85/2021 62808 AML Energy & Environmental Science COMMUNICATION Ore (Ce this: Energy Environ Si 2016.9, 417 plates+ Received Lith dy 2016, Aecepted 13th September 2016 10%: 101059/c6=e01997 won esc orgees Electrolyzers use electricity to convert water into hydrogen fuels that can be stored for prolonged time periods. The large-scale deployment of electrolyzers for energy storage Is hampered by catalyst efficiencies as well as high capital costs. Herein we report 8 rapid prototyping methodology to accelerate the development ‘of complex flow plate architectures as a means of lowering lectrolyzer capital costs. This process involves the electroplating ‘of nickel onto 3D-printed plastic flow cells, which enable the rapid production of experimental flow plates with different flow channel ‘geometries. The test cells containing these metal-coated plastic components matched the performance of conventional metal electrodes. The ability to electroplate 3D-printed parts offers the ‘opportunity to rapidly prototype flow-field geometries to more rapidly evaluate fluid and gas flow through the electrotyzer. Introduction ‘The electrolysis of water into hydrogen and oxygen fuels represents 1 potentially scalable solution for storing renewable electricity." ‘The most widely used commercial eleetrolyzers are alkaline clectrolyzers capable of maintaining high eurrent densities (200 mA em-*) for several years, but few of these units are currently used for energy storage applications.® This scenario has driven a significant effort directed at developing new electrolyzer technologies and architectures that are capable of producing hydrogen more cost effectively."* Proton-exchange ‘membrane (PEM) electrolyzers are an emergent technology * Depart of Chemical & Bnoial ryinecring, The Univers of ri Columbia, 20 Bast Mal Vance, BC WTI, Canada Email: cheringiohemaubca “parent of Chemis, The Univer of rick Columb, 20% Main Mal, vancouver, BC vert, Canada 1 eran supplementary information (Ei) alle: Figures showing photo raph of elects and ofl plates, elearocemieal data (ei vltamno- rams, chonaamperometny, and electrochemical impedance spectroscopy, SEM Images, and energydspersve Xray spectoseopy. Tables showing material ress and data ured in economic aaa Sce BOL: 10.1039)c01978 THs oun © Te Roya Soe oF Cheminy no. sock Vea Rapid prototyping of electrolyzer flow field Jesse R. Hudkins,” Danika G. Wheeler.® Bruno Pefia® and Curtis P. Berlinguette*” Broader context nergy storage Solutions are required wo vemmporaly mach renewable tre ela eed of ter ita hydrogen fuels sone ean of storing renewable elect stseae for etended prin of time, but vances in lets: and tengeerng design are needed a rduce the perting and apt cost ff commercial electrzer hi ty fers a means frp tetng arent designs of eecwoyzer fowl plates. The aviation of thi the eletrofyae, ean be comple and therfore dtc to est efceny We show herein that complex aDprinted architectures can be sleeoplted with a thin coating of «mea oxide eleetocatalyst layer ‘The resultant AD prinfed How field plate diphys leteatae performance tht cin match that of conventional metal components ‘Te ability to quickly prduce there 20-rinfed Hw Geld plates ors 3 reansto ris est dierent fowl plats contigortions tat should ep to accra the development of higher performance elcrlaes, capable of reaching substantially higher current densities (1A cm, but the costs associated with the membrane ‘and acid-stable catalyst layer remain a challenge” The ineorpora tion of a solid electrolyte membrane in alkaline electrolyzers is, also a promising approach for driving up current densities," bt te stability ofthe membranes remains insufficient for broad, deployment." ‘These advances in eleetolyzer eell engineering, in tandem with the global scientific pursuit for new electro- catalysts,""® have nonetheless driven improvements in electro Iyzer efficiencies that make cleetrojytic hydrogen generation increasingly compelling. Notwithstanding, the capital costs remain strikingly high (850-S000SUSD per kW?) This will, inherently suppress technology adoption, particularly forsmaller units where capital costs figure more prominently in the total, cost of ownership. ‘An clectrolyzer stack consists of a membrane electrode assembly (MEA), balance-of-staek, low-field and separator plates, whieh all aecount for approximately half the cost of an electrolyzer.#* tn many advanced electrolyzer technologies, the geometry of the flow field plate plays a critical role in Energy Enron Set, 2016,9, 417-3425 | 3417‘Communication ‘optimizing the delivery of fuid to the electrode and maintain- ing a high rate of product formation. The machining of complex flow field plate architectures can be a challenging and labor-intensive ask that has limited the number of designs. that have been experimentally tested. The 3D printing of plastic clectrodes therefore offers the opportunity to quickly access a wide range of geometries, but the conductivity and catalytic activity of plastics are not currently sufficient to mediate electro- lysis and therefore need a condueting metal layer coating. This concept was first demonstrated by Cronin and workers, whom, coated 3D plastic flow plates with layers of gold and silver.** A logical extension of this work is to eoat 3b-printed plastic with cheaper electrocatalyst coatings that do not suffer from corrosion and delamination during electrolysis. This goal has not yet been documented due tothe challenges associated with depositing pure metal oxide films relevant to electrocatalysis, onto non-conducting substrates. In this manuscript we report for the fist time an electro deposited metal film on 3D-printed plastic, which we have translated into the construetion of a flow-field plate for the clectrobsis of the oggen ewlation reaction (OER). While the electroplating of plastic has been demonstrate on eatbon-doped high density polyethylene,” doing the same on 3D-printed conductive plastic has not been previously done due to the prohibitively high resistivities common to composite polylactic acid (PLA) or acrylonitrile butadiene styrene (ABS) thermo- plastics. We therefore canvassed a series of conductive thermo- plastics and found a PLA filament with suficienty low resistivities (<20 em) to accommodate the electroplating of eatalptic nickel, hydroxide on 3D-printed flowed plates (Fig. 1). This prootof concept study confirmed that the electroplated plastic low-field @ OF hee Fig. 1 Electotyzer test cel (a) Schematic ofan exploded How pate test call used in ths study. Each 3D-prites electrolyzer housing Is made of ‘on-conective ABS plate and decommadetes the titanium busbors and 530-printed flow felt pats that are separated by the membrane electrode ‘assembly (MEA) The MEA conssts of Ni foam gas dflusion layers hot pressed onto each side of a Zifon™ Perl UTP hydronde conducting ‘membrane. [o) Drawings tp) and photographs of 3D-printed PLA flow= fel plate before (ida) and after electroplating (bottom) with nickel The single serpentine ow-fela pattem wae used for al electrochemical tues inthe sty 3418 | Eneray En sci, 2016, 9, 41 Energy 6 Environmental Science Plates yield electrocatalytic performance parameters that are comparable to solid nickel low-field plates, supporting the possibility that plastie may be implemented into clectrolyzers to a greater extent than currently used. Moreaver, the ability to 3D print flow-field plates would be a tremendous asset for rapid prototyping that is otherwise challenging to do with solid metal substrates." Electroplating 3D-printed plastic 3D-printed rectangular blocks (20 mm 10:mm x 2:mm) were ‘constructed from three commercially available conductive com- posite plasties based on ABS and PLA to evaluate the volumetric resistivities of each plastic. The resistance perpendicular to the printing plane was at least five-fold higher than that measured, Parallel to the printing plane in the ease of the ABS substrates, but only tworfold higher for the PLA (Table St, ESt). The resistivities of the ABS substrates were substantial (>500 0 er), hile the PLA filament yielded a resistivity of <20 0 em. These differences in resistivities played a critical role in the elect: plating. of nickel from a Watt's nickel eleetroplating solution: Islands were deposited on planar ABS substrate, while uniform films were achieved on planar PLA substrate (Pig. St, ESI¥)This, ‘outcome is the first documented ease of electrodeposition on a 8p-printed plastic, and leaves open the possibilty that PLA ean, be used as a structural material for conducting plates where ‘current flow ean occur through the conductive shell. ‘We tested the electrocatalytic properties of the planar PLA substrates containing the electroplated nickel by cycle voltam retry at pH 14 over the 1.0-1.8 V range (Fig. 82, ESI). The first 10 seans of the eyelie voltammograms (CVs) recorded at a sean rate of 10 mV s-* revealed a response consistent with electro catalysis and reached a current density of ~30 mA em”? at 1.8 vs, RHE. Chronoamperomeity experiments were conducted to determine the long-term stability of the nickel coating by holding the current density constant at 5 mA em~? for 27 h (Fig, $2, ESI). This stability test shows a reasonably stable response with the small progressive increase in current likely ‘due to evaporation of the electrolyte over the timescale of the ‘experiment. The electroplated nickel coating also showed no visible evidence of degradation or delamination. The thickness ‘of the nickel fllm estimated by measuring the charge passed through the cathodic peak at ~1.3 V (Fig. $3, ESt#)-"°* The thickness of the redoxactive oxyhylroxide layer was estimated to bbe ~19 nm during the first eyele and ~30 nm during the tenth ‘yee (the density of the films was estimated to be 4.1 g em’) “These collective results eonfirmed the viability of depositing an electrocatalytic layer of nickel on 3D printed PLA. 3D-printed plastic flow-field plates “The next stage of our experiments was to design a aD printed flow-field plate that could be used in an electrolyzer test eel ‘There are many different geometries for eleetolyzer flowfield plates to distribute liquid phase reactants and gas phase “Tis Jounal 80 The Royal Sock of Charis 2016AM Energy & Environmental Science Fig. 2 High-angle cross zactional SEM of an electroplated nickel layer on PLA fowls pte products with minimal pressure drop.” The rapid prototyping of diferent low-field plates would therefore be hugely beneficial given the number of parameters that ean be modified (eg, rib width, channel depth, geometric configuration). We therefore produced three 3D printed flow-field plates with different goo metres common to electrolyzers (Fig. 4, ESI), each designed to be easly interchangeable within a single housing unit (Fig. 1). The serpentine low-field plate containing channels 0.1 em wide and 0.1 em deep (Fig. 1) was selected as the superior architecture and used for subsequent electroplating and catalytic testing. The deposition of uniform nickel coatings on the complex geometry was carted out in counterclectrode cage formed from expanded nickel mesh fixed in the Watts electro plating bath so thatthe low-field plate (working eletrode) and the mesh were separated by <3.0 mim to promote a uniform electri field about the flow-eld plate during electroplating The electroplating was carried out on the low-field plate by holding a cathodic current density of 2.5 mA em * fora total of Sh, The resultant nickeF-coated PLA plate (Ni-PLA) was cut with a razor and imaged using a SEM to confirm a uniform coating ~25 jm thick (Pig. 2). This thickness is consonant with the plating thickness of 28.6 um estimated from the charge passed ding the eleetrodeposition process Electrocatalysis with plastic flow-field plates A proof-ofeoncept electrolyzer cell was built inhouse and consists of 3D-printed flow-ficld housing, sD-printed inter- changeable flow-feld plates and a membrane electrode assembly (MEA). This assembly was used to evaluate the electrocatalytic behavior of NEPLA (Fig. 1 and Fig. $5, PSI). The reusable flow- field housing was aD printed from a non conductive ARS plastic, and a titanium busbar was used to interlock the interchangeable fow-field plate to the reusable housing and provide a contact to fan external circuit. The MEA consists of Ni foam anode and cathode gas diffusion layers hot pressed onto a Ziefon™ Perl UTP separator membrane.” The MEA was exchanged before ‘Communication ‘each longterm electrochemical test so degradation of the MEA ‘components could be distinguishable from degradation of the flowfield plates. The electrolyte solution containing 1.0 M KOH, ‘was pumped through the cell using two peristaltic pumps at 2 flow rate of 4.0 mt. min A series of electrochemical tests were recorded on cells containing serpentine flow-ield plates with uniform geometries but differed in the material from which they were constructed. ‘The three types of plates with uniform dimensions tested here were the aD-printed conductive PLA electroplated with nickel (NiPLA) and two floveficld plates used as controls: Machined, Nickel 200 alloy (Ni and 3b-printed conductive PLA (PLA). Stair-ease voltammetry (0.05 ¥ steps from 0 to 2.6 V) mea sured on cells containing Ni showed a progressively increasing current asthe voltage was swept to potentials higher than 1.5 V consistent with electrocatalysis. The same measurements on [NiPLA yielded a response nearly superimposable to that ofthe Nicells, while PLA produced nominal current over this same range. “This data taken together is consistent with clectrobsis occuring at the NiPLA substrate, with litte current arising from the decom Position of the plastic onthe timescale ofthe experiment ‘Chronopotentiometry experiments performed on cells with the control Ni flow-field plate required an applied potential of '~23 V to achieve a constant current density of 100 mA em”? ‘over 1 day of experiments. (The inereasing level of noise in the data collection over the course of the experiments is likely due to build-up of bubbles in our apparatus.) The same experiment with NEPLA required a lower applied potential (~2.2 V) to 2, but this voltage slight) ‘The PEA plate was not capable of electroyss at 100 mA em with a significant applied bias of 10 V. Chronoamperometry ‘experiments held at an external bias of 2.0 V on cells with the different fowficld plates yielded similar results (Fig. S6, FSH) ‘The Ni flow-ield plates produced a steady current density of ~30 mA em’ (the anomalous 5-10 mA em-* reduetions and recoveries observed at >0.5 days is again attributed to the confinement of gas bubbles) Data recorded on NEPLA resulted in current densities that stabilize ata substantially higher value ‘of ~50 mA cmr*, The same experiment using the PLA control resulted in eurrent densities of <5 ma em” ‘The superior performance ofthe cells with the NEPLA plates relative tothe Ni plates can be rationalized by the nature of the ‘metal surface. The surface of the Ni low-field plate needs to be converted to the more active nickel agiydroxide phase during clectrolysis before it can achieve the activity of the electro deposited layer for the NePLA plate." The accessible ‘lectrochemical surface area of the rough, porous nickel layer for NEPLA will also be inherently higher than the smooth, dense surfaces of the Ni plates. This is supported by indepen: dent double-layer capacitance measurements of the plates in a electrode setup, as well as the higher anodie peak at 1.4 V for the NEPLA flowfild plate (Fig. 87, ES). ‘The ohmic cell resistance determined by electrochemical impedance spectroscopy (EIS) of the electrolyzer cell assembly (Fig. $8, EST) before and after 24 h of sustained electrolysis, ruled out differences in ohmic cell resistances being responsible achieve 100 mA c increased with time, 9, zar-s428 | 3419‘Communication Energy 6 Environmental Science @ © 10 20 NPL g2s| [PLA = § nw] & _ z $0 8 of = = Nita |= © 20 w PLA ° ea 100 mAcn® ew) time (ays) Fig. 3 30-printed elactrolyze test cell performance forthe OER in allatine media [a Staircase vtammety recorded on electrolyze ces with flow: fel plates made of bulk Nickel 200 alloy (bie), conductive PLA (range) nice! electoplated PLA red with Squares and rckel electroplated PLA ter 100 h of electrolyzer operation (re with cles) fo] Chronopotentiometry recorded on electolyzer cls wit low-field plates made of bulk Nickel 200 (blue), carbon PLA orange) and nickel electroplated PLA red) at ald current density of 100 mAcm*Elecwochemisty coritons: 10 M KOH Starease voltammogram sample wath = 30s; starcase votammagram sample period = 50s. Data nt corected for ohmic resstance forthe differences in catalytic activity (Ni= 0.11 0; NEPLA=0.47 0), ‘The cell containing the PLA plate had an ohmic resistance (>8 2) too high to expect cleetrocatalysis, thus confirming the need for a conductive coating on the plastic flow-field plate (Fig. 3). We do not claim here that the NHPLA flow plate will be stable forthe time periods necessary for commercial applications, but this proof of principle study does confirm that electrolysis ean occur at eleetto- plated 3D-printed plastie substrates on timescales relevant to flow field plate evaluation. The stability ofthe cells on these shor time periods was further supported by energy-dispersive Xray speetro scopy (EDS) analyses (Fig. $9, ESI) and SEM images (Fig. S10, ESI) that indicate the Ni present on the NEPLA flow-field plate stayed constant before and after 21 h of chronopotentiometry and chroncamperometry. Conclusions ‘This manuscript reports the first demonstration of eleetroplating ‘of any type of metal ona aD-printed plastic. This achievement was used to make nickel coated 3D-printed plastic flow field plates for water electroysis. The 3D printing of flow fieKd plates can replace computerized numerical-control milling of pure metallic compo- nents and enable the rapid prototyping of electrolyzer flow field plates to optimize membrane electrode assemblies (MEAS) ‘The results in this study also demonstrate the possibility of using plastic as a surrogate flow-plate material to lower the materials and manufacturing costs of the electrolyzer unit. Consider that the flow-field and separator plates account for approximately half of the electrolyzer stack cost, nickel and staimless steel accounts for one-quarter of an alkaline electro- lyser unit. A firstprineiples techno-cconomic analysis for both conventional and contemporary “zero-gap” electrolyzers indicates that the raw materials costs of the bipolar plate could be reduced by half if the nickel were replaced by the nickel: coated polylactic acid (NiPLA) used in this study (Table 1). These cost savings are even more substantial in eases where PLA replaces stainless steel components, as well as titanium that may be found in other electrolyzer architectures. More strikingly, replacing the metal plates with metal-coated plasties 3420 | Energy En fon, Sc), 2036,9, 3417-5425 Table 1 Normalized mass and raw materials cost of two common clectrolyzer bipolar plates Zerorgip Planar? Mass Cost Mass Cost Nickel 58 1a 6 19 Stainless sect aa 23 60 25 ‘Titania: 28 28 as 0 * Zero gap flow field plate containing channels to enable Auid flow along the separator. Cros sectional and thickness assumed tobe 25 em fand'2 cm, respoctully* Planar flowfield plate used in conventional Tangescale alkaline elecrolzers. Cros sectional area ard thickness assumed to be 2500 em ands em respectfully. More details provided 4s ESL" Sprinted poblacte avid with electroplated nickel coating (thickness = 35 ua), ‘could reduce the weight of each component by ca. 7-fold, which will inherently lower the shipping and manufacturing costs asso- ciated with the construction of each unit. The rapid preceryping of sD.printed plastic components will likely be an important step Prior to mass manufacturing of such component. Experimental Materials \Nickel(u) sulfate hexahydrate (NiSO,-6110), nickel(n) chloride hhexahydrate(NiCl,-6H,0), and boric acd #,00,) were purchased from Strem Chemicals. Pocassium Hydroxide (KOH) pellets were prchased from Sigma-Aldrich, Reagents were used as received and without further purification. Zefon Peel UTP 500 membranes were sourced from AGFA. Nickel foam gas diffusion electrode material vas purchased from MTT. Ni expanded mesh was purchased from Dexmet Corporation. The PLA filament with added carbon black (1.75 mm diameter “Conductive PLA” by Protoplant) was purchased from Proto-Pasta. ABS filament “Ts joumal BO Te Roa Sock of Charity 2016Energy & Environmental Science (1.75 mm “Natural ABS" by Matterftackers) was purchased from MatterHackers, Filaments were used as received. Nickel 200. sheet with 0.25 in thickness was purchased from onlinemetals.com, aD printing A Rostock MAX v2 desktop 3D-Printer from SceMeCNC running Repetier 0.91 was used to print the flow-field plate and base ‘components. The printer's stock hot end was replaced with an ESD46 with a 0.4 mm nozzle, Mattercontrol 1.22 for Mac (Rostock MAX ¥2 preconfiguration) was downloaded from www.seemecne.com and used to control the printer. Prints were sliced with the Matterslice engine. Flowfield plates were printed with 100% infill, and oriented such that print layers are perpendicular to the raised flow-fleld plate ribs. Conductive PLA was printed with an extruder temperature of 230 °C and heated bed temperature of 60 °C onto a layer of Kapton tape applied to the print bed. ABS was printed with an extruder temperature of 237 °C and a heated bed temperature of 85 °C onto a fresh coating of UHU glue stick applied to the print bed. Flowfields and bases wete designed using SolidWorks 2013 software leetrode preparation and Ni clectroplating ‘The PLA electrode with planar geometry was a sD-printed cuboid (2.0 em x 1.0 em x 0.2 em) with removable (0.5 em x 0.5 em « 0.1 em) on the top face. A serpentine design was chosen for flowfield testing. The aD-peinted flowe-ficlds were 24 em x 2.4 em x 0.5 em with a wetted active area of 4.0 em’. The Mlowficld channels had a width of 1.1 mm, flow. field ribs had a width of 1. mm and a height of 1.2 mim, An additional removable elip (1.1 em % 0.em » 0.2 em) on the top rectangular face was included to attach an alligator eip dri clectroplating. Prior to Nickel electrodeposition, all 3D printed cleetrodes and flow field plates were cleaned by sonication in 1.0 M HCl for 5 min. [A“*Watts" nickel eleeteoplating bath was prepared by adding 27.1 g of NISO,6H.O, 5.4 g of NIClyGHLO, and 3.6 x Of HBO, {0 90 mL of deionized water. The pH of prepared solution was 47. The desited pH for eleetrodeposition was found to be in the range of3.4-3.6 usinga Fisher Scientific Aceumet ABLS0 pH. retet. The pH of the nickel bath inereased dusing electro plating due to hydrogen evolution, therefore after each electro- plating period the pl ofthe bath was adjusted by adding 1.0 M 1,80, dropwise until pH was within the desired range. A nickel counter electrode cage was fashioned from Ni expanded mesh and positioned 1.0-8.0 mm from all submerged outside faces of the sD-printed object. A held cathodic current density of 2.5 mA em”? of submerged surface area was used for lectroplat- ing nickel onto sb-printed objects, using a Keithley 2281A-30-8 triple channel DC power supply. The conductive PLA electrode with planar geometry had a submenged surface area of 5.0 em! Cosaratpunion = 13 mA), Electroplating of conductive PLA electrodes with planar geometries was performed by holding a cathodic current for two periods of 2h for a total electro plating time of 4 h. After each 2 h period, the current was stopped and the pit was adjusted to the desired range before “TH oun The Roya Socey of Chemény ‘Communication the second 2 h electoplating period was initiated, Flow-iekd plates had a submerged surface area of 20.4 em? (sr 51 mA}. Electroplating of flowield plates was performed by holding a eathodie current for 4 periods of 2h for a total clectroplating time of 8 hh. After each 2 h period the current was stopped and the pH was adjusted tothe desired range before the next 2h electroplating period was initiated. The thickness of the Ni coating applied during electroplating was approximated, using the following formula:"” 122041 a 0) where: T is thickness in micrometres (jm); 1 is current in amperes (A) ¢is time in; and A isthe submerged surface area in square decimetzes (de) [MEA preparation Nickel foam gas difusion layers were cut into 2.5 em % 2.5 em squares with geometric surface area of 6.3 em* and were sonicated in acetone for 5 min. The nickel foam was then ‘etched by sonication in 1,0 M H,$0, for 5 min, Substrates were washed with DI water and sonicated in DI water for 5 min. Zirfon™ Perl UTP membranes were cut into 2.8 em x 2.8 em squares with geometric surface areas of 7.8 em. Membrane electrode assemblies were hot-pressed using a Dake hoepress model 44226 at 125 °C and 400 psi for 3 min, MEAS were hydrated in 1.0 M KOH for at east 24 h after hot-pressing. Physical methods Electrochemical data recorded using a C-H Instruments Work station GOOD potentiostat with amp booster. Cyclic voltammo- {grams and chronopotentiometry were recorded on a PLA clectrode and a Niplated PLA electrode with planar geometry in a threeelectrode hcell setup with a Ni Foam counter electrode and a AgiAgcl (sat. KCI) reference electrode. ‘The ‘AgAgGl (Sat. KCI) reference electrode was calibrated in Hy saturated 1.0 M KOH using Pt disk cleettodes as the working and counter electrode. Cyclic voltammograms were acquired at ‘410 mv s-* sean rate. No aggregation was induced on the test ‘el besides that from evolved gaseous produets. Cycle voltamt metry were recorded on Ni (half submerged, geometric surface area of 13.1 em?) and NiPLA (half submerged, geometric surface area of 10.67 em*) in a custom three-eleetrode setup with a Ni foam counter cleettode and a Ag/AgCl (sat. KCl) reference electrode, The Ag/AgC| (sat. KCl reference electrode was calibrated in Hp saturated 1.0-M- KOH using Pt disk electrodes as the working and counter electrode. Data was acquired at a 10 mV s-' scan rate. Reported potentials were not corrected for uncompensated resistance (i). No aggravation was induced on the test cel besides that from evolved gaseous products Eleetrolyzee polarization, chronopotentiometry, chrono: amperomety, alternating eurrent electrochemical impedance spectroscopy (FIS) data were reorded in a tworelectrode setup. ‘The alligator clips from the potentiostat were clipped to Sitanium bus bars designed to insert into the flowfields 16, 9, 417-3425 | 3421‘Communication The anode bus bar had the working electrode alligator clip attached and the cathode bus bar had the reference and counter electrode alligator clip attached. An o-ring is used to create a seal between the MEA and the outside of the flow-field plate, while compression between the two housing halves is provided by four firmly hand tightened bolts. 1.0 M KOH clectrolyte was circulated through the anode and eathodle flow: field channels at a rate of 4 mL min~* using dual peristaltic pumps (Fig. $5, ESI), Polarization data recorded as staircase voltammetry: sample width = 30 s sample period = 50s. EIS data were recorded at open circuit potential between 1 and 10000 Hz with a potential amplitude of 0.005 V. Cell resistance was taken as the xintercept of the Nyquist plot or the value of 2’ at ~Z' = 0. The output electrolyte was filtered through Whatman'™ (Cat. No. 1001 090) filter papers before being recycled through the electrolyzer via a 500 ml. reservoir (Fig. 85, EST). Field emission seanning electron microscopy (PE-SEM) and energy-dispersive X-ray spectroscopy (EDS) data was acquired with a FEI Helios NanoLab 650 microscope. FE-SEM imaging was performed in secondary electron mode (voltage: 3 kV, current: 50 pA, Everhart-Thornley and Elstar in-lens secondary electron detectors) and EDX analysis was carried out at 15 kV and 3.2 nA. Flow-plate techno-economie analysis {A hypothetical zero-zap bipolar flow-feld plate (25 em!) and conventional alkaline bipolar plate (2500 ens") were used for the cconomie analysis found in Table 1. Zero material waste was assumed (Le, 100% of the waste produced during. milli Of flowfield channels is recoverable) and both plates were assumed to be 0.3 em thick. The volumetric prices and densities for all flowfield plate materials were based on values from the (esti) For a 25 em* bipolar flow-ield plate, there would be 25 ribs (0.1 em width, 0.1 em depth) and channels (01 em width). For (one rib + channel the electroplated surface area (4 em") was, determined by multiplying the cleetroplated perimeter of a rib + channel (0.8 em) and flow-field plate side length (5 em). ‘The surface area of the perimeter sides of the flow-feld plate (03 em x 5 em x 4 sides) was added to the coal surface area of ribs and channels (100 cm?) to give the total electroplated surface area (106 cm’). The total electroplated surface area was multiplied by the thickness of the Ni coating (25 jim) to determine the volume (0.265 em’) and mass (2.35 g) of nickel added to the bipolar flow field plate. One-third of the volume of ‘square slab of material (0.3 cm x 25 em’) is removed to create the channels fora bipolar low-field plate resulting in a channeled flowfield plate of 5 em! volume in this example. Density and volumetric cost data were used to tabulate the mass and cost, of flow-field plates of different materials (Table $3, BStt). For conventional electrolyzer a circular eross sectional area of 2500 em? and a plate thickness of 0.3 em was assumed. The total surface area of the plate (2 x 2500 em* + 53.2 em’) was multiplied by the thickness of the Ni coating (25 jm) to determine the volume (12.6 em") and mass (111.9 g) of nickel added by electroplating. The volume of material used in a sources indicated in Table S 3422 | Energy Environ 5 Energy 6 Environmental Science ‘conventional electrolyzer flow: with density and volumetric material cost to tabulate the cost of flowe-field plates made from different materials (Table 83, E Acknowledgements ‘We thank the Canada Foundation for Innovation, Canada Research, Chairs, Canadian Institute for Advanced Research, and the National Seience and Engineering Research Council for financial support. References 1 N.S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sei. U. S.A, 2006, 108, 15729-15735, 2 M. G. Walter, E, L. Warren, J. R. MeKone, 8. W. Boettcher, ‘Q. Mi, E. A. Santori and N. 8. Lewis, Chem. Rev, 2010, 110, 6146-6873, 3 TR Cook, D. K, Dogutan, S. ¥. Reece, ¥. Surendranath, T. ‘Teets and D. G. Nocera, Chem. Rev, 2010, 110, 6474-6502. 4 J.A. Turner, Science, 2004, 305, 972-974. 5 J. Mergel, M. Carmo and D. Fritz, in Transition to Renewable Energy Systems, ed. D. Stolten and V. Scherer, Wiley-VCH, 2013. 6S. Marini, P. Salvi, P. Nell, R. Pesenti, M. Villa, M, Berrettoni, G. Zangari and Y. Kitos, Blectrochim. Acta, 2012, #2, 384-391. 7 P. Vermeiren, J. P. Moreels, A. Claes and H. Beckers, Int. J. Hydrogen Energy, 2009, 34, 9305-9315. 8 P. Millet, R. Ngameni, S. A. Grigoriev, N. Mbemba, F, Brisset, A. Ranjbari and C. Etiévant, Ine J. Hydrogen Energy, 2010, 25, 5043-5052, OM. Carmo, D. L. Fritz, J. Mergel and D. Stolten, Int, J. Hydrogen Energy, 20213, 38, 4901-4934. 10 C, C. Pavel, F, Ceceoni, C. Emiliani, 8, Santicioli, A SeatTidi, Catanorehi and M. Comotti, Angew: Chem, Int. Bd, 2013, 53, 1378-1381. 11 Y, Leng, G. Chen, A.J. Mendoza, P. B, Tighe, M.A. Hiekner and C. ¥. Wang, J. Am. Chem, Soe., 2012, 134, 9054-905 12 M, Manolova, C. Schoeberl, R, Preudenberger, C. Ellwein, J. Kertes, 8. Siypka and B, Oberschachtsiek, Jn. J. Hydrogen Energy, 2015, 40, 11362-11369, 13 J. R. Varcoe, P. Atanassov, D. R. Dekel, A. M. Herring, M.A. Hickner, P. A. Kohl, A. R. Kucernak, W. E. Mustain, K, Nijeijer, K. Scott, T. Xu and L, Zhuang, Energy Environ. Sci., 2014, 7, 3135-3191. A M.W. Kanan and D. G, Nocera, Science, 2008, $21, 1072-1075. 15 M. Dined, Y. Surendranath and D. G. Nocera, Proc. Nat. Acad. Sci. U.S. A, 2010, 107, 10337-10841, 16 B, Zhang, X. Zheng, 0. Voznyy, R. Comin, M. Bajdich, M. Garcia-Melchor, L. Han, J. Xu, M. Liu, L. Zheng, F. P. Garcia de Arquer, ©."T. Dinh, F. Fan, M. Yuan, E. Yassitepe, N. Chen, T. Regier, P. Liu, ¥. Li, P. De Luna, A. Janmohamed, H. L. Xin, H. Yang, A. Vojvodic and EH, Sargent, Science, 2016, 352, 332-237. 17 N.S. Lewis, Chem. Rev, 2015, 115, 12631-12632. seca of Chemisty 20161 2016. Do 108 Oct Energy & Environmental Science 18 BM. Hunter, W. Hieringer, J. R. Winkler, H. B. Gray and A.M. Miller, Bnergy Environ. Sci, 2016, 9, 1734-1743, 19 C. ©, L. MeCrory, S. Jung, J. C. Peters and T. P, Jaramillo, J.Am. Chem. S0., 2013, 138, 16977-16987, 20 CG. L MeCrony, 5. Jung, LM. Ferret, S. M, Chatman, J. C Peters and. P,jaramill,j. Am. Cherm Soc, 2015, 137, 4347-1357 21 R.D.L. Smith, M.S, Prévot, RD. Fagan, S. Trudel and C. P. Berlinguette, J. Am. Chem. Soc, 2013, 135, 11580-11586, 22. R. D. L. Smith, M. 8. Prevot, R. D. Fagan, Z. Zhang, P. A. Sedach, M. K. J. Siu, S, Trudel and ©, P, Berlinguette, Science, 2013, 340, 60-63. 28 J. B. Genovese, K. Harg, M. Paster and J. A. Turner, Current (2009) state-of the-art hydrogen production cost estimate using ‘water electrolysis, U.S. Department of Energy, 2009. 24 G. Saur, Windto-hydrogen project: eletrolyzer capital cst study, [National Renewable Energy Laboratory, 2008, 25K, E. Ayers, EB. Anderson, C. Capuano, B. D. Carter, L. T. Dalton, G. Hanlon, J. Manco and M. Niedzwiecki, ECS Trans., 2010, 33, 3-15, THs oun © Te Roya Soe oF Cheminy ‘Communication 26 G. Chisholm, P. J. Kitson, N. D. Kirkaldy, L. G. Bloor and 1. Cronin, Energy Environ. Sci, 2014, 7, 3026-3032. 27 L, Daniel, US3865699, 1975 24 A. Ambrosi and M. Pumera, Chem. Soc, Rev.. 2016, 45, 2740-2755. 29 ¥.J.Tan and K. ¥. Lim, Sur Coat, Technol, 2003, 167, 255-262. 30 RD. L, Smith and C. P. Betlinguette, J. Am. Chem. Soe, 2016, 138, 1561-1567. 31 L. Trotochaud, J. K. Ranney, K. N. Williams and S. W. Boettcher, J. Am. Chem. Soc, 2012, 134, 17253-17261. 32 X.Liand I Sabir, Ine J. Hydrogen Energy, 2005, 30, 359-371, 33 8. Klaus, M. W. Louie, L. Trotochaud and A. T. Bell, Phys Chem. C, 2015, 149, 18303-18916. 34 M.W. Louie and A.T. Bell, J. Am. Chem. Soc. 2013, 135, 12329-12337. 35 B.S.Yeoand A.T. Bell, J Pips. Chem. C,2012, 116, 8394-8400. 36 S. Klaus, Y. Cal, M. W. Louie, L. Trotochaud and A. T. Bell, J. Phys. Chem. C, 2015, 119, 7243-7254. 37 ‘Nickel Institute, Nickel Electroplating Handbook, 2013. Enevay Environ Sc