& Ligand Effects Dual Metastability in Electroless Plating: Complex Inertness Enabling the Deposition of Composition-Tunable Platinum Copper Alloy Nanostructures Tobias Stohr,[a] Joachim Brçtz,[a] Mehtap Oezaslan,[b, c] and Falk Muench*[a] Abstract: Autocatalytic deposition represents a facile, ver- satile, and scalable wet-chemical tool for nanofabrication. However, the intricate component interplay in plating baths containing multiple metal species impedes alloy deposition. We resolved this challenge in the bimetallic copper-platinum system by exploiting the kinetic stability of platinum complexes, which allows adjusting their ligand sphere and thus reactivity independently from the present copper ions in a preceding, thermally activated ligand exchange step. By using metastable PtIV precursors of varying degrees of complexation, copper-platinum alloys of adjustable atomic ratio were plated from solu- tions of identical composition and concentration, but dif- fering local coordination environment. Due to its excellent conformity and nanoscale homogeneity, the reaction is compatible with ambitious 3D substrate morphologies, as demonstrated in the template-assisted fabrication of nanotubes with high aspect ratio. The ability to generate additional synthetic degrees of freedom by decoupling the metal complex speciation from the solution composi- tion is of large interest for redox-chemical synthesis tech- niques, such as electrodeposition or nanoparticle colloid production. Autocatalytic deposition, also referred to as chemical or elec- troless plating, is an industrially well-established, wet-chemical reaction class used for the conformal metallization of almost arbitrary work pieces.[1] Mechanistically, this process is related to electrodeposition, with the notable distinction that the elec- trons needed for metal-ion reduction are not externally sup- plied, but generated in situ by the oxidation of reducing agents. While the plating baths tend to decompose into ele- mental metals and oxidized reducing agents, this conversion is kinetically suppressed to a large extent. The activation barrier can be overcome by heterogeneous catalytic action, enabling selective deposition on catalytic surfaces and the continued growth of the plated, autocatalytic metal films. Due to its efficiency, autocatalytic deposition is a compelling tool for nanofabrication. High quality nanoscale metal coatings can be obtained by simply submerging substrates in aqueous plating baths at ambient temperatures.[2–8] The outstanding conformity of the reaction allows replicating the shape of intri- cate templating scaffolds,[2–5, 7] providing access to sophisticat- ed nano-architectures such as arrays,[2, 5] networks[4, 5] or multi- scale hierarchical lattices[3] composed of metal nanotubes. Stopping plating at early stages allows producing island-like metal nanoparticle films,[6] and shape-controlled reaction var- iants can be used to directly coat surfaces with nanowires,[8] nanoplates,[8] or nanospikes.[4] Despite their technological importance,[9] alloy nanomaterials are not often produced by autocatalytic deposition, apart from two notable exceptions: (i) combinations of the chemically sim- ilar iron group metals Fe, Co and Ni[10] and (ii) alloys containing heteroatoms like P or B formed as byproducts from from the employed reducing agents.[10] The relative scarcity of autocata- lytic alloy plating reactions in general and comprising noble metals in particular is related to their strong focus on nickel- or copper-based protective and conductive coatings,[1, 10] as well as to their complexity. Compared to electroplating, alloy deposition is considerably more intricate in autocatalytic systems. Here, the electrode po- tential as the pivotal driving force cannot be freely and dynam- ically controlled, but is determined in situ by the interplay of the oxidation and reduction half reactions simultaneously pro- ceeding on the deposit.[1] All included metals must be deposit- ed at comparable rates under these confined conditions, and techniques such as pulsed deposition cannot be used to cross reducibility gaps or to tune the alloy composition.[11] The auto- catalytic qualities of the metallic deposit governing the plating reaction are a strong function of its composition, and thus [a] T. Stohr, Dr. J. Brçtz, Dr. F. Muench Department of Materials and Earth Sciences Technische Universit-t Darmstadt 64287 Darmstadt (Germany) E-mail : muench@ma.tu-darmstadt.de [b] Prof. Dr. M. Oezaslan Institute of Technical Chemistry Technische Universit-t Braunschweig Hagenring 30, 38106 Braunschweig (Germany) [c] Prof. Dr. M. Oezaslan Department of Chemistry Carl von Ossietzky University of Oldenburg 26111 Oldenburg (Germany) Supporting information and the ORCID identification number(s) for the au- thor(s) of this article can be found under : https ://doi.org/10.1002/chem.202000158. T 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA. This is an open access article under the terms of Creative Commons Attri- bution NonCommercial License, which permits use, distribution and repro- duction in any medium, provided the original work is properly cited and is not used for commercial purposes. Chem. Eur. J. 2020, 26, 3030 – 3033 T 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim3030 CommunicationDOI: 10.1002/chem.202000158 http://orcid.org/0000-0001-8545-7576 http://orcid.org/0000-0001-8545-7576 http://orcid.org/0000-0001-5279-0989 http://orcid.org/0000-0001-5279-0989 https://doi.org/10.1002/chem.202000158 http://crossmark.crossref.org/dialog/?doi=10.1002%2Fchem.202000158&domain=pdf&date_stamp=2020-02-18 markedly altered in alloys.[12] Chemical metastability is chal- lenging to achieve, more so simultaneously for multiple spe- cies: Adding ions of a more reactive metal to an autocatalytic deposition solution brings the risk of uncontrolled reduction, resulting in homogeneous nucleation and bath decomposi- tion.[12] While this reactivity can be exploited for seeding bi- metallic nanoparticle colloids by homogeneous nucleation of noble metal cores,[13] it is inappropriate for surface-selective and alloy depositions. Aside the reducer, additives[14] and li- gands can also cause interferences. For instance, while com- plexation represents a powerful strategy for tailoring the elec- trochemical reactivity,[15, 16] limitations arise due to the concur- rent effect of ligands on all present metal ions. Herein, we highlight how this interference can be overcome by exploiting a curious feature of platinum complexes: Their kinetic stability[17] allows predefining the coordination environ- ment of the platinum precursor, effectively decoupling its com- position from the ligands present in the plating solution. This strategy provides independent control over the platinum com- plex reactivity, which remains locked after addition to the plat- ing bath. Using ligand-shell engineering to systematically vary the degree of platinum precursor stabilization, we are able to autocatalytically deposit composition-tunable copper–platinum alloy nanofilms from solutions of identical global composition. Intriguingly, those solutions exhibit dual metastability, with re- spect to both metal reduction and ligand exchange at the plat- inum center. Due to its previous successful implementation in autocata- lytic deposition,[18] we chose ethylenediamine (en) for PtIV com- plexation. Starting from [PtCl6]2@, up to three en units can be attached, each replacing two chloride ions upon chelation (Fig- ure 1 A). UV/Vis spectroscopy was used to confirm the inert- ness of the PtIV species in our reaction system. To this end, four different degrees of complexation were realized by mixing [PtCl6]2@ solutions with 0, 1.3, 2.7, and 4 equivalents of en, spanning the complete range from unaltered to fully ligand-ex- changed PtIV species using a slight en excess. During heating, the color of the initially orange–yellow solutions faded to (pale) yellow, with the extent of color change depending on the amount of added en, indicating different levels of ligand exchange. In the following, we use the number n of en equiva- lents applied in the temperature-facilitated ligand exchange step (Figure 1 B) to denote our experiments. After cooling down to room temperature, each solution was balanced with en to reach a total of four ligand equivalents (Figure 1 B), and mixed with diluted NaOH to mimic the alkalinity of the plating baths the PtIV species are added to later (Supporting Informa- tion, Section S1.5). The distinct variance of the UV/Vis spectra strikingly con- trasts the globally identical composition of the solutions and immediately corroborates the differing nature of the present PtIV complexes (Figure 1 C,D). Notably, it is found after en bal- ancing, providing clear evidence for the PtIV complex inertness in our system: The history of solution preparation counts, not the overall amount of added ligand. Absorption bands are found around 450, 370, and 260 nm, which have been previ- ously assigned to (partially spin-forbidden) d–d transitions (1A1g!1Tg/3Tg) and ligand-to-metal charge-transfer bands (t2u! eg) in [PtCl6]2@.[19, 20] With an increasing fraction of en added at elevated temperature, all absorption bands decrease in intensi- ty, pointing out the increasing substitution of chloride by en. While aquation also must be considered[19] and chloride can be completely replaced by hydroxide—although this process is thermally activated as well[20]—efficient solvolysis can be ruled out in our case: The peak centered around 260 nm, which is characteristic for the chloride-to-metal charge transfer band and disappears in the hydroxo complex,[20] is pronounced in the samples of low en complexation (n = 0, 1.3). Its absence in the most substituted sample (n = 4) verifies total chloride ion replacement by en. Cu–Pt alloy deposition was achieved by adding four differ- ent PtIV solutions (n = 0, 1.3, 2.7, 4) prepared alike to those uti- lized in the spectroscopic studies to alkaline, formaldehyde- Figure 1. A) Scheme showing the different levels of ligand exchange at [PtCl6]2@, which acts as PtIV precursor in our synthesis. B) Preparation Scheme of the solutions for spectrophotometric analysis. UV/Vis spectra of the com- plex solutions: C) diluted 30-fold with water, D) as-prepared (n denotes the equivalents of en used for Pt(IV) complexation). Chem. Eur. J. 2020, 26, 3030 – 3033 www.chemeurj.org T 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim3031 Communication 15213765, 2020, 14, D ow nloaded from https://chem istry-europe.onlinelibrary.w iley.com /doi/10.1002/chem .202000158 by U niversitã¤T s- U nd, W iley O nline L ibrary on [07/04/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense http://www.chemeurj.org based[21] electroless copper plating baths (Supporting Informa- tion, Section S1.4). The final plating baths featured a constant en content of 245 mm. Ion-track etched polycarbonate mem- branes were used as model substrates for the template-assist- ed synthesis of 1D nanostructures.[2, 4] These membranes incor- porate channel-shaped pores of high aspect ratio, and there- fore represent an excellent platform for testing the nanoscale homogeneity and conformity of the deposit. Being intrinsically inactive, polymer membranes have to be decorated with cata- lytic metal nanoparticles like Ag or Pd prior to plating in order to initiate metallization.[2, 4] Opposed to the autocatalytic depo- sition of pure copper [Eq. 1], the Cu–Pt depositions have not been accompanied by strong gas evolution. This hints at the catalytic interference of elemental platinum, onto which form- aldehyde is oxidized together with detached H atoms, prevent- ing H2 recombination [Eq. 2]:[12, 21] HCHOþ 2 OH@ ! HCOO@ þ 1=2 H2 " þH2Oþ e@ ð1Þ HCHOþ 3 OH@ ! HCOO@ þ 2 H2Oþ 2 e@ ð2Þ The optical appearance of the metallized foils ranged from anthracite for the pristine [PtCl6]2@ to coppery for the most substituted PtIV complex, suggesting that increased platinum stabilization translates to reduced incorporation into the de- posit (Figure 2). Energy-dispersive x-ray spectroscopy (EDS) confirms this anticipated behavior, finding at% compositions of Cu79Pt21 (n = 0), Cu87Pt13 (n = 1.3), Cu95Pt5 (n = 2.7), and Cu98Pt2 (n = 4). SEM revealed the successful formation of bimetallic nano- tubes, which endured the template dissolution without notice- able fragmentation and spanned the entire template thickness of &30 mm (Figure 3 A). As a result of efficient nucleation with high areal density,[22] the tubes tightly replicate the cylindrical shape of the template pores and possess compact walls of rel- atively even thickness (Figure 3 B). Investigating the material with TEM exposed the nanoparticulate quality of the deposit, which is composed of fine grains of <5 nm particle size (Fig- ure 3 C, D). In agreement with the small particle size, XRD yielded con- siderably broadened reflexes. Aside from a pronounced back- ground related to the amorphous polycarbonate substrate, all samples exhibit reflexes corresponding to a face-centered cubic (fcc) phase (Figure 3 E). The corresponding lattice con- stants lie in between those of the parent metals, implying that the deposit consists of a disordered Pt–Cu alloy rather than a heterostructure of unalloyed individual metals. The lattice con- stants monotonically increase with increasing Cu content, ap- proaching the values of pure copper in the case of the most Cu-rich sample (n = 4), in agreement with the EDS results showing the presence of only few at % of Cu. In the case of the most Pt-rich sample, the reflexes are closely located to the positions expected for PtCu3 (Supporting Information, Fig- ure S2), again matching the composition estimate by EDS (Pt21Cu79). All samples show signs of oxidation (presence of cuprous oxide), the extent of which decreases with the nobility Figure 2. Photographs of metallized polycarbonate templates obtained with different complexation levels ranging from pristine [PtCl6]2@ (left) to fully chelated PtIV (right). Figure 3. SEM images of template-freed Cu-Pt nanotubes (n = 0), showing A) a drop-coated film and B) a bundle of tube openings. TEM images of a microtome-cut, metallized template (n = 0), displaying C) nanotube cross- sections and D) the nanoparticulate structure of the deposit. E) XRD patterns of template-embedded Cu-Pt nanotubes, complemented with ICDD powder standards. The shaded areas highlight the fcc reflex positions confined by the parent metals. Chem. Eur. J. 2020, 26, 3030 – 3033 www.chemeurj.org T 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim3032 Communication 15213765, 2020, 14, D ow nloaded from https://chem istry-europe.onlinelibrary.w iley.com /doi/10.1002/chem .202000158 by U niversitã¤T s- U nd, W iley O nline L ibrary on [07/04/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense http://www.chemeurj.org of the alloys. The most Pt-rich sample (n = 0) only contains a minor contribution of Cu2O phase (see the hump at the posi- tion of the most intense Cu2O powder reflex, &368). On the contrary, the same Cu2O reflex is dominating the diffractogram of most Cu-rich sample (n = 4). In summary, the XRD analysis confirms the successful deposition of Cu–Pt solid solutions of tunable composition. Particularly the Cu-rich alloys are prone to attack by atmospheric oxygen. XPS with optional sputtering shows that the oxidation is limited to the deposit surface (Sup- porting Information, Figures S3–S5), proving the full reduction of both metal precursors during plating. In this communication, we outlined a reaction system for the highly conformal and autocatalytic deposition of nanocrys- talline Cu–Pt alloy films, the composition of which can be con- trolled by tuning the PtIV precursor reactivity in an initial, tem- perature-assisted ligand exchange step. Plating was performed at low temperatures, at which the predefined PtIV reactivity is preserved in spite of the presence of an almost 50-fold excess of the thermodynamically preferred en ligand (245 mm en versus 5 mm H2PtCl6). Our approach harnesses the kinetic sta- bility of PtIV complexes to generate an additional degree of synthetic freedom, allowing us to adjust the ligand environ- ment of the platinum precursor independently from the overall solution composition. It represents a fully wet-chemical, facile, scalable, and flexible route for fabricating nanostructures of controlled Cu–Pt ratios, which can be employed as electrocata- lysts in, for example, alcohol oxidation[23] or oxygen reduc- tion.[24] Dealloying[23, 25] the as-deposited Cu–Pt nanostructures can be used to further tune the material composition and mor- phology (Supporting Information, Figure S1). Our strategy represents a starting point for devising new au- tocatalytic alloy plating reactions involving other metastable metal centers (e.g. , RuII, OsII, IrIII)[17] and metal combinations. Also, it is interesting for mechanistically related reactions such as electrodeposition,[15] galvanic displacement,[5] or colloidal nanoparticle synthesis,[16] which benefit alike from the ability to independently adjust the reducibility of individual ions in systems the global composition of which is dictated by exter- nal requirements. Acknowledgements T.S. is grateful for the support of the German Federal Ministry of Education and Research (BMBF, contract number 05 K16 RDC, project HI-EXE). In addition, we thank the Prof. Biesalski (Technische Universit-t Darmstadt) for providing access to the spectrophotometer, Prof. C. Trautmann (GSI Helmholtzzentrum fer Schwerionenforschung, Darmstadt) for support with the template preparation and for access to the scanning electron microscope, the Zentraleinrichtung Elektronenmikroskopie and Prof. Dr. P. Strasser (Technische Universit-t Berlin) for their assis- tance, and C. Kipper for synthetic support. Conflict of interest The authors declare no conflict of interest. Keywords: autocatalytic deposition · copper-platinum alloys · kinetic complex stability · ligand effects · template synthesis [1] a) H. Niederprem, Angew. Chem. Int. Ed. Engl. 1975, 14, 614 – 620; Angew. Chem. 1975, 87, 652 – 658; b) Electroless Plating: Fundamentals and Applications (Eds. : G. O. Mallory, J. B. Hajdu), American Electroplat- ers and Surface Finishers Society, Orlando, FL, 1990. [2] S. Papp, G. J#gerszki, R. E. Gyurcs#nyi, Angew. Chem. Int. Ed. 2018, 57, 4752 – 4755; Angew. Chem. 2018, 130, 4842 – 4845. [3] X. Zheng, W. Smith, J. Jackson, B. Moran, H. Cui, D. Chen, J. Ye, N. Fang, N. Rodriguez, T. Weisgraber, C. M. Spadaccini, Nat. Mater. 2016, 15, 1100 – 1107. [4] T. Boettcher, S. Schaefer, M. Antoni, T. Stohr, U. Kunz, M. Duerrschnabel, L. Molina-Luna, W. Ensinger, F. Muench, Langmuir 2019, 35, 4246 – 4253. [5] F. Muench, Catalysts 2018, 8, 597. [6] a) F. Muench, A. Solomonov, T. Bendikov, L. Molina-Luna, I. Rubinstein, A. Vaskevich, ACS Appl. Biol. Mater. 2019, 2, 856 – 864; b) M. D. Susman, Y. Feldman, A. Vaskevich, I. Rubinstein, Chem. Mater. 2012, 24, 2501 – 2508. [7] N. M. K. Kuruppu Arachchige, P. C. Chambers, A. M. Taylor, Z. L. Highland, J. C. Garno, ACS Appl. Nano Mater. 2019, 2, 2193 – 2203. [8] a) F. Muench, R. Popovitz-Biro, T. Bendikov, Y. Feldman, B. Hecker, H. Oe- zaslan, I. Rubinstein, A. Vaskevich, Adv. Mater. 2018, 30, 1805179; b) F. Muench, S. Schaefer, L. Hageleken, L. Molina-Luna, M. Duerrschnabel, H.-J. Kleebe, J. Brçtz, A. Vaskevich, I. Rubinstein, W. Ensinger, ACS Appl. Mater. Interfaces 2017, 9, 31142 – 31152. [9] J.-M. Yan, S.-J. Li, S.-S. Yi, B.-R. Wulan, W.-T. Zheng, Q. Jiang, Adv. Mater. 2018, 30, 1703038. [10] J. Sudagar, J. Lian, W. Sha, J. Alloys Compd. 2013, 571, 183 – 204. [11] B. Yoo, F. Xiao, K. N. Bozhilov, J. Herman, M. A. Ryan, N. V. Myung, Adv. Mater. 2007, 19, 296 – 299. [12] J. E. A. M. van den Meerakker, J. W. G. de Bakker, J. Appl. Electrochem. 1990, 20, 85 – 90. [13] L. Lu, I. Sevonkaev, A. Kumar, D. V. Goia, Powder Technol. 2014, 261, 87 – 97. [14] J. Bielinski, K. Kaminski, Surf. Coat. Technol. 1987, 31, 223 – 233. [15] D. Liang, G. Zangari, Langmuir 2014, 30, 2566 – 2570. [16] N. Ortiz, R. G. Weiner, S. E. Skrabalak, ACS Nano 2014, 8, 12461 – 12467. [17] J. Reedijk, Platinum Met. Rev. 2008, 52, 2 – 11. [18] F. Muench, S. Kaserer, U. Kunz, I. Svoboda, J. Brçtz, S. Lauterbach, H.-J. Kleebe, C. Roth, W. Ensinger, J. Mater. Chem. 2011, 21, 6286 – 6291. [19] E. L. Cox, D. G. Peters, E. L. Wehry, J. Inorg. Nucl. Chem. 1972, 34, 297 – 305. [20] S. Alerasool, D. Boecker, B. Rejai, R. D. Gonzalez, Langmuir 1988, 4, 1083 – 1090. [21] M. Zhao, L. Yu, R. Akolkar, A. B. Anderson, J. Phys. Chem. C 2016, 120, 24789 – 24793. [22] Y. Wang, S. Luo, K. Ren, S. Zhao, Z. Chen, W. Li, J. Guan, J. Mater. Chem. C 2016, 4, 2566 – 2578. [23] X. Zhang, W. Lu, J. Da, H. Wang, D. Zhao, P. A. Webley, Chem. Commun. 2009, 195 – 197. [24] P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C. Yu, Z. Liu, S. Kaya, D. Nordlund, H. Ogasawara, M. F. Toney, A. Nilsson, Nat. Chem. 2010, 2, 454 – 460. [25] M. Oezaslan, F. Hasch8, P. Strasser, J. Electrochem. Soc. 2012, 159, B444. Manuscript received: January 10, 2020 Accepted manuscript online: January 16, 2020 Version of record online: February 18, 2020 Chem. Eur. J. 2020, 26, 3030 – 3033 www.chemeurj.org T 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim3033 Communication 15213765, 2020, 14, D ow nloaded from https://chem istry-europe.onlinelibrary.w iley.com /doi/10.1002/chem .202000158 by U niversitã¤T s- U nd, W iley O nline L ibrary on [07/04/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://doi.org/10.1002/anie.197506141 https://doi.org/10.1002/anie.197506141 https://doi.org/10.1002/anie.197506141 https://doi.org/10.1002/ange.19750871804 https://doi.org/10.1002/ange.19750871804 https://doi.org/10.1002/ange.19750871804 https://doi.org/10.1002/anie.201800954 https://doi.org/10.1002/anie.201800954 https://doi.org/10.1002/anie.201800954 https://doi.org/10.1002/anie.201800954 https://doi.org/10.1002/ange.201800954 https://doi.org/10.1002/ange.201800954 https://doi.org/10.1002/ange.201800954 https://doi.org/10.1038/nmat4694 https://doi.org/10.1038/nmat4694 https://doi.org/10.1038/nmat4694 https://doi.org/10.1038/nmat4694 https://doi.org/10.1021/acs.langmuir.9b00030 https://doi.org/10.1021/acs.langmuir.9b00030 https://doi.org/10.1021/acs.langmuir.9b00030 https://doi.org/10.3390/catal8120597 https://doi.org/10.1021/acsabm.8b00702 https://doi.org/10.1021/acsabm.8b00702 https://doi.org/10.1021/acsabm.8b00702 https://doi.org/10.1021/cm300699f https://doi.org/10.1021/cm300699f https://doi.org/10.1021/cm300699f https://doi.org/10.1021/acsanm.9b00153 https://doi.org/10.1021/acsanm.9b00153 https://doi.org/10.1021/acsanm.9b00153 https://doi.org/10.1002/adma.201805179 https://doi.org/10.1021/acsami.7b09398 https://doi.org/10.1021/acsami.7b09398 https://doi.org/10.1021/acsami.7b09398 https://doi.org/10.1021/acsami.7b09398 https://doi.org/10.1002/adma.201703038 https://doi.org/10.1002/adma.201703038 https://doi.org/10.1016/j.jallcom.2013.03.107 https://doi.org/10.1016/j.jallcom.2013.03.107 https://doi.org/10.1016/j.jallcom.2013.03.107 https://doi.org/10.1002/adma.200600606 https://doi.org/10.1002/adma.200600606 https://doi.org/10.1002/adma.200600606 https://doi.org/10.1002/adma.200600606 https://doi.org/10.1007/BF01012475 https://doi.org/10.1007/BF01012475 https://doi.org/10.1007/BF01012475 https://doi.org/10.1007/BF01012475 https://doi.org/10.1016/j.powtec.2014.04.015 https://doi.org/10.1016/j.powtec.2014.04.015 https://doi.org/10.1016/j.powtec.2014.04.015 https://doi.org/10.1021/la404858j https://doi.org/10.1021/la404858j https://doi.org/10.1021/la404858j https://doi.org/10.1021/nn5052822 https://doi.org/10.1021/nn5052822 https://doi.org/10.1021/nn5052822 https://doi.org/10.1039/c0jm03522j https://doi.org/10.1039/c0jm03522j https://doi.org/10.1039/c0jm03522j https://doi.org/10.1016/0022-1902(72)80390-7 https://doi.org/10.1016/0022-1902(72)80390-7 https://doi.org/10.1016/0022-1902(72)80390-7 https://doi.org/10.1021/la00083a003 https://doi.org/10.1021/la00083a003 https://doi.org/10.1021/la00083a003 https://doi.org/10.1021/la00083a003 https://doi.org/10.1021/acs.jpcc.6b07546 https://doi.org/10.1021/acs.jpcc.6b07546 https://doi.org/10.1021/acs.jpcc.6b07546 https://doi.org/10.1021/acs.jpcc.6b07546 https://doi.org/10.1039/C5TC04151A https://doi.org/10.1039/C5TC04151A https://doi.org/10.1039/C5TC04151A https://doi.org/10.1039/C5TC04151A https://doi.org/10.1039/B813830C https://doi.org/10.1039/B813830C https://doi.org/10.1039/B813830C https://doi.org/10.1039/B813830C https://doi.org/10.1038/nchem.623 https://doi.org/10.1038/nchem.623 https://doi.org/10.1038/nchem.623 https://doi.org/10.1038/nchem.623 https://doi.org/10.1149/2.106204jes http://www.chemeurj.org