Abnormally enhanced terfacial electropolymerization and chemical doping of carbon
thermoelectric transport properties of SWNT/PANI hybrid nanotubes. Compos Sci Technol. 2019;183:107794.
films by the strengthened PANI molecular ordering. Energy 107. Du Y, Shen SZ, Yang W, Donelson R, Cai K, Casey PS. Si-
Environ Sci. 2014;7:3801‐3807. multaneous increase in conductivity and Seebeck coefficient
92. King RCY, Roussel F, Brun JF, Gors C. Carbon nanotube‐ in a polyaniline/graphene nanosheets thermoelectric nano-
polyaniline nanohybrids: influence of the carbon nanotube composite. Synth Met. 2012;161:2688‐2692.
characteristics on the morphological, spectroscopic, electrical 108. Lin YH, Lee TC, Hsiao YS, Lin WK, Whang WT, Chen CH.
and thermoelectric properties. Synth Met. 2012;162:1348‐1356. Facile synthesis of diamino‐modified graphene/polyaniline
93. Wang Q, Yao Q, Chang J, Chen L. Enhanced thermoelectric semi‐interpenetrating networks with practical high thermo-
properties of CNT/PANI composite nanofibers by highly or- electric performance. ACS Appl Mater Interfaces. 2018;10:
ienting the arrangement of polymer chains. J Mater Chem. 4946‐4952.
2012;22:17612‐17618. 109. Hsieh YY, Zhang Y, Zhang L, et al. High thermoelectric
94. Chen J, Wang L, Gui X, et al. Strong anisotropy in thermopower‐factor composites based on flexible three‐dimensional
electric properties of CNT/PANI composites. Carbon. 2017; graphene and polyaniline. Nanoscale. 2019;11:6552‐6560.
114:1‐7. 110. Kim GH, Hwang DH, Woo SI. Thermoelectric properties of
95. Chen J, Wang L, Ren D, et al. Revealing the anisotropy in nanocomposite thin films prepared with poly(3,4‐
thermoelectric transport performances in CNT/PANI com- ethylenedioxythiophene) poly (styrenesulfonate) and gra-
posites. Synth Met. 2018;239:13‐21. phene. Phys Chem Chem Phys. 2012;14:3530‐3536.
96. Yu C, Choi K, Yin L, Grunlan JC. Light‐weight flexible carbon 111. Yoo D, Kim J, Kim JH. Direct synthesis of highly conductive
nanotube based organic composites with large thermoelectric poly(3,4‐ethylenedioxythiophene): poly(4‐styrenesulfonate)
power factors. ACS Nano. 2011;5:7885‐7892. (PEDOT: PSS)/graphene composites and their applications
97. Kim D, Kim Y, Choi K, Grunlan JC, Yu C. Improved thermo- in energy harvesting systems. Nano Res. 2014;7:717‐730.
electric behavior of nanotube‐filled polymer composites with poly 112. Xu K, Chen G, Qiu D. Convenient construction of poly(3,4‐
(3,4‐ethylenedioxythiophene)‐poly(styrenesulfonate). ACS Nano. ethylenedioxythiophene)–graphene pie‐like structure with en-
2010;4:513‐523. hanced thermoelectric performance. J Mater Chem A. 2013;1:
98. Zhang Z, Chen G, Wang H, Li X. Template‐directed in situ 12395‐12399.
polymerization preparation of nanocomposites of PEDOT: 113. Xu K, Chen G, Qiu D. In situ chemical oxidative poly-
PSS‐coated multi‐walled carbon nanotubes with enhanced merization preparation of poly(3,4‐ethylenedioxythiophene)/
thermoelectric property. Chem Asian J. 2015;10:149‐153. graphene nanocomposites with enhanced thermoelectric
99. Hu X, Chen G, Wang X. An unusual coral‐like morphology for performance. Chem Asian J. 2015;10:1225‐1231.
composites of poly(3,4‐ethylenedioxythiophene)/carbon na- 114. Han S, Zhai W, Chen G, Wang X. Morphology and thermo-
notube and the enhanced thermoelectric performance. electric properties of graphene nanosheets enwrapped with
Compos Sci Technol. 2017;144:43‐50. polypyrrole. RSC Adv. 2014;4:29281‐29285.
100. Fan W, Guo CY, Chen G. Flexible films of poly(3,4‐ 115. Zhang Z, Chen G, Wang H, Zhai W. Enhanced thermoelectric
ethylenedioxythiophene)/carbon nanotube thermoelectric property by the construction of a nanocomposite 3D inter-
composites prepared by dynamic 3‐phase interfacial electro- connected architecture consisting of graphene nanolayers sand-
polymerization and subsequent physical mixing. J Mater wiched by polypyrrole nanowires. J Mater Chem C. 2015;3:
Chem A. 2018;6:12275‐12280. 1649‐1654.
101. Lee W, Kang YH, Lee JY, Jang KS, Cho SY. Improving the 116. Wang L, Liu F, Jin C, Zhang T, Yin Q. Preparation of poly-
thermoelectric power factor of CNT/PEDOT: PSS nano- pyrrole/graphene nanosheets composites with enhanced
composite films by ethylene glycol treatment. RSC Adv. 2016; thermoelectric properties. RSC Adv. 2014;4:46187‐46193.
6:53339‐53344. 117. Xin S, Yang N, Gao F, Zhao J, Li L, Teng C. Free‐standing and
102. Hsu J‐H, Choi W, Yang G, Yu C. Origin of unusual thermo- flexible polypyrrole nanotube/reduced graphene oxide hybrid
electric transport behaviors in carbon nanotube filled polymer film with promising thermoelectric performance. Mater Chem
composites after solvent/acid treatments. Org Electron. 2017; Phys. 2018;212:440‐445.
45:182‐189. 118. Du Y, Niu H, Li J, et al. Morphologies tuning of polypyrrole
103. Wang J, Cai K, Shen S, Yin J. Preparation and thermoelectric and thermoelectric properties of polypyrrole nanowire/gra-
properties of multi‐walled carbon nanotubes/polypyrrole phene composites. Polymers. 2018;10:1143.
composites. Synth Met. 2014;195:132‐136. 119. Zhang K, Zhang Y, Wang S. Enhancing thermoelectric prop-
104. Liang L, Chen G, Guo C‐Y. Enhanced thermoelectric perfor- erties of organic composites through hierarchical nanos-
mance by self‐assembled layered morphology of polypyrrole tructures. Sci Rep. 2013;3:3448.
nanowire/single‐walled carbon nanotube composites. Compos 120. Yoo D, Kim J, Lee SH, et al. Effects of one‐and two‐
Sci Technol. 2016;129:130‐136. dimensional carbon hybridization of PEDOT: PSS on the
ZHANG ET AL. | 435
power factor of polymer thermoelectric energy conversion 136. Antar Z, Feller JF, Noël H, Glouannec P, Elleuch K. Ther-
devices. J Mater Chem A. 2015;3:6526‐6533. moelectric behaviour of melt processed carbon nanotube/
121. Wang Y, Yang J, Wang L, Du K, Yin Q, Yin Q. Polypyrrole/ graphite/poly(lactic acid) conductive biopolymer nano-
graphene/polyaniline ternary nanocomposite with high ther- composites (CPC). Mater Lett. 2012;67:210‐214.
moelectric power factor. ACS Appl Mater Interfaces. 2017;9: 137. Hewitt CA, Kaiser AB, Roth S, Craps M, Czerw R,
20124‐20131. Carroll DL. Multilayered carbon nanotube/polymer com-
122. Liang L, Fan J, Wang M, Chen G, Sun G. Ternary thermo- posite based thermoelectric fabrics. Nano Lett. 2012;12:
electric composites of polypyrrole/PEDOT: PSS/carbon na- 1307‐1310.
notube with unique layered structure prepared by one‐ 138. Pang H, Piao Y‐Y, Tan Y‐Q, Jiang G‐Y, Wang J‐H, Li Z‐M.
dimensional polymer nanostructure as template. Compos Sci Thermoelectric behaviour of segregated conductive polymer
Technol. 2020;187:107948. composites with hybrid fillers of carbon nanotube and bis-
123. Wang S, Liu F, Gao C, et al. Enhancement of the thermoelectric muth telluride. Mater Lett. 2013;107:150‐153.
property of nanostructured polyaniline/carbon nanotube compo- 139. Paleo A, Vieira E, Wan K, et al. Negative thermoelectric
sites by introducing pyrrole unit onto polyaniline backbone via a power of melt mixed vapor grown carbon nanofiber poly-
sustainable method. Chem Eng J. 2019;370:322‐329. propylene composites. Carbon. 2019;150:408‐416.
124. Cho C, Stevens B, Hsu JH, et al. Completely organic multi- 140. Liebscher M, Gärtner T, Tzounis L, et al. Influence of the
layer thin film with thermoelectric power factor rivaling in- MWCNT surface functionalization on the thermoelectric
organic tellurides. Adv Mater. 2015;27:2996‐3001. properties of melt‐mixed polycarbonate composites. Compos
125. Martin J, Nolas G, Zhang W, Chen L. PbTe nanocomposites Sci Technol. 2014;101:133‐138.
synthesized from PbTe nanocrystals. Appl Phys Lett. 2007;90: 141. Tzounis L, Gärtner T, Liebscher M, et al. Influence of a cyclic
222112. butylene terephthalate oligomer on the processability and
126. Cho C, Wallace KL, Tzeng P, Hsu JH, Yu C, Grunlan JC. Out- thermoelectric properties of polycarbonate/MWCNT nano-
standing low temperature thermoelectric power factor from composites. Polymer. 2014;55:5381‐5388.
completely organic thin films enabled by multidimensional con- 142. Luo J, Cerretti G, Krause B, et al. Polypropylene‐based melt
jugated nanomaterials. Adv Energy Mater. 2016;6:1502168. mixed composites with singlewalled carbon nanotubes for
127. Cho C, Culebras M, Wallace KL, et al. Stable n‐type thermo- thermoelectric applications: switching from p‐type to n‐type
electric multilayer thin films with high power factor from by the addition of polyethylene glycol. Polymer. 2017;108:
carbonaceous nanofillers. Nano Energy. 2016;28:426‐432. 513‐520.
128. Cho C, Bittner N, Choi W, Hsu JH, Yu C, Grunlan JC. 143. Tzounis L, Hegde M, Liebscher M, et al. All‐aromatic
Thermally enhanced n‐type thermoelectric behavior in com- SWCNT‐polyetherimide nanocomposites for thermal energy
pletely organic graphene oxide‐based thin films. Adv Electron harvesting applications. Compos Sci Technol. 2018;156:
Mater. 2018;5:1800465. 158‐165.
129. Suemori K, Watanabe Y, Hoshino S. Carbon nanotube bun- 144. Krause B, Barbier C, Levente J, Klaus M, Pötschke P.
dles/polystyrene composites as high‐performance flexible Screening of different carbon nanotubes in melt‐mixed
thermoelectric materials. Appl Phys Lett. 2015;106:113902. polymer composites with different polymer matrices for
130. Zhang Y, Zheng D, Pang H, Tang JH, Li ZM. The effect of their thermoelectrical properties. J Compos Sci. 2019;3:
molecular chain polarity on electric field‐induced aligned 106.
conductive carbon nanotube network formation in polymer 145. Krause B, Bezugly V, Khavrus V, Ye L, Cuniberti G,
melt. Compos Sci Technol. 2012;72:1875‐1881. Pötschke P. Boron doping of SWCNTs as a way to enhance the
131. Zhang Y‐C, Huang Y‐F, Dai K, Pang H, Chen C, Li Z‐M. The thermoelectric properties of melt‐mixed polypropylene/
resistivity response of an anisotropically conductive polymer SWCNT composites. Energies. 2020;13:394.
composite with in‐situ conductive microfibrils during cooling. 146. Krause B, Konidakis I, Arjmand M, et al. Nitrogen‐doped
Polym Plast Technol Eng. 2011;50:1511‐1514. carbon nanotube/polypropylene composites with negative
132. Piao M, Na J, Choi J, et al. Increasing the thermoelectric power Seebeck coefficient. J Compos Sci. 2020;4:14.
generated by composite films using chemically functionalized 147. Göldel A, Kasaliwal G, Pötschke P. Selective localization and
single‐walled carbon nanotubes. Carbon. 2013;62:430‐437. migration of multiwalled carbon nanotubes in blends of
133. Yu C, Kim YS, Kim D, Grunlan JC. Thermoelectric behavior polycarbonate and poly(styrene‐acrylonitrile). Macromol
of segregated‐network polymer nanocomposites. Nano Lett. Rapid Commun. 2009;30:423‐429.
2008;8:4428‐4432. 148. Pötschke P, Pegel S, Claes M, Bonduel D. A novel strategy to
134. Kim YS, Kim D, Martin KJ, Yu C, Grunlan JC. Influence of incorporate carbon nanotubes into thermoplastic matrices.
stabilizer concentration on transport behavior and thermo- Macromol Rapid Commun. 2008;29:244‐251.
power of CNT‐filled latex‐based composites. Macromol Mater 149. Zhang Y, Fuentes CA, Koekoekx R, et al. Spreading dynamics
Eng. 2010;295:431‐436. of molten polymer drops on glass substrates. Langmuir. 2017;
135. Moriarty GP, Wheeler JN, Yu C, Grunlan JC. Increasing the 33:8447‐8454.
thermoelectric power factor of polymer composites using a 150. Kim N, Lienemann S, Petsagkourakis I, et al. Elastic con-
semiconducting stabilizer for carbon nanotubes. Carbon. ducting polymer composites in thermoelectric modules. Nat
2012;50:885‐895. Commun. 2020;11:1‐10.
436 | ZHANG ET AL.
AUTHOR BIOGRAPHIES
currently a fellow of the Royal Society of Chemistry.
He has published more than 365 papers and 5 patents
Yichuan Zhang received his bachelor's (H‐index: 72).
degree in 2009 and master's degree in
2012 from Sichuan University, China. Guangming Chen received his BS
Later, he worked in the industry for over degree in 1994 at the Beijing Institute
2 years. Then, he received his PhD of Technology (BIT), MS degree in
degree from KU Leuven, Belgium and PhD degree 1997, and PhD degree in 2000 at
from University of Mons, Belgium, in 2019. He Institute of Chemistry, Chinese Acad-
subsequently joined the Chen group as an Associate emy of Sciences (ICCAS). He did a postdoctoral
Researcher at College of Materials Science and research as STA Fellow in 2000‐2002 at the
Engineering, Shenzhen University. His research National Institute for Materials Science (NIMS),
mainly contains thermoelectric composites, wettabil- Japan. He worked at ICCAS in 2002‐2018. Then, he
ity, and conductive polymer composites. joined Shenzhen University as a Distinguished
Professor. His recent research interests focus on
Qichun Zhang is an Associate Professor organic/inorganic thermoelectric composites and
at the School of Materials Science and their flexible devices.
Engineering at Nanyang Technological
University, Singapore. In 2014, he was
promoted to Associate Professor with
tenure. His research interests include rich‐carbon How to cite this article: Zhang Y, Zhang Q,
conjugated materials and their applications (novel Chen G. Carbon and carbon composites for
organic conjugated materials for optoelectronic and thermoelectric applications. Carbon Energy.
semiconductor devices, novel electrode materials as 2020;2:408–436. https://doi.org/10.1002/cey2.68
well as new design for microbial fuel cell). He is

Carbon Energy - 2020 - Zhang - Carbon and Carbon Composites For Thermoelectric Applications

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Carbon Energy - 2020 - Zhang - Carbon and Carbon Composites For Thermoelectric Applications

Uploaded by

Copyright:

Available Formats

Received: 7 June 2020 | Revised: 24 June 2020 | Accepted: 27 June 2020

Carbon and carbon composites for thermoelectric

Yichuan Zhang1 | Qichun Zhang2,3 | Guangming Chen1

1 | INTRODUCTION very regrettable, and therefore, it is highly and urgently

408 | wileyonlinelibrary.com/journal/cey2 Carbon Energy. 2020;2:408–436.

F I G U R E 7 A, TEM image of polyaniline (PANI)/MWCNT composite. B, C, S and PF of PANI/MWCNT composite. Reproduced

F I G U R E 9 A, CNT is coated by poly(3,4‐ethylenedioxythiophene) (PEDOT):PSS particles, making CNT‐PEDOT:PSS‐CNT

F I G U R E 1 0 A, Schematic illustration of preparation for PEDOT:PSS/MWCNT via a template‐directed in situ polymerization

F I G U R E 1 2 A, Schematic diagram of semi‐interpenetrating networks (SIPNs) constructed by chemically bonded

You might also like