**3. Results and Discussion**

Toward the investigation of increased electrical conductivity of PEDOT:PSS film by using additives, various additives with different chemical structures have been chosen and applied. The electrical conductivity of a solidified PEDOT:PSS film is known to be increased when the stock PEDOT:PSS in a stage of aqueous dispersion is mixed with various additives such as EG, DEG, DMSO, and sorbitol [4,25–28,30,32–34,39]. Especially, EG is a commonly used agen<sup>t</sup> for the conductivity enhancement of PEDOT:PSS since it is a liquid form at ambient conditions and is a mass-produced commercial chemical. To understand the compositional dependence of the electrical conductivity and other properties of PEDOT:PSS films on the structures of additives, we have applied EG and other linear glycol-type additives as the conductivity enhancing agents for PEDOT:PSS complexes (Table 1). These additives share common repeating units of ethylene oxide and chain ends of either hydroxyl or methoxy groups. These additives can be well mixed with water due to their hydrophilic moieties, and thus a homogeneous dispersion of PEDOT:PSS and spin-cast high-quality films can be obtained (Figure S1).


**Table 1.** Chemical structures, boiling points, and density of the additives.

We characterized the electrical properties of the PEDOT:PSS films, which were solutiondeposited after mixing with the additives at different concentrations (Figure 1). The sheet resistances ( *R*sh) of PEDOT:PSS films were measured using a colinear 4-point-probe method. In this simple and straightforward method, the contact resistances at the film/probe interfaces can be excluded by recording the voltage difference between two inner probes under a current applied between two outer probes. The typical sheet resistance of the pristine PEDOT:PSS film was ~70–120 k Ω/sq. Figure 1a shows that the sheet resistances of the PEDOT:PSS films depend on the processing conditions. The sheet resistance of the films was clearly reduced when linear glycols were added, which also agrees to the previous studies (See Table S1) [4,22,28,32–34,39–41]. Even at a low concentration of 0.1 M of TEG, the sheet resistances dropped to 565 Ω/sq and eventually stabilized at higher concentrations. EG and DEG showed the sheet resistance transition points at slightly higher concentrations of 0.3 and 0.5 M, respectively, and the *R*sh values were stabilized at 550–600 Ω/sq at higher concentrations of additives. A much longer glycol, HEG, was also effective in reducing the sheet resistance, although the stabilized sheet resistance was 2 k Ω/sq. On the other hand, EGME, which has one hydroxyl group and one methoxy group at the ends, is relatively ineffective in the resistance reduction. The sheet resistance of EGME-added PEDOT:PSS was 20 k Ω/sq.

**Figure 1.** (**a**) Sheet resistance of poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) films processed with linear glycol-based additives at various concentrations. (**b**) Film thickness of PEDOT:PSS films processed with the additives. (**c**) Electrical conductivity of PEDOT:PSS films as a function of the weight fraction of additives in the stock PEDOT:PSS dispersion. Solid lines in (**c**) are fitting curves of data with a sigmoidal logistic function as shown in Equation (1).

The sheet resistance of a conductive film was also studied with the different film thickness. Figure 1b compares the thickness of the PEDOT:PSS films processed using different concentrations of the additives. All additives decreased the film thickness, compared to the average thickness of 77.7 nm of the pristine PEDOT:PSS. The thickness values are estimated around 55–60 nm at the concentration of 1.0 M. The decrease in the film thickness is similar to the previous report on the thickness of PEDOT:PSS processed with EG as an additive [32]. The decrease in thickness when processed with additives might originate from the liquid-phase additives acting as a diluting agen<sup>t</sup> during the spin-coating procedure.

From the sheet resistance and thickness of the PEDOT:PSS films, we can obtain the electrical conductivity as shown in Figure 1c. The electrical conductivity represents an intrinsic property of the materials, although the sheet resistance is an important parameter of conductive films for applications in transparent electrodes. The electrical conductivity of the PEDOT:PSS films followed a sigmoidal curve as a function of the weight fraction of additives with an inflection point. The sigmoidal dependencies have been previously reported with DEG, EG, DMSO, and sorbitol, and the threshold points have been assigned to be ~0.3–0.6 wt.% [28,34]. The sigmoidal curve is observed when a cumulative effect of a certain probability distribution is important. In the case of the conductivity enhancement in PEDOT:PSS with additives, the sigmoidal behavior can be explained phenomenologically by the probability of molecular interaction between PEDOT:PSS and the additive. Then the concentrations of both polymer and additive are the determining factors for the conductivity enhancement. To quantitatively analyze the data, we have fit the data with a sigmoidal logistic function, which is a cumulative distribution function of the logistic distribution, as in Equation (1):

$$
\sigma\_{dc} = \sigma\_0 + \frac{\sigma\_{\rm sat} - \sigma\_0}{1 + \exp\{-k(w\_{add} - w\_i)\}},\tag{1}
$$

where *σdc* is the electrical conductivity of the sample in a unit of S/cm, σ0 is the base conductivity of untreated PEDOT:PSS, *σsat* is the saturation conductivity, *wadd* is the weight fraction of additives in a unit of wt.% in the PEDOT:PSS dispersion, *wi* is a location parameter indicating the inflection point, and *k* is an inverse of a scale parameter describing the steepness of the curve. Table 2 presents the fitting results of the data with the logistic function (Equation (1)), with the coefficients of determination (*R*2) ranging from 0.911 for EGME to 0.997 for DEG and TEG. We note that other types of sigmoidal curves (e.g., the cumulative distribution function of the normal distribution, which can be expressed with an error function) can also describe the sigmoidal behavior. In this case, the parameters are similar to the values in Table 2 with the differences less than 7%.


**Table 2.** Fitting parameters of a sigmoidal logistic function (Equation (1)).

The saturation conductivities are 271.8, 325.4, and 326.2 S/cm for EG, DEG, and TEG, respectively. These polar solvents can strongly interact with PSS from the PEDOT:PSS complexes, resulting in an effective phase separation in the films and/or elimination of the insulating components from the films [26,28,29,39,45]. The saturation values are much lower for HEG (95.9 S/cm) and EGME (11.0 S/cm). We believe that the low conductivity of HEG-added PEDOT:PSS films is due to the residual additives remaining in the films. It has been previously reported that the residual solvents can limit an efficient charge transport in the polymer complexes [39]. The case of EGME is interesting, considering that the structural difference between EG and EGME is one end group (hydroxyl group vs. methoxy group). Although EGME is also polar and can enhance the conductivity, the affinity with PSS is not strong enough to result in a large degree of changes. The sheet resistance of EGME-processed PEDOT:PSS films was reduced by a factor of only ~2.4 although the post-treatment was very effective in the resistance reduction by a factor of ~200 in the previous reports [5,30]. The PEDOT:PSS films have similar transition points (*wi*) of 0.6–1.0 wt.% of the linear glycol series as additives, which show close agreemen<sup>t</sup> to the values in the previous reports [28,34]. The coefficient *k*, representing how steep the curve is near the point of inflection, is the smallest for DEG. On the other hand, EG and HEG showed sharp increase in the conductivity as the additive concentration passed through the transition point. In these cases, an additive concentration of 0.8 wt.% is enough to show a saturation in the conductivity.

Such a sigmoidal dependence of the electrical conductivity on the amount of glycol additives suggests that a fraction of the components in PEDOT:PSS are influenced by the additives. The probability density function of a logistic distribution shows a symmetric curve peaked at the location parameter (i.e., *wi* in Equation (1)). In this case, *wi* may reflect the quantity of additives with the highest probability of interaction between the additive molecule and PEDOT:PSS. With a small amount of additive (*wadd* < *wi*), only a small fraction of the components in PEDOT:PSS, either in an unbound form or in a complexed form, are affected and participate in molecular redistribution and reorganization. The conductivity sharply increased as the weight fraction of additives increased to near the inflection point (*wadd* ~ *wi*) because many additive molecules can interact with PEDOT and PSS. When the fraction of additives increased further to the saturation points (*wadd* > *wi*), there are only little molecules left to interact with. The points of inflection and saturation depend on the polymer compositions, the molecular weights, the polymerization conditions, and the degree of complexation, as well as the structure of additives.

We investigated the optical properties of PEDOT:PSS films using UV-Vis-NIR spectrophotometer, since transmittance is a critical factor for applications in transparent electrodes and the spectra can also provide information on the doping states of the PEDOT composites (Figure 2) [3,8,22,46,47]. Transmittance (*T*) spectra of the PEDOT:PSS films made from dispersions with the additives (1 M), regardless of the structural variations, showed a fairly high optical transparency of *T* > 90% in the visible region with a peak value at around 450 nm. These films can be used for transparent coating layers that pass through the visible light. The transmittance *T* at 550 nm, which is often used as a benchmark condition for visually transparent coatings, was >97% obtained by addition of DEG. TEGME and EGDME follow close second and third ones, respectively.

**Figure 2.** (**a**) Transmittance spectra, (**b**) representative transmittance values at 550 nm, and (**c**) absorption coefficients of the PEDOT:PSS films processed with various solvent additives at the concentrations of 1 M.

Typical PEDOT:PSS films show a large absorption in the near-infrared (NIR) region and relatively high transmission in the visible region. Indeed, the absorption coefficient (*α*) in the NIR region reaches ~10<sup>4</sup> cm<sup>−</sup>1, while those in the visible region were an order of magnitude lower (Figure 2c). When the additives were mixed in the PEDOT:PSS dispersion, the absorbance of films slightly decreased compared to the pristine PEDOT:PSS. However, there was no noticeable change in the shape of the absorption spectra. Although the absorption coefficients vary depending on the additives, the ratios between the absorbance at NIR to that at visible regions remain similar, suggesting that the doping states are rather invariant.

From the sulfur 2p signals of X-ray photoelectron spectroscopy (XPS), we extracted the molar ratios of sulfonate to EDOT and sodium sulfonate to sulfonic acid (Figure 3). The XPS signals for S 2p at 163–170 eV in the PEDOT:PSS films should be deconvoluted, because: (a) there are multiple peaks for the aromatic S in PEDOT with a binding energy peak at 163–166 eV and the sulfonate form of PSS at 167–170 eV, (b) each of S 2p spectra shows spin-split signals (i.e., S 2p1/2 and S 2p3/2) with the corresponding ratio of 1:2, (c) the sulfur in PEDOT can be either in a neutral form (S) or in a cationic form (S+), and (d) the sulfonic acid (i.e., PSSH) and sodium sulfonate (i.e., PSSNa) also have a difference of ~0.4 eV in the binding energy [26,45,48,49]. Changes in the ratios, examined by XPS, have been correlated with the conductivity increase [32,45,50]. In our samples, we found that the sulfonate-to-EDOT ratio was 2.4 for pristine PEDOT:PSS films, similar to the nominal ratio of the stock dispersion. The ratios slightly decreased to 2.29, 2.18, 2.27, 2.31, and 2.16 for EG, DEG, TEG, HEG, and EGME, respectively, as shown in Figure 3g. Because the ratios are rather close to each other, it is not clear at this moment if a change in the ratio at the surface of PEDOT:PSS films by additives is a necessary condition or not for the conductivity increase. The PSSNa/PSSH ratios dropped from 4.3 for the pristine sample to 1.6 for EG, 1.7 for DEG, 1.2 for TEG, 1.4 for HEG, and 1.9 for EGME, as shown in Figure 3h. Sodium ions can be eliminated by attractive interaction with the oxygen-rich additives.

**Figure 3.** XPS S(2p) spectra of PEDOT:PSS films processed with various additives at 1 M: (**a**) Without additive and with (**b**) EG, (**c**) DEG, (**d**) TEG, (**e**) HEG, and (**f**) EGME. (**g**) Ratios of sulfonate to EDOT monomer using the areal ratios of deconvoluted peaks of the XPS signals. (**h**) Ratios of PSSNa/PSSH.

We investigated, parameter wise, how effective these additives are for applications in transparent conducting films by comparing two representative figures of merit. Sheet resistance (*Rsh*) and transmittance (*T*) of transparent conducting films are two important parameters and often combined to evaluate the figures of merit of transparent electrodes. One method has been proposed to use Equation (2):

$$
\phi\_{Haackc} = \frac{T^{10}}{R\_{sh}} = \sigma\_{dc} d \, \exp(-10ad),
\tag{2}
$$

to evaluate a figure of merit in a unit of Ω−<sup>1</sup> and to quantitatively compare the performance of transparent electrodes [51]. Alternatively, the two parameters, *T* and *Rsh*, are often correlated, at least for a thin metallic film placed in a free space, with Equation (3):

$$T = \left(1 + \frac{Z\_0}{2R\_{\rm sl}} \frac{\sigma\_{op}}{\sigma\_{\rm dc}}\right)^{-1} \,\, \,\, \,\tag{3}$$

where *<sup>σ</sup>*op is the optical conductivity and *Z*0 is the impedance of free space (*Z*0 = 4π/*c* = 4.19 × 10−<sup>11</sup> s/cm in cgs unit, which corresponds to 337 Ω in SI unit) [46,52]. In this case, a unitless ratio of *<sup>σ</sup>dc*/*<sup>σ</sup>op* is often used as an alternative figure of merit of transparent conductive films. It has been suggested that the ratio of >10 is required for touch panels [53]. We applied these approaches to compare the effectiveness of the additives (Figure 4). We selected the wavelength of 550 nm for comparison and found that DEG and TEG are equally effective as an additive to result in the PEDOT:PSS-based transparent electrodes. The Haacke figures of merit (*ΦHaacke*) of the PEDOT:PSS films, calculated from Equation (2), are 1.6 × 10−<sup>3</sup> Ω−<sup>1</sup> for DEG and 2.0 × 10−<sup>3</sup> Ω−<sup>1</sup> for TEG. EG has 1.2 × 10−<sup>3</sup> <sup>Ω</sup>−1, marking the third place among the additives of interest. For the ratios of *<sup>σ</sup>dc*/*<sup>σ</sup>op* shown in Equation (3), we ge<sup>t</sup> the 15.3, 32.3, and 27.3 for EG, DEG, and TEG, respectively. As expected from the electrical conductivity of the polymer films, HEG and EGME are rather ineffective as an additive for PEDOT:PSS. We note that the DEG- and TEG-added PEDOT:PSS films are more stable than the pristine PEDOT:PSS, as the films do not show significant changes in the electrical and optical properties after aging in air for several days (Figure S2).

**Figure 4.** Figures of merit of the PEDOT:PSS films processed with additives. (**a**) The parameter proposed by Haacke. (**b**) The ratios of electrical conductivity to optical conductivity, at the wavelength of 550 nm.
