Demand-side flexibility enables cost savings in a Reversible pH-swing electrochemical process for Oceanic Co2 Removal

Raw datasets and codes used in our manuscript of the same title. Figure numbers correspond with the figure numbers in the manuscript.

Figure 2: Process Overview(a) Steps of the proposed, reversible, oceanic CDR process, starting with acidification of the ocean water, followed by vacuum stripping of the CO2, basification of the oceanwater, and return to the ocean to equilibrate with the atmosphere. (b) Corresponding effects on pH and dissolved inorganic carbon (DIC) at each step in the oceanic CDR process. Oceanwater pH decreases from 8.1 during acidification to produce a shift in speciation from HCO3 todissolved CO2, which is phase-separated between step 2-3 to capture pure CO2 which decreases the DIC. Basification negates these changes through the addition of the same concentration of OH− ions as H+ ions added to ocean water during acidification, resulting in pH > 8.1 at the end of basification. CO2-starved ocean water will re-equilibrate with the atmosphere and get buffered back to a neutral-pH to re-enter acidification and the cycle continues.

Figure 3: Predicted Performance from 0-D Equivalent Circuit Modeling (Sec. Prediction of current-volatge behavior) for (a,c) acidification and (b,d) basification to determine (a,b) polarization behavior and (c,d) Faradaic efficiencies; including (c) power density produced during acidification.Different mass-transfer conditions and therefore limiting currents are modeled by varying the liquid-side boundary layer thicknesses, (δl). Two different pH-shift conditions are modeled for the low and high pH values at the end of acidification and basification, that are respectively constrained by the concentration of H+ and OH− additions to be equal during both steps – (pHlow, pHhigh): (4,10.7) and (6,9) as described in Sec. S3. Baseline parameter values listed in Table S1 are used in the equivalent circuit models with consideration of desired and competing reactions. Note the differences in the current and potential ranges plotted between (a) acidification and (b) basification steps.

Figure 4: Energy Intensity Comparisons (a-c) Predictions for electrochemical energy intensity normalized by the amount of CO2 captured (Eq. (14)) in (GJ/tCO2 ) as a function of the current density when jbase = jacid (a) without and (b) with competing reactions modeled. Dashed lines on the plot indicates a baseline cell resistance of 2.18 Ω cm2 and the shaded areas are indica-tive of this ohmic resistance in the range of 0 - 5 Ω cm2; the shades of blue are indicative of theliquid-side boundary layer thicknesses, δl, modeled. Electrochemical energy intensities from (a)and (b) where δl = 50 μm are shown in (c), and compared with state-of-the-art ocean water CDRprocess intensities reported by Digdaya et al.,15 Kim et al.,23 and Yan et al.22 The gray filled areashows possible operational points with varying extents of competing reactions with the maximumbounded by competing reactions occurring at mass-transfer limited rates. (d) Overall energy inten-sity (GJ/tCO2 ) including electrochemical and parasitic loads (as described in Sec. Energy IntensityCalculations) for the proposed process without competing reactions, calculated at an industrially-relevant current density of 100 mAcm−2, the baseline cell resistance, and pH shift of 6 to 9.6,compared with electrochemical and thermally-driven direct air capture (DAC) processes with data obtained from Singh et al.;24 error bars for the thermal DAC processes are from Viebahn et al.

Figure 5: Operating Energy Profiles and Effects of Load Shifting (a) Integration of solar and wind electricity production in California, CAISO 2022-2023,40 with predicted energy demand for the proposed oceanic CO2 removal process as a function of varying extents of yearly CO2 captured, 0 – 16 MtCO2,year. (b) Hourly power demand to capture 1 tCO2 per day. The gray curve indicates the process shown in (a), while the red curves shows a constant power demand process occurring at an equivalent energy intensity, i.e., the area underneath both curves are the same. (c) Hourly cost to capture 1 tCO2 per day. The gray curve indicates the hourly cost associated with variable energy process shown in (b) if all produced electricity is sold at market value, while the red curve compares it to the hourly cost associated with constant power consumption process shown in (b). Operating conditions of the proposed, variable process shown in (a-c) include jacid = 75 mA cm−2 and jbase = 150 mA cm−2, while still constraining the same volume of ocean water being treated over a 24-hour day in the individual steps (i.e. tbase = 0.5tacid). The process switches between acidification (occurring early mornings and evenings/nights when electricity price is high) and basification (occurring mid-day when electricity price is low) based on the supply and the cost of electricity purchased from CAISO (Sec S8.5). (d) Electricity cost savings for the proposed process operation compared to a process with constant energy usage at different ratios of tbase to tacid, assuming a time-weighted average operating current density of 100 mA cm−2. Different line styles indicate different proportions of market value at which produced electricity can be sold (0%, 50%, and 100%). For all cases: competing reactions are not considered, a thin boundary layer is assumed (δl = 10 μm), and the pH shift is from 6 to 9.6. Parasitic energy costs are included for CO2 phase separation, but not for pumping as we assume co-location with a desalination plant. Parasitics are assumed to only operate during the basification process, at high enough power to remove all CO2 treated over the day.