The effect of thermal annealing on dopant site choice in conjugated polymers

Highlights

• Transparent conductive polymer S–P3MEET can be p-type doped by F4TCNQ.

• Dopant additive competes for doping sites covalently attached on the polymer.

• F4TCNQ doped S–P3MEET shows increased melting temperature of the side chains.

• F4TCNQ binds much more strongly under thermal stress in S–P3MEET than in P3HT.

• Polar side-chains in polymers increases the thermal stability of molecular dopants.

Abstract

Solution-processed organic electronic devices often consist of layers of polar and non-polar polymers. In addition, either of these layers could be doped with small molecular dopants. It is extremely important for device stability to understand the diffusion behavior of these molecular dopants under the thermal stress and whether the dopants have preference for the polar or the non-polar polymer layers. In this work, a widely used molecular dopant 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4TCNQ) was chosen to investigate dopant site preference upon thermal annealing between the polar thiophene poly(thiophene-3-[2-(2-methoxy-ethoxy)ethoxy]-2,5-diyl) (S–P3MEET) and non-polar thiophene poly(3-hexylthiophene) (P3HT). F4TCNQ is able to p-type dope both P3HT and S–P3MEET. Further doping studies of S–P3MEET using near edge X-ray absorption fine structure spectroscopy, conductivity measurements and atomic force microscopy show that the F4TCNQ additive competes for doping sites with the covalently attached dopants on the S–P3MEET. Calorimetry measurements reveal that the F4TCNQ interacts strongly with the side-chains of the S–P3MEET, increasing the melting temperature of the side-chains by 30 °C with 5 wt% dopant loading. Next, the thermal stability of doping in the polar/non-polar (S–P3MEET/P3HT) bilayer architectures was investigated. Steady-state absorbance and fluorescence results show that F4TCNQ binds much more strongly in S–P3MEET than P3HT and very little F4TCNQ is found in the P3HT layer after annealing. In combination with reflectometry measurements, we show that F4TCNQ remains in the SP3MEET layer with annealing to 210 °C even though the sublimation temperature for neat F4TCNQ is about 80 °C. In contrast, F4TCNQ slowly diffuses out of P3HT at room temperature. We attribute this difference in binding the F4TCNQ anion to the ability of the ethyl-oxy side-chains of the S–P3MEET to orient around the charged dopant molecule and thereby to stabilize its position. This study suggests that polar side-chains could be engineered to increase the thermal stability of molecular dopant position.

Introduction

Organic electronic materials and devices have recently been extensively studied because they show the potential to replace inorganic components for some applications, including organic photovoltaics (OPVs) and organic light emitting diodes (OLEDs) [1], [2], [3]. Organic electronic devices often consist of multiple thin (5–100 nm) layers, each of which has a specific electronic and/or optical function. It is tempting to assume that the device will operate as if each layer is discrete, pure, and has sharp interfaces. In practice, however, it is difficult to control the composition and morphology of these layered materials because of chemical reactions at layer interfaces and diffusion of mobile species through the interface [4], [5]. Since the thermal stability of the layers and interfaces are essential in determining the quality and stability of the devices, the characterization of organic layer properties in connection with thermal stress is imperative in understanding the mechanisms involved in device degradation [6], [7].

Many solution-processed organic electronic devices incorporate a polar (hydrophilic) layer and a non-polar (hydrophobic) layer because the use of orthogonal solvents allows layer stacking (Fig. 1). In fact, the most widely used material in solution-processed organic electronic devices is the transparent conductive polymer poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS) because of this layering ability [8], [9], [10], [11]. However, even covalently bound ionic dopants have been shown to cause inadvertent doping of adjacent layers. Excess polymeric dopant (PSS) has a tendency to phase segregate into a surface layer on top of the PEDOT:PSS mixture, and after heat treatment mix with and dope the adjacent non-polar semiconducting polymer layer [12], [13], [14], [15]. A similar effect has also been observed in mixtures of another polymeric dopant, perfluorinated ionomer (PFI), with poly(thiophene-3-[2-(2-methoxy-ethoxy)ethoxy]-2,5-diyl) (S–P3MEET). It was shown that PFI tends to phase separate out of the polymer due to its low miscibility in conjugated polymers, forming a PFI-rich skin between S–P3MEET and poly(3-hexylthiophene) (P3HT) layers. In this case, the location of the dopant remains stable under considerable thermal stress, which implies that reduced mixing can hinder undesired doping [16]. This relationship between ionic dopant miscibility and contact doping across a polar/non-polar interface was recently quantified for a variety of self-doped polymers [17].

By comparison, molecular dopants are more traditionally used in small molecule evaporated devices. Since the dopant is a separate molecule, it can be co-evaporated with a hole or electron transport material and causes doping by forming charge transfer salts in-situ in specific layers [18], [19]. This dopant type is used to tune the work function and/or induce selective n-type or p-type charge transport [20]. A disadvantage, however, is that these dopants are not covalently bound into position and thereby can diffuse into other layers within the device under thermal or electronic stress. The weaker van der Waals or Coulombic forces at play in these samples are not sufficient to bind the dopant into place [21], [22]. Li et al. recently showed that 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4TCNQ) sublimates out of the N,N,N′,N′-tetrakis(4-methoxy-phenyl)benzidine (MeO-TPD) mixed layer at ∼150 °C and diffuses at room temperature [23]. F4TCNQ has also been investigated as a p-type dopant for polymer samples using solution-processing with many of these studies focused on the non-polar polymer P3HT [24], [25], [26]. More recently, F4TCNQ was used to control the solubility of P3HT for patterning purposes and is likely therefore to be incorporated into solution-processed devices in the near future [27]. Unfortunately, little attention has been paid to the question of whether the molecular dopants have preference for the polar layer or for the non-polar layer, which is of high interest from the device stability perspective.

In this study, conjugated polymers, S–P3MEET and P3HT (Fig. 1), will be used to fabricate the polar and non-polar layers, respectively. S–P3MEET and P3HT are chosen because 1) both materials have similar polythiophene backbones; 2) P3HT is the most widely used conjugated polymer while S–P3MEET has been shown as a promising alternative to PEDOT:PSS [28], [16], and 3) S–P3MEET forms a thermally stable bilayer with P3HT, so we can study the transport of F4TCNQ without interference from mixing between the layers. The first part of this paper explores whether F4TCNQ is able to dope S–P3MEET. A comprehensive study on the electronic, morphological and thermal properties of the F4TCNQ doped S–P3MEET layer is performed. Next, a combination of absorbance/fluorescence spectroscopy and X-ray reflectometry techniques are used to investigate the location of F4TCNQ upon thermal annealing of bilayers of polar S–P3MEET and non-polar P3HT.