Decarbonizing the power grid requires the widespread adoption of renewable (wind/solar) power sources (1,2). Mitigating the unpredictable and intermittent nature of renewable power presents a grand challenge, driving the unprecedented restructuring of the power grid. The multiscale and spatio-temporal behaviors from this restructuring will require the power grid to harness flexibility in diverse forms and enable a reliable and sustainable power supply to society (3,4). Lack of power grid flexibility is currently manifesting as large amounts of renewable power that remain stranded (cannot be absorbed by the grid) or even the curtailment of renewable electricity generation (5,6); stranded power is projected to grow as more renewable power is injected into the grid and hinders the ability to decarbonize the grid(4). Independent system operators (ISOs) operate different electricity markets to harness and remunerate flexibility provided at different times and network locations (7). For instance, in markets such as the day-ahead market (DAM) and real-time market (RTM), electricity prices are updated every hour or 5-15 minutes to promote balancing of power supply and demand (Fig. 1A). The frequency regulation (FR) market is a critical ancillary service market that operates at a 2 second time resolution, and is designed to secure balancing at fast timescales (7). The need for FR services continues to increase as more renewable power is absorbed in the grid. The price signals from these markets, particularly negative prices resulting from stranded power, indicate when and where resources are most valuable and provide an economic incentive for technologies that can use/supply power (8). Therefore, flexible technologies that can participate in markets at different times/locations can help the grid absorb more wind/solar power and economically monetize this service (9).
Various technologies have been developed to handle power fluctuations in the grid, such as hydroelectric storage (PHES) (10,11), compressed air energy storage (CAES) (12,13), battery systems (14,15), and power to gas (P2G) (16,17) (Fig. 1B). These technologies could strategically convert excess power from the grid (or power at low or negative prices) into potential or chemical energies, and then use stored energies during power scarcity (or power at high prices). Compared to PHES or CAES with long storage/operation timescale, electrochemical technologies with a short storage/operation timescale, especially P2G, can lead to spatially distributed power demands and new sources of flexibility. Recent results show that incorporating proton exchange membrane electrolysis capable of flexible participation in DAM and RTM to produce hydrogen in methanol synthesis processes can lower the hydrogen cost by 31% from that via the conventional steammethane reforming process (18). However, a fundamental issue of such an electrochemical system with coupled half-reactions is that the slow reaction constrains the system response; from a power grid perspective, this can limit the system's ability to respond to signals that are faster than the limiting time constant. The mismatch between the timescales of paired half-reactions can also lead to system inefficiencies (parasitic reactions) (19,20). Instead of the coupled electrochemical system, decoupling paired reactions using redox-active materials (21)(22)(23)(24)(25)(26) or redox reservoirs (RRs) (27)(28)(29) allows for flexible pairing of different half-reactions potentially to participate in different electricity markets. In principle, using such a strategy, half-reactions can be performed at desired rates and under various timescales to provide more design flexibility for participation in different electricity markets and process modularization of chemical supply chains.
Here, we present a highly rate-mismatched modular electrochemical synthesis (ModES) process that can effectively participate in DAM, RTM, and FR markets (Fig. 1C) and reduce electricity costs. We decouple the conventionally tightly coupled hydrogen evolution reaction (HER) with fast kinetics and persulfate production reaction (PSR) with slow kinetics and use a redox reservoir (RR) electrode with fast kinetics to separately pair with these two half reactions.
The electricity in spikes of a few seconds at a low or even negative price can be used to produce H2 and charge the RR electrode quickly. The electrons/ions stored in the RR electrode enable slower persulfate (S2O8 2-) production over a few minutes. A proton battery material that shows fast proton transport and storage via the Grotthuss mechanism, copper hexacyanoferrate (CuHCF), is demonstrated as the RR electrode with excellent rate capability and cycling stability even in mismatched charging/discharging conditions. The developed computational modeling framework shows that this highly rate-mismatched ModES system operating under a flexible power load could reduce electricity costs by 30-40% compared to the baseline and even generate revenues by participating in the FR market. These results confirm the promise of modular electrosynthesis for dynamic electricity markets and show how flexible electrochemical systems can open new revenue opportunities.
We design a highly rate-mismatched ModES system to produce value-added chemicals under different timescales to participate in dynamic electricity markets. The key to such a system is the use of solid-state redox materials that can achieve fast storage/release of electrons and ions and decouple electrochemical half-reactions with different intrinsic kinetics. As the initial demonstration, we choose HER, one of the fastest electrochemical reactions that produces green hydrogen fuel, as the model reaction to participate in the FR market, and PSR, a widely used electrosynthesis with slow intrinsic kinetics (estimated market size of $750 million in 2027)(30), as the other model reaction to participate in DAM and RTM. Compared to conventional persulfate production (fig. S1), here flexible HER and PSR with different reaction rates are achieved in a membrane-free, decoupled fashion mediated by a RR electrode (Fig. 1C): proton-consuming HER pairs with the oxidation of RR that releases protons at a fast reaction rate. Then the reduction of RR that uptakes protons pairs with the slower PSR in the other electrochemical cell. The RR electrode can transport protons from the PSR cell (Cell S 2 O 8 2-) to the HER cell (Cell H 2 ) to maintain proton balance. Because dynamic electricity markets exhibit price fluctuations at different timescales across seconds to hours, stable RRs with fast kinetics and selective proton transport must be developed for highly rate-mismatched electrochemical productions.
The rate capability of energy storage materials is strongly influenced by the charge carriers.
Compared to metal ions, protons have a smaller mass and ionic radius and exhibit fast diffusion kinetics via the unique Grotthuss mechanism (Fig. 2A) (31,32), where proton jumps from one water molecule to the next through the hydrogen-bonding network (33). Moreover, redox materials with defects (34,35) or large tunnels(36) may form inner hydrogen-bonding networks and achieve fast Grotthuss conduction inside solids. Therefore, we choose Prussian Blue analog copper hexacyanoferrate (CuHCF) with an open-framework structure that could be filled with a crystal water network (37), which showed promise for fast redox kinetics (38).
Nanostructured CuHCF material (Fig. 2B) was synthesized via a co-precipitation method (see Methods for details). The characteristic (200), (220), and (400) powder X-ray diffraction (PXRD) peaks of the CuHCF material match well with the standard pattern of the cubic PBA phase (JCPDS No. 52-1907) (fig. S2). Thermogravimetry analysis (TGA, fig. S3) and inductively coupled plasma optical emission spectroscopy (ICP-OES, table S1) reveal the water content and element compositions of the as-synthesized CuHCF material, which yield the chemical formula of K0.06Cu1.5[Fe(CN)6] • 5.5 H2O.
We then studied the electrochemical behaviors of the CuHCF material in 2 M H2SO4 solution. Cyclic voltammograms (CVs) showed four pairs of reversible redox peaks, where O1/R1 was attributed to Cu II /Cu I and other peaks were from Fe III /Fe II (Fig. 2C) (39,40). The peak current in CVs, I, was analyzed based on equation I = av b , where a was the coefficient and v was the scan rate. The b values for the anodic peaks (Fig. 2D) and cathodic peaks (fig. S4) were close to unity, indicating strong capacitive behaviors with fast kinetics (41,42). The CuHCF electrodes exhibited excellent capacity retention of 75.3 % at even 1500 C rate (a short (dis)charge period < 2 s) (Fig. 2E and table. S2). Note that the nC rate means that the (dis)charge period of the electrode is 1/n hours, and the 1 C rate is defined as 65 mA g -1 based on the theoretical capacity of CuHCF (65 mAh g -1 ). The linearly increasing potential hysteresis of CuHCF with increasing current densities (fig. S5) revealed that its rate capability was limited by uncompensated resistance in the electrochemical cell rather than the kinetics of proton storage/release. Galvanostatic cycling tests confirmed the long-term stability of CuHCF at 1 A g -1 (15 C) with 0.006 % decay per cycle (fig. S6A) and at 5 A g -1 (75 C) with 0.005 % decay per cycle (fig. S6B).
We further evaluated the capacity and Coulombic efficiency (CE) of CuHCF at mismatched charging/discharging rates. No matter how fast the charging rate was, CuHCF showed CE above 90 % when the discharging rate was above 1.5 C (Fig. 2F and fig. S7). Reduced CE might result from the self-discharge of the electrode at a high charging rate (fig. S8). The CuHCF electrode also showed capacity retention above 78% (compared to the capacity at 1 C rate) and good cycling stability at mismatched rates if the ratio of the charging rate/discharging rate stayed below 500 (Fig. 2G and fig. S9). These results suggest that rate-mismatched production of persulfate and hydrogen with a time ratio from 1:1 to 100:1 should be achievable with a high CE and capacity retention using the CuHCF electrode (Fig. 2H). Specifically, when the time ratio of PSR to HER is 100:1 with an HER rate of 1000 C and a PSR rate of 10 C, the HER reaction would finish in 2 s to enable its participation in FR market to receive flexibility incentives. The excellent long-term stability and fast proton transport suggest the suitability of CuHCF for rate-mismatched electrochemical productions.
We first separately investigated HER and PSR in 2 M H2SO4 solution considering their fast kinetics and high efficiency in acidic solutions (43). Pt electrode, widely studied for acidic HER (44), was chosen as the catalytic electrode for H2 production. The linear sweep voltammetry (LSV) curve of the Pt electrode shows an overpotential as low as 43 mV to achieve the current density of 10 mA cm -2 (fig. S10A). Chronoamperometry curves at different applied potentials show nearly constant currents without noticeable decay (fig. S10B), suggesting excellent catalytic stability.
To minimize the competitive oxygen evolution reaction (OER) and possible chemical corrosion (43,45,46), we chose a stable boron-doped diamond (BDD) electrode with a high overpotential for OER and a good activity for PSR (see details on electrode fabrication in Materials and Methods and fig. S11). The LSV curve of the BDD electrode shows an increased anodic current density with increasing applied potential and a high current density of 85 mA cm -2 at 3.10 V vs. standard hydrogen electrode (SHE) (fig. S10C). As the applied potential increases, the Faradaic efficiency of PSR [FE (S2O8 2-)] increases and reaches above 80 % when the applied potential is above 2.84 V vs. SHE (figs, S10C, D, and 12). Moreover, stable bulk PSR electrolysis over the BDD electrode can accumulate S2O8 2-(fig. S10D), confirming the excellent activity and stability of the BDD electrode.
We then demonstrated a prototype rate-mismatched ModES production of H2 and S2O8 2- mediated using a CuHCF RR electrode in two electrochemical cells (Fig. 3A, see Materials and Methods for details). During rate-mismatched ModES cycles (Fig. 3B and fig. S13), in Cell H 2 , HER at the Pt electrode accompanies the oxidation of RR that releases protons; then in Cell S 2 O 8 2-, PSR at the BDD electrode accompanies the reduction of RR that uptakes protons. Here, we assume the FE of HER to be 100 % since the only possible product during the cathodic reaction is H2. The FE (S2O8 2-) was calculated based on the measured concentration of the produced persulfate in Cell S 2 O 8 2-. During ModES cycles, the average potential difference between the catalytic and RR electrodes determines the average cell voltages (Vcell). To describe the energy utilization of the ModES, we use the voltage efficiency (VE) defined for decoupled aqueous electrosynthesis (27,29) based on the following equation:
where VDIR is the potential difference between the HER at the Pt electrode and PSR at the BDD electrode, respectively; ∆V represents the difference between the average potential of the RR electrode in Cell H 2 and Cell S 2 O 8 2-.
We started ModES production of H2 and S2O8 2-from equal rates of HER /PSR and finally achieved a high mismatched production rate (100:1) of HER /PSR: the PSR rate was kept at 10 C based on the capacity of the CuHCF RR electrode, and the HER rate increased from 10 C to 1000 C. During the first 5 cycles with equal rates, the average RR potentials (ERR) in both cells were nearly 1.0 V vs. SHE (Fig. 3C), and the ∆V was only 20 mV (fig. S14), confirming the fast proton kinetics of CuHCF. The corresponding Vcell showed repeatable values of 1.00 V and 1.91 V with fluctuations less than 10 mV in Cell H 2 and Cell S 2 O 8 2-. Moreover, an average CE of RR of 99.4 %, VE of 99.7 %, and FE (S2O8 2-) of ~75 % (Fig. 3D) indicate the high energy efficiency of the ModES system. During cycles 6 to 10 when the HER rate increased to 100 C, ERR and Vcell in Cell H 2 increased by about 20 mV and 90 mV (Fig. 4C and fig. S15), respectively, which could be attributed to the iR drop of the RR electrode and overpotential for HER. The ERR and Vcell in Cell S 2 O 8 2-showed nearly no change. The average CE of RR, VE of the system, and FE (S2O8 2-) were like those seen in cycles 1-5, confirming that RR enabled a stable decoupled operation at a 10:1 ratio of the HER rate/PSR rate.
We then performed highly rate-mismatched ModES production at the HER rates of 500 C and 1000 C and the PSR rate of 10 C, using a smaller CuHCF RR electrode. When the ratios of mismatched rates were 50:1 and 100:1, fast HER can be achieved in 4.77 s (Fig. 3B and fig. S16) and 1.87 s (Fig. 3B and fig. S17), respectively, in agreement with the rate-mismatched cycling results of the CuHCF electrode (Figs. 2Fand G). The ERR decreased by about 70 mV in Cell S 2 O 8 2-and increased only about 170 mV in Cell H 2 with the increased HER rates (Fig. 3C). The Vcell did not change in Cell S 2 O 8 2-and increased by nearly 350 mV in Cell H 2 (Fig. 3C). Although the HER rate increased dramatically, the average CE of RR remained above 99.0 %, the FE of PSR was ~70 %, and the VE of the ModES system was above 91.8 % (Fig. 3D). These results confirm the tolerance and robustness of the decoupled ModES system operated at highly mismatched reaction rates. Consequently, this ModES system could effectively produce H2 within 2s and persulfate within 5 min, which makes it feasible to participate in different dynamic electricity markets.
We developed a computational modeling framework to simulate different electrochemical systems (coupled and decoupled), operating modes (flexible and inflexible power loads), and market participation strategies (Fig. 4A, see more details in figs. S18-26, and supplementary text). This framework allows us to quantify the economic benefits of connecting a flexible, decoupled ModES system to electricity markets operating at different timescales (DAM, RTM, and FR) and at different geographical locations in the PJM (Pennsylvania, Jersey, Maryland) market and in the ERCOT (Electric Reliability Council of Texas) market. All systems and operating modes compared in this study produce the same total amount of chemicals. Fig. 4B shows that the decoupled ModES system that operates under a flexible power load, particularly in PJM and Houston, TX (ERCOT), can reduce electricity costs by as much as 30-40% compared to the baseline system (a coupled system operating under an inflexible/constant power load). In areas with a strong incentive for market flexibility (high frequency of negative prices), such as the RTM in Panhandle, TX (ERCOT), operating under a flexible load can offset or even reverse electricity costs (generating revenue), highlighting the significant economic benefits of providing flexibility to the power grid. The decoupled ModES system has the ability to participate in the fast FR market to generate revenue (due to decoupling of reaction timescales), a flexibility feature that coupled systems cannot provide. To give more perspective on the benefits of participating in FR, we note that allowing the flexible ModES system to simultaneously participate in the DAM and in the FR market can generate up to 0.37 million USD/yr in revenue from the FR market and reduce the total electricity cost by 65% (relative to a flexible ModES system that does not participate in the FR market). A substantial spatial variation in electricity costs was also observed. For instance, in the PJM market, the annual electricity costs reach up to 3.5 million USD/yr, whereas responding to FR signals in Panhandle, TX results in revenue generation. These results not only demonstrate that the flexibility of the decoupled ModES system can unveil new revenue opportunities, but also that such opportunities are strongly affected by electricity market conditions.
We also note that cost reductions require the synchronization of chemical production (anodic and cathodic) with market signals (full details in figs. S23-26). For instance, Fig. 4C shows the chemical production schedule of a coupled system operated with constant power loads. In contrast, in the production schedule for a coupled system operated under flexible loads for the DAM in Houston, TX (Fig. 4D), the anode and cathode production schedules follow the slow dynamics of the DAM market. In the production schedule of the decoupled ModES system (Fig. 4E), it becomes apparent how persulfate production tracks the RTM market dynamics, while hydrogen production tracks the fast FR market dynamics. This separation in production schedules is key for enabling cost reductions.
We developed a modular electrochemical process with highly mismatched reaction rates that can participate in dynamic electricity markets to reduce the cost of electricity consumption for chemical production. Such a strategy is enabled by a fast proton battery material with excellent rate capability and electrochemical stability that delivers high Coulombic efficiency and capacity retention even under highly mismatched charging/discharging rates. The rate-mismatched ModES system achieves the co-production of hydrogen gas and persulfate with reaction rate ratio from 1:1 to 100:1 (when hydrogen production takes less than 2 s) and shows excellent Coulombic and energy efficiency ~ 100%. Furthermore, the developed computational framework shows that the ModES system operating under a flexible power load reduces electricity cost by 30-40% compared to the traditional coupled system operating under a constant power load. Such a decoupled ModES system can even generate revenue by participating in the FR market, which is impossible in the traditional coupled system. These results provide a new conceptual strategy to help decarbonize the power grid by connecting flexible electrochemical synthesis to dynamic electricity markets and show how the decoupled electrosynthesis can open new revenue opportunities. Flexible technologies with different storage/operation timescale, such as pumped hydroelectric storage (PHES), compressed air energy storage (CAES), battery systems, power to gas (P2G), and decoupled electrosynthesis, could participate in multiple markets and at different times/locations to help the grid absorb wind/solar power and monetize this service. (C) The novel decoupled electrosynthesis could effectively participate in the FR market, which is not possible for traditional systems with long storage/duration time.