Accurate determination of catalyst loading on glassy carbon disk and

4 hours ago - Thin film-rotating disc electrode (TF- RDE) experiment provides a fast research platform for screening of newly developed electrocatalys...
0 downloads 0 Views 3MB Size
Subscriber access provided by Kaohsiung Medical University

Article

Accurate determination of catalyst loading on glassy carbon disk and its impact on thin film rotating disk electrode for oxygen re-duction reaction Muralidhar G. Chourashiya, Raghunandan Sharma, and Shuang Ma Andersen Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02697 • Publication Date (Web): 08 Nov 2018 Downloaded from http://pubs.acs.org on November 9, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Accurate determination of catalyst loading on glassy carbon disk and its impact on thin film rotating disk electrode for oxygen reduction reaction. Muralidhar Chourashiya$, Raghunandan Sharma and Shuang Ma Andersen* Department of Chemical Engineering, Biotechnology and Environmental Technology, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark. ABSTRACT: Thin film-rotating disc electrode (TF- RDE) experiment provides a fast research platform for screening of newly developed electrocatalysts for oxygen reduction reaction (ORR) activity, however, precise estimation of their performance parameters is necessary to avoid wastage of resources in the testing of otherwise unpromising electrocatalyst in actual fuel cells. Here we show the importance of the accurate amount of catalyst (e.g. Pt) on glassy carbon (GC) disk of RDE in TF-RDE experiment by characterizing the commercial catalysts for their electrocatalysis performance (electrochemical surface area and ORR activity) values. The Pt loadings used to calculate these performance values were obtained using two schemes, namely, using the literature based (conventional) scheme and an X-ray fluorescence (XRF) based scheme. A parameter called ‘catalyst-density-ofthe-ink’ is used to correlate the variations observed in performance values and the amount of Pt on GC disk of RDE obtained using both the schemes. The investigation suggests that the actual Pt loading on GC disk of RDE varies with the ink-conditions, which is considered constant in the conventional scheme and might be one of the reasons of irreproducibility of the data obtained by TFRDE experiments. The XRF based scheme, which is simple and direct, can have the potential to replace conventional scheme for accurate catalyst loading estimation, improve experimental reproducibility and open many other possibilities (e.g. postmortem analysis of catalyst) in electrocatalysis studies.

The current polymer electrolyte membrane fuel cell (PEMFC) research status indicates that it will be playing a very crucial role in imminent revolutions in automotive and renewableenergy industries. Nano-sized platinum particles dispersed on high surface area carbon substrates have been used as the standard electrocatalyst as anode and cathode electrocatalysts in PEMFCs. At the cathode, an oxygen reduction reaction (ORR) takes place and due to their sluggish kinetics causes a large overpotential across the cell, leading to corrosion reaction in the carbon support. To overcome this restriction, a relatively high amount of Pt is used at the cathode and therefore are the major contributor (approximately 80% of total catalyst cost) towards the high cost of PEMFCs. This residual issue of electrocatalyst cost, specifically for ORR catalyst, in PEMFC, has motivated the scientist in the labs, worldwide, for a focused research on its development. The laboratory-scale electrocatalyst is typically synthesized in extremely small quantities, e.g. few tens of micrograms and demand a rapid screening for their ORR activity prior to scaleup and subsequent evaluation in low-temperature fuel cells. 1,2 The electrochemical characterization using thin film rotating disk electrode (TF-RDE) technique has proven a highthroughput research platform for evaluation of ORR activity. The scientific community has been engaged in the standardization of TF-RDE technique with elaborative research on various parameters involved in the technique which can affect the estimation of “true” electrocatalysis performance parameters such as electrochemical surface area (ESA), surface specific activity (SSA) and mass specific activity (MSA). 1,3-7 It is expected that following these

standardized protocols for TF-RDE experiments, one can fairly produce trustworthy and reproducible data that allows the comparison of data from different laboratories. Although according to Masa et al, RDEs coated with layers with uncontrolled porosity/roughness more often cause irreproducible/misleading results in TF-RDE and therefore should be taken into account, for example, by modelling the porosity/roughness in the coated layer prior the analysis of RDE data.8 Limitations occurring in TF-RDE technique for ORR, based on ref.8-10, are briefly discussed in a recent article by Sokolov et al.11 and they also proposed a methodical approach to overcome those limitations while developing a relatively simple method. The general procedure followed by researchers to evaluate the newly developed electrocatalyst for electrocatalysis performance parameters prior to scale-up 12,13 is outlined in supporting information (SI) and figure S-1. This procedure is also extended to benchmark the TF-RDE system with standard/commercial electrocatalyst. Usually, the homogeneous ink of electrocatalyst (with known/TGAestimated 14,15 Pt loadings) is prepared using a standardized recipe. 16 This ink is then drop-casted onto the glassy carbon (GC) disk of RDE and upon drying are characterized in a TFRDE experiment with the standardized protocols 1. The evaluation scheme seems straightforward (figure S-1), however, the results for similar electrocatalyst obtained using similar procedure reported in the literature varies widely. 17 There are several factors which affect such a spread and are well-reviewed in the literature. 12,13 However, the observations under this investigation indicate that the inaccuracies in the

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 16

estimation of the actual weight of Pt on GC disk of RDE affect the estimated ESA/MSA values severely and hence irreproducibility (or spread) of data in the literature. According to Gasteiger et al, for quantitative measurements using TF-RDE, the amount of Pt dispersed onto the GC disk of RDE must be known precisely, while at the same time very low loadings (on the order of ≪100 μg catalyst/cm2) need to be applied. 17 The actual Pt amount/weight on GC disk of RDE than the expected/calculated amount of Pt using Pt loading in composite (either by TGA, based on precursor amount or manufacturer supplied data) leads to changes in the calculated electrocatalysis performance parameters, even for the same electrocatalyst/ink. Diversion of the calculated weight of Pt from the actual weight of Pt on GC disk of RDE can raise from different stages in the scheme and few of them can be, • errors in weighing of different components for ink, • ultrasonication treatment for homogenizing of ink (detachment of Pt from carbon support), • inhomogeneity of ink while casting on GC disk of RDE, • error in pipetting a drop amount (ink-residue in pipetting tip) or errors in weighing of drop weight, etc. In fact, the TGA data, which is considered as a working tool for estimation of Pt loading in carbon supported Pt electrocatalyst, might also need consideration of following points to avoid any inconsistency in the estimation of Pt loadings. During TGA, volatile but very unstable PtO2 forms at high temperature. This unstable PtO2 decomposes into Pt and O2 at a lower temperature. The yield of forming unstable PtO2 is very low and therefore almost 100% of the Pt metal will be recovered at a lower temperature. Usually, the Pt loading is considered to be equal to that of the residue weight % in TG at high temperature. Therefore, to avoid the error of considering the residue weight at high temperature as Pt loading, which will also include the weight of PtO2, one should consider the residue weight % in TG at near room temperature (after decomposition at higher temperature) as actual Pt loading. Additionally, it is also possible that along with the Pt metal in the residue, some ashes, initially present in the carbon support, are also present. The amount of ashes depends on the type of carbon support used. For example, the carbon black and MWCNTs as support produce ashes amounting 2.44 wt.% and 8.74 wt.% in the residue, respectively. 18 Such a high amount of ashes in the final residue will lead to an estimation of incorrect Pt wt.% in electrocatalyst under investigation. In good quality MWCNTs, the ash content is significantly lower and must be taken in to account when using it as supports. To investigate the importance of Pt weight present on GC disk of RDE, we adopted an approach in which electrocatalysis performance parameters were calculated using Pt loading by directly exposing coated GC disk of RDE to a portable X-ray fluorescence (XRF) spectrometer in addition to the scheme shown in figure S-1b (with manufacturer provided data). A correlation between the observed electrocatalysis performance and the direct estimation of Pt loadings on GC disk of RDE estimated using XRF spectrometer allowed to point out the shortfalls of the literature based standardized protocol/scheme shown in figure S-1b.

EXPERIMENTAL Catalysts and XRF calibration. Two types of commercial electrocatalysts were investigated, namely, 20 wt.% Pt/C and 55 wt.% Pt/C identified as BASF (procured from BASF) and HiSpec9100 (Johnson Matthey, 55.5 – 57.8 wt.% Pt/C; for sake of simplicity 55 wt.%Pt/C is considered for calculations) catalysts, respectively. Portable XRF spectrometer was employed to measure the catalyst loading on GC disk directly by exposing the Pt/C catalyst on GC disk of RDE with X-rays such that the exposure area (circular opening) remained larger than the GC disc (to ensure X-ray signal from the whole catalyst on GC) but smaller than the outer PTFE jacket of the RDE. The instrument was calibrated using known catalyst loadings of electrocatalyst (here 20 wt% Pt/C, BASF, electrocatalyst) in isopropanol (solid content: 3.0 wt. %) catalyst ink (drop casted on a glass slide (~0.1 mm thickness) and dried at 80°C for 30 min before weighing with a microbalance (METTLER TOLEDO, ± 2 µg accuracy). X-ray peak intensity for a known catalyst (Pt) loading was obtained by exposing the glass slide to X-ray beam such that the catalyst coating faced the beam directly while an uncoated GC was kept on the other side of the glass slide to obtain a background close to that of GC. Linear fitting to the XRF intensities for different catalyst loadings was used to generate a calibration line. As shown in figure 1, the XRF signal (XRF readings) varies linearly (R2 = 0.99438) with the Pt loading on GC tip for a broad range of the later (0-80 µg of Pt), suitable for real RDE experiments involving the typical Pt loadings well below 20 µg of Pt on GC tip. Further experiments (data not shown here) suggest a linear behavior for Pt loadings as high as 1 mg/cm2.

Figure 1. Calibration of XRF spectrometer for Pt/C catalyst (BASF) loading on glassy carbon disk of RDE.

In TF-RDE experiments, it is highly important to know the actual Pt-loadings on RDE as it is used while computing ESA/MSA values to decide the performance of catalyst under investigation. Therefore, the XRF technique which is sensitive to a thickness of layer (not problematic for TF on RDE) as well as the atmosphere of element (of interest) in the coating

ACS Paragon Plus Environment

Page 3 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

can be well suited for TF-RDE experiments. Although the calibration of XRF technique was carried out, however, when coatings are dried on RDE a porosity/roughness formation in layered structures are possible which is found to affect the XRF readings, making absolute estimation difficult/misleading. However, the change in original (i.e. relative) loadings in coating can readily be achieved and therefore all the comparison of loading valued done here is of relative nature. Table 1. The defining parameters and IDs of inks studied in this investigation. ID

Electro-catalysts

BASF_O1 BASF_O2 BASF _F1

Three months old BASF catalyst (20%Pt/C) Freshly prepared

BASF_F2 HiSpec_O1 HiSpec_O2 HiSpec_F1 HiSpec_F2

Ink age

Three months old HiSpec9100 catalyst (55%Pt/C) Freshly prepared

The calibration of XRF spectrometer was carried out with BASF and the same calibration was used for HiSpec9100 catalyst, as well. Ink preparation and drop casting. Inks were prepared using the standard ink-recipe 16, in which, 10 mg of the catalyst is mixed with 5 mL of a stock solution. The stock solution is prepared by mixing isopropyl alcohol (HPLC grade), 5 wt% Nafion solution (Ion Power, Dupont D521) and Milli-Q water (18 MΩ.cm) in the proportion of 20: 0.4: 79.6, respectively. The catalyst mixed with the stock solution is then ultrasonicated using a Hielscher UP200St ultrasonic homogenizer (50 W, high-energy ultrasonication) for a minute. In literature1,5, ink preparation recipe includes an ultrasonication of catalyst/solvent mix for 15-20 min in a bath. However, our optimization with an above-mentioned instrument (which is used by immersing ultrasonication transducer rod directly in the ink) showed more efficient homogenization and was able to achieve the similar homogenization in one minute of treatment. Two types of inks for each of the catalysts were investigated, one of them was aged for 3 months and one was freshly prepared (see the design of experiment in SI). The aged ink showed catalysts sedimentation, observed as a thick layer of the catalyst at the bottom of the container, and to re-disperse it in the solution a simple shaking of the container was enough. Both the type of inks, i.e. the freshly prepared and aged ink were characterized for their electrochemical activity. These two types of inks (aged and fresh) of each catalyst (BASF and HiSpec9100) were further kept on a mixer, (a vial rotator, at 36 rpm), for 1h and characterized again for their electrochemical activity. An aliquot of these homogenized ink (10μl) is then casted on GC disk of RDE and dried the cast of the ink by rotating the coated RDE in the air at 700 RPM for 30 minutes. The conditions of inks are summarized in table 1 along with their IDs. Estimation of Pt loadings. For sake of simplicity, Pt loading obtained using scheme shown in figure S-1b is called as conventional (in short Conv.) scheme while that of using the portable XRF spectrometer is called as XRF based scheme

#of ink

Condition(s)

#O1

Aged ink – re-dispersed

#O2

The ink of #O1 + 1h on the mixer

#F1

immediately after high-energy ultrasonication

#F2

Ink of #F1 + 1h on mixer

#O1

Aged ink – re-dispersed

#O2

The ink of #O1 + 1h on the mixer

#F1

immediately after high-energy ultrasonication

#F2

Ink of #F1 + 1h on mixer

(figure S-2). Under the conventional scheme, the weight of drop was recorded while casting the drop onto the GC disk of RDE. This drop weight was then converted using the scheme shown in figure S-1b to dried Pt loadings on GC disk of RDE. Under XRF based scheme, dried catalyst coated GC disk of RDE is directly exposed to XRF spectrometer and the recorded reading is converted to the Pt loadings using calibration data. Electrochemical measurements. For electrochemical characterization, the catalyst inks were casted on GC disk of RDE from PINE® Research Instrumentation, Inc. (attached to PINE® MSR rotator) and used as a working electrode, in a three-electrode cell. A coiled Pt wire as a counter electrode and HydroFlex® (Gaskatel GmbH) electrode as a reversible hydrogen electrode (RHE) was used. The catalyst coated GCRDE was then characterized using electrochemical workstation (ZAHNER® – Zennium Potentiostat) using cyclic voltammetry (CV in Argon saturated electrolyte), linear sweep voltammetry (LSV) firstly in argon (background) and then in oxygen saturated electrolyte (with the rotator at 1600 RPM) for electrocatalytic activity. Perchloric acid (0.1M, HClO4) was used as an electrolyte for all the electrochemical characterizations. The CV for ESA was carried in potential window of 20mV to 1.2V at 20mV/s for 3 cycles and LSVs for ORR activity were carried out in potential window of 20mV to 1.03V at 20mV/s for 3 cycles. For adequate saturation of electrolyte with a desired gas, the (Ar/O2) gases were purged through the electrolyte for more than 30 min prior any measurements and kept purging during the measurements. Background and iR corrected LSVs (20mV to 1.03V) were evaluated to estimate MSA/SSA (at 0.9 V) while the hydrogen absorption region of CV was evaluated to estimate the ESA. A constant of 210 μC/cm2, the charge of full coverage for clean polycrystalline Pt, is used as the conversion factor19. The method used to estimate ESA/MSA/SSA is well established in the literature and in this work, we followed a standard methodology as outlined in ref.16.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 16

RESULTS AND DISCUSSION Reproducibility studies. Figure 2a shows estimated Pt loadings using Conv. and XRF based scheme along with drop weight of drops casted on GC disk of RDE. The difference between Pt loadings estimated by XRF based and the Conv. scheme is increased with mixing for fresh inks (F1  F2) while decreased for aged inks (O1  O2) for either of the catalysts, suggesting a stabilization of inks with mixing time. To clarify this further, a parameter termed as ‘catalyst-densityof-the-ink’ defined as estimated loading (mg/cm2) divided by measured drop weight (mg), for either scheme and is shown in figure 2b. According to the conventional scheme, for a specific ink (i.e. BASF _O# or _F# and HiSpec _O# or _F# are same inks), the ‘catalyst-density-of-the-ink’ remains constant. However, when the ‘catalyst-density-of-the-ink’ for XRF based scheme is plotted, a clear variation can be seen and suggests that even though the same ink is investigated after simple mixing for 1hour, casted Pt loading on GC disk of RDE differs. This

Figure 2. (a) Pt loading estimated using conventional (Figure S-1b, using manufacturer provided data) and XRF based scheme (Figure S-2). Drop weights of the drops used to cast the ink on GC disk of RDE are also shown for reference. (b) Variation of loading/drop wt. (i.e. ‘catalyst-density-of-ink’ defined as estimated loading divided by measured drop weight). ty. In Figure 3a (left side), where the Pt loading was estimated difference might lead to ambiguous results and can hamper the by (conventional scheme) considering the non-changing reproducibility of data, even with the same setup and ink. The catalyst-density-of-the-ink, observed changes in ESA values cyclic voltammograms (in SI, figure S-3) of these inks show therefore can generally be attributed only to the quality of the slight differences and can be attributed to the quality of coating on GC disk of RDE. However, as observed in Figure coating and the corresponding interface structure, which is 3b (right side), where the Pt loading was estimated by XRF another factor leading to irreproducibility of data and is based scheme showing change in catalyst-density-of-ink with beyond the scope of the current study. In literature, there are ink conditions, the observed changes in ESA values not only articles which deal with this aspect in great details and can be attributed to the quality of coating but also to the changed Pt found in the literature. 19-21 loading on GC disk of RDE. The estimated loading values using both the schemes were In Figure 3b, the computed ESA values are inversely further used to obtain electrocatalysis performance parameters, proportional to loadings obtained by XRF based scheme i.e. ESA and MSA (at 0.9V) values and are presented in (Figure S-2) which was not observed in figure 3a. This Figure 3. As per the definition of ESA and MSA, these observed proportionality/dependence of Pt loading on ESA parameters are computed by dividing the hydrogen adsorption highlights the importance/correctness of Pt loadings in TFcharge and kinetic current by Pt loading values, respectively. RDE experiments. Therefore, it is obvious that the slight error in estimation of Pt loadings on GC disk of RDE can cause a significant change in Similarly, in Figure 3a, changes in MSA values can be the computed ESA and MSA values. Effect of coating quality arbitrarily attributed to the coating quality along with factors on CV is documented in ref.16 and they estimated that the affecting the kinetic current (such as, diffusivity of fresh O2 change in shape of CVs is more probably related to the coating species, O2 saturation of electrolyte, RPM of RDE, porous quali structure of coatings, amount of accessible Pt for O2 i.e. dispersion

ACS Paragon Plus Environment

Page 5 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 3. Calculated ESA (m2/g) values and MSA (A/mg) values of investigated inks, based on Pt loading estimated using (a) conventional (Figure S-1b) and (b) XRF based scheme (Figure S-2). The plot also includes respective loading/drop weight (see figure 2b). According to the conventional scheme, for a specific ink (i.e. BASF _O# or _F# and HiSpec _O# or _F# are same inks), the ‘catalyst-density-of-the-ink’ (loading/drop weight) remains constant. The ‘catalyst-density-of-the-ink’ for XRF based scheme varies even for the same ink suggesting, after a simple mixing of 1-hour, casted Pt loading on GC disk of RDE differs. Figure 4 shows a CV recorded before, during and at the end of of Pt, etc.), but not to the Pt loadings as in conventional durability (AST) measurements (using standard protocols 1) scheme the catalyst-density-of-the-ink is considered constant. for a BASF catalyst and the estimated ESA from respective Interestingly, in figure 3b, the MSA values obtained shows a CVs, by following conventional as well as XRF based scheme. different trend than that of observed for ESA values. For aged The percentage-wise degradation obtained by either scheme is inks (BASF _O# or HiSpec _O#), MSA values were inversely the same. As mentioned above, the ability of direct estimation proportional to Pt loadings estimated by XRF based scheme of Pt amount on GC disk of RDE enabled the quantification of while for freshly prepared inks (BASF _F# or HiSpec _F#) Pt after the durability measurement. MSA values were directly proportional to the Pt loadings estimated by XRF based scheme. Apart from this observation, In the conventional scheme, the degradation obtained in ESA one can also see that the aged BASF inks showed an during support focused AST is attributed to the corrosion of improvement in MSA values and decreased catalyst-densitycarbon support, as it is considered (due to impossibility in the of-ink (loading/drop wt.) upon mixing for 1h, while the estimation of post-AST Pt loading) that the Pt loading remains reverse was true for aged HiSPec9100 ink The observed unchanged (figure 4b). Comparing the pre-AST (100%) and variation of MSA values and catalyst-density-of-the-ink are post-AST (86%) Pt loading values obtained by XRF based interlinked and can be attributed to agglomeration properties scheme, almost 14% of Pt was lost during the durability of Pt in the ink (in SI and figure S-4). Though. the MSA and measurement, which can severely affect the measured ESA of ESA values varied for Pt loadings estimated by conventional the catalyst under investigation. Therefore, due to the virtue of and XRF based schemes, the SSA values, as expected (by XRF based scheme, one can attribute the degradation of ESA definition), were independent of Pt loading estimation during support focused AST not only to the corrosion of schemes (in SI and figure S-5). The study shows that the carbon support and/or agglomeration of Pt particles but also to importance of actual Pt weight on GC disk of RDE to estimate the physical loss or dissolution of Pt from the GC disk of reproducible electrocatalysis performance parameters using RDE. A detailed discussion of AST protocols and related TF-RDE experiments and the portable XRF spectrometer as a attribution/effect on observed results are discussed in the ref. simple/direct tool to estimate the same. Though the XRF 23. spectrometer requires a careful calibration for estimation of Similarly, the XRF based scheme showed differences in the absolute amount of element under investigation, the estimated MSA values (figure 5). In figure 5, the LSVs for estimation of the relative percentage of elements under ORR activity estimation is shown along with MSA values investigation is readily obtainable. obtained by the conventional and XRF based scheme. The Before and after tests. Addition to the advantage of the percentage change in MSA values (figure 5-inset) obtained by simple/direct tool to estimate the loadings, the portable XRF conventional scheme amounts to 63% of its original value spectrometer can also be used to determine the physical while by XRF based scheme the change in MSA value (relative) loss of platinum from GC disk of RDE before and amounts to 57%. The difference observed in percentage after ORR activity or durability (accelerated stress test – AST) change of MSA values can be attributed to the physical loss of measurements of electrocatalysts. This is not possible if the Pt from GC disk of RDE. conventional scheme is followed. The relative changes Outreach. Following this concept, the implementation of observed before and after ORR activity or AST measurements XRF in TF-RDE experiments opens a range of possibilities in can provide important information such as dissolution or electrocatalysis. More examples include investigation on detachment (physical loss) of platinum particles.22 degree of alloy or element leaching from catalysts after various electrochemical treatments; quantitative estimation of

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

non-crystalline materials; moreover, in a situation where the electrocatalysts to be investigated are directly synthesized/sputtered on GC disk of RDE (where the deposited film thickness/loading relation is impossible to estimate), the XRF

Page 6 of 16

For a comparison purpose, normalization of ESA/MSA values of different catalyst must be done using the absolute loadings on RDEs. Authors believe that such insights can and will assist the scientific community to assign different mechanisms for the observed phenomenon in TF-RDE experiments. Additionally, attribution to appropriate mechanisms will preserve the reproducibility and reliability of electrocatalysis performance parameters of the catalyst under investigation and improve the confidence in choosing the electrocatalyst for testing in actual fuel cells.

Figure 5. LSVs recorded before (PreAST ORR), and after (PostAST ORR) durability measurement of a BASF catalyst. The durability measurement was carried out using a standard protocol reported in the literature.1 Inset: Estimated MSA values using pre/post-AST LSVs and the Pt loadings obtained by the conventional and XRF based scheme. Percentage change observed in MSA values for the schemes are highlighted.

CONCLUSIONS Figure 4. (a) CVs recorded before (0 stress cycles – AST_0000), during (after 250 stress cycles – AST_0250) and at the end (after 5250 stress cycles – AST_5250) of durability measurement of a BASF catalyst. The durability measurement was carried out using a standard protocol established in the literature.1 (b) Estimated ESA values from respective CVs and loading values obtained by conventional (orange bars) and XRF based scheme (green bars). At bottom of the chart, the Pt loadings (µg/cm2) obtained by conventional (in red text) and XRF based scheme (in blue text) are mentioned. As XRF based scheme allowed the estimation of Pt after the degradation measurement, the post-AST Pt-loading is shown. based scheme can be the only practical option for estimation of the amount of element under investigation. Apart from that, there are few film deposition techniques which can provide the film thickness relation to obtain the loading on the substrate 24,25. Although the XRF based scheme can avoid the ink-based discrepancies, the proposed way of estimation of catalyst loading using XRF is most-useful when the relative comparison is to be made. These characteristics are desired during the estimation of the durability of catalysts. XRF technique can readily estimate the relative loadings, although with careful calibration for a particular-catalyst one can extend its capability to estimate the absolute loadings in the coatings and thereby to estimate its actual/accurate ESA/MSA values.

TF-RDE experiments provide a fast research platform for screening of newly developed electrocatalysts for oxygen reduction reaction (ORR). Accurate estimation of their performance parameters is a necessity to reliably transfer the information to the next development stage. This investigation shades light upon the importance of Pt weight on glassy carbon (GC) disk of rotating disk electrode (RDE) used in thin film RDE experiments. The investigation suggests that the actual Pt loading of GC disk of RDE varies with the inkconditions and can hamper reproducibility and reliability of performance parameters. The term defined as catalyst-densityof-the-ink (ratio of estimated loadings and drop wt.) for a specific ink was varied when the loading is estimated using XRF based scheme with mixing time while it was considered constant when the loading is estimated using conventional scheme. This observation suggests that the conventional scheme lacks this aspect and can cause irreproducibility of data. XRF based scheme is simple and direct (requires onetime calibration), can, therefore, have the potential to replace the conventional scheme for catalyst loading estimation and possibilities for post-mortem analysis. The methodology has a profound impact on electrocatalysis, reaction mechanism understanding and new materials development.

ASSOCIATED CONTENT Supporting Information

ACS Paragon Plus Environment

Page 7 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

General procedure followed by researchers to evaluate the newly developed electrocatalyst for electrocatalysis performance parameters prior to scale-up is described and outlined (Figure S1). XRF based scheme (Figure S-2). Design of experiment. Comparison of CVs of the inks (Figure S-3). TEM images of BASF and HiSpec9100 catalyst (Figure S-3). Calculated ESA, MSA (and SSA values of investigated inks (Figure S-5).

AUTHOR INFORMATION Corresponding Author(s) * Shuang Ma Andersen: +45 6550 9186 [email protected] $ Muralidhar G. Chourashiya: [email protected] * ORCID: Shuang Ma Andersen: 0000-0003-1474-0395 $ ORCID: Muralidhar G. Chourashiya: 0000-0001-9248-0029

Author Contributions Designing and execution of experiments and analysis of results presented in the manuscript were carried out by MGC. XRF calibration experiment was carried out by RS. The manuscript was written by MGC and discussed with RS and SMA. All authors have given approval to the final version of the manuscript.

(17) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Applied Catalysis B: Environmental 2005, 56, 9-35. (18) Molina-Garcia, M. A.; Rees, N. V. RSC Advances 2016, 6, 94669-94681. (19) Ke, K.; Hiroshima, K.; Kamitaka, Y.; Hatanaka, T.; Morimoto, Y. Electrochim. Acta 2012, 72, 120-128. (20) Shinozaki, K.; Morimoto, Y.; Pivovar, B. S.; Kocha, S. S. J. Power Sources 2016, 325, 745-751. (21) Ke, K.; Hatanaka, T.; Morimoto, Y. Electrochim. Acta 2011, 56, 2098-2104. (22) Sharma, R.; Andersen, S. M. ACS Catalysis 2018, 34243434. (23) Sharma, R.; Andersen, S. M. Applied Catalysis B: Environmental 2018, 239, 636-643. (24) Sarapuu, A.; Kasikov, A.; Laaksonen, T.; Kontturi, K.; Tammeveski, K. Electrochim. Acta 2008, 53, 5873-5880. (25) Jukk, K.; Kozlova, J.; Ritslaid, P.; Sammelselg, V.; Alexeyeva, N.; Tammeveski, K. J. Electroanal. Chem. 2013, 708, 31-38.

ACKNOWLEDGEMENT Authors thank Innovation Fund Denmark (IFD) for a financial support through project UpCat (ForskEL J.No 2015-1-12315), 4M Centre (J.No 12-132710) and VILLUM FONDEN blokstipendier.

REFERENCES (1) Kocha, S. S.; Shinozaki, K.; Zack, J. W.; Myers, D. J.; Kariuki, N. N.; Nowicki, T.; Stamenkovic, V.; Kang, Y.; Li, D.; Papageorgopoulos, D. Electrocatalysis 2017, 8, 366-374. (2) Shinozaki, K.; Zack, J. W.; Pylypenko, S.; Richards, R. M.; Pivovar, B. S.; Kocha, S. S. Int. J. Hydrogen Energy 2015, 40, 16820-16830. (3) Gojković, S. L.; Zečević, S. K.; Savinell, R. F. J. Electrochem. Soc. 1998, 145, 3713-3720. (4) Gloaguen, F.; Andolfatto, F.; Durand, R.; Ozil, P. J. Appl. Electrochem. 1994, 24, 863-869. (5) Schmidt, T. J.; Gasteiger, H. A.; Stäb, G. D.; Urban, P. M.; Kolb, D. M.; Behm, R. J. J. Electrochem. Soc. 1998, 145, 23542358. (6) Paulus, U. A.; Schmidt, T. J.; Gasteiger, H. A.; Behm, R. J. J. Electroanal. Chem. 2001, 495, 134-145. (7) Higuchi, E.; Uchida, H.; Watanabe, M. J. Electroanal. Chem. 2005, 583, 69-76. (8) Masa, J.; Batchelor-McAuley, C.; Schuhmann, W.; Compton, R. G. Nano Research 2014, 7, 71-78. (9) Treimer, S.; Tang, A.; Johnson, D. C. Electroanalysis 2002, 14, 165-171. (10) Zhou, R.; Zheng, Y.; Jaroniec, M.; Qiao, S.-Z. ACS Catalysis 2016, 6, 4720-4728. (11) Sokolov, S. V.; Sepunaru, L.; Compton, R. G. Applied Materials Today 2017, 7, 82-90. (12) Shinozaki, K.; Zack, J. W.; Richards, R. M.; Pivovar, B. S.; Kocha, S. S. J. Electrochem. Soc. 2015, 162, F1144-F1158. (13) Shinozaki, K.; Zack, J. W.; Pylypenko, S.; Pivovar, B. S.; Kocha, S. S. J. Electrochem. Soc. 2015, 162, F1384-F1396. (14) Hafez, I. H.; Berber, M. R.; Fujigaya, T.; Nakashima, N. Scientific Reports 2014, 4, 6295. (15) Jukk, K.; Kongi, N.; Tammeveski, K.; Arán-Ais, R. M.; Solla-Gullón, J.; Feliu, J. M. Electrochim. Acta 2017, 251, 155166. (16) Garsany, Y.; Baturina, O. A.; Swider-Lyons, K. E.; Kocha, S. S. Anal. Chem. 2010, 82, 6321-6328.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 16

Table of Contents artwork:

ACS Paragon Plus Environment

8

Page 9 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

218x178mm (300 x 300 DPI)

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

205x126mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 10 of 16

Page 11 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

209x139mm (300 x 300 DPI)

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

383x154mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 12 of 16

Page 13 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

234x175mm (300 x 300 DPI)

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

253x196mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 14 of 16

Page 15 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

233x168mm (300 x 300 DPI)

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

631x316mm (96 x 96 DPI)

ACS Paragon Plus Environment

Page 16 of 16