Polymer semiconductor

crystals
One of the long-standing challenges in the field of polymer
semiconductors is to figure out how long interpenetrating and entangled
polymer chains self-assemble into single crystals from the solution
phase or melt. The ability to produce these crystalline solids has
fascinated scientists from a broad range of backgrounds including
physicists, chemists, and engineers. Scientists are still on the hunt for
determining the mechanism of crystallization in these information-rich
materials. Understanding the theory and concept of crystallization of
polymer semiconductors will undoubtedly transform this area from an
art to an area that will host a bandwagon of scientists and engineers.
In this article we describe the basic concept of crystallization and
highlight some of the advances in polymer crystallization from crystals to
nanocrystalline fibers.
Jung Ah Lim, Feng Liu, Sunzida Ferdous, Murugappan Muthukumar, and Alejandro L. Briseno*
Polymer Science and Engineering Department, University of Massachusetts, Amherst, MA 01003, USA
*E-mail: [email protected]

From a fundamental viewpoint, polymer semiconductor single

highly organized solids will one day help answer the age-long

crystals are critical for understanding the physics of polymer

question regarding the mechanism of charge transport in

crystallization. They are also important tools for elucidating

conjugated polymers1. From a technical point of view, polymer

macromolecular interactions and solid-state packing in benchmark

crystals may well play an important role in consumer electronic

materials, i.e., poly(3-hexylthiophene)(P3HT), and in newly

applications such as in flexible displays based on high mobility

synthesized polymer semiconductors. Knowledge in this area has

transistors, solar cells, gadgets and toys, or possibly in technical

flourished to a point where structure-property studies have taught

applications not yet realized by us.

and enabled chemists to design materials with tunable properties

14

Organic single crystals grown from small-molecule semiconductors

for a variety of applications. Polymer single crystals are also

have attracted much attention in recent years because of their

important for studying intrinsic charge transport and determining

utilization in both fundamental and applied science2,3. However,

the performance limitation of the material in question. These

compared to organic crystals, there is very limited literature on polymer

MAY 2010 | VOLUME 13 | NUMBER 5

ISSN:1369 7021 Elsevier Ltd 2010

Polymer semiconductor crystals

REVIEW

semiconductor single crystals or highly crystalline self-assembled

breakthroughs in polymer crystallization, at least within the limitation

nanostructures. These solids are not to be confused with the well

of this context.

established conventional polymer crystals fabricated from soft chain

polymers such as polyethylene, which have been recorded in literature

Frustration in polymer crystallization

for well over 50 years. Polydiacetylenes, a unique family of acetylene-

The requisite of morphological purity of polymer crystals fabricated

containing polymer chains, are the only known examples of conjugated

for harnessing the desired electronic properties is commonly faced

polymers although it can easily form macroscopic single crystals by

with the difficult task of polymer crystallization itself. The primary

polymerization of the crystal of diacetylene monomers4,5. This paper,

challenge arises from the length of polymer chains that self-assemble

however, focuses on conjugated polymer semiconductors that form

into crystals. When collections of interpenetrating chains try to

highly crystalline solids. To date, a detailed description and mechanism

assemble together, the free energy landscape is very rich and there

of polymer semiconductor crystallization is still not reported in

are multiple free energy barriers due to the formation of contacts

literature although polymer crystallization has been described by

of monomers (i.e. polymer chains) without significant registry of

various models and theories for well over 70 years6-8. This is because

chain orientations6,8. Fig. 1a gives a illustrative description of the

most of the current work on this topic is concentrated on materials

free energy landscape for the polymer crystallization process. The

such as regioregular poly(3-alkythiophenes), which introduces new

troughs correspond to the metastable state in the crystallization

parameters that significantly affect the physics of crystallization. For

process. They come from the competition of several nuclei for the

example, these conjugated polymers contain electron-rich backbones

acquisition of monomers or chain segments from polymer chains not

and aliphatic chains that have unique handles in crystallization kinetics.

yet incorporated into the crystalline phase. The peaks correspond to

The challenge is to get bundles of interpenetrating and entangled

the energy barrier that separate two metastable states. These barriers

conjugated chains to order into a crystal. However, the problem is

are due to the free energy cost involved in the rearrangement of chain

that polymer crystallization is frustrated by the large free energy

conformations originally distributed among many nuclei into fewer

barriers associated with the reorganization of polymer conformations

nuclei with greater crystalline order. There is a hierarchy of metastable

into ordered

states8.

Nevertheless, the theory and principles of

states before the final step of crystallization is reached. This entropic

crystallization should be quite applicable to that of these modern

part is further modulated by energy contributions arising from

polymer semiconductors.

stacking, hydrophobic interactions among the aliphatic chains, and

In this article, we discuss a basic concept on polymer semiconductor

the skeletal bond energies from the local conformation of the chain

crystallization. We also highlight some of the recent advances in

backbone. Further, since the self-assembled crystal is present in the

polymer crystallization with emphasis on single crystals, highly ordered

solution growth medium, there are interfacial energies between the

1-dimensional self-assembled nanostructures, and highly crystalline

various facets of the crystal and the medium. In order to illustrate the

films. Because the poly(3-alkylthiophenes) (P3ATs) are among one

key concepts underlying the spontaneous selection of the morphology

of the most promising and interesting polymer semiconductors

of a single crystal grown from solutions, let us consider a crystal

to crystallize, we pay close attention to literature work in this

prism of lengths L1, L2, and L3, as sketched in Fig. 1b. The molecular

article. Since this area of research has produced a large quantity

interpretation of this sketch is illustrated in Fig. 1c, where a labeled

of publications, it is not possible to examine all the interesting

chain can fold several times into a lamella-like organization typical

(a)

(b)

(c)

Fig. 1 (a) Sketch of a free energy landscape for polymer crystallization, (b) a 1-D crystal prism labeled with the lengths (L) and interfacial surface energies (), and
(c) an illustration of the possible self-organized chain packing in a P3HT single crystal labeled with inter- and intramolecular energies ().

MAY 2010 | VOLUME 13 | NUMBER 5

15

REVIEW

Polymer semiconductor crystals

of polymer crystals. The energy of a bond along the conjugated chain

Cho and coworkers9 reported that poly(3-hexylthiophene) (P3HT)

axis is 1, while in the orthogonal direction along the crystal axis

(molecular weight (MW)=54 000 g/mol, regioregularity (RR)=98.5%)

the energy per a pair of monomers is 3, due to stacking. In the

1-D microwire single crystals can be grown by the self-seeding method

other orthogonal direction to the chain axis, the energy per a pair of

where a dilute P3HT solution in CHCl3 (40C) is slowly cooled to

monomers is 2, due to the hydrophobic interactions. Let the interfacial

a crystallization temperature of 10C over three days. P3HT single

energy of the folded surfaces be 3, which arises from both the

crystals grown by this method exhibited well-defined facets and a

entropic part associated with the chain folding and the energetic part

rectangular cross-section (Fig. 2a). Selected area electron diffraction

associated with the polymer-solvent interaction. The corresponding

(SAED) patterns of crystals showed Bragg diffraction spots, not

interfacial energies for the end faces (containing aromatic groups,

arc-shaped or ring-type diffraction which are often observed from

with surface area L2L3) and the side faces (containing the aliphatic

nanostructures with preferred orientation or polycrystalline ordering.

tails, with surface area L1L3) are 1 and 2, respectively. Theoretical

The fact that diffraction spots are observed signify that the solid is of

arguments7 lead to the following result for the equilibrium anisotropy

single-crystalline nature. It is noteworthy to mention that the P3HT

of the crystal as dictated by the above parameters and the following

chain axis within the single crystal is perpendicular to the long wire axis

expression:

and - stacking direction is along the crystal growth direction. What

is not certain is if the P3HT chains are extended or folded. Preferential

L1
L2
L3

=
=

3
2

1 + 2 + 3 + 1
2
2
2

- stacking in 1-D P3HT single crystals enable charge carrier transport

along the wire axis, resulting in high conductivity along the wire
direction. In one report, single crystals of poly(3-butylthiophene)
(P3BT) (MW=16 000 g/mol, RR=97%) were obtained from solvent

Any aspect ratio of the single crystal can thus, in principle, be tuned

mixtures containing a good solvent to provide chain movement

by designing the chemistry in obtaining the pertinent values of the

and a poor solvent to enhance molecule-molecule interactions10. In

parameters. The kinetics of the assembly can similarly be controlled by

comparison with the P3HT single crystal, P3BT chains within the single

navigating through the free energy landscape, which is also calculated

crystal packed normal to the substrate and alkyl side chains were along

by using the above energy parameters and the chain entropy. We also

the crystal growth direction (data not shown). Poly(3-octylthiophene)

note that the above parameters can be normalized to accommodate

(P3OT) (MW=120 000 g/mol, RR=98.5%) single crystals were also

energies associated with templating substrates (e.g. SiO2).

obtained by a solvent vapor annealing process11. It was shown that

P3OT chains in a single crystal pack with the polymer backbone parallel

Polymer semiconductor single crystals

to the length of the crystal axis and - stacking perpendicular to

In general, the observed morphology of single [nano] crystals can

crystal growth direction (Fig. 2b). The crystal structure was verified by

be correlated to the molecular structure by electron diffraction

TEM electron diffraction.

studies so long as one can acquire the lattice parameters- which

16

More recently, it was shown that self-assembly from dilute

can be determined by x-ray crystallography. The geometry of the

solutions can produce single crystal nanowires from a rigid rod polymer

TEM electron diffraction pattern (i.e. reciprocal lattice points) should

semiconductor, a poly(para-phenylene ethynylene) derivative with

in principal correspond to the geometry of the crystals unit cell

thioacetate end groups (TA-PPE) (MW=51 318 g/mol)12. TA-PPE

(e.g., orthorhombic for P3HT) when viewing down the zone axis. In

nanowires exhibited single-crystal diffraction spots and the polymer

other words, one should be able to see that the geometry of the

backbone of TA-PPE chains within the nanowire packed parallel to

reciprocal unit cell resembles the geometry of a rotated real space

the nanowire long axis with their side chains vertical to the substrate

unit cell. For example, if the d-spacing of the (010) and (001) planes

(Fig. 2c).

can be experimentally determined, then the in-plane unit cell angle

Based on these results for polymer semiconductor single crystals, it

() can also be easily measured. From these three parameters, one

is proposed that polymer chains in a single crystal have fully extended

can calculate the d-spacing from the equation rd=L; where r is the

conformations. However, the packing orientation, crystal morphologies,

distance between diffraction spots, d is the spacing between lattice

and relation between crystal growth direction and -stacking directions

planes, L is the camera length, and is the electron wavelength.

appear to be significantly different among polymer materials. This

Moreover, the b and c unit cell values can be determined by the

suggests that molecular weight, side chain length, rigidity of chain

following expression: |b|=1/(d010sin()) and |c|=1/(d001sin()). The

backbone, and crystal growth conditions such as choice of solvent,

consistency of measured values to that of the bulk unit cell dimensions

all have a profound effect on the crystallization kinetics. Further

suggest that the molecular or chain packing within the crystal is nearly

investigation on the single crystal growth mechanism including origin

the same as the bulk crystal structure and therefore the long axis of a

of nucleation, crystallization kinetics, and interaction between polymer

1-D polymer crystal can be accurately elucidated.

chains will be required. An important question is whether the c-axis of

MAY 2010 | VOLUME 13 | NUMBER 5

Polymer semiconductor crystals

(b)

(a)

REVIEW

(c)

Fig. 2 Field-emission SEM images and SAED patterns for single crystals of (a) poly(3-hexylthiophene) (Reprinted with permission from9. 2006 Wiley-VCH), (b)
poly(3-octylthiophene) (Reprinted with permission from11. 2009 American Chemical Society), and (c) thioacetate substituted poly(para-phenylene ethynylene)
(Reprinted with permission from12. 2009 American Chemical Society). TEM images for individual crystals along with corresponding SAED analyses are shown in the
insets. Schematic illustrations of polymer chain packing in the respective single crystals are also represented below the images.

the polymer chain undergoes chain folding or it remains extended in

interaction between the solvent and aromatic main chain16,17. In

the solid state. Further investigations will await a better understanding

a seminal paper by Ihn and coworkers18, they reported that P3HT

of the nature of chain packing in polymer semiconductor single

crystallized into nanowhiskers from dilute solution in a poor solvent

crystals.

(i.e. cyclohexanone, n-decane), in which the polymer dissolved only at

elevated temperatures. The P3HT nanowhiskers exhibited an average

One-Dimensional (1-D) crystalline nanowires

width of 15~20 nm, a length of 10 m, and a thickness of 5 nm which

Polymer semiconductors can often have a semi-crystalline nature

corresponds to two or three P3HT chain layers (Fig. 3a). X-ray and

due to a relatively rigid -conjugated backbone. Compared to the

electron diffraction results confirmed that the P3HT chains pack normal

crystallization of typical insulating polymers such as polyethylene (PE)

to the long axis of the whisker. Interestingly, P3HT chain folding in

which form two-dimensional (2-D) lamellar crystals, -conjugated

a nanowhisker structure was expected because of the considerably

polymer semiconductors generally crystallize into one-dimensional

greater contour length of P3HT (65 nm for MW 29 000 g/mol) than

(1-D) nanostructures, such as nanowires via - interactions and via

the whisker width (15nm), although the relatively rigid backbone of

hydrophobic interactions associated with the aliphatic chains13. In

P3AT assumed to hamper chain folding. In fact, it has been shown by

studies on crystallization kinetics of P3ATs, reported by Nandi and

scanning tunneling microscopy (STM) that the polythiophene backbone

coworkers14, P3ATs isothermally crystallized into 1-D whiskers, and

can be folded in a planar conformation19. The charge carrier mobility in

the Avrami

exponent15

varied in the range of 0.6~1.4, indicating

field-effect transistors from a single P3HT nanowire has been reported

one-dimensional heterogeneous nucleation with linear growth. The

as high as 0.06 cm2/Vs which is comparable to that of thin films

Avrami equation, x(t) = 1-exp(-ktn), is a quantitative description of

(~10-3 cm2/V-1s-1)20.

the isothermal crystallization kinetics, where x(t) is the crystallinity

Results of more recent work demonstrated that dimensions of

developed at time t, k is the overall rate constant, and n is the Avrami

P3HT nanowires can be controlled by varying the molecular weight

exponent denoting the nature of the nucleation and growth process15.

and concentration of solutions21. When the molecular weight of P3HT

In general, self-assembly of -conjugated polymers in a dilute

was less than a critical molecular weight (number average molecular

solution under limited solubility can afford highly crystalline 1-D

weight (Mn) ~10 kDa), the width of P3HT nanowire corresponded

nanowires. At partially soluble conditions, conjugated molecules tend

to the length of the extended polymer chain. Whereas for higher

to aggregate in a face-to-face stacking fashion (via - interaction

molecular weight, the width of nanowire remains constant about

between aromatic backbones) in order to decrease unfavorable

15 nm, regardless of increase in Mn, (Fig. 3b). This result suggested that

MAY 2010 | VOLUME 13 | NUMBER 5

17

REVIEW

Polymer semiconductor crystals

(a)

(b)

Fig. 3 (a) TEM image of P3HT nanowhiskers crystallized from a cyclohexanone solution and the corresponding schematic representation of the chain arrangement
in a nanowire. (inset: corresponding SAED pattern) (Reprinted with permission from18. 1993 American Chemical Society) (b) TEM images of P3HT nanowires and
schematic illustration of the extended and folded P3HT backbone packing prepared from low Mn (left) and high Mn (right). (Inset: SAED pattern for P3HT nanowire
at molecular weight of 35 kDa) (Reprinted with permission from21. 2009 American Chemical Society).
(b)

(a)

Fig. 4 (a) TEM image of a shish-kebab nanofiber of P3HT and a schematic illustration of the shish-kebab chain packing. The P3HT chain is oriented parallel to the
fiber axis in shish, and perpendicular to the nanofiber in kebab. (Reprinted with permission from22. 2009 Wiley-VCH) (b) TEM image of an individual BBL nanobelt
(inset: the corresponding SAED pattern) and the chain packing of BBL polymer within a nanobelt. (Reprinted with permission from28. 2008 American Chemical
Society).

18

MAY 2010 | VOLUME 13 | NUMBER 5

Polymer semiconductor crystals

REVIEW

morphologies were prepared by epitaxial crystallization in a mixture of

(a)

(b)

1,3,5-trichlorobenzene (TCB) and a poor solvent of pyridine, as shown

in Fig. 4a. In a shish fiber, P3AT chains were oriented parallel to the
long axis of TCB needles that were preferably formed in pyridine. In a
kebab fibril, folded P3AT chains packed with the pi-stacking direction
parallel to the fibril axis.
Formation of 1-D nanowires for various polymer semiconductors,
such as polyphenylenevinylenes (PPV), polyfluorenes (PF), poly[(9,9
-dioctylfluorenyl-2,7-diyl)-co-(bithiophene)] (F8T2), and poly(3,3didodecyl quarter thiophene) (PQT-12), have also been achieved by

(c)

(d)

self-assembly in a solution with controlled solubilities23-27. A recent

example of poly(benzobisimidazobenzophenanthroline) (BBL) 1-D
nanobelts prepared by self-assembly in solution showed that highly
planar and rigid chains pack with fully stretched structures28. Unlike
polythiophenes, BBL polymer chains preferably packed face-to-face
perpendicular to the nanobelt axis without chain folding, as shown in
Fig. 4b.
Solvent vapor treatment in a film formation process produces 1-D
polythiophene nanowires29. P3HT thin films spun-cast under solvent

(e)

vapor pressure exhibited the nanowire structure (widths of 20~22

nm and heights of 4-5 nm) and nanowire length was controlled
by solvent vapor pressure (0~56.5 kPa). It was reported that the
crystalline structure and morphology of polythiophene nanowires can
be changed by carbon disulfide (CS2) solvent vapor annealing (Fig. 5)30.
Pristine P3BT thin films were composed of typical nanowire crystals
in which the polythiophene chains pack parallel to the substrate
and perpendicular to the long wire axis. However, upon exposure of
pristine films to CS2 vapor, the morphologies of the films significantly
evolve from nanowire crystals to spherulites which consist of whiskers

Fig. 5 (a) TEM image of pristine P3BT film spin-coated from an

o-dichlorobenzene solution; (b) Polarized light microscopy (PLM) image
showing the P3BT film composed of closely contacted spherulites upon
CS2 vapor treatment; (c) TEM image demonstrating the fine structure of
the P3BT spherulites and (d) the SAED pattern from the area marked in (c).
(e) Schematic illustration of the parallel and perpendicular alignments of
P3BT backbones to the substrate, corresponding to edge-on and flat-on
orientations in the conventionally prepared and CS2-vapor related thin film,
respectively. (Reprinted with permission from30. 2007 Wiley-VCH).

radiating from the center. Interestingly, in this whisker crystal, the

P3BT chains (c-axis) are oriented perpendicular to the substrate
and the whisker grows along the - stacking direction (b-axis), as
schematically illustrated in Fig. 5e.
Polymer semiconductor copolymers or blends combining different
components can be one efficient route to generate self-assembly of
polymer semiconductor into 1-D crystalline nanostructures. Diblock
copolymer of poly(3-hexylthiophene)-b-poly(styrene), synthesized
by McCullough and coworkers, formed well-defined nanowires with a

conformation of P3HT chain in nanowhisker begins to change from

30-40 nm width that corresponded to the fully extended polythiophene

fully extended to folded at a critical Mn. In addition, for fully extended

block, as shown in Fig. 6a31. For the possible mechanism of this

chains (i.e. at low Mn), P3HT nanowires can transform into nanoribbon

nanowire formation, the core-shell architecture with the polythiophene

structures with increased concentration.

component constituting the core was expected due to the immiscibility

Appropriate choices of solvents can also induce unique 1-D

of P3HT and poly(styrene). Nanofibrillar structures have also been

nanocrystalline structures. Kiriy and coworkers17 demonstrated

found in diblock copolymers of P3HT and rubbery amorphous

that the addition of hexane into P3AT/chloroform solutions,

poly(methyl acrylate)32. The transistors based on these nanofibrillar

which is the selective solvent for alkyl side chain but poor solvent

structures exhibited good charge carrier mobilities approaching that of

for polythiophene backbones, results in 1-D aggregation of P3ATs

P3HT thin film devices. Jenekhe and coworkers33 recently demonstrated

with helical conformation of main chain. Recently, Brinkmann and

that crystalline-crystalline diblock co-poly(3-alkylthiophene)s self-

coworkers22 reported that the P3AT nanofibers showing shish-kebab

assembled into crystalline nanowires, as shown in Fig. 6b. The detailed

MAY 2010 | VOLUME 13 | NUMBER 5

19

REVIEW

Polymer semiconductor crystals

(b)

(a)

Fig. 6 (a) Tapping mode AFM image (phase) for a poly(3-hexylthiophene)-b-poly(styrene) copolymer thin film prepared from toluene (P3HT fraction: 37%).
(Reprinted with permission from31. 2002 Wiley-VCH ) (b) TEM image and schematic illustration of lamellar structure of poly(3-butylthiophene)-b-poly(3octylthiophene) copolymer (P3BT fraction: 50%) (Reprinted with permission from33. 2009 American Chemical Society)

chain packing within the nanowires remains to be investigated,

trapping of charges at the boundaries of rod-like crystalline domains.

however, they speculate that melt-phase self-assembly of diblock

For the high MW films, it was suggested that crystalline domains are

co-poly(3-alkylthiophene)s into lamellar structure has two crystalline

interconnected by the long polymer chains and thus enable more

domains of the two different side chains. For a blend system, it was

efficient charge transport to occur across grain boundaries. In a separate

reported that polythiophenes/poly(styrene) blend films prepared

study, Neher and coworkers38,39 suggested that P3HT crystalline

from selectively marginal solvent for polythiophene component show

whiskers from low MW P3HT (Mn<3 kDa) are embedded within a

the polythiophene nanowires embedded in amorphous polystyrene

disordered matrix (figure not shown). Neher also proposed that high

matrix34,35.

MW P3HT (Mn>27 kDa) films consist of partially ordered domains in

which the chains either fold back, interconnect to neighboring domains,

Nanocrystalline thin films

or extend into a disordered phase. McCullough and coworkers40 also

Polymer semiconductor thin films prepared by solution cast processes

demonstrated that the charge carrier mobilities in transistors fabricated

are composed of both polycrystalline and amorphous domains. Since

from well-ordered P3HT films increase exponentially with nanofiber

charge transport in polymer semiconductor thin films is limited

width which corresponds to the weight average contour length of

by amorphous regions, considerable efforts have been devoted to

polymer chains with Mn ranging from 2 to 13 kDa. The propensity of

enhancing crystalline domains. It has been reported that polymer

molecular weight dependency on crystallinity and charge transport

crystallization is strongly depended on molecular weight during the

properties has also been observed for poly(2,5-bis(alkylthiophen-2-

film forming process. There is a speculation that low MW chains find

yl)thieno[3,2-b]thiophene) (PBTTT) synthesized by McCulloch and

an equilibrium position more easily and produce ordered structures

coworkers41,42. Higher molecular weight PBTTT (Mn~30 kDa) exhibited

compared to high MW chains because high MW chains have slow

a more three-dimensional microstructure with larger crystalline domains

crystallization kinetics due to higher viscosity or chain entanglement.

and resulted in one of the highest charge carrier mobilities reported to

McGehee and coworkers36,37 reported that low MW P3HT (Mn<4 kDa)

date (0.2-0.6 cm2V-1s-1).

films form rod-like crystalline nanostructures, while high MW

20

A variety of polymer semiconductors have solubilizing side chains

P3HT (Mn>30 kDa) films have isotropic nodule structures with low

attached to the infusible conjugated backbone to improve solution

crystallinity, as shown in Fig. 7a. It is interesting, however, that the

processability. The crystallization of these polymer semiconductors can

field-effect mobilities in transistors increase with increasing MW. It was

be affected by side chain structures. Bao and coworkers43 and later

suggested that the low mobility of low-MW films is probably due to the

Sirringhaus and coworkers44 reported that P3HT chain orientation

MAY 2010 | VOLUME 13 | NUMBER 5

Polymer semiconductor crystals

REVIEW

post treatment of thin film deposition. For instance, rapid evaporation

(a)

of solvents can reduce the time to enable highly order films.

Sirringhaus and coworkers46 reported that by using higher boiling
point solvents, i.e. 1,2,4-trichlorobenzene (213C), P3HT films with
highly ordered crystalline fibrils were obtained. Furthermore, Yang
and coworkers47 reported that the orientation of P3HT crystallites
were dependent on boiling point of solvents, solubility of polymer to
solvent, and specific deposition conditions. When P3HT films were
prepared from marginal (partially soluble at room temperature) low
(b)

boiling point solvents such as methylene dichloride (40C), highly

crystalline nanofibrillar structures with edge-on orientation were
observed (Fig. 7b). These nanostructured films were prepared by
both rapid (spin-coat) and slow (drop-cast) solvent drying methods.
Further enhancement of structural order in polymer semiconductor
thin films can be obtained through post-deposition annealing. In
particular, for polymer semiconductors with liquid crystalline (LC)
characteristics, such as PBTTT, PQT-12, poly(didodecylquaterthiophenealt-didodecylbithiazole), annealing the films above the liquid crystal

(c)

transition temperature forms large crystalline domains42,48,49. A

representative example of annealing effects of PBTTT thin films is
shown in Fig. 7c42. Nodule-like structures of as-prepared films were
transformed into large polycrystalline domains with 200 nm diameter
grains by annealing at 180C. More recently, it was reported that
PBTTT annealed at the second transition temperature (240C) exhibited
unique crystal shape of terraced nanoscale ribbons50. In the case of
poly(9,9-n-dioctylfluorene-alt-bithiophene) (F8T2) deposited on a
Fig. 7 (a) Tapping mode AFM images (phase) of P3HT thin films with different
molecular weight; 3.2 kDa (left) and 31.1 kDa (right). Films were prepared
by spin-casting from chloroform. (Reprinted with permission from36. 2003
Wiley-VCH) (b) AFM topographs and grazing incident x-ray diffraction patterns
(GIXRD) for P3HT thin films spin-casted from chloroform (left) and methylene
chloride (right). GIXRD pattern for chloroform indicates the face-on chain
conformation, while for the methylene it shows the edge-on conformations
against the substrate. (Reprinted with permission from47. 2007 American
Institute of physics) (c) AFM images (phase) of poly(2,5-bis(alkylthiophen-2yl)thieno[3,2-b]thiophene) (pBTTT) before (left) and after (right) annealing
at 180C. (Inset: chemical structure of pBTTT (R=C12H25) ) (Reprinted with
permission from42. 2006 Nature Publishing Group).

and coworkers53 found that octadecyltrichlorosilane (OTS) treatment

in crystallites was affected by regioregularity of hexyl side chains. In

of silicon dioxide surfaces yields a preferred orientation of P3HT

P3HT films with high regioregularity (>91%), the crystallites were

crystallites for in-plane charge transport at the buried interface

preferentially oriented with (100) axis normal to the film and the (010)

compared to hexamethyldisilizane (HMDS) treated substrates which

axis in the plane of the film, often referred to as edge-on. In contrast,

yields misoriented crystallites (Fig. 8). Rocking curve analysis was used

the chain orientation in low regioregular P3HT films (81%) showed the

to determine that crystallites nucleate at substrate-polymer interfaces

(100) axis in the plane and the (010) axis normal to the film, referred

and also to determine the degree of misorientation of crystallites

to as face-on44. There are also reported cases where sufficiently

from both HMDS and OTS treated substrates. In several reports,

low side chain attachment densities (e.g., PBTTT, PQT-12) permit

increased charge carrier mobilities in transistors from several polymer

strategic interdigitation of side chains. DeLongchamp and coworkers45

semiconductors, such as P3HT, PQT-12, PBTTT, F8T2, were investigated

suggested that side chain interdigitation of these materials provides a

by various SAM treatments among which OTS treated substrate gave

three dimensional chain ordering, and results in high charge transport

the highest mobilities54-58. For PQT-12, this enhanced charge mobility

properties in the films.

was attributed to an edge-on orientation of PQT-12 crystalline

The crystalline structures in polymer semiconductor thin films

significantly depend on how films are prepared, solvent choice, and

mechanically rubbed polyimide, annealing of the film in the LC phase

attributed to the chain alignment along the rubbing direction of the
underlying polyimide51.
Self-assembly of polymer semiconductor thin films can be
governed by the interplay between polymer-polymer interactions
and polymer-substrate interactions. Several groups have reported
that treating the substrate surface with self-assembled monolayers
(SAMs) can significantly vary the crystallite orientation52-58. McGehee

domains induced by the OTS layer57. Edge-on orientation gives rise to

two-dimensional charge transport caused by efficient - stacking in

MAY 2010 | VOLUME 13 | NUMBER 5

21

REVIEW

Polymer semiconductor crystals

(a)

(b)

(c)

(d)

Fig. 8 Rocking curves at (100) diffraction peaks for the spin-casted P3HT films with low molecular weight (a) before and after annealing for both HMDS and OTStreated substrates and (b) for various film thicknesses on OTS-treated substrate. Proposed schematic of low molecular weight P3HT film with packing arrangements
at the buried interface (c) for the film on HMDS treated-substrate; grain boundaries contact with a (100) face and (d) for the film on OTS treated substrate; grain
boundaries contact with (010) and (001) faces. (Reprinted with permission from53. 2006 Nature Publishing Group).
(a)

(b)

Fig. 9 (a) STM images (200 200 ) of a short range ordering of P3HT on HOPG. The chain-to-chain distance corresponds to 13-14 . The molecular packing
corresponding to white square is shown under the image. (Reprinted with permission from19. 2000 Wiley-VCH) (b) STM image of alkyl chain interdigitation of
PQT-12 on HOPG and Proposed schematic of chain folding resulting for the U-shaped segments. (Reprinted with permission from60. 2008 Wiley-VCH).

22

MAY 2010 | VOLUME 13 | NUMBER 5

Polymer semiconductor crystals

REVIEW

(b)

(a)

Fig. 10 (a) SEM image of the friction-transferred P3HT film on silicon substrate and schematic of packing arrangement. (Reprinted with permission from64. 2003
American Chemical Society) (b) Dark field TEM image of P3HT film oriented by directional solidification method obtained by selecting the intense 020 reflection. The
crystalline lamellae appear bright. (inset: corresponding electron diffraction). Schematic model for organization of the semicrystalline P3HT is represented under the
image. (Reprinted with permission from67. 2006 Wiley-VCH).

a parallel direction to the substrate. In contrast, for PBTTT, crystalline

Improved chain orientations have been achieved by other methods

domain have similar molecular packing structure on both OTS-treated

such as friction transfer, mechanical rubbing and directional epitaxial

and untreated SiO2 surfaces. However, OTS treated surface gives larger

solidification. P3HT and polyfluorenes (PF) have been employed on

domain size and reduce grain boundaries compared to that on SiO2

friction transferred poly(tetraflouroethylene) (PTFE) surfaces with

surfaces58.

uniaxially aligned nanoscale ridges62,63. The polymer chains aligned

Two-dimensional crystalline domains of polymer semiconductors

along the PTFE friction transfer directions. Friction transferred P3AT

have also been studied on highly oriented pyrolytic graphite (HOPG)

films onto substrates exhibited a well-ordered chain backbone along

surfaces19,59,60. The HOPG surface facilitates epitaxial growth of these

the transfer direction with alkyl side chains lying in the film plane

polymers via interaction with alkyl chains. The driving force for this is

(Fig. 10a)64. Analogous to this approach, mechanical rubbing of P3HT

the alignment of alkyl chains by van der Waals interaction in a planar,

interdigitated, zigzag way with the HOPG crystallographic axes, hence

films aligned the chain backbone along the sliding direction65,66.

Brinkmann and coworkers67 demonstrated directional epitaxial

guiding the placement of the main chain. Two-dimensional (2-D)

solidification of P3HT by using a crystallizable solvent, 1,3,5-

crystallization of P3ATs at solution/HOPG interface was observed

trichlorobenzene (TCB). Crystallization of P3HT occurred at the TCB

by means of STM19. In this study, conjugated chain folding of P3ATs

was clearly visualized, as shown in (Fig. 9a). Calculations based on

crystal surface where the chain direction (cP3HT) is parallel to long

needle axis of TCB crystals (cTCB) and forms fibrous morphology

this image confirmed that a folded chain has cis-conformation. STM

(Fig. 10b). Using P3HT films fabricated by this technique, Salleo and

analysis of P3HT films casted on HOPG exhibited that crystalline

coworkers68 demonstrated that the energy barrier for charge transport

mono domains were interconnected by folded

chains59.

More recently,

at crystallite boundaries along a fibril (cP3HT) is relatively small

interdigitation of alkyl side chains of polythiophene semiconductor

compared to fiber-fiber grain boundaries. More recently, Samitsu

was directly observed for self-assembly of PQT-12 on HOPG surface

and coworkers69 reported that unidirectional alignment of PQT-12

(Fig.

9b)60.

Studies on crystallization of poly(dioctylbithiphene-alt-

nanofibers is obtained by anisotropic crystallization in a nematic liquid

fluorenone) (PDOBTF) on HOPG exhibited less chain folding due to a

crystal matrix. Liquid crystal ordering induces polymer chain alignment

more rigid backbone compared to P3ATs61.

along the nematic direction.

MAY 2010 | VOLUME 13 | NUMBER 5

23

REVIEW

Polymer semiconductor crystals

Summary and outlook

towards this goal are becoming available, as highlighted from the

Polymer semiconductor crystallization has been a major challenge for

elegant studies exposed in this review.

several years now. The phenomenon of crystallization is significantly

There is only a hand full of reports that have demonstrated true

different than that of small molecules and from traditional soft

growth of polymer single crystals. Perhaps the use of low molecular

polymers such as polyethylene. Frustration in polymer semiconductor

weight polymers may yield better crystal growth success since the

crystallization is attributed to the multiple free energy barriers

crystallization kinetics of low molecular weight polymers may very

associated with the cost of polymer conformational changes to

well resemble the free energy landscape of small molecule crystallization

reach ordered states. A simple formulation has been described in

where fewer energy barriers exist and frustration is essentially non-

this paper which takes into account the energy contributions from

existent. This is a particularly exciting time for researchers in the field of

stacking, skeletal bond energies, hydrophobic interactions, and

polymer semiconductors to apply the facts and theories of crystallization

interfacial energies from solution growth mediums. By combining

and formulate a protocol for growth of polymer crystals.

this mathematical expression with high level molecular modeling and

24

with the modern availability of sensitive experimental techniques,

Acknowledgement

one could piece together the mechanism of polymer semiconductor

The authors acknowledge support by the NSF Materials Research Science

crystallization. Such a heroic effort will require collaboration between

and Engineering Center on Polymers at UMass Amherst (DMR-0820506),

physicists, chemists, and engineers. Fortunately, most of the results

and the Center for Hierarchal Manufacturing (CMMI-0531171).

REFERENCES

35. Qiu, L. Z., et al., Adv. Mater. (2009) 21 (13), 1349.

1. Bredas, J. L., and Street, G. B., Acc. Chem. Res. (1985) 18 (10), 309.

36. Kline, R. J., et al., Adv. Mater. (2003) 15 (18), 1519.

2. de Boer, R. W. I., et al., Phys. Status Solidi a-Appl. Res. (2004) 201 (6), 1302.

37. Kline, R. J., et al., Macromolecules (2005) 38 (8), 3312.

3. Briseno, A. L., et al., Nature (2006) 444 (7121), 913.

38. Zen, A., et al., Adv. Funct. Mater. (2004) 14 (8), 757.

4. Young, R. J., et al., Faraday Discuss. Chem. Soc. (1979) 68, 509.

39. Zen, A., et al., Macromolecules (2006) 39 (6), 2162.

5. Thakur, M., and Meyler, S., Macromolecules (1985) 18 (11), 2341.

40. Zhang, R., et al., J. Am. Chem. Soc. (2006) 128 (11), 3480.

6. Muthukumar, M., Eur. Phys. J. E (2000) 3 (2), 199.

41. McCulloch, I., et al., Adv. Mater. (2009) 21 (10-11), 1091.

7. Rice, S. A., Advances in Chemical Physics, John Wiley & Sons, Inc. (2004), 128, Ch. 1.

42. Mcculloch, I., et al., Nat. Mater. (2006) 5 (4), 328.

8. Muthukumar, M., Lecture Notes in Physics (2007) 714, 1.

43. Bao, Z., et al., Appl. Phys. Lett. (1996) 69 (26), 4108.

9. Kim, D. H., et al., Adv. Mater. (2006) 18 (6), 719.

44. Sirringhaus, H., et al., Nature (1999) 401 (6754), 685.

10. Ma, Z. Y., et al., Polymer (2007) 48 (1), 31.

45. Kline, R. J., et al., Macromolecules (2007) 40 (22), 7960.

11. Xiao, X. L., et al., J. Phys. Chem. B (2009) 113 (44), 14604.

46. Chang, J. F., et al., Chem. Mater. (2004) 16 (23), 4772.

12. Dong, H. L., et al., J. Am. Chem. Soc. (2009) 131 (47), 17315.

47. Yang, H. H., et al., Appl. Phys. Lett. (2007) 90 (17), 172116.

13. Briseno, A. L., et al., Materials Today (2008) 11 (4), 38.

48. Ong, B. S., et al., Adv. Mater. (2005) 17 (9), 1141.

14. Malik, S., and Nandi, A. K., J. Polym. Sci. Part B-Polym. Phys. (2002) 40 (18), 2073.

49. Kim, D. H., et al., J. Am. Chem. Soc. (2009) 131 (17), 6124.

15. Cheng, S. Z. D., and Wunderlich, B., Macromolecules (1988) 21 (11), 3327.

50. DeLongchamp, D. M., et al., ACS Nano (2009) 3 (4), 780.

16. Goto, H., et al., Macromolecules (2002) 35 (12), 4590.

51. Sirringhaus, H., et al., Appl. Phys. Lett. (2000) 77 (3), 406.

17. Kiriy, N., et al., Nano Lett. (2003) 3 (6), 707.

52. Kim, D. H., et al., Adv. Funct. Mater. (2005) 15 (1), 77.

18. Ihn, K. J., et al., J. Polym. Sci. Part B-Polym. Phys. (1993) 31 (6), 735.

53. Kline, R. J., et al., Nat. Mater. (2006) 5 (3), 222.

19. Mena-Osteritz, E., et al., Angew. Chem. -Int. Ed. (2000) 39 (15), 2680.

54. Majewski, L. A., et al., Appl. Phys. Lett. (2006) 88 (22), 222108.

20. Merlo, J. A., and Frisbie, C. D., J. Polym. Sci. Part B-Polym. Phys. (2003) 41 (21), 2674.

55. Sirringhaus, H., et al., Science (1998) 280 (5370), 1741.

21. Liu, J. H., et al., Macromolecules (2009) 42 (24), 9390.

56. Salleo, A., et al., Appl. Phys. Lett. (2002) 81 (23), 4383.

22. Brinkmann, M., et al., Adv. Funct. Mater. (2009) 19 (17), 2759.

57. Wu, Y., et al., Appl. Phys. Lett. (2005) 86 (14), 142102.

23. Samitsu, S., et al., Macromolecules (2008) 41 (21), 8000.

58. Chabinyc, M. L., et al., J. Am. Chem. Soc. (2007) 129 (11), 3226.

24. Samitsu, S., et al., Thin Solid Films (2008) 516 (9), 2478.

59. Grevin, B., et al., Adv. Mater. (2003) 15 (11), 881.

25. Samori, P., et al., Opt. Mater. (1998) 9 (1-4), 390.

60. Keg, P., et al., Macromol. Rapid Commun. (2008) 29 (14), 1197.

26. Grenier, C. R. G., et al., Angew. Chem. -Int. Ed. (2007) 46 (5), 714.

61. Brun, M., et al., Adv. Mater. (2004) 16 (23-24), 2087.

27. Simmonds, H., et al.,J. Vac. Sci. Technol. B (2009) 27 (2), 831.

62. Amundson, K. R., et al., Thin Solid Films (2002) 414 (1), 143.

28. Briseno, A. L., et al., Chem. Mater. (2008) 20 (14), 4712.

63. Brinkmann, M., et al., Macromolecules (2009) 42 (21), 8298.

29. Kim, D. H., et al., Macromol. Rapid Commun. (2005) 26 (10), 834.

64. Nagamatsu, S., et al., Macromolecules (2003) 36 (14), 5252.

30. Lu, G. H., et al., Adv. Mater. (2007) 19 (21), 3594.

65. Heil, H., et al., J. Appl. Phys. (2003) 93 (3), 1636.

31. Liu, J. S., et al., Angew. Chem. -Int. Ed. (2001) 41 (2), 329.

66. Derue, G., et al., J. Am. Chem. Soc. (2005) 127 (22), 8018.

32. Sauve, G., and McCullough, R. D., Adv. Mater. (2007) 19 (14), 1822.

67. Brinkmann, M., and Wittmann, J. C., Adv. Mater. (2006) 18 (7), 860.

33. Wu, P. T., et al., Macromolecules (2009) 42 (7), 2317.

68. Jimison, L. H., et al., Adv. Mater. (2009) 21 (16), 1568.

34. Lu, G. H., et al., Macromolecules (2007) 40 (18), 6579.

69. Samitsu, S., et al., Macromolecules (2009) 42 (13), 4366.

MAY 2010 | VOLUME 13 | NUMBER 5