Correction to “Zeta Potentials of Cotton Membranes in Acetonitrile Solutions”

The original paper reported ion-specific zeta potentials of cotton membranes in acetonitrile and a possible electroviscous effect. Here, we correct two independent issues in the analysis and measurements. First, the theoretical curves in Figure were calculated using the zeta potential in millivolts rather than volts, resulting in substantially overestimated peaks. This error affected only the theoretical prediction. Second, the input resistance of the voltmeter used previously (Sanwa PC773, 10–100 MΩ) was also insufficient. We therefore repeated the measurements using a high-input-resistance voltmeter (Keithley SMU 2450, >10 GΩ), and all corrected experimental quantities, including the zeta potentials and fitted adsorption parameters, are based on these new measurements.

7.

Normalized pressure difference as a function of the Bu₄NPF₆ concentration. The solid lines show the theoretical electroviscous enhancement calculated for effective pore radii of R = 1, 2, and 8 μm. The values at the highest concentration in each data set were used as Δp ₀ for normalization.

Because the finite input resistance allows leakage current, current conservation gives, L ₂₁Δp + L ₂₂Δψ = −Δψ/R ₀, where L ₂₁ is the electro-osmotic coefficient, L ₂₂ is the conductance of the solution in the membrane, and R ₀ is the input resistance of the voltmeter. Thus, the streaming potential including the effect of the finite input resistance is given by

$$\mathrm{\Delta }\psi =-\frac{L_{21}}{L_{22}+{R_0}^{-1}}\mathrm{\Delta }p$$ 1

The estimated solution resistance in membrane increases from 25 MΩ at 1 mM to 25 GΩ at 0.001 mM, assuming that the area fraction equals the volume fraction. Thus, in the previous measurements, the input resistance of the voltmeter (10–100 MΩ) was not sufficient. At the lowest concentrations, the estimated solution resistance in membrane becomes comparable to 10 GΩ, and therefore a residual input-resistance effect cannot be completely excluded.

Figure of the original paper should be replaced by the figure given here. Although the signs of the streaming potentials do not change from the previous measurements, the streaming potentials in the low salinity significantly increase in magnitude compared to the previous measurements. The induced pressure difference was about three times smaller than that in the previous measurements, probably reflecting differences in the packing state.

2.

(a) Streaming potentials and (b) induced pressure differences as functions of Bu₄NPF₆ concentration.

Panel (a) of Figure a should be replaced by the image given here; panel (b) remains unchanged. The corrected Bu₄NPF₆ zeta potentials reach approximately 50 mV in magnitude and are described by the same Langmuir adsorption isotherm with α_– = −7.0 and Γ₀ = 0.009 nm^–2. Because the corrected streaming potentials are larger than those reported previously, the fitted adsorption energy becomes more negative in order to reproduce the revised zeta potentials. In contrast, the adsorption site density remains of the same order as that reported originally.

4.

(a) Zeta potentials as a function of Bu₄NPF₆ concentrations.

Panels (a)–(c) of Figure should be replaced by the panels shown here; panel (d) remains unchanged.

5.

(a) Zeta potentials, (b) streaming potentials, and (c) induced pressure differences as functions of the LiBF₄ concentration.

Figure should be replaced by the following figure. Because the predicted electroviscous enhancement is only a few percent and is smaller than the experimental variation, the previous interpretation should be withdrawn.

In summary, the corrected theoretical calculation eliminates the previously reported electroviscous enhancement, whereas the repeated streaming-potential measurements revise the quantitative zeta potentials and the fitted adsorption parameters. The main conclusion regarding ion-specific charging remains unchanged. The authors regret these errors and any confusion they may have caused.