RSC Advances View Article Online PAPER View Journal | View Issue ‘Green’ Cr(III)–glycine electrolyte for the This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Cite this: RSC Adv., 2019, 9, 25762 production of FeCrNi coatings: electrodeposition Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. mechanisms and role of by-products in terms of coating composition and microstructure† Enrico Bertero, *ab Cristina V. Manzano, a Eva Pellicer, c Jordi Sort, cd Robert M. Ulfig,e Stefano Mischler,b Johann Michlera and Laetitia Philippea The electrodeposition of stainless steel-like FeCrNi alloys for miniaturised devices is appealing as it would allow combining excellent material properties (e.g. corrosion resistance, hardness, biocompatibility) at low-cost. However, conventional baths often contain hazardous hexavalent chromium. Cr-based alloys electrodeposited from environmentally friendly trivalent chromium electrolytes are crucial for industrial application for facilitating the transition towards sustainable and ecological production and processing. Nevertheless, this process has not been comprehensively studied or understood in depth: especially the role of organic agents (common additives for improving Cr(III)-based plating; e.g. glycine) in terms of material properties of the electrodeposits. The aim of this work was to investigate the electrodeposition of FeCrNi coatings from a ‘green’ Cr(III)–glycine electrolyte. Novel information was attained by analysing films developed under various conditions and characterising them using a combination of advanced techniques. The evolution of microstructure (from amorphous to nanocrystalline) in correlation with film composition (i.e. metals ratio and presence of impurities) and elemental 3D spatial distribution was achieved for coatings produced from different anode materials and thermal post-treatment. The influence of Cr(III) and glycine in terms of coating atomic contents (i.e. Fe–Cr–Ni–O–C–N–H) was evaluated for films in which both the applied current density Received 6th June 2019 Accepted 2nd August 2019 and electrolyte composition were varied. These results, together with a thorough analysis on metals speciation/complexation allowed us to propose various Cr(III)-based electroreduction mechanisms, and to DOI: 10.1039/c9ra04262h observe, upon annealing, segregation and distribution of impurities, as well as of oxides and metals with rsc.li/rsc-advances respect to microstructure variation, providing an explanation for the amorphisation process. many outstanding properties of Cr coatings (hardness, wear 1. Introduction and corrosion resistance, etc.).1 However, the commonly used Electrodeposition of chromium-based lms has been widely hexavalent chromium electrolyte must be replaced, as it is toxic used in various industrial sectors, due to several process and its waste poses an environmental threat. Conversely, the advantages (e.g. versatility, scalability, cost-effectiveness) and trivalent chromium plating electrolyte constitutes a ‘green’ and effective alternative to Cr(VI) baths.2 Unfortunately, Cr coatings obtained from trivalent chromium electrolytes do not yet match a Empa – Swiss Federal Laboratories for Materials Science and Technology, Laboratory the outstanding characteristics of the lms produced using for Mechanics of Materials and Nanostructures, Feuerwerkerstrasse 39, 3602, Thun, Cr(VI) baths. The reactions taking place during trivalent chro- Switzerland. E-mail: enrico.bertero@empa.ch; cristina.vicente@empa.ch; johann. michler@empa.ch; laetitia.philippe@empa.ch; Fax: +41 58 765 69 90; Tel: +41 58 mium electrodeposition are very complex and have yet to be 765 63 93 thoroughly investigated. Organic compounds are usually added b Ecole Polytechnique Fédérale de Lausanne, Tribology and Interfacial Chemistry to the electrolyte to form Cr(III)-complex molecules, which Group, Materials Institute Station 12 (SCI-STI-SM), 1015 Lausanne, Switzerland. improve the deposition process. However, they are a source of E-mail: stefano.mischler@ep.ch c unwanted impurities, e.g. carbon. Many studies have focussed Departament de Fı́sica, Facultat de Ciències, Universitat Autònoma de Barcelona, on Cr(III) deposition mechanisms without investigating the 08193 Bellaterra, Spain. E-mail: eva.pellicer@uab.cat; jordi.sort@uab.cat d Institució Catalana de Recerca i Estudis Avançats (ICREA), Pg. Lluı́s Companys 23, pathways of how the main impurities are incorporated in the 08010 Barcelona, Spain lms and how they can affect the material properties of such e CAMECA Instruments Inc., 5470 Nobel Drive, Madison, WI 53711, USA. E-mail: coatings. Additionally, more complex electrolytes including robert.ulg@ametek.com other metallic ions (e.g. transition metals like Fe and Ni) are † Electronic supplementary information (ESI) available. See DOI: seldom mentioned in the literature. FeCrNi electrodeposits 10.1039/c9ra04262h 25762 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances similar to austenitic stainless steel are of great interest for The use of Fe and Ni in a Cr(III)-complex electrochemical creating state-of-the-art micro-nano components, which could system can be found in the literature for FeNiCr20–24 and be used in the bio-medicine/micro-robotic sector. Therefore, FeNiCrMo25 electrodepositions. However, these works mainly a study of the mechanisms occurring in Cr(III)-based electrolytes focus on the inuence of deposition parameters on lm containing other metal ions and organic complexing agents is features, overlooking the role of the electrochemical reactions essential for the development of this important process. on the coating's composition and properties. Only one of our The electroreduction mechanism of the trivalent chromium previous publications26 tackled the electrodeposition of FeCrNi electrolyte was shown to follow a step-wise mechanism with the from a Cr(III)–glycine electrolyte and put forward an explanation formation of rather stable intermediate Cr(II) compounds:3–6 for the source of impurities incorporated in the lms. According This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. to this study, the use of an inert platinum anode was the cause of Cr(III) / Cr(II) E0 ¼ 0.41 V vs. SHE (1) Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. the oxidation of glycine, which resulted in the formation of by- products (e.g. formaldehyde, formic acid). These were, in turn, Cr(II) / Cr(0) E0 ¼ 0.91 V vs. SHE (2) the main source for carbon inclusion and consequent amorphisation of the deposit. By obtaining nanocrystalline low- Step (2) has been accepted to be the limiting reaction in the carbon (0.6 wt%) FeCrNi coatings on a double-cell set-up with Cr(III) to Cr(0) reduction process.3 The Cr(III) ion forms chro- separated anode and cathode compartments, it was thought mium–aqua complexes, such as [Cr(H2O)6]3+, which are char- that the amorphisation mechanism was caused by inclusion of acterised by low kinetic activity due to the close-ordered hexa- interstitial carbon, present in a supersaturated state. Nonethe- water structure that prevents the central Cr3+ ion to exchange less, that preliminary work did not propose an electrochemical electrons with the cathode surface.4,7,8 For this reason, a com- pathway for carbon incorporation. Additionally, the fact that plexing agent (e.g. glycine, formate, oxolate, acetate or citrate) is carbides were not detected differs from many similar studies on oen introduced into the electrolyte to destabilise the chro- Cr(III)–formate electrolytes.10–12 Only Protsenko et al.11 suggested mium–aqua complex and enable chromium deposition. Addi- a possible mechanism for carbon inclusion from a trivalent tionally, pH buffers (e.g. boric acid) are added to hinder chromium bath containing formic acid as the complexing agent. hydrolysis and olation (and, ultimately, polymerisation). When On one hand, the presence of carbides increases the hard- in close proximity to the cathode surface, chromium-complexes ness of stainless steels.27 On the other, high amounts of carbon/ typically undergo these reactions when the pH has increased carbides are undesirable as they can decrease corrosion resis- because of hydrogen evolution reaction (HER).8 tance and, therefore, the biocompatibility of Cr-based coat- Glycine (NH2CH2COOH) has been used in trivalent chro- ings.28 Other impurities, such as nitrogen and hydrogen are also mium deposition, as it is both a strong complexing agent and known to have an impact on the morphology and properties of exhibits pH buffering properties, allowing to avoid olation alloys containing chromium like stainless steels.27 According to reactions.7 McDougall et al.9 stated that optimal deposition Gabe,29 a local rise of pH at the cathode results in the formation, results (i.e. high deposition rate, good surface morphology and adsorption and incorporation of Cr hydrides (Cr–H). H incor- uniformity) could be obtained from an electrolyte with poration is thought to play a key role in the coating's porosity Cr(III) : glycine molar ratio of 1 : 1, resulting in monoligand and embrittlement, lowering deposition efficiency and Cr(III)–glycine as the main compound present inside the elec- producing amorphous lms. trolyte. Still the electroreduction mechanism involved are not There is a lack of thorough research conducted to explain or completely understood, in fact, the vast majority of works using to better comprehend the possible mechanisms involved in complexed chromium electroplating (e.g. formate,4,5 acetate,5 complexed Cr(III)-based alloys electroplating and how they are glycine7,9) propose that the complexing ligands are completely linked to the obtained lms' characteristics. This is particularly released during the reduction process (and stabilised by cations true regarding the pathways by which impurities are incorpo- present in the electrolyte). However, in many studies10–12 the rated in the deposits, as they are as important as the main presence of impurities (mainly carbon) and the formation of metallic elements in determining the coatings' material prop- compounds, such as carbides, provide evidence of the partici- erties. It is crucial to investigate how the morphology, micro- pation of these organic molecules and the incorporation of their structure and lm composition are affected by deposition moieties during Cr(III)-complex deposition process. parameters and post-processing for this systems. This is The purpose of electrodepositing multicomponent alloys a necessary step for further improving Cr(III) electrodeposition, containing Cr is to obtain a combination of properties that considering that the process is of great importance for envi- cannot be achieved by pure Cr coatings. For example, ferro- ronmentally friendly and sustainable Cr-based applications. magnetic elements such as Fe, Ni and Co are utilised to confer The objective of this work is to further investigate the a magnetic response to the electrodeposited lms.13,14 In considered system consisting of Cr(III)–glycine complexes, Fe general, magnetic, corrosion and mechanical properties of and Ni ions, in terms of morphology, microstructure and electrodeposited coatings are dictated by the amount and composition of the nal electrodeposits. nature of the transition metals and alloying elements.15 The inuence of impurities and deposition parameters on Biocompatibility as a material property can be obtained from such material's properties is achieved by rst varying the anode very few electrodeposited materials, mainly CrCo and stainless material and then the electrolyte composition, in combination steel such as FeCrNi-based alloys.16,17 with post-treatments (i.e. thermal annealing) and several This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25763 View Article Online RSC Advances Paper advanced characterisation techniques. In particular, X-ray Table 1 Bath composition for FeCrNi electrodepositiona diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) Components Concentration (mol L1) are utilised to obtain information on the correlation between coating microstructure and composition, while taking into CrCl3$6H2O 0.4 account the chemical states of Fe, Cr, Ni, O, C and N. Addi- Glycine (NH2CH2COOH) 0.4 tionally, He elastic recoil detection analysis (He ERDA) is used to FeCl2$4H2O 0.03 accurately determine the hydrogen content throughout the NiCl2$6H2O 0.2 NH4Cl 0.5 coating's thickness. Furthermore, 3D atom-by-atom elemental H3BO3 0.15 reconstruction is implemented using cutting-edge atom probe This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. NaCl 0.5 tomography (APT). This XPS complementary technique is used a pH: 1, temperature: 22  C. Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. to precisely determine which elements/compounds are present and how they are spatially distributed within the FeCrNi coatings at the nano-scale, providing information regarding material inhomogeneity (i.e. clustering, phase separation, precipitates) constituents were mixed by following the procedure described and coating elemental evolution during microstructural and by Hasegawa et al.26. The Cr(III) and glycine solution was rst material composition variations (before and aer annealing complexed at 80  C for 30 minutes, to which a second solution, process). The inuence of glycine in terms of metals complexa- comprising of all the other elements, was added. The electro- tion and information on Cr oxidation species within different Cr- deposition was carried out in galvanostatic mode using based electrolytes were evaluated by means of chemical equi- a potentiostat (PGSTAT 302N, Metrohm Autolab B.V.) controlled librium diagrams and UV-vis spectra, respectively. by NOVA (version 2.1) soware. The used electrolytes were aged Based on all of the obtained results, some mechanisms are galvanostatically for a few hours at a constant current density to proposed to explain the following processes involved in the obtain lms with good surface morphology and uniformity. amorphisation of the studied FeCrNi electrodeposits: Two investigation methods were performed to better (1) Complexed Cr(III) electroreduction. understand the deposition mechanisms involved in the studied (2) Complexing agent (i.e. glycine) side-reactions. Cr(III)-based FeCrNi electrolyte. Particular emphasis was placed (3) Inuence of transition metals (i.e. Fe, Ni) on cathodic on the role of Cr(III), additives (i.e. glycine) and ferromagnetic reduction processes. elements (i.e. Fe and Ni) on material composition and micro- structural changes of FeCrNi electroplated lms. The rst method involved studying the inuence of the anode 2. Experimental methods material by electrodepositing FeCrNi alloys using a platinum FeCrNi lms were deposited using a standard three electrode anode (Pt anode) and a pure nickel anode (Ni anode). The applied electrochemical cell equipped with a water jacket for tempera- deposition parameters were the same for both cases (current ture control. The temperature of the electroplating bath was set density ¼ 80 mA cm2, total deposited charge density z 200 C to 22  C  0.5  C using a temperature controlled circulator cm2) resulting in average deposition potentials of 1.44 V vs. (Julabo, F12-ED). Portions of silicon wafer covered with Ag/AgCl and 1.41 V vs. Ag/AgCl for the Pt anode and Ni anode a sputter-coated Au layer (100 nm) deposited on top of a Cr cases, respectively. Next, the samples were annealed using adhesion layer (5 nm) were used as substrates. Prior to each a rapid thermal annealing (RTA) system (MILA-5050, ULVAC) electrodeposition experiment, the substrate was cleaned in under an Ar controlled atmosphere (99.9999% purity) for 1 hour a freshly prepared Piranha solution (30% H2O2 : H2SO4 ¼ 1 : 3) at 600  C (ramp up and ramp down of 10  C min1). and thoroughly rinsed in deionized water (18.2 MU cm). In the second method, the inuence of Cr(III)–glycine Subsequently, the substrate was partially masked with plastic complexation was determined by using three different electro- tape to expose a plating area of 1.5 cm  1.5 cm. A saturated Ag/ lytes: Standard (as described by Table 1), No Cr–Glycine (Stan- AgCl electrode was used as the reference electrode. Depending dard without Cr and glycine) and No Cr (Standard without Cr). on the experiment, a Pt mesh (80 mesh, 25 mm  35 mm, ALS FeCrNi electrodeposition was performed galvanostatically at Co., Ltd) or pure Ni pellets inside a platinised titanium basket 60 and 80 mA cm2 (total deposited charge density z 50 C (PtTi ne mesh, H  W  D: 65 mm  50 mm  15 mm) were cm2) with the Standard electrolyte resulting in average depo- used as a counter-electrode. The bath composition for FeCrNi sition potentials of 1.30 V vs. Ag/AgCl and 1.44 V vs. Ag/AgCl, electrodeposition is shown in Table 1. This electrolyte compo- respectively. For the No Cr–Glycine and No Cr baths electrode- sition was found to be the most adequate for obtaining coatings position at current density of 80 mA cm2, average deposition with the right balance of Fe–Cr–Ni similarly to metallurgical potentials were 1.55 V vs. Ag/AgCl and 1.39 V vs. Ag/AgCl, austenitic stainless steel (18Cr–11Ni).30 The difference in respectively. A Pt mesh was used as the counter electrode in concentration between Fe and Ni salts is due to the so-called all experimental variations. anomalous co-deposition, which favours Fe electroreduction. The surface morphology of the FeCrNi samples was observed High amounts of chromium salt are necessary, as a large via eld-emission scanning electron microscopy (FE-SEM, reduction potential is needed for the deposition of this metal. Hitachi S-4800, Hitachi High-Technologies Corporation). All the chemicals were of reagent grade (Sigma-Aldrich) and Crystal structures were characterised by means of in situ Bragg– were used as-received without any further purication. The bath Brentano X-ray diffraction (BB-XRD, Brucker) with a Cu Ka 25764 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances radiation source (40 kV, 40 mA). Diffractograms were recorded Scanning XPS Microprobe System{ (to determine the impact of in the 2q range between 15 and 100 . An offset of 1 was used to Cr(III)–glycine complexation on deposit properties by varying the avoid the signal from the Si substrate. Moreover, possible electrolyte composition). uorescence signals associated with the presence of Fe in the XPS spectra were analysed using CasaXPS soware32 (version lms were ltered by using a larger lower discriminator. The 2.3.19). The spectra were charge corrected by shiing all mean size of the crystallite s (more specically, the coherent binding energies (BE) with respect to adventitious carbon (C 1s; diffraction length) was calculated using the Debye–Scherer C–C, C–H) at 284.8 eV. The regions of interest and corre- equation: sponding relative sensitivity factors (R.S.F.) for elemental Kl quantication and tting were 1s for oxygen, carbon and This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. s¼ nitrogen, and 2p3/2 for Fe, Cr and Ni elements. Core level XPS b cos q peak tting was performed using an asymmetric line shape Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. where: K is the shape factor (0.9), l the radiation wavelength, dened in CasaXPS as LF(a, b, w, m), where a and b set the b is the broadening of full width half maximum (FWHM) of the spread of the tail on each side of the Lorentzian component. selected peak and q the corresponding Bragg angle. The used The w parameter is the damping factor, which gives the inte- peaks and Bragg angles were (110) peak at 2q ¼ 45 for the a-Fe gration limit for tail reduction and m is the width of the phase and (111) peak at 2q ¼ 44 for the g-Fe phase. Fitting of Gaussian convoluted with the Lorentzian. All other components the XRD peaks was performed by using the Crystallographic were tted using a Gaussian (Y%)–Lorentzian (X%) prole Open Database – COD.31 dened in CasaXPS as GL (X), where Y ¼ 100  X and the X value X-ray uorescence (XRF, Fischerscope® X-RAY XDV®-SDD, was varied depending on the analysed element (from 25 to 50). Fischer Technology) was used to estimate the Fe–Cr–Ni wt% Moreover, a standard Shirley background was used for all and thicknesses along the sample area. The measurements were spectra. performed at 25 uniformly distributed points on each spec- Additionally, H ratio in at% with respect to all metal species imen. The lm was characterised by an uneven composition (Fe, Cr and Ni) was determined at depths up to 200 nm by from the edges to the centre of the sample (Cr variation helium elastic recoil detection analysis (He-ERDA, ETH Zurich) z15 wt%), caused by the primary current distribution during using a 2 MeV He beam and the absorber foil technique.33 electrodeposition. Therefore, the averages of 9-matrix central Atom probe tomography (APT, EIKOS X,34 CAMECA points on each sample were chosen as representative concen- Instruments Inc. – Madison – USA) was performed on the trations. Approximate current efficiencies (C.E., %) from XRF FeCrNi electrodeposited lms. Li-out specimens attached thickness measurements were calculated by knowing the weight to W posts were sharpened into tip geometry with radius of percentage of each metal element, assuming the density of 50 nm using focused ion beam annular milling. Atom probe stainless steel to be rSS ¼ 7.93 g cm3 and using Faraday's law. analysis was conducted in laser pulsing mode with wave- The elemental composition (at%) and chemical binding length of 532 nm and laser pulse energy of 13–20 nJ operating energies (BE) of the metal species (Fe, Cr and Ni) and impurities at 200 kHz with a specimen base temperature of 50 K and (O, C and N) were evaluated via X-ray photoelectron spectros- detection rate set to 0.005–0.02 atom per pulse. The typical copy (XPS) on the coatings' surface and proling at different dataset size of each analysis is around 10–50 M ions. IVAS 3.8 depths from the surface down to approximately 500 nm. This soware from CAMECA was used for data reconstruction, was achieved by sputtering the material with Ar+ ions (2 or 4 keV detailed analysis for local chemical identication/ sputtering energy). XPS data was acquired using various composition (deduced from mass-over-charge ratio) and equipment: Physical Electronics (PHI) 5500 Multi-technique distribution of impurities. System‡ (to investigate the inuence of anode type on as- Chemical equilibrium diagrams and concentration of deposited samples), Physical Electronics (PHI) Quantum 2000 species were evaluated using Hydra-Medusa soware.35 Addi- Scanning ESCA Microprobe System§ (to investigate the effect of tionally, UV-vis absorbance spectra were measured in the thermal annealing) and Physical Electronics (PHI) VersaProbe II wavelength range between 300 to 700 nm using a UV-vis spec- trophotometer (UV-vis Lambda 900 UV, PerkinElmer) for ‡ Monochromatic Al Ka X-ray radiation source (hn ¼ 1486.6 eV) of 350 W power different electrolytes: freshly prepared and galvanostatically with a typical beam diameter of 800 mm. Hemispherical capacitor aged (using Pt and Ni anodes) FeCrNi solutions diluted 10 times electron-energy analyser equipped with a multichannel plate detector. Electron in deionised water, trivalent decorative chromium and hard take-off angle of 45 and analyser operated in the constant pass energy mode at hexavalent chromium commercial electrolytes (riag 23.50 eV. Compensation of eventual surface charging with built-in electron neutralizer. Base pressure of the system below 1  108 Pa. Binding energy calibrated using Ag 3d5/2 at 368.21 eV with FWHM of 0.8 eV. § Monochromatic Al Ka X-ray radiation source (hn ¼ 1486.7 eV) of 29.7 W power with a typical beam diameter of 150 mm. Hemispherical capacitor { Monochromatic Al Ka X-ray radiation source (hn ¼ 1486.7 eV) of 25.2 W power electron-energy analyser equipped with a channel plate and a position-sensitive with a beam diameter of 100 mm. Spherical capacitor electron-energy analyser set detector. Electron take-off angle of 45 and analyser operated in the constant at 45 take-off angle respect to sample surface. Analyser operated in the constant pass energy mode at 29.35 eV. Compensation of eventual surface charging with pass energy mode at 46.95 eV. Compensation of eventual surface charging with built-in electron and argon ion neutralizers. Base pressure of the system below built-in electron and argon ion neutralizers. Base pressure of the system below 5  107 Pa. Binding energy calibrated using Cu 2p3/2, Ag 3d5/2 and Au 4f7/2 at 5  107 Pa. Binding energy calibrated using Cu 2p3/2, Ag 3d5/2 and Au 4f7/2 at 932.62 eV, 368.21 eV and 83.96 eV, respectively to within 0.1 eV. 932.62 eV, 368.21 eV and 83.96 eV, respectively to within 0.1 eV. This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25765 View Article Online RSC Advances Paper Oberächentechnik AG, Switzerland) diluted in deionised water and Ni anode (Fig. 1d) thermally treated samples show larger 10 and 20 K times, respectively. cracks with respect to the as-deposited counterparts. At high temperatures, unstable Cr hydrides easily decompose to metallic Cr, therefore producing more cracks. The FeCrNi 3. Results electrodeposits present nodule-like features, which are undis- 3.1. Anode role investigation tinguishable in terms of composition with respect to the smooth surface, as observed in our previous work.30 3.1.1. Surface morphology. The electrodeposit from Nevertheless, the difference in anode material does not a nickel anode (Ni anode) differs slightly in terms of appearance particularly affect the surface morphology of the studied This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. from that obtained using a platinum anode (Pt anode). Semi- coatings. bright grey coatings were achieved from both anodes, however Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. 3.1.2. Crystal structure. Ex situ Bragg–Brentano XRD the one produced using a nickel anode is less uniform, as the measurements of the electrodeposited FeCrNi Pt anode (Fig. 2a) borders appear to be duller and more fragile. The differences and Ni anode (Fig. 2b) lms are shown in Fig. 2, both as- between the samples are more pronounced aer annealing: deposited and annealed under an Ar controlled atmosphere. blue-violet semi-bright coloured surface from the Pt anode Diffraction peaks at 2q z 38 , 69 and 82 correspond to face- sample and dull-grey from the Ni anode one (pictures of the centered cubic fcc Au (111), Si (100) and fcc Au (222) crystal samples in ESI†). The colour variation from grey to blue could planes (substrate contributions), respectively. The XRD dif- be caused by a change in the thickness of the passive oxide fractogram for the as-deposited Pt anode coating presents a very layer, based on stainless steel tempering colours and oxide broad 2q peak at approx. 45 , which highlights that the amor- growth theory.27 phous state has a tendency towards (110) reection of the bcc a- Optical microscopy and SEM observations show that the Ni Fe structure. In contrast, the as-deposited Ni anode lm shows anode sample (Fig. 1c) has more cracks than the Pt anode coating (Fig. 1a). As reported in previous similar works,26,30 the presence of cracks is mainly dependant on the thickness of the lm (crack-free samples obtained when coating thickness is less than 5 mm). However, crack-free coatings (thickness # 5 mm) started to crack immediately or a few days aer deposition, when the Cr content in the coating exceeded a certain amount (Cr $ 28 wt%). This behaviour seems consistent with the incorporation mechanism of Cr hydrides within Cr(III)-based electrodeposited alloys.29 High overpotentials (therefore increase in pH) result in the formation and incorporation of Cr hydrides (hexagonal close-packed, hcp), which are unstable and decompose into metallic Cr (body-centered cubic, bcc). This crystal phase transformation leads to approx. 15% volume shrinkage and, in turn, crack formation. The annealed samples conrm the above-mentioned mech- anism. In fact, SEM observations for both the Pt anode (Fig. 1b) Fig. 1 Surface morphology of electrodeposited FeCrNi films: Pt anode Fig. 2 BB-XRD measurements of FeCrNi as-deposited and annealed case (a) as-deposited and (b) after thermal treatment, Ni anode case (c) (Ar atmosphere, 1 h at 600  C) electrodeposits for the (a) Pt anode and as-deposited and (d) after annealing process. (b) Ni anode coatings. 25766 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances a narrower peak in the same position as previously stated, of the samples. In the same table, H content ratios with respect revealing an ultrane-grained a-Fe material with a calculated to the total metal species obtained from He-ERDA are also listed crystallite size of approx. 5 nm. (the content inside the coating was determined aer 20 min of The annealed samples (Ni anode and Pt anode) possess Ar sputtering). similar diffractograms. However, there is a clear difference The unsputtered surfaces of both as-deposited lms are rich between the as-deposited and annealed lms. The thermally in oxygen and carbon, with relatively low metal species contri- treated samples are characterised by many new peaks at approx. butions, except for the Ni anode lm, where 10.2 at% Fe was 44 , 51 , 75 , 91 and 95 corresponding to the (111), (200), measured. The amount of oxygen in-depth of the coating is (220), (113) and (222) crystal planes, respectively, which are relatively large regardless of the anode type used (more than 15 This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. reections of the g-Fe fcc phase. The mean crystallite size at%). The nitrogen concentration is very low and close to the calculated from the most intense g-Fe peak is in both cases in Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. detection sensitivity limit of the XPS instruments (z1 at%), the range of 20–25 nm. Several small peaks are discernible in suggesting that there is no direct incorporation of Cr(III)–glycine the diffractograms corresponding to iron oxides36 (Fe3O4 – complexes or uncomplexed glycine molecules inside the magnetite) and chromium oxide (a-Cr2O3 – eskolaite). More- coating. As expected, the carbon content is more pronounced over, in the case of the Pt anode annealed coating, although the for the Pt anode lm (z16 at%), whereas for the Ni anode predominant crystalline structure is g-Fe, a distinct a-Fe peak is coating it reaches 4 at%. Interestingly, the ratio of hydrogen to also visible at 45 , therefore showing that the lm is a mixture of total metal atoms is quite large (above 12 at%), and slightly g-Fe and a-Fe phases. The Ni anode annealed sample, in the higher for the Pt anode derived sample. other hand, does not exhibit a marked a-Fe contribution. It is In contrast, annealed coatings exhibit lower amounts of both worth mentioning that all diffractograms present a high back- C and H in-depth of the lm at the expenses of oxygen and ground at low angles, which could be attributed to the presence chromium, which at% contents increase. Fe and Ni composi- of amorphous oxides. tion variation is thought to be linked with diffusion mecha- It could be expected that the higher amount of Ni (see nisms and the elemental transformation of compounds due to Section 3.1.3) in the Ni anode samples would cause g-Fe to be the high temperatures involved in the annealing process. the predominant phase in the electrodeposited lms (nickel Moreover, the composition is not perfectly homogeneous along retains austenitic phase in stainless steel27). However, from one the thickness of the coatings, especially in the case of the Ni of our previous studies,30 the microstructure of FeCrNi lms was anode (variation of Cr peaks intensity and oxidation states in found not to be inuenced by Cr and Ni concentrations. Fig. 3b and d). The carbon content is reduced by approx. three Therefore, the microstructure diversity is anode-dependant and times with respect to the as-deposited counterparts and the most probably related to impurity variations. hydrogen-to-metals ratio is below 4 at% for the Pt anode lm 3.1.3. Elemental composition, oxidation states and 3D and undetectable for the Ni anode sample. Thus, it appears that atom-by-atom reconstruction. XRF was used to estimate the the majority of C and H incorporated within the coatings are composition in wt% and current efficiency of the FeCrNi elec- weakly bonded and diffuse and outgas upon annealing. More- trodeposited coatings. When a platinum anode was used (Pt over, not only the large carbon concentration, but also the high anode case), the composition of the resulting coating was Fe57– amount of incorporated hydrogen is responsible for the amor- Cr28–Ni15 (6.2% current efficiency, C.E.), whereas Fe52–Cr26– phous character of as-deposited FeCrNi lms. Ni22 (C.E. 8.3%) lms were obtained from using a nickel anode Detailed information about the oxidation states of the (Ni anode case). The higher nickel content in the latter may be different elements from XPS spectra can be found in the ESI.† caused by higher amounts of nickel ions being released in the Overall, the tting results of both the Pt anode and Ni anode as- electrolyte during anode oxidation throughout the deposition deposited coatings are very similar in terms of binding energies process. (BE) of the deconvoluted bands. The surface is mainly charac- The elemental compositions in at% for all samples (Pt anode, terised by carbon and oxygen contribution, with little amounts Ni anode: as-deposited, annealed) including both metallic of chromium oxide and hydroxide (respectively, 2p3/2 BEs at species (Fe, Cr and Ni) and impurities (O, C and N) were eval- 576.2 eV and 577 eV; Fig. 3a surface: 0 min sputtering time), as uated by means of XPS and are depicted in Table 2, both outer well as traces of iron in the oxide phase in the Pt anode lm. surface and in-depth (achieved by sputtering for overall 50 min) Similar results were obtained for the Ni anode lm (little Table 2 Elemental composition of Fe–Cr–Ni–O–C–N (XPS) and of H (He-ERDA) of as-deposited and annealed electrodeposited FeCrNi obtained from platinum and nickel anodes, analysed both on the surface and in-depth of the coating Fe (at%) Cr (at%) Ni (at%) O (at%) C (at%) N (at%) H (at%)/metals Pt anode as-deposited Surface 3.2 2.7 0.2 36.4 56.2 1.3 — In-depth 37.8 15.8 11.7 16.8 16.4 1.5 16.0 Ni anode as-deposited Surface 10.2 2.6 3.2 49.7 33.3 1.0 — In-depth 35.2 15.5 21.0 22.2 4.2 1.9 12.0 Pt anode annealed In-depth 36.0 23.3 10.3 22.0 6.5 1.9 3.7 Ni anode annealed In-depth 26.5 24.7 14.6 32.6 0.8 0.8 0.1 This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25767 View Article Online RSC Advances Paper This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. Fig. 3 Cr 2p XPS spectra in-depth profiling (0 to 70 min sputtering time) of electrodeposited FeCrNi films with corresponding oxidation states composition graphs: Pt anode (a) as-deposited and (b) annealed cases, Ni anode (c) as-deposited and (d) annealed ones. Secondary plane at 573.4 eV corresponds to chromium metallic peak. chromium oxide; Fig. 3b surface: 0 min sputtering time), For some elements, annealing led to a variation in the however, with a more intense iron oxide contribution with oxidation states as a function of the coatings' depth. For the Ni respect to the previous case. The elemental evolution from anode lm, if the nickel content is not affected by thermal surface to depth shows that Fe and Ni are mainly in metallic treatment (likewise in the Pt anode sample), then iron exhibits state, regardless of anode type. Interestingly, the clear carbon C a rise in the oxide to metallic ratio. Aer annealing, chromium 1s peak at around 282.5 eV could be associated with chromium shows an increase in total contribution (oxide plus carbide-like) carbide (e.g. Cr3C2).37 However, it is rather difficult to distin- with respect to the metallic one (Fig. 3c and d in-depth) for both guish between the oxides, carbides and nitrides from the anode cases. Additionally, when observing other signals (C 1s chromium tting, as their spectral lines are close together. For and O 1s), the thermally treated lms show a decrease in this reason and as the XRD diffractograms do not show any carbide-like state in carbon from 12.5 to 6.0 at% for the Pt anode presence of carbides, it seems more accurate to refer to the coating and from 3.9 to 0.5 at% for the Ni anode one. Simulta- tted peaks as chromium carbide-like (Cr–C) for chromium and neously, an increase in Cr oxide contribution in oxygen is carbide-like for carbon. Nevertheless, chromium has a rather observed from 12.4 to 16.7 at% for the Pt anode lm and from constant distribution along the depth of the coating for the Pt 15.8 to 26.1 at% for the Ni anode sample (Table S1 in ESI†). anode lm (Fig. 3a in-depth: 20 to 70 min sputtering time), i.e. These results show that annealing in general enhances a larger metallic phase with respect to both oxide and carbide- chromium oxide transformation, due to the rather large like contributions. For the Ni anode coating (Fig. 3b in-depth: amount of oxygen present in the coating, especially in the Ni 20 to 70 min sputtering time) the peaks' intensity for Cr anode case. metallic is lower compared to the overall oxidation state Additionally, atom probe tomography (APT) was performed, contribution (Cr2+/3+). resulting in 3D elemental reconstructions of nanometre tips for 25768 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. Fig. 4 APT results of the FeCrNi tip sample from the Pt anode as-deposited coating: (a) 3D elemental reconstruction, (b) isosurface rendering of Cr–O and CO2H2 compounds and (c) atom maps for different elements/compounds. electrodeposited FeCrNi lms. Both the as-deposited Pt anode distributed within the material (dark-gold colour) correspond- (Fig. 4a) and Ni anode (Fig. 5a) samples seem to have a similar ing to a mass-over-charge peak identied and labelled as CO2H2 homogenous elemental distribution. Statistical analysis of the (more generally a carboxyl molecule). Based on the atom maps collected data produced some clustering, as depicted in the (Fig. 4c and 5c), it appears that all the analysed elements have renderings (Fig. 4b and 5b). The magenta-coloured clusters no preferential distribution along the tips (except for the top (30% concentrated isosurfaces) are most probably chromium most part in the Pt anode as-deposited lm). oxide (termed Cr–O). Except for a large cluster on the top-most The data from FeCrNi annealed lms gives a more complex part of the Pt anode as-deposited case (passive oxide layer at the reconstruction (Fig. 6a and 7a) than from as-deposited ones. surface of the lm), these Cr–O clusters are infrequent and not Isosurface renderings from clustering analysis (Fig. 6b and 7b) uniformly distributed within both case tips. Instead, low show that Cr–O regions are phase separating (30% concentrated concentration (2% isosurfaces) precipitates are heterogeneously isosurfaces) in a convoluted pattern inside the material. These Fig. 5 APT results of the FeCrNi tip sample from the Ni anode as-deposited coating: (a) 3D elemental reconstruction, (b) isosurface rendering of Cr–O and CO2H2 compounds and (c) atom maps for different elements/compounds. This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25769 View Article Online RSC Advances Paper This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. Fig. 6APT results of the FeCrNi tip sample from the Pt anode annealed film: (a) 3D elemental reconstruction, (b) isosurface rendering of Cr–O and CO2H2 compounds and (c) atom maps for different elements/compounds. Cr–O clusters have a chemical microstructure region size of rich zones (mainly for the Pt anode annealed coating), whereas around 20–25 nm. CO2H2 precipitates are also present in both the latter is present rather uniformly inside the tips. annealed samples with a slightly higher concentration prole (5% isosurfaces) and are associated with the highly dense Cr–O regions. Atom maps for single elements/compounds (Fig. 6c 3.2. Investigation of Cr(III) complexation and 7c) show that some elements are correlated with each other. 3.2.1. Surface morphology. The surface of the studied C and O rich regions follow Cr rich one in the Pt anode lm. samples was mainly semi-bright grey for the Standard and No Cr Instead, the annealed Ni anode sample has an overall lower electrolytes. Conversely, the coating produced from the No Cr– carbon concentration with respect to the Pt anode counterpart Glycine bath was bright and light-grey. and it seems not to have a preferential spatial arrangement. In From both optical and scanning electron microscopies, no both cases, CO2H2 is present in close proximity to Cr dense morphological dissimilarities were observed among the lms as areas and at the boundaries of Cr–O clusters. Ni and Fe are not depicted in SEM top-view images for the Standard (Fig. 8a and segregating, the former mainly distributed outside of the Cr–O b), No Cr–Glycine (Fig. 8c), No Cr (Fig. 8d) coatings. It should be Fig. 7 APT results of the FeCrNi tip sample from the Ni anode annealed coating: (a) 3D elemental reconstruction, (b) isosurface rendering of Cr– O and CO2H2 compounds and (c) atom maps for different elements/compounds. 25770 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. Fig. 8 Surface morphology of electrodeposited FeCrNi films from the Standard electrolyte with current density of (a) 60 mA cm2 and (b) 80 mA cm2, from the (c) No Cr–Glycine bath at 80 mA cm2 and from the (d) No Cr bath at 80 mA cm2. noted that differently to the coatings described in Section 3.1, without chromium (No Cr bath), shows a crystalline structure (‘Anode role investigation’), no cracks were observed. In fact, in similar to ultrane-grained FeCrNi lms (Fig. 9b), with a minor these experiments, a lower total deposited charge was used (i.e. peak at 2q z 45 corresponding to a-Fe phase and a calculated lower deposition time). Such behaviour is consistent with the crystallite size of approx. 10 nm. previously mentioned crack formation mechanism. 3.2.3. Elemental composition, oxidation states and speci- 3.2.2. Crystal structure. The crystalline structure studied by ation. XRF measurements give the Fe–Cr–Ni composition means of ex situ Bragg–Brentano XRD measurements of FeCrNi in wt% and current efficiency of the electrodeposited samples electrodeposits is depicted in Fig. 9a for the No Cr–Glycine lm depending on the electrolytes: the Standard bath samples for and in Fig. 9b for the No Cr sample. The visible peaks at 2q of 60 and 80 mA cm2 current densities are Fe65–Cr17–Ni18 38 , 69 and 82 are linked to fcc Au (111), Si (100) and fcc Au (C.E. 3.6%) and Fe58–Cr28–Ni14 (C.E. 6.9%), respectively. The (222) crystal planes (substrate contributions), respectively. No Cr–Glycine bath coating is Fe75–Ni25 (C.E. 34.7%) and, the The diffractograms for the Standard electrolyte samples are No Cr bath lm is Fe64–Ni36 (C.E. 3.5%). the same as for the as-deposited Pt anode coating in Section 3.1: The compositions in at% of Fe–Cr–Ni–O–C–N for the studied the lack of peaks or a very broad hump at approx. 45 demon- electrodeposits from various electrolytes were obtained via XPS strates that the material is amorphous with a tendency to form in-depth of the coating (50 min sputtering time) (values listed a-Fe phases. When neither chromium nor glycine are present in on Table 3). In the same table, H atomic percentage with respect the electrolyte (No Cr–Glycine solution), therefore without Cr to the total metal atoms is given from He-ERDA analysis in- content in the nal coating, the XRD plot (Fig. 9a) clearly shows depth (20 min sputtering time). that the material is g-Fe (2q peaks at 44 , 52 , 76 and 92 ) with The Standard electrolyte coatings show that increasing the an estimated mean crystallite size of 16 nm. Interestingly, the current density (therefore the overpotential) from 60 mA cm2 coating obtained from the electrolyte containing glycine and to 80 mA cm2 causes an increase in Cr content from 12.6 at% Fig. 9 BB-XRD measurements of FeCrNi electrodeposits from the (a) No Cr–Glycine and (b) No Cr electrolytes. This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25771 View Article Online RSC Advances Paper Table 3 Elemental composition of Fe–Cr–Ni–O–C–N (XPS) and of H (He-ERDA) of as-deposited FeCrNi electrodeposits obtained from the Standard, No Cr–Glycine and No Cr electrolytes and analysed in-depth i (mA cm2) Fe (at%) Cr (at%) Ni (at%) O (at%) C (at%) N (at%) H (at%)/metals Standard electrolyte 60 40.9 12.7 18.1 16.2 11.2 0.9 13.8 80 41.3 17.1 11.9 17.2 11.1 1.4 18.2 No Cr–Glycine electrolyte 80 28.5 0.0 68.8 2.1 0.6 0.0 0.6 No Cr electrolyte 80 42.7 0.0 39.1 15.0 2.6 0.6 10.4 This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. to 17.1 at%, similarly as measured by XRF. This is in agreement the latter can be related to the amorphisation process, similarly with previous results,30 where for high overpotentials, chro- as mentioned in a previous work from our group26 on FeCrNi mium ions are in the mixed mass/charge-transfer region coatings: glycine oxidises at the platinum anode, producing differently to Ni and Fe ions, which are diffusion limited. It mainly formaldehyde and formic acid, which are sources of should be noted that there are no contents variations for all the carbon atoms incorporated interstitially within the lm, leading non-metallic elements with changes in current density, except to an amorphous microstructure. However, the deposition for hydrogen, where the ratio to total metal species increased mechanism of both Cr and impurities in that work was not clear. from 13.8 at% to 18.2 at%, suggesting there is a correlation In the present research, carbon XPS spectra show a peak in between chromium deposition and hydrogen inclusion. the C 1s signal, which can be associated with chromium Furthermore, removing both chromium and glycine from the carbide. However, as the peaks in the Cr 2p signal corre- electrolyte (No Cr–Glycine bath) cause that no impurities are sponding to oxides, carbides and nitrides are near each other, it incorporated in the electrodeposits (C, N and H below 1 at%), is impossible to completely rule out other possible contribu- oxygen presence is very low (z2 at%) and that there is more tions. A photo-emission study of the interaction of chromium nickel than iron. By removing chromium and keeping glycine in with polymers containing carbonyl groups (i.e. C]O) states the the electrolyte (No Cr bath), the composition of the obtained formation of Cr–C carbide-like bonds, with a visible peak at lm completely changes: iron content is again higher than 283 eV in the C 1s signal,38 similarly as in the outcome of this nickel, oxygen and hydrogen (z10 at%) return to high atomic work. For this reason, carbide-like terminology has been used. percentages similar to the Standard electrolyte coating, and the Furthermore, the chromium to carbon stoichiometric ratio carbon concentration is no longer negligible (z3 at%). (Cr : C) was evaluated for the carbide-like tted contributions From the oxidation point of view, detailed descriptions are (see ESI†). In the as-deposited coatings, the high amount of present in the ESI.† Nevertheless, the Standard electrolyte cases carbon atoms linked to chromium (e.g. Cr : C 1 : 3 for the Pt described in this section are identical to the Pt anode lm anode sample) is not consistent with the presence of carbides described in Section 3.1.3. The increase in current density (i.e. Cr3C2, Cr7C3, Cr23C6). This fact seems correlated with mainly leads to a rise in the metallic chromium contribution. By carbon and carbon moieties (i.e. carboxyl groups) being in comparing No Cr–Glycine and No Cr electrolyte coatings, it supersaturated state, probably bonded to Cr and to other could be concluded that the addition of glycine to the bath is molecules such as Cr oxide. Aer thermal annealing, the Cr : C linked to an increase in both Fe metallic and oxide contribu- stoichiometric ratio increases in both anode cases, however, tions, together with a signicant decrease in Ni content. XRD diffractograms and APT results do not show any evidence To better understand the possible reactions taking place towards the presence of carbides. inside the electrolyte, as well as the speciation/complexation of From comparing as-deposited and annealed coatings, it was metal ions obtained from both chemical equilibrium diagrams observed that C and H contents decreased aer thermal treat- and UV-vis absorbance spectra, additional literature on acid ment in all samples. This strengthens the concept that some of dissociation and formation/stability constants has been the carbon within the coating is linked directly to hydrogen, reviewed (see ESI†). possibly in a carboxyl manner. The presence of Fe and Cr oxides aer annealing is conrmed by both XRD and XPS. Diffractograms show a low 4. Discussion signal-to-noise ratio, associated with possible uorescence due to the presence of Fe within the coatings. The Fe and Cr oxide In this work, a comparison among FeCrNi coatings electro- phases are most probably present in amorphous state due to the deposited from a Cr(III)–glycine bath was pursued by using inert high background at low angles. FeCrNi electrodeposits can be platinum and pure nickel anodes. The use of a Ni anode reduces seen as either completely amorphous (as-deposited) or nano- the amount of carbon incorporation, conrming that glycine crystalline with an oxide amorphous background (aer reactions at the anode are hindered by nickel preferential annealing). oxidation to Ni(II) ions. Additionally, from APT measurements, results of as- In terms of microstructure of as-deposited coatings, the Pt deposited coatings suggest that all the elements in the lm anode sample is amorphous whereas the Ni anode coating is are randomly space-distributed (identical results for different ultrane-grained. The three-fold decrease in carbon content in 25772 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances analysed sections of the lm), with the exception of carboxyl (No Cr electrolyte) is characterised by a low C.E. (less than 4%), molecules (labelled as CO2H2) which seem to be directly is rich in O and H, but contains a low amount of C. In the rst incorporated within the lm as a precipitate. Depending on the case (NiFe lm), the higher amount of Ni is due to a much larger used anode, the two studied coatings have very similar APT Ni ions concentration within the bath with respect to Fe (Ni : Fe outcomes, except for the lower carbon concentration for the Ni molar ratio > 5), even though iron–nickel is a well-known anode lm. APT reconstructions of the annealed samples illus- anomalous co-deposition system. However, the main result is trate more interesting details about the FeCrNi electrodeposits the inuence of uncomplexed glycine inside the bath. Glycine at the nano-scale, which were not possible to observe from did not seem directly responsible for impurity incorporation. microscopic XPS bulk measurements. Cr–O isosurfaces show However, its presence or, more probably, the presence of its by- This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. a complex chromium oxide pattern, which has a chemical products (i.e. formaldehyde, formic acid and/or carboxyl group Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. region size in the same range as the estimated mean crystallite molecules) clearly favours metal hydroxides formation and size using XRD. These Cr–O isosurfaces are associated with incorporation. Moreover, competing hydroxides/hydrides carboxyl precipitates (also detected in the as-deposited lms). reduction processes decrease the efficiency of FeCrNi alloy Based on atom maps, there is a clear correlation between C, O electrodeposition. XRD results show that without glycine in the and Cr rich regions, as well as Cr–O clusters. In terms of bath (NiFe coating without impurities) the lm is nanocrystal- elemental uniformity of the annealed coatings, APT data depict line, and by adding glycine into the electrolyte (FeNi coating that homogeneity is not maintained at the nano-scale. However, with high hydrogen and oxygen contents and low amount of analysis performed at different sections of the same analysed carbon) the material tends to become amorphous again. sample always show identical distributions, demonstrating that Atomic weights in terms of oxidation state (Table S2 in ESI†) the annealed coating is reasonably uniform at the micro-scale. depicts the role of glycine as an intermediator for impurity In summary, as a result of anode investigations via XRD, XPS incorporation. By comparing the resulting coatings from the No and APT, it can be stated that using an inert platinum anode Cr and No Cr–Glycine electrolytes, it can be seen that the addi- leads to the oxidation of glycine, probably into formaldehyde tion of glycine signicantly increases the Fe metallic contribu- (CH2O), next formic acid (HCOOH), followed by the direct tion, while simultaneously increase the H and O contents incorporation of a certain amount of carboxyl molecules into within the produced lm. the coating. It was observed that it is possible to avoid/decrease Based on the literature and on the results from speciation/ glycine-anode reactions by using a nickel anode. This was complexation studies (see ESI†), Cr(III) is complexed with conrmed by surface morphology variations between the glycine in the FeCrNi Standard bath as veried by UV-vis produced coatings. Elemental analyses depict that carbon and absorbance spectra. Moreover, in the absence of Cr(III) and carbon-based moieties contents decrease when using a nickel glycine in the electrolyte (No Cr–Glycine case), Fe and Ni ions are anode. Moreover, heat treatment leads to different processes. almost certainly following the anomalous co-deposition via Firstly, the diffusion of impurities, i.e. carbon, oxygen and hydroxides mechanisms.18,19 When glycine is added (No Cr carboxyl precipitates, within the deposit (probably directly electrolyte), the most probable and stable glycine complex to be incorporated during the plating) distribute in close proximity to present in the bulk electrolyte is Ni(II)–glycine (conrmed by chromium/chromium oxide rich regions at grain boundaries. chemical equilibrium diagrams). However, due to high local pH Secondly, annealing seems to be also responsible for the out- near the cathode surface, the presence of carboxylate and gassing of carboxyl and –H compounds weakly bonded to hydroxo ions can strongly affect the metals' complexation. Fe(II)- chromium. However, this does not fully explain the chemical complex molecules possess a greater tendency to be penetrated reactions and the mechanisms responsible for the amorphous due to their lower stability with respect to Ni(II)–glycine mole- structure of electrodeposited FeCrNi. cules. This explains the increase in Fe deposition with respect to The study of Cr(III)–glycine complexation shows that both Ni even when the Ni : Fe molar ratio is greater than 5. The Cr(III) and glycine electrolyte addition strongly inuence the amorphous/ultrane-grained microstructure may be linked to composition variation and microstructural changes of the the high hydrogen, oxygen and carbon contents brought about electrodeposited FeCrNi lms. Based on the collected XPS data, by the co-deposition of carboxylate–hydroxo complexes. Addi- it can be seen that an increase in Cr content (Standard electro- tionally, it seems that when Cr(III)–glycine complexes are lytes) is linked to higher hydrogen incorporation within the present in the bath (Standard case), the correlation between the coatings, with no other metal or impurity (C and N) content increase in chromium concentration, as well as hydrogen and variations. This demonstrates that the reduction of complexed carbon contents can only be explained by an incomplete Cr(III) is inuenced by a hydroxides and/or hydrides deposition reduction at the cathode. The carboxylate groups are still process4,29 and probably without direct glycine incorporation, attached to chromium during the electroreduction process, due to low nitrogen content. XRD diffractograms show that the leading to the presence of metalorganic or organometallic samples are amorphous, as expected from the rst part of this compounds within the coatings. study. By comparing the two chromium-free baths, it can be Although further investigations, including in situ character- observed that the absence of glycine (No Cr–Glycine electrolyte) isation, would be necessary to learn more about this system, ex leads to an impurity-free NiFe coating with very low O content situ analysis presented in this work and previously reported and a high current efficiency (C.E. z 35%). The FeNi electro- information on the studied Cr(III)–glycine based electrolyte, deposited lm obtained from the electrolyte containing glycine allow to propose the following reactions (Fig. 10): This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25773 View Article Online RSC Advances Paper turn, the crystallinity. Amorphous and signicantly contami- nated (high C and H contents) coatings were obtained when using an inert platinum anode, whereas using a nickel anode provided ultrane-grained lms with a lower impurity content. Annealing at 600  C under an Ar controlled atmosphere produced nanocrystalline g-Fe coatings (estimated mean crys- tallite size z 20–25 nm) with substantially decreased carbon and hydrogen contents, showing that these elements are weakly bonded within the material and released during thermal pro- This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. cessing. Additional analysis by varying the bath composition Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. depicts that both Cr(III) and glycine are responsible for the large quantity of impurities. First, glycine oxidises at the anode and forms carboxyl group molecules, such as formic acid. These by- products, together with a local pH increase at the cathode, are the major factors for side-reactions, direct carboxyl moieties inclusion and lower deposition efficiency. Moreover, substan- Fig. 10 Proposed mechanisms in Cr(III)–glycine based FeCrNi elec- tial impurity incorporation (C 11 at% and H 18 at%) in the trodeposition: (top) anodic/cathodic reactions and (bottom) illustra- presence of Cr(III)–glycine complexes suggests that this elec- tion of microstructural variations due to annealing. trodeposition process can undergo:  An incomplete metal–complex reduction, leading to carboxyl molecules weakly bonded to chromium and - Anode glycine oxidation: in the presence of an inert anode, chromium oxide present within the lm, the oxidation of uncomplexed glycine (rst into formaldehyde  Cr–hydroxides/hydrides: formation, adsorption on the and later formic acid) becomes relevant. cathode's surface and inclusion within the coating. - Cathodic reactions of Cr(III): Impurities from these processes are the main contributor to  Cr(III)–glycine to metallic chromium: a two-step mech- lm amorphisation. anism with the release of glycine anions. These Complexation/speciation studies conrm Cr is complexed uncomplexed glycine molecules are free in the electro- with glycine. Additionally, both Ni(II)–glycine stability and the lyte and can react at the anode or with other metal ions. tendency of Fe(II) to form carboxyl/hydroxo complexes,  Cr(III)–glycine hydroxides/hydrides: the presence of enhances anomalous Fe–Ni co-deposition. strong hydrogen evolution and high local pH at the Cr(III)–glycine FeCrNi electroreduction mechanisms were cathode favours the formation of adsorbed Cr- also veried by comparing the as-deposited and thermally complex-OH (via hydrolysis reaction) and/or Cr- treated samples using XPS and APT. The electrodeposits were complex-H molecules at the surface. a mixture of Fe, Cr, Ni, as well as large quantities of Cr oxide,  Cr(III)–glycine incomplete deposition: based on anneal- hydrogen, carbon and carbon-based impurities. Films obtained ing results, the drastic decrease in C and H shows that using a nickel anode exhibited lower carbon and carboxyl these two elements are weakly bonded to Cr or CrOx moieties contents than when using a platinum anode. Further (most probably via a carboxyl group) within the coating. annealing demonstrated that CrOx preferentially segregated at More precisely, incomplete glycine molecules form the grain boundaries and was associated to carbon-based weak carbide-like and nitride-like bonds with Cr mole- precipitates. cules. High temperatures destabilise both weakly This study provides guidelines on future electrodeposition of bonded carboxyl and –H group molecules, which are FeCrNi coatings by using a ‘green’ Cr(III) electrolyte, paving the then released from the Cr compound and free to outgas way for a cost-effective, versatile and non-toxic chromium-based from the lm. deposition technique, which could be used for many advanced - Carboxyl group incorporation: direct inclusion from the scopes (e.g. stainless steel-like FeCrNi micro-nanocomponents electrolyte as precipitate. for bio-medical applications). - Cathodic reactions of Fe(II) and Ni(II): anomalous co- deposition is enhanced due to the high stability of complexed Conflicts of interest Ni(II)–glycine and to the fact that Fe(II) ions have a higher tendency to be complexed by carboxylate/hydroxo ions. The authors declare no conicts of interest. 5. Conclusions Acknowledgements The electrodeposition of FeCrNi from a ‘green’ Cr(III)–glycine This work was partially supported by the European project based electrolyte was investigated. Electroreduction mecha- SELECTA-ETN (H2020-MSCA-ITN-2014 no. 642642) and the e- nisms were proposed correlating microstructural and compo- MINDS COST Action (no. MP1407) from the EU Framework sition variations. The anode affects the impurity content and, in Program Horizon 2020. Partial funding from the 2017-SGR-292 25774 | RSC Adv., 2019, 9, 25762–25775 This journal is © The Royal Society of Chemistry 2019 View Article Online Paper RSC Advances project from the Generalitat de Catalunya, and the MAT2017- 16 H. Hermawan, D. Ramdan and J. R. P. Djuansjah, in 86357-C3-1-R project (co-nanced by the Fondo Europeo de Biomedical engineering-from theory to applications, InTech, Desarrollo Regional, FEDER) from the Spanish Ministerio de 2011. Economı́a y Competitividad (MINECO) is also acknowledged. Dr 17 M. Niinomi, Metals for biomedical devices, CRC Press, 2010. Eva Pellicer is grateful to MINECO for the “Ramon y Cajal” 18 H. Dahms and I. M. Croll, J. Electrochem. Soc., 1965, 112, 771. contract (RYC-2012-10839). The authors thank Centres Cientı́f- 19 W. C. Grande and J. B. Talbot, J. Electrochem. Soc., 1993, 140, ics i Tecnològics de la Universitat de Barcelona (CCiTUB) – UB, 669–681. Laboratory for Nanoscale Materials Science – Empa Dübendorf 20 L. Philippe, C. Heiss and J. Michler, Chem. Mater., 2008, 20, and Molecular and Hybrid Materials Characterization Center 3377–3384. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. (MHMC) – EPFL for the XPS measurements and Dr Max Döbeli 21 M. R. El-Sharif, A. Watson and C. U. Chisholm, Trans. Inst. from Laboratory for Ion Beam Physics – ETH Zürich for ERDA Open Access Article. Published on 16 August 2019. Downloaded on 7/27/2021 3:45:19 PM. Met. Finish., 1988, 66, 34–40. characterisation. Additionally, we are grateful to Dr Albert Serrá 22 M. Willibald and G. M. F. Mahm El, Electro-depositing Ramos (albert.serraramos@empa.ch) of Empa Thun for fruitful stainless steel coatings on metal surfaces, US Pat., discussions on chemical equilibrium diagrams and electro- US3093556A, 1963. reduction mechanisms, to Dr Xavier Maeder (xavier.maeder@- 23 M. J. Law, Electrodeposition of chromium and chromium empa.ch) of Empa Thun for signicant support on XRD bearing alloys, US Pat., US4610763A, 1986. analysis, as well as Caroline Hain (caroline.hain@empa.ch) of 24 P. Polinski and J. Paulina, Stainless steel electrolytic coating, Empa Thun for English corrections. US Pat., US20060257683A1, 2003. 25 A. G. Dolati, M. Ghorbani and A. Afshar, Surf. Coat. Technol., References 2003, 166, 105–110. 26 M. Hasegawa, S. Yoon, G. Guillonneau, Y. Zhang, C. Frantz, 1 P. Morisset, J. W. Oswald, C. R. Draper, R. Pinner and C. Niederberger, A. Weidenkaff, J. Michler and L. Philippe, R. A. Ehrhardt, J. Electrochem. Soc., 1955, 102, 143C–144C. Phys. Chem. Chem. Phys., 2014, 16, 26375–26384. 2 V. S. Protsenko and F. I. Danilov, Clean Technol. Environ. 27 G. E. Totten and M. A. H. Howes, Steel Heat Treatment Policy, 2014, 16, 1201–1206. Handbook, CRC Press, 1997. 3 F. I. Danilov and V. S. Protsenko, Prot. Met., 2001, 37, 223– 28 S. Ghaziof, M. A. Golozar and K. Raeissi, J. Alloys Compd., 228. 2010, 496, 164–168. 4 V. Protsenko and F. Danilov, Electrochim. Acta, 2009, 54, 29 D. R. Gabe, J. Appl. Electrochem., 1997, 27, 908–915. 5666–5672. 30 E. Bertero, M. Hasegawa, S. Staubli, E. Pellicer, 5 Y. B. Song and D.-T. Chin, Electrochim. Acta, 2002, 48, 349– I. K. Hermann, J. Sort, S. Mischler, J. Michler and 356. L. Philippe, Surf. Coat. Technol., 2018, 349, 745–751. 6 S. Surviliene, O. Nivinskiene, A. Češuniene and A. Selskis, J. 31 S. Gražulis, D. Chateigner, R. T. Downs, A. F. T. Yokochi, Appl. Electrochem., 2006, 36, 649–654. M. Quirós, L. Lutterotti, E. Manakova, J. Butkus, P. Moeck, 7 A. Baral and R. Engelken, J. Electrochem. Soc., 2005, 152, A. Le Bail and IUCr, J. Appl. Crystallogr., 2009, 42, 726–729. C504–C512. 32 N. Fairley, CasaXPS: Spectrum Processing Soware for XPS, 8 N. V. Mandich, Plat. Surf. Finish., 1997, 84, 108–115. AES and SIMS, ©CasaXPS Ltd., 2005, vol. 2. 9 J. McDougall, M. El-Sharif and S. Ma, J. Appl. Electrochem., 33 M. Nastasi, J. W. Mayer and Y. Wang, Ion beam analysis: 1998, 28, 929–934. fundamentals and applications, CRC Press, 2014. 10 O. V. Safonova, L. N. Vykhodtseva, N. A. Polyakov, 34 K. P. Rice, Y. Chen, R. M. Ulg, D. Lenz, J. Bunton, M. van J. C. Swarbrick, M. Sikora, P. Glatzel and V. A. Safonov, Dyke and D. J. Larson, Microsc. Microanal., 2017, 23, 42–43. Electrochim. Acta, 2010, 56, 145–153. 35 I. Puigdomenech, HYDRA (Hydrochemical Equilibrium- 11 V. S. Protsenko, V. O. Gordiienko and F. I. Danilov, Constant Database) and MEDUSA (Make Equilibrium Electrochem. Commun., 2012, 17, 85–87. Diagrams Using Sophisticated Algorithms) soware, Royal 12 J. H. O. J. Wijenberg, M. Steegh, M. P. Aarnts, K. R. Lammers Institute of Technology, Stockholm, 2006, http:// and J. M. C. Mol, Electrochim. Acta, 2015, 173, 819–826. www.kemi.kth.se/medusa/. 13 O. Ergeneman, K. M. Sivaraman, S. Pané, E. Pellicer, 36 F. S. Yen, W. C. Chen, J. M. Yang and C. T. Hong, Nano Lett., A. Teleki, A. M. Hirt, M. D. Baró and B. J. Nelson, 2002, 2, 245–252. Electrochim. Acta, 2011, 56, 1399–1408. 37 K. Petkov, V. Krastev and T. Marinova, Surf. Interface Anal., 14 N. V. Myung, D. Y. Park, M. Schwartz, K. Nobe, H. Yang, 1992, 18, 487–490. C. K. Yang and J. W. Judy, Proc. - Electrochem. Soc., 2000, 38 J. L. Jordan, C. A. Kovac, J. F. Morar and R. A. Pollak, Phys. 20, 29. Rev. B: Condens. Matter Mater. Phys., 1987, 36, 1369–1377. 15 A. Brenner, Electrodeposition of alloys: principles and practice, Academic Press, 1963. This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 25762–25775 | 25775