3.2.2. EQCM Measurements

While the EQCM can more reliable distinguish between mass changes from viscoelastic effects at lower film thicknesses, it is also known that very thin films have different ion transport properties. Hence, it is important to find a compromise. PEDOT was deposited potentiostatically here at different polymerization potentials that were equal to those used in the aforementioned actuation studies (EP values of 1.0 V, 1.2 V, and 1.5 V) for actuator films. The mass changes accompanying the CV-driven processes for the PEDOT films of different formation potentials are presented in Figure 4. Charge–potential curves are presented in Figure S1c, again revealing that the charging/discharging behaviors were in balance.

EQCM analysis allows one to distinguish between mass balance changes in the various regions of a redox cycle. Upon oxidation from −1.0 V to 0.0 V, all the PEDOT films here presented a slightly negative slope, which was virtually invisible for Ep 1.0 V, and the slopes became steeper for higher polymerization potentials. This means that the PEDOT films lost weight in this potential range, as observed before by Kvarnström et al. [37] in a solution of TBAPF6 in acetonitrile. As expected, the bending displacement of PEDOT-BL in the same potential range (Figure 3b) showed contraction. In particular, in the case of the films where the EP was 1.0 V, the slight mass loss in the beginning of the oxidative scan may (in addition to cation involvement) also be explained by the current still remaining negative at these potentials (Figure 3c), corresponding to ongoing reduction, even as the oxidative scan had started. The next phase started for the potentials corresponding to the oxidation peak in the CV and progressed all the way to 1.0 V, thereby resulting in a significant mass increase for all PEDOT films (Figure 4a). This meant that ions and solvent molecules had moved in, as demonstrated above by the expansion of the PEDOT-BL (Figure 3b). In the reverse scan, from 1.0 V until the reduction peak at −0.4 V, a loss of mass was recorded (negative slope) for all materials, which correlates with the contraction of PEDOT-BL as shown in Figure 3b. In the case of the PEDOT films where EP = 1.5 V (Figure 4c), a breaking point can clearly be observed at around −0.2 V, which is where the slope changed, indicating a change in the anion-cation participation ratio.

**Figure 4.** Cyclic voltammetry-driven (±1.0 V, scan rate of 10 mV s<sup>−</sup>1, current denoted by the red curve) EQCM measurements (mass change denoted by the black curve) for a TBAPF6-PC electrolyte with PEDOT films polymerized at different polymerization potentials: (**a**) EP = 1.0 V; (**b**) EP = 1.2 V; (**c**) EP = 1.5 V. The arrows indicate the direction of the scan (starting point of −1.0 V).

Plotting the frequency change Δ*f* against the charge Δ*Q*, as in Equations (2) and (3), the slopes of the curves (Figure S2a–c) of the different regions give the molecular weights of the compensating charge species, i.e., *MCCS* (Equation (4), dimensionless). The obtained *MCCS* values indicate the amounts and kinds of ions with or without a solvent (molar mass of PC of 102.09 g·mol<sup>−</sup>1) that were incorporated or expelled during voltammetric cycling. For TBAPF6, the molecular weight of the anion PF6 - was 144.96 g·mol<sup>−</sup>1, and that of the cation TBA+ was 242.47 g·mol<sup>−</sup>1. In a case where *MCCS* is greater than the molecular weight of the anion or cation, solvent molecules are likely transported, while negative values smaller than the anion or cation molecular weights hint to mixed processes of simultaneous anion and cation exchange, potentially even in combination with solvent molecules. With the EQCM methodology, the exact determination for which ions move in or out in a mixed activity scenario can only be estimated.

Figure 5a,b show the values of *MCCS* as a function of the cycle potential for the PEDOT samples polymerized at different potentials, with the values separated for oxidation and reduction. For each polymerization potential, at least three independent measurements

were taken and the results represent mean values with the error bars of the calculated *MCCS* values.

**Figure 5.** PEDOT deposited at different polymerization potentials where EP = 1.0 V (··-··), EP = 1.0 V (··•··), and EP = 1.5 V (····), showing charge compensation for species *MCCS* in (**a**) oxidation against the potential direction for −1.0 V to +1.0 V and (**b**) reduction against the potential direction of 1.0 V to −1.0 V. The included dotted green line represents the molar mass for PF6 - anions.

It is immediately clear that the trends for the three polymerization potentials were all different. Upon oxidation where EP = 1.0 V, the *MCCS* values (Figure 5a) for the potentials starting from −1.0 V were negative, i.e., −18 ± 2, followed by −38 ± 4 for EP = 1.2 V, and lower still for EP = 1.5 V with −57 ± 6. The following slopes at a potential of −0.5 V revealed positive values for EP = 1.0 V (67 ± 7) and EP = 1.2 V (26 ± 3), while the values were still negative when EP = 1.5 V. The *MCCS* values were nowhere near the molecular weights of the ions, where mixed ion actuation takes place from −1.0 V to −0.5 V where cations are expelled and anions are incorporated. This process is likely accompanied by solvent molecules. Vandesteeg et al. [22] likewise proposed cation expulsion instead of anion insertion during oxidation in EDOT-based polymers, but without further investigations. At a potential of 0.0 V the *MCCS* values were all positive (but still under the molecular weight of PF6 - anions corresponding to mixed activity). With increasing polymerization potential, the *MCCS* values decreased (higher cation involvement). This corresponds to the switching region in Figure 3a,b, where contraction changed to expansion. Upon further oxidation, (slopes at 0.5 V and 1.0 V), the films where EP = 1.0 V and EP = 1.2 V had *MCCS* values in a range above the molecular weight of PF6 - , corresponding to dominant anion incorporation, along with solvent molecules (as seen in Figure 3a,b with the primary expansion upon oxidation). In the case where EP = 1.5 V, the *MCCS* value of 136 ± 14 is still slightly lower than the molecular weight of PF6 - , where the dominant process was still anion incorporation but with some mixed characteristics. It has been proposed that the hysteresis of the cycle might be caused by a variable number of trapped ions in PEDOT network in the considered timeframe [38].

As usual, a reduction scan can be more informative when starting from a fully oxidized and conductive state. For the reverse scan (Figure 5b), in the case where EP = 1.0 V, the *MCCS* values of −166 and −152 in a potential range from 1.0 V to −0.5 V show a clear indication of (solvated) anion expulsion, thus corresponding to contraction in PEDOT-FF and PEDOT-BL (Figure 2a,b). In the case where EP = 1.2 V, the picture is different. The reduction scan *MCCS* values do not reflect those for the oxidation scan. While it might be expected that the same solvated PF6 - anions left the film, the *MCCS* values of −66 ± 6 do not describe that. Similarly, that discussed above can be explained (in addition to the unlikely cation involvement) by the fact that in the beginning of the reduction scan, the potential

decrease did not immediately stop incomplete oxidation, and some anions still entered the film before beginning to leave. A further reduction from 0.5 V to 0.0 V showed a *MCCS* result of −150 ± 15, which already corresponds to anions with solvent molecules leaving the PEDOT film. By reaching a potential of −1.0 V, PEDOT-BL and PEDOT-FF (Figure 3a,b) showed small expansion, and the mixed ion involvement was reflected here by a *MCCS* result of −66 ± 6, as anions were mainly leaving and cations were already moving in (with solvent molecules involved in both processes). The polymerization potential of EP = 1.5 V differed the most, where, during the whole reduction scan, the *MCCS* values never reached the anion molecular weight value, corresponding to a mixed ion process throughout, with cation incorporation dominating as the potential dropped, corresponding to a higher strain/displacement for PEDOT-FF and PEDOT-BL upon reduction. The polymerization potentials chosen here have demonstrated significant effects regarding the dominant actuation direction, along with the accompanying ion flux.

To our knowledge, this is the first attempt to analyze the effect of polymerization potential in regard to the anion flux behind the linear and bending actuation modes of PEDOT actuators using EQCM measurements, especially in terms of addressing potential regions individually, thus allowing the tuning of the actuator response by selecting potential windows for desired outcomes. In summary, higher formation potentials of 1.2 V and 1.5 V brought about increased cation involvement in redox processes, while solvent uptake only played a minor role, as discussed before [24]. While some earlier works have shown significant solvent effects [33], their omission of any possible cation involvement does not allow a clear comparison. As seen from the structural description, with an increasing deposition potential, the polymer matrix becomes denser and less permeable for ions. As such, solvation shells are unlikely to remain intact as ions enter the material.
