A flexible small-scale electrochemical flow cell Authors: Thomas R Goodwin, Trent A Weiss, Alexander H Quinn, Fikile R Brushett Published: 2026 ABSTRACT Graphical abstractKeywords: Electrochemistry, Flow cell, Redox flow battery, Repeatability, Battery cycling Specifications table Hardware nameGen 2 Flow CellSubject area•Engineering and materials scienceHardware type•Mechanical engineering and materials scienceClosest commercial analogThe closest commercial analog is a flow cell developed by Fuel Cell Technologies, Inc. The cell can be found at the following link: https://fuelcelltechnologies.com/single-cell-hardware/cell-hardware/5-square-centimeter-cell-hardwareOpen-source licenseDocumentation / Illustrations / Drawings / Figures•CC BY-NCHardware•CERN-OHL-PCost of hardware$300Source file repositoryNone •Engineering and materials science Engineering and materials science •Mechanical engineering and materials science Mechanical engineering and materials science •CC BY-NC CC BY-NC •CERN-OHL-P CERN-OHL-P Hardware in context Electrochemical flow systems hold promise for many applications including power generation, energy storage, water remediation, material separation and recovery, and chemical synthesis [1], [2], [3]. Exploring the electrochemical cell alone for any of these applications might involve considering different electrolyte chemistries, electrodes, membranes, flow fields, and operational conditions. While combinatorial explosion will often preclude exploration of all possibilities, screening campaigns generally benefit from multiple experiments. As such, understanding the influence of the different cell components and materials on performance and durability can inform system design, scale-up potential, and economic viability. Having well-defined, flexible flow cells can facilitate higher-throughput experimentation. Existing commercial flow cell platforms are either expensive or relatively inflexible in material selection and component type. Accordingly, academics frequently design and manufacture bespoke cells to support their specific research and educational objectives. For these cells, the documentation practices vary—rarely do they facilitate ease-of-reproduction. Motivated by expanding availability and accessibility of flow cell hardware, several kits have been published in peer-reviewed journals or in databases with the intent of disseminating designs to be reproduced [4], [5], [6], [7], [8], [9]. Furthermore, the Flow Battery Research Collective (FBRC), has published open-source designs for a 4-cm2 single-cell and a 175-cm2 modular stack [10]. Amongst these efforts, there exist opportunities to broaden the offerings of flow cells by extending the range of device sizes and incorporating additional design features. For instance, many commercial and open-source flow cells exist at scales greater than 10-cm2, requiring more material to fabricate constituent components (e.g., flow fields) and larger balance of plant (e.g., reservoirs, pumps) to enable cell operation. Decreasing the geometric area enables use of smaller electrodes and membranes in the cell and lower electrolyte volumes; this, in turn, alleviates bottlenecks in materials production from expensive precursors or laborious fabrication. As a result, higher-throughput testing of membrane, electrode, and electrolyte materials can be done without committing to larger-scale synthesis when compared to other open-source flow cell systems. Finally, well-documented experimental apparatus can improve inter-system comparisons across the growing literature, leading to a portfolio of reliable device options in support of different chemistries and testing goals [11]. The flow cell presented here is a small area (2.55 cm2) device designed to be flexible in componentry and applicable to many electrochemical systems [12], [13]. It is relatively easily manufactured and simple to assemble. The chosen materials are compatible with a wide range of aqueous and non-aqueous chemistries. The compactness of the cell facilitates a relatively small footprint in the laboratory and enables the use of small electrolyte volumes. The design incorporates interchangeable flow fields which require minimal graphite, often a material with expensive stock and machining cost. Due to its small size, this cell is minimally influenced by spatial heterogeneities (i.e., it is akin to a “differential” element in a larger cell). The cell, often referred to as the Gen 2 cell, has been used for a wide variety of purposes in our group including, but not limited to, redox flow batteries [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], carbon dioxide (CO2) electrolyzers [35], [36], and diagnostic flow cell configurations targeting electrode or membrane performance [37], [38], [39]. It has also been adopted by several different research groups whose published studies provide new users with additional datasets to compare against [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50]. Moreover, this cell has the potential to accelerate research in other emerging areas of electrochemical science and engineering (e.g., critical materials recovery, electrofuel production, CO2 capture) and is well-suited for early-stage exploratory investigations. Our intent is to document the features and operation of this flow cell to facilitate its use within the broader research community. Hardware description Design overview This flow cell accounts for five primary design considerations: 1) small requisite sample volume, 2) chemical stability, 3) scalable performance, 4) flow field variety, and 5) low cost. The flow field design is loosely based on a previously published subscale cell (25 cm2) from the former United Technologies Research Center (UTRC, now RTX Technology Research Center) [51], [52]. Basic performance metrics of our cell were compared against those of the 25 cm2 cell and showed consistent values, confirming that this cell design is representative of larger cells [53]. The half stack of components consists of an electrolyte diffuser with ports for tube connections and divots for O-rings, an interchangeable flow field, gaskets with rectangles cut out for the electrodes to sit in, and a separator or membrane; a full stack mirrors this description. Alignment pins (0.125″ in diameter) are used to position stack components and bolts are used to ensure constant compression of each assembly. With a geometric electrode active area of 2.55 cm2 and an internal cell volume on the order of 1 mL per half-cell, this compact design requires minimal material for testing. As shown in Fig. 1, there are alternative component options to expand the capabilities of this flow cell. The ⅛” NPT fittings enable an assortment of means to connect to soft (e.g., using barbed fittings) and hard tubing (e.g., using swaged or compression fittings) of various dimensions. The standard #014O-rings (0.489″ ID, 0.629″ OD) are made in a broad range of materials to suit different chemistries. Our default material choice for O-rings in many aqueous and non-aqueous systems is ethylene propylene diene monomer (EPDM) rubber. We have also employed Viton® for acidic solutions. The selection of ¼-28 bolts and nuts is due to their common use in the United States of America. However, M6 metric hardware, whose bolts are slightly smaller in diameter, can also be employed. 316 stainless steel is a good general-purpose material for the washers, bolts, and nuts. Finally, alignment pins should be solid, ionically and electronically insulating, ideally rigid and straight, and slightly undersized relative to their holes to ensure free motion of the gaskets and flow fields along the pin for assembly—polypropylene rods are a common choice. Exploded and assembled views of the Gen 2 flow cell. Diffusers The function of the diffuser is to compress the cell stack, distribute fluid to the backside of the flow fields, and provide two NPT ports for a variety of fluidic connectors. The diffuser also situates the alignment pins and the bolts to constrain the cell components during and after assembly. Polypropylene, our material of choice, was selected for its accessibility, inexpensiveness, relative ease to machine, and chemical compatibility with a broad range of aqueous and nonaqueous electrolyte chemistries. Polypropylene has a mechanical disadvantage in this design: at room temperature it is above its glass-transition temperature meaning the diffuser tends to deform near the bolt holes when the cell is compressed [54]. In practice, this is not a concern for shorter-duration experiments or when employing compressible gaskets (discussed below) that allow the flow cell to seal better. However, for weeks-long cycling experiments, this creep can cause a loss in compression at the gasket and lead to failure via leaking. There are several remedies including a change of the diffuser material (e.g., materials with higher tensile strength, such as polyether ether ketone, polycarbonate, acrylic, or polyoxymethylene), thickening the dimension of the diffuser (at the expense of machining effort and requiring longer hardware), or employing a backing plate based on a rigid material (e.g., stainless steel) [55]. Material choice should be considered in the context of the experimental system, focusing on the required mechanical strength and chemical compatibility. Other design options may exist for minimizing diffuser complications. Flow fields The flow fields distribute electrolyte throughout the porous electrode and behave as electronic current collectors. Our system design allows for interchangeable flow fields, enabling different flow patterns and materials. Some canonical examples are the flow through flow field (FTFF), interdigitated flow field (IDFF), serpentine flow field (SFF), and parallel flow field (PFF), shown in Fig. 2. Such flow field designs are inspired by prior developments and investigations in fuel cells, and can serve variable purposes [56], [57]. For instance, SFFs and PFFs are often favored for use in gas flow systems, whereas IDFFs and FTFFs are more common in flow batteries. IDFFs of variable channel and rib dimensions for this flow cell have been explored in the literature [44], [58]. This architecture is compatible with custom flow field designs as well, given the design follows the geometric constraints (e.g. inlet and outlet locations, flow path area). As for material, we often use a resin-impregnated graphite because of its impermeability and broad chemical compatibility; however, other materials such as copper, stainless steel, titanium, and aluminum can and have been used. In principle, electronically insulating materials (e.g., plastic) could also be employed. For example, insulating flow fields may be useful for high-throughput pressure drop measurements of novel flow field designs, for which 3D printing can accelerate iterative design and fabrication. Additionally, transparent plastic can enable optical interrogation of the flow, while the current collector is connected to electrode separately. Illustration of patterned flow fields that are commonly used in this cell. Black solid arrows indicate possible paths for fluid flow based on channel configuration. Gaskets The gaskets are designed to provide a pocket between the membrane/separator and flow field for the electrode to sit in, dictate the electrode compression, and seal each half cell from the external environment and from each other. A good seal prevents electrolyte crossing between half-cells (i.e., seals around the membrane) and prevents electrolyte from leaking out of the flow cell. A suitable gasket should be chemically compatible with the electrolytes of both half-cells. The mechanical properties will influence compressibility and sealing capability. “Incompressible” gaskets, those which minimally deform upon compression, may not seal as well as “compressible” ones but allow for greater control of electrode compression. “Compressible” gaskets, those which substantially deform under load, seal better but complicate estimation of the electrode/membrane stack compression. Candidate compressible gasket materials include silicone, ethylene propylene diene monomer (EPDM), buna-N, Viton, and other rubbers, where the choice will often be determined by chemical compatibility. In our build instructions—and, to date, in routine laboratory practice—we employ skived polytetrafluoroethylene (PTFE) gaskets given its excellent chemical compatibility, good reusability, and availability in a range of sheet thicknesses. The electrode compression can be controlled by changing the gasket thickness or stacking multiple gaskets together. While this approach functions well for aqueous systems, we have observed that low surface tension solvents (e.g., acetonitrile) readily leak at the gasket interfaces, often evinced by dried active species or solvent accumulating on the gasket periphery within a few hours). Expanded PTFE (ePTFE—e.g., Gore tape), by contrast, provides improved sealing for such solvents due to its compressibility; however, expanded PTFE gaskets cannot be reliably reused after compression. Creep deformation, like with the polypropylene diffuser, might be concerning for longer-duration operation of a flow cell [59]. While the above-proposed backing plate helps maintain effective sealing by limiting the degree of polypropylene diffuser creep, the PTFE gaskets themselves may deform with time and cause leaks days or weeks into operation. Longer operation may be achieved with more compressible and lower durometer materials such as ePTFE or EPDM. Electrodes and separators/membranes The design choice of stackable gaskets accommodates a wide range of electrode thicknesses, enabling testing of both novel constructs and commercial standards. For this flow cell, porous graphitic electrodes, such as carbon papers, felts, and cloths, which can be functionalized with catalysts through a variety of methods, are mostly commonly implemented. Metal foams and meshes are also compatible with this architecture. In addition, the flow cell readily supports small-area, nonstandard electrodes, such as those fabricated via non-solvent induced phase separation (NIPS) and electrospinning [22], [34]. A wide variety of membranes or separators can be utilized. The membrane or separator is positioned between two gaskets; because there is no dedicated cavity for this component, sealing is achieved by either (1) compression on the membrane itself such that electrolyte does not pass through its pores at the gasket-membrane interface or (2) deformation of the gasket stack around the region not covered by the membrane. Consequently, the maximum usable thickness of the membrane or separator depends on its compressibility and that of the gasket to conform around it. In practice, this flow cell accommodates separators and membranes ranging in thickness from ca. 10 μm to 1 mm, including commercial offerings such as Celgard, Daramic, Nafion, and Fumasep. This platform is also compatible with bespoke or emerging separators, membranes, and composites (e.g., ceramic composite membranes) [39], [60]. Critical balance of plant At minimum, the flow cell must be connected to tubing, a pump, an electrolyte reservoir, and a potentiostat. While there are many pumps available to which satisfy different operating requirements (e.g., pulsing behavior, flow rates, efficiency, pressure envelopes), peristaltic pumps are frequently employed due to their ease of use and simplicity in exchanging the tubing components that are wetted. Diaphragm pumps are another common choice, which may be preferred for certain applications because of the different mechanism of operation (e.g., avoiding oscillations at low flow rates, reducing mechanical wear on tubing). Tubing should be chemically compatible with the electrolyte(s). For small electrolyte volumes (on the order of 10 mL and smaller), long lengths should be avoided and, if possible, the diameter restricted (e.g., < 1/16″ ID) to minimize holdup. While it is possible to employ a wide range of tubing for peristaltic pumps, those not designed for such purposes may disintegrate faster and lead to leaks. Still, even well-designed tubing can leak. To facilitate easy replacement at the most-likely failure point (i.e., the section of tubing compressed in the peristaltic pump), tubing is frequently divided into multiple sections, with one section dedicated for the peristaltic pump head (e.g., norprene tubing), and the other sections of (often less-expensive) tubing selected for completing the fluidic circuit (e.g., teflon or polypropylene tubing). Electrolyte reservoirs vary in size and function. Critical considerations might include volume, material(s) of construction (for chemical compatibility, thermal tolerance), pressure rating, aspect ratio (as large diameter:height ratios may be challenging), and options for ports (i.e., inlet/outlet, instrumentation connections). Often, we employ glass reservoirs with open lids for acidic or neutral systems and polypropylene reservoirs for alkaline systems. Using hole punches, gaskets can be cut from materials compatible with electrolyte (e.g., EPDM) to accommodate tubing for the inlet and outlet flow paths as well as for inert gases (vide infra); it is often best to place the inlets on the opposite side of the reservoir from the outlets to avoid bypassing the bulk volume (e.g., inlet at bottom, outlet at top). Additionally, stir bars can be implemented to ensure homogeneous concentrations within the reservoir. Improved sealing apparatuses are sometimes needed for chemistries employing solvents that readily evaporate. For non-aqueous systems that require sealing, we often employ 10 mL volume molded perfluoroalkoxy (PFA) jars (Savillex, 102–0010-01). More specialized reservoirs can be used to address advanced challenges with flow cell use (e.g., removing headspace for gas to prevent volume imbalance). Finally, equipment to control the current or voltage supplied to the cell is required. Our device of choice is a potentiostat, which can reliably perform many electrochemical methods useful in the analysis of cell performance. Increased precision, higher current capability, access to a broader range of techniques, more mature software interfaces, and integration with dedicated instrument ecosystems tend to correlate with higher costs. Alternatively, more affordable equipment, like a battery cycler or a power source, can be used depending on the application of the flow cell. Alternative and additional ancillaries The preceding sections describe the basic needs for operation. In certain cases, additional equipment is required for experiments—we briefly discuss some of these here to ensure adequate consideration of necessary componentry for a distribution of experiments and to support more advanced campaigns. Gases are often supplied to the head space of the reservoirs or sparged into the electrolyte. Humidifiers, vessels containing the same solvent to that of the electrolyte and a gas sparger, can be used to pre-saturate the gas with solvent vapor to minimize evaporation from the electrolyte tank(s). Sensors installed in reservoirs or inline to tubing can offer information about the electrolyte during operation [61]. Such measurements might include conductivity, spectroscopy, or mass-transport limited concentration measurement (e.g., via ultramicroelectrode), amongst others. The specific need and applicability of each of these techniques depends on the target research question, the chemistry, and sometimes the scale of the system (e.g., smaller-volume systems can be more difficult to instrument). Exchange of certain components and/or modifications to materials or functionality can also expand the type of and/or elevate the quality of data obtained from the cell. For instance, variants of the diffuser design have been developed to implement reference electrodes and to facilitate elevated temperature operation by using a combination of thermocouples, heating elements such as patch or cartridge heaters, and a feedback temperature controller [20], [62]. Further, the inlet and outlet ports might be replaced with alternative fittings to ensure gas-tight seals and facilitate elevated-pressure operation [63]. The flow cell might be placed in dedicated enclosures to keep it in an inert environment (e.g., glovebox). Filters can be used to prevent accumulation of solids in the system, which might damage the flow cell, pumps, and/or impede flow. Other choices are described in earlier sections, and additional possibilities extend beyond those discussed here. Design files summary In Table 1, we summarize engineering drawings and computer aided design (CAD) SolidWorks 2025 files for the diffuser and flow fields (IDFF, FTFF, PFF, and SFF). In the Supplementary Information (SI), additional SolidWorks design files are included primarily to visualize the parts together in an assembly (Assembly Gen2.SLDASM). Some parts are SolidWorks files retrieved online from McMaster-Carr (including nuts, bolts, washers, O-rings, and NPT fittings). Other visualized components are approximations of parts that do not have drawings or CAD files readily available online (alignment pins, Pomona connectors). Note that these are not necessarily accurately designed. Instead, they are intended for visualization in the SolidWorks assembly file. All design files can be found in the Open Science Framework (OSF) repository for this article. Table 1Summary of the cell design files.Design file nameFile typeOpen-source licenseLocation of the fileGEN2_Electrolyte_Diffuser.pdfPDFCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAGEN2_Flow_Thru.pdfPDFCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAGEN2_IDFF.pdfPDFCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAGEN2_Parallel.pdfPDFCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAGEN2_Serpentine.pdfPDFCERN-OHL-PDOI:10.17605/OSF.IO/8WBKADiffuser_1_original.SLDPRTSolidWorks Part FileCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAFF_FTFF.SLDPRTSolidWorks Part FileCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAFF_IDFF.SLDPRTSolidWorks Part FileCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAFF_PFF.SLDPRTSolidWorks Part FileCERN-OHL-PDOI:10.17605/OSF.IO/8WBKAFF_SFF.SLDPRTSolidWorks Part FileCERN-OHL-PDOI:10.17605/OSF.IO/8WBKA Summary of the cell design files. GEN2_Electrolyte_Diffuser.pdf •This file contains the engineering drawing for the electrolyte diffuser, detailing its four screw holes for cell compression, two NPT ports for fluidic-loop connections, two fluid paths, and divots for O-rings. This file contains the engineering drawing for the electrolyte diffuser, detailing its four screw holes for cell compression, two NPT ports for fluidic-loop connections, two fluid paths, and divots for O-rings. GEN2_Flow_Thru.pdf •This file contains the engineering drawing for the FTFF, detailing the flow field neck with a banana plug port and two holes with dimensions of 13 mm × 1 mm for the fluid inlet and outlet. This file contains the engineering drawing for the FTFF, detailing the flow field neck with a banana plug port and two holes with dimensions of 13 mm × 1 mm for the fluid inlet and outlet. GEN2_IDFF.pdf •This file contains the engineering drawing for the IDFF, detailing the flow field neck with a banana plug port, three 1 mm-diameter fluid outlet holes, four 1 mm-diameter fluid inlet holes, and seven fluid channels with dimensions 16 mm × 1 mm. This file contains the engineering drawing for the IDFF, detailing the flow field neck with a banana plug port, three 1 mm-diameter fluid outlet holes, four 1 mm-diameter fluid inlet holes, and seven fluid channels with dimensions 16 mm × 1 mm. GEN2_Parallel.pdf •This file contains the engineering drawing for the PFF, detailing the flow field neck with a banana plug port, seven 1 mm-diameter fluid outlet holes, seven 1 mm-diameter fluid inlet holes, and seven fluid channels with dimensions 16 mm × 1 mm. This file contains the engineering drawing for the PFF, detailing the flow field neck with a banana plug port, seven 1 mm-diameter fluid outlet holes, seven 1 mm-diameter fluid inlet holes, and seven fluid channels with dimensions 16 mm × 1 mm. GEN2_Serpentine.pdf •This file contains the engineering drawing for the SFF, detailing the flow field neck with a banana plug port, one 1 mm-diameter fluid outlet hole, one 1 mm-diameter fluid inlet hole, and one winding fluid channel with a width of 1 mm. This file contains the engineering drawing for the SFF, detailing the flow field neck with a banana plug port, one 1 mm-diameter fluid outlet hole, one 1 mm-diameter fluid inlet hole, and one winding fluid channel with a width of 1 mm. Diffuser_1_original.SLDPRT •This file contains the SolidWorks part file for the diffuser, depicted in the GEN2_Electrolyte_Diffuser.pdf drawing. This file contains the SolidWorks part file for the diffuser, depicted in the GEN2_Electrolyte_Diffuser.pdf drawing. FF_FTFF.SLDPRT •This file contains the SolidWorks part file for the FTFF, depicted in the GEN2_Flow_Thru.pdf drawing. This file contains the SolidWorks part file for the FTFF, depicted in the GEN2_Flow_Thru.pdf drawing. FF_IDFF.SLDPRT •This file contains the SolidWorks part file for the IDFF, depicted in the GEN2_IDFF.pdf drawing. This file contains the SolidWorks part file for the IDFF, depicted in the GEN2_IDFF.pdf drawing. FF_PFF.SLDPRT •This file contains the SolidWorks part file for the PFF, depicted in the GEN2_Parallel.pdf drawing. This file contains the SolidWorks part file for the PFF, depicted in the GEN2_Parallel.pdf drawing. FF_SFF.SLDPRT •This file contains the SolidWorks part file for the SFF, depicted in the GEN2_Serpentine.pdf drawing. This file contains the SolidWorks part file for the SFF, depicted in the GEN2_Serpentine.pdf drawing. Bill of materials summary Table 2 is a bill of materials (BOM) for a complete set of flow cell components. Note that manufacturing costs for the flow fields and diffusers may vary based on stock material and supplier. Due to the cost of graphite stock, building ca. 10 cells is more economic than a single one. Thus, the “minimum cost” describes the minimum purchase required for building a single cell, whereas the “cost per cell” is the price of each component for building multiple cells. The cost of a cell has an upper bound of $834 (total minimum cost) and a lower bound of $297 (total cost per cell). Table 2Bill of materials. Currency is in 2025 USD.Used in this work?DesignatorComponentQuantity per cellMinimum costCost per cellSource of materialMaterial type✔GEN2_Electrolyte_Diffuser.pdfDiffuser, manufactured externally2 diffusers$120.00 /diffuser$240.00Adaptive Engineering, Inc.Polymer✔GEN2_IDFF.pdfInterdigitated Flow Field, milled in house from G347B isotropic graphite sheet2 flow fields$605 /graphite sheet$33.61MWI Inc.,https://mwi-inc.com/online-ordering/OtherGEN2_IDFF.pdfInterdigitated Flow Field, manufactured externally2 flow fields$50.00 /flow field$100.00Adaptive Engineering, Inc.Other✔ScrewStainless steel screw, 1/4″-28 thread, 2–1/4″ length, 10 per pkg4 screws$16.32 /pkg$6.53McMaster-Carr,https://www.mcmaster.com/92196a340/Metal✔NutStainless steel hex nut, 1/4″-28 thread size, 100 per pkg4 nuts$4.10 /pkg$0.16McMaster-Carr,https://www.mcmaster.com/91845A105/Metal✔WasherStainless steel washer, 1/4″ screw size, 100 per pkg8 washers$7.63 /pkg$0.61McMaster-Carr,https://www.mcmaster.com/90107A029/MetalGasket0.005″ Teflon Gasket Material (12″ × 12″ sheet)2 or 4 gaskets$7.12 /sheet$0.80McMaster-Carr,https://www.mcmaster.com/8569K38/PolymerGasket0.010″ Teflon Gasket Material (12″ × 12″ sheet)2 or 4 gaskets$11.72 /sheet$1.32McMaster-Carr,https://www.mcmaster.com/8569K41/Polymer✔Gasket0.003″ Teflon Gasket Material (12″ × 12″ sheet)2 or 4 gaskets$4.21 /sheet$0.48McMaster-Carr,https://www.mcmaster.com/8569K36/PolymerGasket0.002″ Teflon Gasket Material (12″ × 12″ sheet)2 or 4 gaskets$3.14 /sheet$0.36McMaster-Carr,https://www.mcmaster.com/8569K34/PolymerGasket0.001″ Teflon Gasket Material (12″ × 12″ sheet)2 or 4 gaskets$3.67 /sheet$0.60McMaster-Carr,https://www.mcmaster.com/1192N11/PolymerGasket0.010″ GORE® ePTFE Gasket Material (2″ × 50′ roll)2 or 4 gaskets$585.52 /roll$7.81Gallagher Fluid Seals,https://tinyurl.com/mtwrvrp6Polymer✔Banana plugBanana plug2 plugs$3.70 /plug$7.40Mouser Electronics,https://tinyurl.com/njtf3cbwMetalO-ringViton O-ring, 1/16 fractional width, 100 per pkg4 rings$9.38 /pkg$0.36McMaster-Carr,https://www.mcmaster.com/9464K19/Polymer✔O-ringEPDM O-ring, 1/16 fractional width, 100 per pkg4 rings$15.22 /pkg$0.61McMaster-Carr,https://www.mcmaster.com/9557K466/Polymer✔Barbed fittingPolypropylene barbed fitting, 1/16″ tube ID × 1/8″ male NPT, 10 per pkg4 fittings$4.81 /pkg$1.92McMaster-Carr,https://www.mcmaster.com/5116K116/PolymerBarbed fittingPolypropylene barbed fitting, 3/32″ tube ID × 1/8″ male NPT, 10 per pkg4 fittings$4.80 /pkg$1.92McMaster-Carr,https://www.mcmaster.com/5116K119/PolymerBarbed fittingPolypropylene barbed fitting, 1/8″ tube ID × 1/8″ male NPT, 10 per pkg4 fittings$4.82 /pkg$1.92McMaster-Carr,https://www.mcmaster.com/5116K137/PolymerBarbed fittingPolypropylene barbed fitting, 3/16″ tube ID × 1/8″ male NPT, 10 per pkg4 fittings$6.78 /pkg$2.72McMaster-Carr,https://www.mcmaster.com/5121K381/PolymerSwagelok fittingPFA Swagelok tube fitting, 1/8″ tube OD × 1/8″ male NPT4 fittings$29.15 /fitting$116.60Swagelok,https://tinyurl.com/27jfepxjPolymer✔Teflon tapeTeflon sealant tape, 48′ length × 1/2″ wide, 0.0028″ thick1 ft tape$2.71 /roll$0.06McMaster-Carr,https://www.mcmaster.com/4591K12/PolymerTubingMasterflex Norprene tubing (A60 G), L/S 14, 50′6 ft tubing$48.30 /roll$5.80Cole-Parmer,https://tinyurl.com/32tfyuvhPolymer✔TubingMasterflex Norprene tubing (A60 G), L/S 16, 50′6 ft tubing$50.40 /roll$6.05Cole-Parmer,https://tinyurl.com/4ju9vrdbPolymer Bill of materials. Currency is in 2025 USD. Build instructions The reusable diffuser, flow fields, and gaskets need to be manufactured. While the gaskets can be prepared in any laboratory stocked with common tools, the diffusers and flow fields can also be produced in-house by facilities equipped with an adequate computer numerical control (CNC) mill—we provide additional commentary on the production methods below. Electrode and membrane preparation approaches are also briefly documented; however, as fabrication of these components is often the focus of dedicated research efforts, we omit detailed discussion and only comment on methods applicable to commonly available materials often employed in flow cell embodiments. Thereafter, we document the complete cell assembly process at a level of detail necessary to achieve replicable data. Fabrication of diffuser and flow fields Several components are most-easily outsourced to a third-party manufacturer (e.g., Adaptive Engineering, Protolabs, Xometry)—for instance, we advise that the diffusers be fabricated by an external vendor, as achieving the dimensional accuracy required for the fine features necessary for repeatable experimentation can be challenging. While these parts can be machined on a small CNC mill by individuals with appropriate experience, we recommend that such work be undertaken only by those with appropriate expertise. It is also possible to manufacture the diffusers using 3D printing techniques (e.g., resin or fused deposition modeling (FDM) printing), but distinct challenges with chemical compatibility, dimensional tolerance, part warping, and sealing may arise depending on the chosen filament materials and printing equipment. Flow fields—typically fabricated from graphite due to its chemical compatibility with chemistries commonly utilized in the flow battery community, high electronic conductivity, and relatively low cost—may be machined in-house or outsourced. Graphite can be a difficult material to mill because machining generates conductive dust that can short equipment, poses respiratory hazards, and often necessitates specialized cleaning systems; additionally, graphite’s abrasiveness leads to severe tool wear, and its brittleness increases the risk of chipping during production. However, in our experience, the fine features required for high experimental accuracy are more readily fabricated in-house using thin, resin-impregnated graphite as opposed to other materials sets (e.g., stainless steel, copper, aluminum). If graphite is milled in-house, additional precaution should be taken to minimize exposure to the graphite dust generated (e.g., employ hermetic enclosures, outfit machining equipment with vacuum systems designed to capture graphite particulate, wear a respirator during handling and cleanup). Accordingly, outsourcing may therefore present as the optimal choice; vendors similar to those used for diffuser fabrication may be suitable, although alternative materials (e.g., copper, stainless steel, titanium) may need to be selected based on machine shop capabilities as well as the intended chemistry. Gasket preparation Gaskets are typically fabricated from PTFE of shore durometer 52D-55D with varying thickness which ensures compatibility with the electrolyte, simultaneous leak prevention, and known stack heights to control and quantify electrode compression, though we note that other materials including, but not limited to, EPDM and expanded PTFE can also be used. They are prepared by cutting the material into squares with four 0.129 in diameter holes for insertion of alignment pins (vide infra) and a 1.7 cm × 1.5 cm rectangular hole for the electrode stack. The gaskets are more readily fabricated in a die (e.g., a leather cutting die) rather than by hand to expedite production and generate repeatable and uniform pieces; however, alternative methods (e.g., CO2 laser cutter, vinyl cutter) may also be employed. The die template can be constructed from steel blades secured with a wooden backing; however, because precise production of robust templates can be challenging without specialized expertise, we recommend their manufacture be outsourced (e.g., Millennium Die Group). The blades are designed to punch gaskets from a sheet (e.g., of PTFE) in a manner that leaves holes for the alignment pins and an area in the center for electrode placement as shown in Fig. 3a. Photographs of components and tools necessary and/or recommended for flow cell assembly. (a) Die cutter (Millennium Die Group) designed to cut two gaskets simultaneously. (b) Vinyl cutter (Cameo 3, Silhouette) used to prepare an array of equivalently-sized electrodes. (c) Image of all components and tools necessary to assemble the flow cell where the individual components are listed in the bill of materials and in each step in the cell assembly. The numbered tools in the photograph are (1) ½” socket wrench for fittings, (2) 2× ¼” wrenches for banana plugs, (3) 7/16″ wrench for nuts, (4) Teflon tape, and (5) torque screwdriver for bolts. The remaining items are flow cell parts: (6) diffusers, (7) flow fields, (8) barbed fittings, (9) gasket stacks, (10) O-rings, (11) bolts, (12) nuts, (13) alignment pins, (14) washers, (15) membrane/separator, and (16) electrodes. Note that there may be alternatives for the tools in this image. The die is operated by placing a PTFE sheet of a desired thickness over the blades of the die, placing a plate of ¼” thick soft plastic (e.g., acrylic) over the top for the blades to punch into, inserting the assembly into a mechanical press, and applying pressure until the die has fully cut the gaskets. While the applied load will vary based on the sharpness of the blade and the gasket material, it has been internally found that a load of 3000 lbf on a 6 in × 6 in mechanical press is sufficient (i.e., approximately 100 psi). We recommend that pressures do not exceed 200 psi to avoid cracking the acrylic plate and dulling the blades from excessive wear. Electrode preparation Electrodes for redox flow cells comprise a variety of chemical (including carbon/graphite, nickel, copper, etc.) and physical identities (including papers, foams, felts, etc.) [17], [18], [64], [65]. Furthermore, a range of pretreatments can be applied to these materials, including acid soaking and thermal oxidation [21], [66]. The choice of electrode and pretreatment depends on the electrolyte chemistry and testing of interest. While extensive discussion of potential approaches is beyond the scope of this article, we emphasize that this selection should be deliberate and intentional to avoid unnecessary complexities that hinder hypothesis testing. Electrodes should be cut into rectangular pieces with dimensions approximately matching the gasket cavity size. In many cases, including this work, electrodes are cut to 1.7 cm × 1.5 cm, corresponding to the inner dimension of the gaskets to ensure consistent and tight sealing. At a minimum, electrodes must fully cover the flow field features to force convective flow through the electrode and reduce electrolyte bypass. Slightly smaller electrodes (e.g., 1.6 cm × 1.4 cm) can also be used, as these still cover the flow field features while lowering the probability of gasket overlap caused by electrode shifting during cell assembly, which can compromise compression uniformity or cell sealing. However, when electrodes smaller than the gasket inner dimension are utilized, care must be taken to limit positional drift during assembly to ensure adequate coverage of the flow field features to limit bypass, especially for low permeability electrodes. For certain electrodes and/or experimental studies, adequate repeatability in electrode dimensions can be achieved by cutting against a simple stencil pattern of appropriate size. CNC cutting tools, such as vinyl or laser cutters, can provide higher accuracy than manual cutting; however, their suitability depends on both electrode material and tool characteristics. For example, laser cutting may damage electrode edges—an effect whose relevance depends on the application—while certain vinyl cutters may entangle woven or fibrous electrodes and may be limited in depth of cut. An example of using a vinyl cutter (Cameo 3, Silhouette) to cut multiple carbon paper electrodes simultaneously is shown in Fig. 3b. To obtain a target electrode thickness, stacking multiple layers of material may be required. If different types of electrode materials are being used on each side of the cell, care should be taken to avoid unequal compression due to differences in intrinsic compressibility. In such cases, empirical measurements are required to determine gasket stack thicknesses that yield the intended compression on both sides of the cell, noting that certain compression targets may not be readily attainable. This consideration is addressed further in the cell assembly section. Membrane or separator preparation Similar to electrodes, membranes (or separators) studied in redox flow cells span a variety of chemical identities (e.g., polymers, ceramics, etc.) and may be subject to a variety of pretreatments (e.g., acid soaking) [67], [68], [69]. Selection again depends on the electrolyte chemistry and specific testing objectives. As before, while a comprehensive discussion of the options and reasoning of selection is beyond the scope of this work, membrane/separator choice should be deliberate and intentional to avoid introducing complexities that confound hypothesis evaluation. Membranes should be cut into square pieces measuring 2.5 cm × 2.5 cm. These dimensions enable full coverage of the electrodes and sufficient overlap of the gaskets to ensure sealing between the half-cells without obstructing the alignment pins. As above, consistent membrane dimensions can be obtained using an appropriately-sized stencil; alternative cutting methods (e.g., vinyl, laser) may also be used. In general, precise dimensional accuracy of the membrane is less critical than ensuring adequate size for cell sealing. If mechanical robustness or species selectivity are a concern, multiple membranes can be stacked to reduce the risk of puncture-induced shorting or to mitigate crossover. Cell assembly Upon preparation of all components and materials listed above, the cell employed in electrochemical experiments is assembled in a linear fashion. The materials and tools required for assembly are shown in Fig. 3c. Below, we list the general nature of each step in bold, and include commentary and assembly details where appropriate. We do not include instructions regarding incorporation of componentry beyond what is necessary for basic operation discussed in Alternative and additional ancillaries above. 1.Two diffusers are prepared with O-rings (Fig. 4a) by slotting them into the diffuser grooves. The O-ring type should be appropriate for the cell chemistry and the anticipated experimental conditions. Two diffusers are prepared with O-rings (Fig. 4a) by slotting them into the diffuser grooves. The O-ring type should be appropriate for the cell chemistry and the anticipated experimental conditions. 2.The flow fields are prepared with banana plugs (Fig. 4b) by inserting one plug into the hole of each flow field tab such that it points outward and away from the flow field face. A metal spacer and a nut are placed on the threaded side, and the nut is secured with a wrench. Care should be taken to tighten the nut such that good electrical contact is made without physically damaging the tab. The flow fields are prepared with banana plugs (Fig. 4b) by inserting one plug into the hole of each flow field tab such that it points outward and away from the flow field face. A metal spacer and a nut are placed on the threaded side, and the nut is secured with a wrench. Care should be taken to tighten the nut such that good electrical contact is made without physically damaging the tab. 3.Four bolts with washers are inserted (Fig. 4c) through the back of one of the diffusers, which is then laid flat with the O-rings facing upward. The bolt lengths should be selected such that they can reach through the entire stack and still be tightened down—in cases where thicker stacks are employed, longer bolts may be required. Four bolts with washers are inserted (Fig. 4c) through the back of one of the diffusers, which is then laid flat with the O-rings facing upward. The bolt lengths should be selected such that they can reach through the entire stack and still be tightened down—in cases where thicker stacks are employed, longer bolts may be required. 4.Alignment pins are then inserted (Fig. 4d) through the top into their predestined holes and pushed through the entirety of the diffusers to maximize rigidity. Alignment pins are then inserted (Fig. 4d) through the top into their predestined holes and pushed through the entirety of the diffusers to maximize rigidity. 5.The first flow field plate is added (Fig. 4e) by sliding it onto the alignment pins, where the flow field tab should be positioned over the notch in the diffuser body and the flow field design faces up and away from the diffuser. The first flow field plate is added (Fig. 4e) by sliding it onto the alignment pins, where the flow field tab should be positioned over the notch in the diffuser body and the flow field design faces up and away from the diffuser. 6.The gasket(s) are then added (Fig. 4f) in the same manner by sliding them over the alignment pins. The total thickness of the gasket stack is intentionally selected in flow cell experiments to achieve a target electrode compression (e.g., 20%). This is accomplished by first measuring the uncompressed electrode(s) thickness(es) and then stacking gaskets of varying thicknesses such that the final stack thickness corresponds to the remaining fraction of the electrode thickness (e.g., 80%). As stated in Electrode preparation (vide supra), if different types of electrode materials are utilized on either side of the cell that have different compressibility, simply stacking gaskets to a final fractional thickness of the uncompressed electrode(s) may be insufficient to achieve the target compressions. The gasket(s) are then added (Fig. 4f) in the same manner by sliding them over the alignment pins. The total thickness of the gasket stack is intentionally selected in flow cell experiments to achieve a target electrode compression (e.g., 20%). This is accomplished by first measuring the uncompressed electrode(s) thickness(es) and then stacking gaskets of varying thicknesses such that the final stack thickness corresponds to the remaining fraction of the electrode thickness (e.g., 80%). As stated in Electrode preparation (vide supra), if different types of electrode materials are utilized on either side of the cell that have different compressibility, simply stacking gaskets to a final fractional thickness of the uncompressed electrode(s) may be insufficient to achieve the target compressions. 7.The electrode(s) are positioned inside the hole in the gasket(s) (Fig. 4g). The electrode should be centered within the gasket and cover the entirety of the flow field grooves. The electrode(s) are positioned inside the hole in the gasket(s) (Fig. 4g). The electrode should be centered within the gasket and cover the entirety of the flow field grooves. 8.The separator(s) / membrane(s) is then centered over the electrode(s) and gasket(s) (Fig. 4h) such that the entirety of the electrode(s) is/are covered. Care should be taken to avoid disturbing the position of the electrode(s) within the gasket(s) below. The separator(s) / membrane(s) is then centered over the electrode(s) and gasket(s) (Fig. 4h) such that the entirety of the electrode(s) is/are covered. Care should be taken to avoid disturbing the position of the electrode(s) within the gasket(s) below. 9.Step 6 is repeated (Fig. 4i), adding the remaining gasket(s), again selected to achieve a target electrode compression. Step 6 is repeated (Fig. 4i), adding the remaining gasket(s), again selected to achieve a target electrode compression. 10.Step 7 is repeated (Fig. 4j), centering the electrode(s) within the gasket(s). Step 7 is repeated (Fig. 4j), centering the electrode(s) within the gasket(s). 11.Step 5 is repeated (Fig. 4k), where the flow fields should face inside toward the electrode stack. The tab of both flow fields should be pointing toward the same direction but horizontally offset. Step 5 is repeated (Fig. 4k), where the flow fields should face inside toward the electrode stack. The tab of both flow fields should be pointing toward the same direction but horizontally offset. 12.The second diffuser is then slid over the alignment pins (Fig. 4l) with the O-rings facing inward toward the stacked cell components. The second diffuser is then slid over the alignment pins (Fig. 4l) with the O-rings facing inward toward the stacked cell components. 13.Washers and nuts are added to the bolts (Fig. 4m) and threaded until finger tight. Washers and nuts are added to the bolts (Fig. 4m) and threaded until finger tight. 14.The bolts are tightened using an adjustable torque screwdriver (Fig. 4n) after standing the assembly upright. This is done in a star pattern to ensure equal compression across the cell which helps to minimize leaking. We recommend increasing the torque incrementally to achieve uniform compression—in our experience, progressing from 3 lb-in, to 7 lb-in, then 10 lb-in, and finally 14 lb-in yields consistent results, though the final torque required may vary depending on the materials used. The bolts are tightened using an adjustable torque screwdriver (Fig. 4n) after standing the assembly upright. This is done in a star pattern to ensure equal compression across the cell which helps to minimize leaking. We recommend increasing the torque incrementally to achieve uniform compression—in our experience, progressing from 3 lb-in, to 7 lb-in, then 10 lb-in, and finally 14 lb-in yields consistent results, though the final torque required may vary depending on the materials used. 15.Four polypropylene barbed fittings are then prepared with Teflon tape (Fig. 4o) that is wrapped clockwise with the barb oriented downward to create two layers—we have found that the resulting thickness helps minimize electrolyte leaking. Four polypropylene barbed fittings are then prepared with Teflon tape (Fig. 4o) that is wrapped clockwise with the barb oriented downward to create two layers—we have found that the resulting thickness helps minimize electrolyte leaking. 16.The fittings are screwed into the holes of the diffuser bodies (Fig. 4p) until hand-tight and then further tightened using a socket wrench until the fittings are roughly 8 threads (for standard ⅛” NPT fittings) or ca. 5/16″ deep into the holes, such that the fittings are not flush with the diffuser body. Over-tightening has been empirically found to induce leaking. The fittings are screwed into the holes of the diffuser bodies (Fig. 4p) until hand-tight and then further tightened using a socket wrench until the fittings are roughly 8 threads (for standard ⅛” NPT fittings) or ca. 5/16″ deep into the holes, such that the fittings are not flush with the diffuser body. Over-tightening has been empirically found to induce leaking. Photographs depicting each step of the cell assembly process. (a) Preparation of diffusers with O-rings. (b) Addition of banana plugs to flow fields. (c) Preparation of a diffuser with bolts. (d) Alignment pins inserted into the diffuser. (e) Flow field placed onto diffuser. (f) Addition of first gasket stack. (g) Incorporation of electrodes within the gasket(s). (h) Placement of the membrane centrally over the first gasket stack. (i) Addition of a second gasket stack. (j) Incorporation of electrodes within the gasket(s). (k) Flow field placed facedown onto gasket stack. (l) Second diffuser placed facedown onto the flow field. (m) Washers and nuts are added to the bolts. (n) Bolts tightened with a torque wrench. (o) Compression fittings prepared with Teflon tape. (p) Prepared compression fittings are screwed into the diffusers. Connection to tubing and reservoirs For aqueous chemistries, barbed fittings paired with Masterflex tubing provide a cost-effective and straightforward connection strategy. For low to moderate flow rates (e.g., <20 mL min−1) with Masterflex peristaltic pumps, 1/16″ tube ID × 1/8″ male barbed fittings in combination with L/S 14 tubing are suitable. At higher flow rates, clamps may be used to secure tubing to barbed connections, or straight pipe fittings may be used as an alternative. In contrast, nonaqueous systems often require PTFE or Norprene tubing with PFA fittings (see Table 2) to mitigate leakage when working with low-viscosity or volatile solvents (e.g., acetonitrile). These components may be interfaced with the Masterflex tubing used in the pump head via stainless steel Swagelok unions. Note that this guidance is intended as a heuristic, as optimal tubing or fitting selections will depend on chemical compatibility and pump specifications. When cutting tubing, identical lengths should be used on both sides of the cell to ensure equivalent pressure drops. Failure to do so can result in volume imbalances in the reservoirs in symmetric or full-cell configurations (vide infra) over extended operation due to pressure-driven flow across the membrane/separator. The tubing length should also be minimized to reduce pressure-drop related complications, particularly when using small diameter tubing, and limit liquid holdup, which may influence measured conversion. In the configuration described herein, two segments of ca. 20 cm and two segments of ca. 70 cm of Masterflex L/S 14 tubing are connected to the lower and upper barbed fittings, respectively; note that the required lengths may vary depending on spatial constraints and experimental layout. The longer tubing segments are routed into glass reservoir(s) and should extend to the bottom and remain fully submerged in the electrolyte after filling. The number of reservoirs is dependent on the operating configuration: in single-electrolyte configurations, all inlet and outlet tubing is connected to a single, common reservoir (Fig. 5a), whereas in a symmetric cell or full-cell configurations, each diffuser or half-cell is supplied by an independent reservoir (Fig. 5b) [13]. In the single-electrolyte configuration, the same solution is circulated through both half-cells, such that oxidation and reduction of the same species occur simultaneously on opposing electrodes. As both chambers receive identical electrolyte composition and state-of-charge (SoC), concentration gradients across the membrane are minimized, thereby suppressing crossover effects and maintaining steady-state operation. This configuration is particularly well-suited for polarization and impedance measurements (vide infra), as it avoids any overpotential dynamics associated with shifts in SoC and minimizes sensitivity to membrane selectivity. However, because no net compositional difference is sustained between each half-cell, this configuration does not permit cell cycling. By contrast, in the symmetric cell configuration, both half-cells have independent reservoirs that contain the same electrolyte composition, which may be initialized at different SoCs, thereby affording cell cycling while maintaining chemical symmetry. This arrangement avoids complications associated with pairing two distinct redox couples, as in a full-cell configuration, and reduces sensitivity to active-species imbalance or membrane crossover due to the relatively small electrochemical potential gradients between half-cells. It should be noted that maintaining two independent fluidic loops can reduce pressure imbalances across electrodes and half-cells, as well as potential SoC discrepancies arising from asymmetric inlet concentrations. Schematic depicting the connection diagram for tubing into the cell and peristaltic pump for the (a) single-electrolyte configuration and (b) symmetric cell configuration. Photographs of these respective setups are shown in (c) and (d), which also include a sparging assembly (larger media jar, left) configured to provide an inert gas blanket over the electrolyte. The other ends of the longer tubing segments are connected to the lower diffuser fittings. The shorter segments are routed into the reservoir(s) but remain elevated just above the electrolyte level such that they are not submerged but still minimize splashing, and the opposing ends are connected to the upper diffuser fittings. We note that certain chemistries may require alternatives to glass reservoirs; in such cases, Savillex jars may be used in conjunction with the PTFE tubing (vide supra). For water- or oxygen-sensitive chemistries, the assembled cell can be transferred to a glovebox or other inert environment for operation. When using barbed fittings, completing tubing connections inside the glovebox may be preferable. In contrast, for PFA fittings requiring wrench tightening, all connections to the cell should be completed outside of the glovebox prior to transfer. Operation instructions Pump calibration In order to ensure the most accurate results, it is recommended that the pump utilized for experimentation is calibrated prior to each use. The calibration procedure may vary by pump type, manufacturer, and model; accordingly, detailed instructions are not provided here, and consultation of the pump manual or manufacturer is advised. We note that, in general, pumps can either be directly calibrated to reset the readings for volumetric output, or may require construction of a calibration curve if resetting is not possible. Connection to pump and filling cell with electrolyte When a Masterflex peristaltic pump is utilized, the central section of the longer tubing segments should be seated in the pump head in accordance with the manufacturer’s instructions, after which the head is closed. Empirical observations indicated that these pumps or comparable models can impart sufficient friction to draw the tubing through the pump head during operation. Wrapping tape around each tubing segment near the inlet and outlet of the pump head can mitigate this effect (Fig. 5b); however, care should be taken in applying the tape so as not to constrict the tubing. Application of silicone grease to the pump head as recommended by the manufacturer may provide a similar benefit and help protect against corrosion in the event of electrolyte contact; however, it should be applied properly so as to minimize tubing slippage, accumulation of dirt and dust, or degradation of the lubricant. Electrolyte solutions are then formulated in accordance with the experimental protocol (e.g., deterministically prepared 50% SoC solution). For large glass reservoirs, electrolyte volumes can simply be measured with a graduated cylinder and dispensed directly into each container. When alternative reservoir designs are utilized, backfilling may be required (e.g., Savillex jars). This process is performed by submerging the tubing ends connected to the upper fittings into a vial of prepared electrolyte and operating the pump in a “reverse flow,” in which the direction of pumping is opposite that used during experiments, to draw electrolyte into the system. In configurations employing two reservoirs containing identical electrolytes (i.e., symmetric cell), volumes are measured out into two separate vials with a pipette to ensure equivalence. After the electrolyte has been fully transferred, the pump should be stopped, the tubing connected to the reservoir(s), and the direction of pump operation returned to the experimental flow direction. Electrical connections When fabricated from an electrically conductive material such as graphite, the flow fields described above function as current collectors and are directly interfaced with a potentiostat or battery cycler. Electrical connections to the flow cell may be established using alligator clips or Pomona cables. When Pomona cables are connected to banana plugs, even pressure should be applied to both sides of the graphite tab to avoid mechanical damage. The flow fields are compatible with two-point measurement configurations, in which the current-carrying electrodes (i.e., working and counter) are electrically joined with their respective voltage-sensing leads (i.e., working sense and reference) (Fig. 6a). For example, in standard four-lead potentiostat arrangements, the working and working-sense leads are combined to form the working electrode, while the counter and reference leads are combined to form the counter electrode, with each pair connected to a corresponding flow field. However, four-point configurations have been empirically found to yield lower measurement noise and inductance, particularly for sensitive techniques such as electrochemical impedance spectroscopy (EIS). Such setups may be implemented by connecting the voltage-sensing leads to the bottom of the flow field tab using alligator clips (Fig. 6b). Accordingly, four-point configurations are preferred when experimental conditions permit. (a) Schematic and associated photograph depicting the connection diagram for potentiostat leads to the flow fields in a two-point electrode configuration. (b) The same as (a) for a four-point electrode configuration. Cell priming and burn-in Once all connections are made and the cell is prepared for operation, the pump speed is set to that desired for the experiment. When the cell is assembled as described, the electrolyte is drawn from the bottom of the reservoir, enters the flow cell through the lower fittings, and exits through the upper fittings. Bottom-to-top flow through the cell has been observed to facilitate removal of entrained air bubbles. The cell and balance-of-plant components should be inspected for rapid leaks, particularly at fittings or tubing junctions, as these locations are most susceptible to mechanical instability. If no leaks are observed, the flow cell should be rotated side to side, gently shaken, or lightly tapped on a rigid surface to dislodge air trapped within the porous electrodes and promote complete wetting. During operation, the system should be monitored for salt deposits or electrolyte loss, which may indicate the presence of slow leaks or insufficient sealing. Under certain conditions—such as specific material sets or electrolyte compositions—polarization can induce material “activation” through chemical modification of the electrodes or changes in membrane composition [70]. These effects may be significant depending on experimental objectives. Various approaches can be employed to induce such activation or “burn-in,” including application of electrochemical conditioning protocols, electrolyte exchange, or sustained electrolyte circulation. Definitions of burn-in vary as the distinction between short-term conditioning processes and longer-term equilibrium or degradation-driven changes in material performance is not always clearly delineated. Burn-in can strongly influence data repeatability and temporal consistency; although specific procedures are not detailed here, validation of repeatability across multiple experiments is strongly advised [11], [71]. Electrochemical testing protocol A range of electrochemical tests may be conducted to assess the performance of redox chemistries or component materials; among these, the most prevalent are EIS, polarization, and cell cycling. EIS is often applied to gather electrode and cell phenomena over a wide range of timescales (frequencies) using small voltage or current perturbations. At the cell level, it is commonly used to deconvolute resistive contributions into ohmic, kinetic, and mass-transport components, probe temperature-dependent behavior, and monitor the SoC and/or state-of-health (SoH). At the electrode scale, it can be used to estimate electrochemical surface area (ECSA) or other capacitive properties, while also capturing structural or compositional heterogeneities across the electrode(s). More broadly, EIS can be employed for system-level diagnostics, including identification of contact resistances, parasitic inductive elements, and other non-idealities. However, this technique necessitates specialized hardware beyond that standard for common battery cyclers. Rigorous application of EIS also requires careful consideration of analytical approaches (e.g., distribution of relaxation times, equivalent circuit fitting), validation criteria such as Kramers-Kronig consistency, and experimental constraints including linearity and time-invariant system behavior. Given this complexity, we encourage the interested reader to consult the rich and growing literature on best practices in EIS analysis [72], [73]. Polarization measurements evaluate the steady-state voltage response of a system over a range of applied current densities (or, less commonly, the current response to stepped potentials) [74], [75], [76]. By progressively increasing the imposed load and allowing the cell to approach steady operation at each step, polarization curves capture the cumulative impact of ohmic, kinetic, and mass-transport losses under practical operating conditions. From these data, key performance metrics including the area-specific resistance (ASR), limiting current density, onset of mass-transport limitations, and optimal operating windows can be extracted. Polarization is advantageous for flow cell systems because it can be performed rapidly, requires only standard potentiostatic/galvanostatic control without specialized hardware, and affords straightforward data analysis and interpretation. It should be noted that meaningful comparison across datasets requires careful control of flow rates, equilibration times, SoC swings, and thermal conditions to ensure consistent steady-state assumptions. Cell cycling monitors the temporal evolution of cell voltage (or current) during repeated charge and discharge processes under galvanostatic (or potentiostatic) control [76]. Unlike polarization, which probes performance at instantaneous operating points, cycling evaluates dynamic behavior over extended operation and directly reflects electrolyte utilization, accessible capacity, and degradation processes. From cycling data, coulombic, voltaic, and energy efficiencies can be quantified, as well as capacity fade rates and round-trip performance under defined current densities and voltage limits (or vice versa). Extended cycling further enables characterization of redox species stability, membrane selectivity, and electrode durability. Because cycling integrates electrochemical, transport, and crossover phenomena over practical timescales, it provides system-relevant performance indicators for redox flow batteries and related technologies. As with impedance analysis, rigorous cycling studies require careful definition of cutoff criteria, preconditioning protocols, and electrolyte management to ensure that observed trends reflect intrinsic material behavior rather than experimental artifacts. As noted above, electrochemical protocols are a common approach to induce burn-in. They may also be used to detect its presence—for example, repeated measurements can track the consistency of results over time. These approaches are most effective when the timescales of equilibrium processes—such as decomposition of active species or degradation of constituent components that influence test outcomes—are significantly longer than those typically associated with burn-in, such as changes in electrode surface composition. Given the breadth of possible protocols, their extensive coverage in the literature, and the dependence of implementation on system-specific considerations, no distinct protocols are prescribed here. Instead, electrochemical protocols should be selected and tailored to the target performance metrics and executed accordingly. In this context, techniques such as impedance spectroscopy, polarization, and galvanostatic cycling provide complementary insights and, when applied judiciously, can collectively enable robust and interpretable evaluation of system performance. Cleaning and disassembly The extent of flow cell cleaning may vary depending on the chemistries employed, conditions explored, and subsequent experimental plans. The following instructions outline procedures found to effectively clean the cell while minimizing hazards, contamination, or spills. Note that more rigorous cleaning procedures may be necessary if future experimental plans involve chemistries highly sensitive to contaminants or incompatible with the prior chemistry. The potentiostat leads should first be carefully disconnected from the flow fields. All of the electrolyte should then be returned to the reservoir(s) using the pump—this may involve either lifting the tubing from the bottom of the reservoir or reversing the pump direction, depending on the setup utilized. In a manner similar to the cell priming step, the flow cell should be turned side to side several times, gently shaken, or lightly tapped on a rigid surface to expel any remaining electrolyte trapped in the porous electrodes. The reservoir(s) are then emptied into an appropriate waste container, and the tubing and cell are rinsed with an appropriate solvent such as that utilized in the electrolyte. One example of this rinsing step is to refill the reservoir(s) with approximately 10 mL of solvent and circulate this through the flow cell system for roughly 5 min and a flowrate of 10 mL min−1. At the end of this period, the pump is used to return all of the fluid to the reservoir(s), and the reservoir(s) are again emptied into an appropriate waste container. The tubing is then disconnected from the flow cell fittings. For chemistries with toxic volatile components, the aforementioned steps should be conducted in a ventilated hood. The remaining flow cell components are then disassembled sequentially in reverse order from that listed in Cell assembly and rinsed using an appropriate cleaning solvent (e.g., water for aqueous chemistries, acetone or isopropyl alcohol for nonaqueous chemistries). Bolts, nuts, and banana plugs should be cleaned only as-needed, as these parts do not contact the electrolyte and repeated solvent exposure may induce corrosion, adversely affecting both contact resistance and assembly integrity. Components fabricated from inert materials (e.g., polypropylene or PFA fittings, diffusers, and alignment pins, depending on material selection) may be submerged and sonicated in an appropriate cleaning solvent to remove lingering residues. PTFE gaskets and graphite flow fields should not be sonicated, as this can cause wrinkling or mechanical degradation of the gaskets and cracking of the graphite plates. When these cleaning procedures are followed with suitable solvents, all flow cell components and balance-of-plant materials (e.g., tubing) are generally reusable without performance degradation; however, depending on the chemistry and materials employed, this may not extend to electrodes and membranes/separators. Common challenges In our preceding instructions, we described several performance-verification steps (e.g., checking for leaks under flowing electrolyte), but there remain ample opportunities for flow cells to misbehave fluidically, electrically, and mechanically due to errors in assembly that manifest during operation. Here we highlight a few commonly observed issues and discuss corrective actions, while acknowledging this is an incomplete coverage of the possible under-performance and failure modes. Fluid leaks, which can spawn at connections, damaged tubing, O-rings, or gasketing are a persistent challenge in flow cells. Leaks at gaskets or O-rings can be caused by out-of-tolerance flow fields/diffusers, insufficient or uneven bolt tightening, or diffuser warping. These issues can be often addressed by re-tightening or further tightening the cell, employing backing plates that evenly distribute pressure across the gaskets, and/or replacing gasket materials with more compressible alternatives. NPT connections at the diffuser bodies rely on PTFE tape to seal against the threads, where the quality and/or number of wraps can influence sealability. Connections such as Swagelok and IDEX require proper technique to ensure sealing. Finally, gas-tight cell sealing can be difficult to guarantee. Several standard tests (e.g., ASTM E2930-13) exist for verifying gas tightness, which, in turn, is a reasonable check for liquid tightness [77]. One test involves pressurizing the flow cell attached to a sealed reservoir, sealing the system, and observing the pressure drop over time using a pressure gauge. Another such test might involve submerging the flow cell system (including all connections) underwater to seek bubbles while the system is under pressure. Ohmic issues can occur at multiple points within the electrochemical–electronic circuit comprising the flow cell components, external wiring, and potentiostat interfaces. Lower-than-expected ohmic resistance might arise from internal shorts (e.g., due to electrode materials puncturing the membrane/separator or dendritic growth). Short resistances can vary considerably, resulting in partial or near-complete loss of flow cell function. Impedance spectroscopy, particularly of a dry cell (i.e., an assembled cell without electrolyte injected), can be used to identify electronic shorts. Relatedly, an atypically low OCV might imply substantial crossover of active species due to membrane failure or underperformance (e.g., lack of selectivity). Conversely, higher-than-expected resistances can arise from contact or interfacial resistances. Contact resistances commonly arise at electrical junctions (e.g., leads, connectors, battery terminals) due to factors such as corrosion, often caused by inadvertent electrolyte exposure, or roughening associated with mechanical degradation, as a result of repeated cleaning or extended use. In contrast, interfacial resistances originate at material-material boundaries within the cell (e.g., flow field-electrode, electrode-membrane) and may result from the formation of passivating surface films, changes in surface chemistry, or physical degradation that reduces effective contact area. Routine testing of leads can identify contact resistances which may not be immediately apparent, for instance using a potentiostat (e.g., using impedance or measuring DC resistance), whereas interfacial resistances may require EIS, microscopic, or spectroscopic characterization to be fully elucidated. Volume imbalance between reservoirs can arise due to osmotic differences between half-cell electrolytes, electro-osmotic drag of solvent during operation, and/or pressure differences across the membrane (e.g., due to unequal electrolyte pumping rates, unequal tubing lengths in the fluid circuits, partial or complete clogs, or differences in heights of the two reservoirs relative to the cell). If osmotic in causation, volume imbalance may be a thermodynamic inevitability only slowed by low permeability membranes. Electrolyte compositions can also be altered or the system allowed to equilibrate. When pressure differences arise for other reasons, pressure management might involve active or passive controls (e.g., valving). Validation and characterization Materials and methods Herein, we demonstrate the performance of a simple, common aqueous electrolyte of 0.1 M potassium ferricyanide (K3[Fe(CN)6)], 99%, Sigma-Aldrich) / 0.1 M potassium ferrocyanide (K4[Fe(CN)6)], 99%, Sigma-Aldrich) in 1 M potassium chloride (KCl, 99%, Sigma-Aldrich) and deionized (DI) water (18.2 MΩ cm, Milli-Q), which were all used as-received. The solution was deterministically prepared at 50% SoC—defined by an equal concentration of Fe2+ and Fe3+ species—at room temperature in a volumetric flask. Our cell consisted of components readily available and commonly used in the flow battery community: 1 × of heat-treated Freudenberg H23 (Fuel Cell Store, 450 °C for 6 h in air (Thermo Scientific Thermolyne F47925) and then stored in ambient conditions) surrounded by a 0.003 in PTFE gasket cut from sheets (McMaster-Carr) using a custom die (Millennium Die) to achieve a ca. 20% electrode compression per half-cell, a single Daramic 175 (175 μm nominal thickness, Daramic LLC) separator, and an IDFF (standard size, 1 mm rib, 1 mm channel width) per half-cell. To isolate the contribution of the cell itself, we minimized variation attributable to component materials by using consistent batches of electrodes, membranes, and gaskets. The electrolyte was blanketed with humidified nitrogen (99.999%, Linde, passed through DI water) continuously during experimentation. Reservoirs were covered with aluminum foil to minimize exposure to UV radiation and associated degradation. We operate the pump at a constant flow rate of 10 mL min−1. We then developed an experimental protocol spanning both single-electrolyte and symmetric cell configurations and interrogated the system using EIS, polarization, and cell cycling. To quantify repeatability more rigorously, we introduce a framework distinguishing marginal and total repeatability, drawing on precedent in the prior literature that differentiates related concepts of repeatability, replicability, and reproducibility [11], [71]. Within this taxonomy, both experimental designs considered here fall under the definition of repeatability; the proposed distinction is therefore intended to refine this category, not to introduce a new classification. Marginal repeatability captures variability attributed to rebuilding the same cell using identical components: after each measurement, the cell was disassembled, cleaned, and reassembled to probe the intrinsic repeatability, independent of material- or operator-level differences. We note that the experiments detailed herein do not perfectly control all contributing factors and thus do not strictly measure ideal intrinsic repeatability—which is not practically attainable—but do provide a close approximation. Total repeatability, by contrast, captures the full variability introduced by different operators as well as minor geometric, fabrication, or chemical differences between nominally identical parts. This framework both quantifies and bounds the minimum variance achievable by a single operator and the maximum variance expected across different users. As such, we distinguish between repeatability and replicability, as each experimental trial was conducted by all three operators under identical laboratory conditions. Specifically, in this set of experiments, we control as closely as reasonably possible for materials differences (e.g., by obtaining electrodes and membranes from the same lots, conducting the heat treatment for all electrodes in the same environment simultaneously, and using electrolyte from the same batch), environmental differences (e.g., by running experiments for each operator simultaneously such that laboratory temperature fluctuations occur at identical times), and operational differences (e.g., by building cells with the same set of tools, with the same procedure, and testing with the same potentiostat). In addition to enabling robust comparison between single-electrolyte and symmetric cell measurements, the total repeatability experimental design also provides insight into phenomena such as cell burn-in. We employ the following protocol for each experiment, which was conducted simultaneously by three operators on different channels of the same potentiostat (BioLogic VMP3). The protocol comprised a single-electrolyte configuration measurement sequence and a symmetric-cell sequence, followed by cell reassembly. For the single-electrolyte configuration, measurements were performed in triplicate for each trial using 50 mL of electrolyte. Experiments began with a 10-min open-circuit voltage (OCV) hold. EIS was then conducted with a 10-mV amplitude around OCV versus the counter electrode. The frequency range spanned 1.0 MHz to 5.0 mHz with logarithmic spacing of 10 points per decade, as this was determined to be a suitable range for assessment of ohmic, kinetic, and mass transport resistances. A wait time of 0.10 periods was imposed between frequency changes, and five repeats were recorded per frequency. Following EIS, a second 10-min OCV hold was applied. Polarization via chronopotentiometry was subsequently performed by applying a sequence of current densities from 0 to ± 150 mA cm−2 in 10 mA cm−2 increments, alternating polarity at each step. Each current density was held for 1 min, with voltage recorded every 0.5 s and voltage limits set to − 0.5 V and 0.5 V. Due to the rapid equilibration of the ferri-/ferrocyanide system, all data from each hold were averaged for polarization analysis. Additionally, this chemistry allowed for shorter galvanostatic holds with no rest between applied currents. A final 10-min OCV hold concluded the sequence. This protocol was repeated two additional times for each cell, yielding three full measurements per trial. While we include impedance in our protocol to exemplify a typical procedure (where impedance is often used), we avoid discussing its interpretation due to the additional complexities of analysis covered extensively in other texts [72], [73]. Following completion of the single-electrolyte measurements, the electrolyte was removed, the cell was flushed with DI water and dried with nitrogen, and fresh electrolyte was introduced. The cell was then reconfigured to a dual-reservoir symmetric-cell setup where each reservoir was prepared with equivalent 50 mL aliquots of electrolyte. Experiments were conducted under the same conditions as the single-electrolyte protocol: a 10-min OCV hold, EIS under identical parameters, followed by another 10-min OCV hold and chronopotentiometry using the same current sequence. In contrast to the single-electrolyte configuration, the subsequent 10 min OCV hold was proceeded by galvanostatic cycling with potential limitation for 20 cycles. Charge and discharge currents were set to ± 25 mA cm−2, with potential limits of 0.5 V and − 0.5 V and time limits of 6 h per half-cycle, adjusted for theoretical capacity. Chronopotentiometry at 400 mA was then applied for 1 h with a 0 V potential limit, followed by chronoamperometry at 0 V for 6 h with a minimum current limit of 1 μA to return the solutions to 50% SoC. An OCV hold of 1 h was then implemented. Potentiostatic EIS was repeated under the previous conditions, followed by a 10 min OCV hold, and a final chronopotentiometry sequence using the same current protocol. The full protocol was repeated three times, with cell disassembly, cleaning, and reassembly using the same materials (electrodes and membrane) between trials, enabling evaluation of both marginal and total repeatability by avoiding confounding material effects. Single-electrolyte cell measurements Polarization performance across three different single-electrolyte cells is shown in Fig. 7. As described above, each user completed three repeats with the same cell components, with disassembly, cleaning, and reassembly performed between trials according to the protocol outlined above. To mitigate potential confounding effects associated with initial cell or component burn-in, only the third polarization measurement from each repeat is utilized in the plot. Notably, though not shown here, minimal differences were observed between datasets constructed using either the first or third polarization measurement from each repeat, indicating that any burn-in effects were limited under these conditions. All measurements across users were conducted in parallel to control for environmental effects (e.g., temperature, humidity). Polarization curves collected using identical single-electrolyte flow cell setups by three different users. The average voltages and error bars (one standard deviation) represent three repeat experiments per user. Between experiments, each user disassembled, cleaned, and reassembled the cell using the same components as the prior experiment (including the electrode and membrane). Electrolyte prepared from the same batch was also used. The flow cell builds each include one heat-treated FH 23 electrode per half-cell and a Daramic 175 separator. The electrolyte is 100 mM K3Fe(CN)6, 100 mM K4Fe(CN)6, 1000 mM KCl in DI H2O (50 mL). The flow rate is 10 mL min−1. In this system, the major contributions to the observed overpotentials are anticipated to be ohmic and mass transport resistances, as the ferri-/ferrocyanide redox couples are reported to exhibit facile reaction kinetics on carbon electrodes [78]. The magnitude of ohmic overpotential depends on the resistances of a cell (contact, electrolyte, membrane, etc.), so the cell overpotential values between systems will differ, even with the same electrolyte chemistry. However, flow cell systems in literature with similar electrolyte, cell components, and flow rates show comparable overpotentials to those in Fig. 7 [58]. Additionally, the polarization curves show relatively linear behavior, implying that 150 mA cm−2 is well below the limiting current density for this cell and experimental conditions, which represents the mass-transport-limited flux of the active species to the electrode surface. With respect to the plotted polarization results, marginal repeatability is described by the error bars for each individual user, whereas total repeatability is conceptually shown by observing the range of voltage values across the three independent users. We note that several statistical frameworks can be used to assess repeatability and measurement-system variability, including ANOVA Gage R&R and evaluating the measurement process (EMP) methods. ANOVA Gage R&R typically requires repeated measurements by multiple users across multiple independent parts to decompose variability into operator-, part-, and repeatability-associated contributions. As this study does not include measurements on multiple independent flow cells by each user, this framework cannot be applied rigorously here. EMP methods focus on evaluating measurement stability over time, with over 20 repeated measurements in a single time series, which is not the focus of this variability study. However, such approaches provide useful guidance for future experimental designs aimed at quantifying sources of variability in flow cell measurements. Instead, repeatability was quantified using coefficients of variation (CoVs), which enable direct comparison of variability among measurements with significantly different magnitudes, as encountered across the range of overpotentials collected in polarization testing. CoVs were calculated at each current density and subsequently averaged across all current steps to yield a single metric representing the overall variability in each user’s polarization results. The CoVs for each step and associated CoVavg is calculated as follows: where σpotential [V] is the standard deviation for the potential at a given current, μpotential [V] is the mean potential at that current, and nsteps [-] is the number of current steps applied in a polarization experiment. In defining CoV in this manner, we assume that all datasets share a common reference point (open circuit voltage (OCV), assumed to be zero for all polarization curves) and that the standard deviation scales approximately linearly with the mean. Visual inspection of the data in Fig. 7 suggests that both assumptions are reasonably satisfied [79]. To quantify marginal repeatability, the pooled CoV, CoVpooled, was calculated from the three individual user values: To quantify the total repeatability, the average potential at each current step for each user is treated as a single observation. For each step, the total mean and standard deviation are then computed across each of these observations (i.e., across each of the three user averages)—this gives the CoV at each step, which can be averaged as in Equation (1) to give the CoVtotal. It should be noted that because CoVtotal is calculated from the user averages, it does not account for repeat-to-repeat variability and thus does not exactly fit the definition of total repeatability introduced earlier. However, CoVtotal still captures the operator-dependent variability and thus can be compared to CoVpooled to assess the relative contribution of marginal repeatability to total repeatability. Table 3 contains the CoV values previously described. For reference, data with CoV values less than 1 is considered to have low variability [79]. The total CoV was greater than 2 times the magnitude of the marginal CoV, showing that the marginal repeatability has lower error than the total repeatability of the single-electrolyte cell configuration. The higher error of the total repeatability suggests that differences in cell components and user experience influence the cell performance more significantly than the cell building process. These results also highlight the importance of validating results across batches when studying individual component performance within the flow cell system (e.g., assessing active species chemistry across different batches of electrodes). Table 3Coefficients of variance for the polarization results.Cell configurationCoVavg,1CoVavg,2CoVavg,3CoVpooledCoVtotalSingle-electrolyte9.47 × 10-25.10 × 10-27.26 × 10-27.49 × 10-215.4 × 10-2Two-reservoir symmetric7.92 × 10-26.20 × 10-24.01 × 10-26.25 × 10-217.0 × 10-2 Coefficients of variance for the polarization results. Two-reservoir symmetric cell measurements Symmetric cell polarization performance is shown in Fig. 8. As described above, following completion of the single-electrolyte experiments, each operator flushed the cells and injected fresh electrolyte prior to initiating the symmetric cell measurements, keeping all internal cell components the same as the single-electrolyte configuration. This configuration was evaluated in triplicate, with disassembly, cleaning, and reassembly performed between repeats. The magnitude of overpotentials of the symmetric configuration averaged 5% lower than those of the single-electrolyte configuration. Assessing the repeatability, the total and marginal trends are consistent with the results of the single-electrolyte configuration in Fig. 7. Based on the values in the last row of Table 3, the ratio of total CoV to marginal CoV is greater than 2, again suggesting that the cell components and different operators have greater influence on repeatability than the building process. Comparing repeatability values between single-electrolyte and symmetric cell configurations, there are slight differences in the magnitude of CoVs. Changes in repeatability and overpotential magnitudes could be due to a variety of experimental reasons and the extent of difference between configurations can greatly depend on the chosen material sets and operating conditions. Ultimately, both configurations can be evaluated, as variability within each system may be dependent on the experimental conditions employed. Polarization was also performed before and after cell cycling to determine if cycling affected cell performance. For all three users, results show negligible changes between pre- and post-cycling polarization, suggesting that the cell components remain unchanged from the cycling procedure. Pre- and post-cycling polarization curves collected using identical two-reservoir symmetric flow cell setups by (a) user 1, (b) user 2, and (c) user 3. The average voltages and error bars (one standard deviation) represent three repeat experiments per user re-using the electrodes and membrane, and using freshly prepared electrolyte for each repeat. The flow cell build includes one heat-treated FH23 electrode per half-cell and a Daramic 175 separator. The electrolyte is 100 mM K3Fe(CN)6, 100 mM K4Fe(CN)6, 1000 mM KCl in DI H2O (50 mL per reservoir). The flow rate is 10 mL min−1. Charge–discharge cycling results for the two-reservoir symmetric cell are shown in Fig. 9. The average electrolyte utilization of the symmetric cell configuration reported here is ca. 84%; this is comparable to reported values which range from 80 to 90% for ferri-/ferrocyanide-containing cells at similar current densities and electrolyte concentrations [80], [81]. The ca. 16% of unused capacity is partially due to the galvanostatic nature of this cycling procedure. Specifically, by setting a constant current density, the potential reaches the user-defined voltage limits due to mass transport overpotential contributions before the cell can reach 0% or 100% SoC [82]. The coulombic efficiency values in Fig. 9, ranging from 99.4% to 99.9%, are also comparable to reported values, approaching 100% for every cycle [80], [81], [82]. Flow cell cycling data collected using identical two-reservoir symmetric flow cell setups by three users. The average properties (discharge capacity, coulombic efficiency) and their error bars (one standard deviation) represent three repeat experiments per user. The flow cell build includes one heat-treated FH23 electrode per half-cell and a Daramic 175 separator. The electrolyte is 100 mM K3Fe(CN)6, 100 mM K4Fe(CN)6, 1000 mM KCl in DI H2O (50 mL per reservoir). The flow rate is 10 mL min−1. Capacity and coulombic efficiency are highly consistent across all users and repeats, shown by the small coefficients of variance in Table 4. Quantifying this consistency, the ratios of total to marginal CoVs for capacity and coulombic efficiency data, are 0.17 and 0.97, respectively. These values suggest that total repeatability does not contribute more to overall variability than marginal repeatability, in contrast to the polarization results discussed earlier. As noted above, repeat-to-repeat variability is not captured in CoVtotal and may dominate in this case, leading to ratios below unity. Furthermore, these results suggest that cell cycling as a technique has high total and marginal repeatability, at least for this specific experimental setup and these operating conditions. However, it should be noted that the 20 cycles were performed over the course of three days. When assessing longer-term cycling performance for a system with minimal capacity fade and/or degradation, it is unclear if these repeatability trends will hold; further repeatability tests should be conducted in those cases. Table 4Coefficients of variance for the cycling results.MetricCoVavg,1CoVavg,2CoVavg,3CoVpooledCoVtotalDischarge Capacity2.25 × 10-22.32 × 10-21.06 × 10-21.97 × 10-23.44 × 10-3Coulombic Efficiency1.99 × 10-37.75 × 10-41.19 × 10-31.41 × 10-31.37 × 10-3 Coefficients of variance for the cycling results. Additional performance characterizations Using this identical flow cell setup, additional experiments can be performed to gain insight into material properties and/or elucidate device performance. Below, we briefly discuss four examples: pressure drop across components and cells, electrochemical surface area, bulk electrolysis, and full-cell cycling. Because of complexities associated with each of these techniques, we elect to link to relevant literature [15], [18], [21], [34], [83] and simply highlight high-level possibilities for this cell architecture. Pressure drop measurements of electrodes can help estimate fluid dynamic losses due to pumping. Placing a pressure sensor upstream of the cell and another downstream of the cell can be used to measure the pressure drop across a half-cell—comparison to that for an empty cell enables extraction of the pure component data. Relationships between pressure drop and variable flow rates can be established and physical properties, such as electrode permeability, can be estimated. Tradeoffs between the performance gains and energetic losses associated with different pumping rates are an important point of comparison for different flow field and electrode choices. Electrochemical surface area (ECSA) can be estimated with capacitance measurements which offers insight into the electrode area available to support redox reactions. In brief, an electrolyte that does not have redox active species should be chosen (e.g., one only containing the supporting salt). Cyclic voltammetry measurements of the cell are then collected in a potential region in which neither electrode participates in faradaic reactions. The CVs can be related to the capacitance, typically by the slope of the current magnitude against the scan rate at a particular potential. This cell capacitance can be translated to a single-electrode capacitance, assuming equivalent capacitances at two electrodes, after accounting for contributions made by the cell components. This value is then divided by an equivalent value obtained for a smooth planar substrate of similar composition. Note the choice of electrolyte, potential region, electrode (and its pretreatment), and more can affect the measured capacitances. In addition, the surface area estimated may not necessarily be fully accessible to the particular redox-active species of interest, for instance due to solvation and size exclusion effects. Bulk electrolysis is often performed using reticulated vitreous carbon (RVC) electrodes in stirred batch cells comprising two distinct reservoirs separated by an ion-conducting membrane or glass frit. In such configurations, mass transport is typically governed by diffusion and convective mixing induced by stirring, which can introduce spatial concentration gradients within the porous electrodes. In comparison, small flow cells such as that described herein can perform the same electrochemical function while providing enhanced and more uniform mass transport of the electrolyte species to the porous electrodes through controlled convection, thereby reducing concentration polarization and improving electrode utilization. Further, flow cells enable precise control over hydrodynamic parameters (e.g., volumetric flow rate, linear velocity), enabling systematic interrogation of mass-transport limited behavior and greater control over the experimental process. The well-defined geometric area and controlled electrode compression also facilitates a more reproducible environment relative to batch configurations. The flow cell described here can, of course, also be employed in a full-cell configuration where each half-cell electrolyte is distinct in redox chemistry. The above demonstration of this flow cell was conducted with diagnostic single-cell and symmetric-cell configurations because they are relatively easy to replicate and can often be informative of single redox couples without complicating effects of two distinct redox couples. Full-cell investigations will evoke additional complexities including crossover (e.g., capacity loss, additional undesired reactions) and SoC imbalance (e.g., due to unequal side reactions at the positive and negative electrodes) that must be managed for successful battery operation. Ultimately, alternative and additional characterizations using this flow cell can deepen the analysis of different components and electrolytes. Conclusion In this work, we present the design, assembly, and validation of a modular laboratory-scale flow cell platform intended to enable a wide array of reproducible measurements. Detailed fabrication and operational guidance are provided to lower the barrier to adoption: we explicitly document key construction considerations (e.g., gasket selection, contact resistance mitigation, tubing configuration) to minimize sources of variability that often complicate the interpretation of flow-cell data. To assess experimental variability, we introduce a framework distinguishing marginal and total repeatability, expanding upon previously documented terminology surrounding repeatability, replicability, and reproducibility [11], [71]. Application of this framework demonstrates that polarization and galvanostatic cycling provide stable and interpretable metrics within the present architecture. Repeatability is quantified across independent cell builds and users, validating that the platform produces consistent results under both conditions, particularly for cycling protocols. The ability to systematically vary flow rate, electrode compression, and configuration (e.g., single-electrolyte, symmetric, etc.) in this embodiment enables quantitative investigation of mass transport, interfacial behavior, and performance evolution, including phenomena such as electrode conditioning or burn-in. Although we demonstrate the platform using a relatively straightforward redox flow battery diagnostic test, we discuss the adaptability to additional measurements (e.g., pressure drop, ECSA) and related systems (e.g., full-cell configurations). The design further accommodates integration of supplementary components (e.g., patch heaters, reference electrodes), expanding the experimentally accessible parameter space. This work provides a versatile and repeatable flow cell foundation that can be adapted to various systems (e.g., CO2 capture, electrolyzers) through tailored design, enabling innovation in methods for electrochemistry. By articulating both the capabilities and limitations of the platform, we aim to contribute to improved experimental rigor and cross-laboratory comparability within the redox flow battery community and related electrochemical fields. Ethics statements Human and animal rights No human or animal studies were conducted in this work. CRediT authorship contribution statement Thomas R. Goodwin: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Trent A. Weiss: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing. Alexander H. Quinn: Writing – original draft, Visualization, Validation, Resources, Methodology, Investigation, Data curation, Conceptualization. Fikile R. Brushett: Writing – review & editing, Supervision, Project administration, Funding acquisition. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors express gratitude to the many prior members of the Brushett group who designed, iterated, and refined the Gen 2 flow cell providing a foundation for this manuscript. In particular, we acknowledge John L. Barton, Jarrod D. Milshtein, Jesse J. Hinricher, and Katharine V. Greco who all conceptualized early drafts of a manuscript that partly influenced this one. The authors also thank Mike Perry, Rob Darling, and James Saraidaridis for inspiration and extensive feedback on the design of the flow cell. Lastly, we gratefully acknowledge the Joint Center for Energy Storage Research (JCESR) for previous funding support that enabled the original development of this device. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the author(s) used ChatGPT in order to edit and consolidate sentences for increased clarity and brevity of communication. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. Biographies Thomas Goodwin is a PhD candidate in Chemical Engineering at MIT, studying membrane science and electrochemical energy storage. His research focuses on the intersection of these two disciplines, specifically on the design of thin-film composite membranes for nonaqueous flow batteries, as well as standardization of experimental methods for evaluating membrane performance in flow batteries. Trent Weiss is a PhD candidate in Chemical Engineering at MIT, where his research focuses on electrochemical energy storage. His work centers on understanding the fundamental processes governing redox-active materials in redox flow and lithium-ion batteries, alongside the development of experimental platforms for evaluating these electrochemical architectures. Alexander Quinn recently earned his PhD in Chemical Engineering from MIT, specializing in electrocatalytic processes for energy storage and conversion. He leveraged traditional 3-electrode systems and flow cell reactors to provide insights into durational catalytic performance, developed diagnostic reactor configurations, and assessed experimental reproducibility. Fikile Brushett is the Ralph Landau Professor of Chemical Engineering Practice at MIT, where he leads a research group focused on advancing the science and engineering of electrochemical systems for energy storage, chemical manufacturing, and environmental remediation. Ultimately, he aims to develop guiding principles for the design of materials, processes, and devices that harness electrochemical phenomena. Footnotes Supplementary materials to this article can be found online at https://doi.org/10.1016/j.ohx.2026.e00834. Supplementary materials The following are the Supplementary materials to this article: Supplementary Data 1 The supplementary materials for this article consist of design files for the described hardware. Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6 References Associated Data Supplementary Materials Supplementary Data 1 The supplementary materials for this article consist of design files for the described hardware. Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6