Análisis de la viabilidad del magnetron sputtering como técnica de fabricación de catalizadores alternativa

Análisis de la viabilidad del magnetron sputtering como técnica de fabricación de catalizadores alternativa Antía Villamayor Dono Directoras de tesis: Dra. Victoria Laura Barrio Cagigal Dra. Eva Gutiérrez Berasategui Bilbao, noviembre 2024 (cc) 2024 Antía Villamayor Dono (cc by-nc 4.0) i Agradecimientos Hace ya cinco años que empecé este camino largo y tortuoso pero muy enriquecedor que por fin llega a su final. En primer lugar, me gustaría darles las gracias a mis directoras Laura y Eva, por guiarme, escucharme y ayudarme siempre que lo he necesitado. A Tekniker por brindarme todos los recursos necesarios para poder formarme y hacer esta Tesis. A mis compañeras de trabajo de Tekniker, siempre dispuestas a colaborar y arrimar el hombro. A Lucía, por su apoyo y energía incansable. A Sergio, Dmitry y la gente del ICP-CSIC, por haberme aportado muchísimo conocimiento y ayuda. A mis amigas de Euskadi, Oihane, Ángela, Alonso, Iñaki, Joseba, Patri, Fer, Jago, Eider, Mikel, Ane (y todas las demás), gracias por hacerme reír, acogerme como a una más y hacerme sentir como en casa desde el principio. Especialmente, a Meri, por tus mil consejos y por ser mi apoyo incondicional durante todos estos años. Grazas a ti Gonzalo, por ser un amigo íncrible, o mellor compañeiro que podría ter e sempre estar ó meu lado. Quero agradecerlles tamén ás miñas amigas de Galiza, en especial a María, que aínda que estes lonxe é como se nunca marcharas. Tamén á xente de Barcelona, por facelo todo tan divertido. E finalmente, á miña familia, en especial a meus pais, por ser os mellores que podía ter e por todas as facilidades que me deron. Á miña irmá Mariña, que dende sempre podemos contar una ca outra, es a mellor. Moitísimas grazas a todos! Eskerrik asko guztioi! ii iii Resumen Dentro del marco actual de emergencia climática, la descarbonización de la industria de procesos cobra una especial relevancia debido al impacto de esta en el medioambiente. Un factor común a todos los procesos es el uso de catalizadores, que son fundamentales para poder llevar a cabo reacciones eficientes y así reducir la energía externa necesaria para llevar a cabo un determinado proceso. Para ello no solo es fundamental que estos catalizadores sean eficientes, además son muy importantes su coste de producción y la técnica de fabricación empleada. Con lo cual es necesario contar con una técnica de fabricación versátil, que permita elaborar estrategias de mejora, reducir el material empleado y simplificar y automatizar el proceso de síntesis. En esta Tesis Doctoral se ha estudiado el magnetron sputtering como técnica de fabricación de catalizadores, con la intención de mostrar su versatilidad y validez al desarrollar catalizadores diferentes para diversas aplicaciones. Este objetivo está alineado con la mejora de la eficiencia, la reducción de material y la optimización del proceso de síntesis. En el Capítulo 1 se incluye una visión general sobre la postura actual europea en relación a la emergencia climática y el marco legislativo para llevar a cabo la transición energética. En esta transición energética, los catalizadores juegan un papel muy relevante y por ello, necesitan ser eficientes, asequibles y de fabricación sencilla. Los métodos empleados en la actualidad fallan a la hora de cumplir esos requisitos, con lo cual, ya que la técnica de fabricación va a influir en estos tres factores, se propone como técnica alternativa el magnetron sputtering. Este método de deposición física en fase vapor (PVD), posee ciertas características iv que posibilitarían mejorar la eficiencia, reducir material, bajar costes y simplificar el proceso de síntesis de los catalizadores. En base a lo expuesto anteriormente, en el Capítulo 2 se marca el objetivo general de esta Tesis Doctoral, el análisis de la viabilidad y versatilidad del magnetron sputtering como técnica de fabricación de catalizadores alternativo a los métodos químicos actuales, enfocado a reducir la cantidad de material catalítico, mejorar la eficiencia catalítica y optimizar el proceso de fabricación. Para ello en el Capítulo 3 se estudia la mejora de la eficiencia para la degradación de azul de metileno de un fotocatalizador de TiO2 mediante su nanoestructuración y la formación de una heterounión p-n con el NiO. La incorporación de una capa superficial de NiO se tradujo en un aumento de la constante y porcentaje de degradación tanto en comparación con el TiO2 puro como con una muestra de vidrio Pilkington Activ. El porcentaje de degradación del azul de metileno disminuye después del primer ciclo catalítico, pero se mantiene estable durante otros 9, mostrando la estabilidad de este fotocatalizador. En el Capítulo 4 se establece como objetivo la reducción de la cantidad de Pt empleada como electrocatalizador para la reacción de evolución de H2 (HER) en electrólisis PEM. Se depositaron nanopartículas de Pt sobre un soporte de carbono microporoso, con tres cargar de Pt menores de 0.1 mg cm-2, las cuales se caracterizaron electroquímicamente en una celda de tres electrodos. De esta caracterización inicial, se pudo observar que en condiciones ácidas la cantidad de Pt no influye ni en la actividad catalítica ni en la estabilidad, cobrando mayor relevancia la disponibilidad de sitios activos. Estos electrodos demostraron ser altamente activos para la HER y estables tanto a corto como a largo plazo. Al evaluar en monocelda el electrodo de v Pt con carga de 0.1 mg cm-2 y compararlo con uno comercial de 0.3 mg cm-2, se pudo corroborar que la reducción de material no afecta ni a la actividad ni a la estabilidad de la celda, siendo la respuesta de voltaje constante durante 170 h. Finalmente, en el Capítulo 5 se busca demostrar la versatilidad del magnetron sputtering como técnica de fabricación de catalizadores, mediante el desarrollo de fotoánodos de BiVO4 y BiVO4/WO3 para un sistema fotoelectroquímico (PEC) de reducción de CO2 a metano. En este trabajo se probó la mejora tanto de la producción de metano como de la selectividad de la reacción con la adición de una capa de WO3. Esta capa forma una heterounión de tipo II con el BiVO4, y viéndose incrementada la movilidad electrónica y reducida la tasa de recombinación. Esta versatilidad se confirma por la posibilidad de elaborar estrategias de mejora de la eficiencia como es la formación de heterouniones y la deposición de dos materiales diferentes en un mismo proceso, para una aplicación muy novedosa y de elevado interés medioambiental. En resumen, en esta Tesis Doctoral se ha planteado el desarrollo de tres tipos de catalizadores para diferentes aplicaciones empleando el magnetron sputtering como técnica de fabricación. En los tres casos planteados se han cumplido los objetivos de mejora de eficiencia, reducción de material y optimización del proceso de síntesis, lo cual contribuye a una reducción de costes de fabricación y la versatilidad del magnetron sputtering para la fabricación de un amplio rango de catalizadores. Finalmente, cabe destacar, que siendo el magnetron sputtering una técnica fácilmente escalable, esto podría añadir un impulso extra al desarrollo de tecnologías que potencien la descarbonización de la industria de procesos. vi vii Listado de abreviaturas AEM Alkaline exchange membrane AFM Atomic force microscopy AWE Alkaline water electrolysis CB Conduction band (banda de conducción) CCE Catalyst coated electrode CCM Catalyst coated membrane CCU Carbon capture and utilisation CP Chronopotentiometry CV Cyclic voltammetry DC Cirect current ECSA Electrochemical surface area EDLP Estrategia de Descarbonización a Largo Plazo EIS Electrochemical impedance spectroscopy ETF Energy to fuel FE Faradaic efficiency FESEM Field emission scanning electron microscopy GDE Gas diffusion electrode GDL Gas diffusion layer GSA Geometric surface area 2 eficiencia energética y fijar la cuota de energías renovables al 40% [3]. Para conseguir estos objetivos, dentro del paquete de propuestas legislativas se incluye el fomento del uso de las energías renovables, mejorar la eficiencia energética en la industria, aumentar la presencia de los procesos industriales de conversión directa o el impulso en la producción y el uso del hidrógeno como sustituto de los combustibles fósiles. Sin duda, uno de los puntos clave que marca la UE para la descarbonización de la industria y de la energía, es el uso del hidrógeno, ya que permitiría una descarbonización transversal, potenciando una energía circular basada en el hidrógeno verde, lo cual se ve reflejado en la Estrategia Europea del Hidrógeno. En el marco de este Pacto Verde Europeo, se incluye La Estrategia de Descarbonización a Largo Plazo (EDLP) 2050 del Gobierno de España, en el cual se marcan los mismos objetivos que a nivel europeo para la descarbonización y reducción de emisiones de sectores como el industrial, el energético o el tratamiento de residuos (Figura 1) [4]. Figura 1. Diagrama de las principales líneas de actuación dentro de la EDLP. 3 En la EDLP también se indica como objetivo fundamental el uso de energías renovables y la potenciación de la producción de hidrógeno, siendo ambas fundamentales para una descarbonización completa del sector industrial. Figura 2. Principales magnitudes de la EDLP. El plan de emergencia climática que se ha puesto en marcha, en los últimos años en Europa, responde a una necesidad y reclamo social que urgentemente había que abordar; una transición energética que rompa con el uso de combustibles fósiles, que respete nuestro medioambiente y, que sea justa y asequible para todos los ciudadanos, para poder así parar el cambio climático y los efectos devastadores que tiene para los ecosistemas del planeta y la sociedad. 1.2. Transición energética y catálisis Dentro de este marco de transición energética, actualmente, existen tecnologías muy ligadas al uso de energías renovables, que están captando la atención por su bajo impacto ambiental. Estas tecnologías o sistemas se pueden encontrar en campos tan diversos como puede ser 4 la producción y almacenamiento de energía, la síntesis de productos químicos de valor añadido e incluso para la limpieza del aire o del agua mediante procesos de degradación de partículas contaminantes. Cabe destacar que, todos estos procesos tienen en común que siempre tienen lugar mediante una reacción química que permite que se lleven a cabo; y estas reacciones químicas, necesitan un catalizador para ocurrir de forma rápida, selectiva y eficiente. En un contexto de crisis energética y climática, la potencial capacidad de un catalizador para reducir tanto el coste energético de un proceso, como el tiempo o la reducción de productos de reacción secundarios, los convierte en un elemento clave para contribuir a la mejora de la eficiencia energética de multitud de procesos industriales [5]. La importancia de los catalizadores a la hora de contribuir a la mejora de esa eficiencia energética está reflejada en la SRIA y desde la Unión Europea se propone que estos catalizadores puedan obtener energía de forma directa de fuentes renovables, en forma de fotones y electrones, para llevar a cabo las reacciones en las que participan [6]. Dentro de este tipo de catalizadores, hay dos que generan especial interés: • Los fotocatalizadores que tienen la capacidad de absorber la luz solar y almacenarla en enlaces moleculares o llevar a cabo reacciones en su superficie. • Los electrocatalizadores que pueden llevar a cabo reacciones redox cuando se les aplica un determinado voltaje externo. Ambos se están empleando e investigando actualmente en campos como la generación y combustión de hidrógeno, la reducción de CO2 a hidrocarburos como el metano, o la fijación de nitrógeno. Sin embargo, la aplicación industrial de estos catalizadores presenta diversos retos, principalmente su costo, eficiencia y síntesis [7]. Muchos de estos catalizadores son compuestos metálicos u óxidos metálicos, algunos comunes y asequibles, como pueden ser el Fe, Ni, 5 Zn o el Cu y, otros más escasos y costosos, generalmente aquellos que se encuentran en lo que de denomina metales del grupo del platino (PGM) donde se engloban el Ir, Ru, Rh o el propio Pt [8–16]. Los procesos catalíticos que dependen de los PMGs tienen un costo elevado, con lo cual la reducción de la carga catalítica o la optimización del proceso de síntesis son dos puntos clave para reducir el coste de estos procesos. La síntesis de estos catalizadores se realiza normalmente mediante métodos de impregnación química (ejemplos en la Tabla 1), muy estudiados y que poseen múltiples ventajas, entre ellas la posibilidad de otorgar un área activa superficial muy alta a estos compuestos, lo que es fundamental para obtener catalizadores de alto rendimiento. Sin embargo, son rutas sintéticas que requieren de varios pasos, difíciles de escalar a nivel industrial y, donde se acaba perdiendo material catalítico durante el proceso. Además, el uso de múltiples reactivos da lugar a la formación de productos secundarios de reacción que, en su gran mayoría no son aprovechables y son perjudiciales para el medioambiente. El proceso de síntesis también tiene impacto en la eficiencia de los catalizadores, ya que este tiene que ser lo suficientemente versátil para poder llevar a cabo estrategias de mejora de esta eficiencia. Además, una falta de eficiencia catalítica se traduce en una limitación del rendimiento de la reacción o en la necesidad de emplear más material para no comprometer ese rendimiento. 6 Tabla 1. Métodos químicos más comunes empleados en síntesis de catalizadores. Método Características principales Impregnación En este proceso, se sumerge el sustrato en una disolución en la cual se encuentra un precursor que se adhiere al sustrato. A continuación, hay que evaporar el disolvente y, por lo general, se necesita de un paso de calcinado o activación para obtener el compuesto deseado. Síntesis hidrotermal Los reactivos se introducen dentro de un crisol que se sella. Para llevar a cabo este tipo de síntesis se necesitan temperaturas y presiones bastante elevadas. Una vez acabada la reacción, se pasa por una fase de enfriamiento y extracción del catalizador sintetizado y finalmente, se adhiera a un sustrato determinado. Sol-Gel Esta técnica implica la formación de un "sol", es decir, una disolución coloidal, la cual se somete a procesos de hidrólisis y condensación para desarrollar enlaces covalentes entre las especies disueltas. Cuando esta disolución se transforma en un gel, se realiza un paso de envejecimiento o maduración para completar el proceso de gelificación. Finalmente, se suele hacer un tratamiento térmico para eliminar todo el disolvente y, también, inducir cambios estructurales. 7 Las implicaciones ambientales, la baja eficiencia de los materiales y la dificultad de escalar la producción catalizadores pueden representar un desafío para el rápido crecimiento de la industria verde en Europa. Con lo cual es necesario la incorporación de técnicas de fabricación de catalizadores automatizables que permitan la fabricación de estos materiales a gran escala de manera competitiva, evitando la formación de productos secundarios potencialmente contaminantes. Estas técnicas de fabricación de catalizadores son críticas para la implementación a gran escala de tecnologías verdes que contribuyan a la descarbonización de la economía. 1.3. Características y mejora de la eficiencia de los catalizadores En cualquier proceso industrial que implique una reacción química el objetivo es aportar la mínima cantidad de energía externa posible y que esta reacción sea lo más rápida posible, con la finalidad de aumentar al máximo la eficiencia de este proceso. Esto se consigue mediante el uso de catalizadores. Cualquier reacción química no espontánea requiere una determinada energía de activación para que ocurra. El papel de un catalizador en una reacción es reducir esa energía de activación, acelerando el tiempo de reacción y disminuyendo así el aporte externo de energía necesario. 8 Figura 3. Diagrama del efecto de un catalizador sobre la energía de activación de una reacción [17]. Para que un catalizador sea activo y eficiente para un determinado proceso químico deben tenerse en cuenta distintos factores: • Actividad catalítica: para que un compuesto pueda catalizar una reacción debe tener una determinada estructura y/o composición para que ese sirva como centro activo para llevar a cabo la transformación química, con una energía de enlace adecuada para la fácil adsorción de los reactivos y desorción de los productos. • Selectividad: es la capacidad para obtener un producto de reacción determinado, evitando reacciones secundarias y aumentando la eficacia del proceso al producir un mayor rendimiento de un producto determinado. • Estabilidad: es fundamental que el catalizador no se desactive o degrade con el número de ciclos de reacción. 9 • Área superficial: a mayor área superficial, más sitios activos, con lo cual mayor capacidad para transformar reactivos en productos. • Condiciones de reacción: factores como la temperatura, presión, concentración de reactivos y el pH pueden afectar la actividad catalítica. Con lo cual, obtener un catalizador eficiente y estable es muy complejo, pero al mismo tiempo, se pueden explorar muchas rutas para intentar mejorar el rendimiento de este. Desde el punto de vista industrial, la mejora en el desarrollo de catalizadores podría seguir las siguientes estrategias: 1. Mejorar la eficiencia: todo lo relacionado con la capacidad propia del catalizador para llevar a cabo una reacción de forma rápida y eficiente. Entrarían todos los factores mencionados con anterioridad como la selectividad, estabilidad… 2. Optimización de los materiales empleados: en muchos casos, el catalizador idóneo para una determinada reacción es un material escaso y/o caro, siendo una gran parte del coste de un proceso industrial. Con lo cual reducir su cantidad o buscar un sustituto sin perder eficiencia es una estrategia fundamental para reducir costes. 3. Mejora de las técnicas de síntesis de catalizadores: normalmente, estos compuestos se sintetizan mediante procesos químicos que constan de muchas etapas de reacción y dan lugar a subproductos y rendimientos bajos, además de ser difícilmente escalables a nivel industrial. La búsqueda de un método de desarrollo de catalizadores que permita solventar estos tres hándicaps es esencial para poder mejorar estos 10 materiales y, a su vez los procesos químicos que estos catalizan. Como se ha comentado con anterioridad, los métodos químicos tradicionales mediante los cuales se sintetizan comúnmente los catalizadores sí que son versátiles a la hora de desarrollar nuevas estructuras o composiciones a la vez que aporten un área superficial activa elevada, pero presentan muchas limitaciones a nivel industrial. Muchos de estos nuevos catalizadores presentan grandes problemas de estabilidad a pesar de su buena actividad catalítica, y para obtenerlos las rutas de síntesis suelen ser de varios pasos de reacción, dando lugar a rendimientos muy pequeños, con lo cual la eficiencia de estas rutas sintéticas es muy baja. Esto también se traduce en una elevada pérdida de material catalítico a lo largo de la síntesis, sumándole todos los disolventes y subproductos que se forman, generando una alta cantidad de residuos perjudiciales para el medioambiente. Por último, aunque se está avanzando, con técnicas como el esprayado, su automatización y el escalado a nivel industrial, es complejo. Se requiere de muchas etapas, la síntesis de laboratorio a menudo es difícil de llevar a escalas mayores y, además, necesitan de un buen plan de gestión de residuos, ya que muchos de ellos son tóxicos. El magnetron sputtering es una técnica de deposición física mediante fase vapor que permite la evaporación de casi cualquier compuesto inorgánico y llevar a cabo procesos de deposición tanto no reactivos (metales puros, aleaciones) como reactivos (óxidos, carburos) que pueden dar lugar a compuestos de interés catalítico que se usan en un gran número de procesos. Además, es una técnica completamente automatizada que se realiza en un solo paso y sin la emisión de ningún tipo de producto secundario contaminante. El uso del magnetron sputtering como proceso de fabricación de catalizadores a gran escala tiene el potencial para impulsar excepcionalmente la digitalización, 11 automatización, reducción de costo y la eficiencia de tecnologías clave para la descarbonización de la industria de procesos. En esta Tesis Doctoral se estudia y evalúa la viabilidad del magnetron sputtering, tecnología industrial basada en plasma para la fabricación de recubrimientos, como método sintético vérsatil para la fabricación de catalizadores de bajo coste y eficientes. Estos catalizadores participan en reacciones de degradación de compuestos orgánicos disueltos en agua mediante fotocatálisis, producción de hidrógeno por electrólisis PEM o fotoelectrorredución de CO2 para obtención de hidrocarburos, todas ella tecnologías claves para la descarbonización. 1.4. Desarrollo de catalizadores por magnetron sputtering El magnetron sputtering se engloba dentro las técnicas de deposición física en fase vapor (PVD, physical vapor deposition), utilizadas para depositar películas delgadas de diversos materiales, generalmente compuestos inorgánicos, metálicos, óxidos metálicos, y nitruros sobre un sustrato, mediante la evaporación o el bombardeo de un material dentro de una cámara de deposición. Entre las técnicas de PVD que se emplean en el desarrollo de catalizadores se encuentran la evaporación térmica, la deposición con láser pulsado, además de la pulverización catódica (sputtering). El sputtering es un método de deposición de capas finas completamente automatizable, el cual empieza a suscitar mucho interés debido a sus potenciales ventajas para el desarrollo de catalizadores, como, por ejemplo: • La versatilidad del proceso de deposición permitiendo un control preciso sobre las propiedades, tanto estructurales como fisicoquímicas, del recubrimiento a depositar. • Es un proceso que se opera de forma automatizada a larga escala y a nivel industrial. 18 Aunque es una técnica poco explorada, se pueden encontrar algunos trabajos de revisión muy completos e interesantes sobre sobre las tendencias y avances actuales además de los desafíos de esta misma [30], la aplicación de esta técnica en campos como la catálisis o los sensores [31], sobre las opciones de escalado industrial existentes o también sobre modelizado para la deposición de nanopartículas mediante síntesis en fase gas [32]. Fuera de estos trabajos de revisión y ya dentro del campo de la catálisis, la aplicación de este método es aún muy reducida, principalmente debido a su novedad ya que el primer trabajo publicado sobre esta tecnología es de 1991 [27]. Dentro de los pocos casos que podemos encontrar, cabe resaltar los enfocados al desarrollo de catalizadores para la reacción de reducción de oxígeno (ORR) en pilas de combustible y para la reacción de evolución del oxígeno (OER) en electrolizadores de membrana polimérica (PEMEC), por la relevancia de su aplicación y, además, por lo novedoso del método. Estos trabajos confirman que esta técnica de síntesis de partículas es muy prometedora para el desarrollo de catalizadores de procesos químicos que pueden ayudar a la descarbonización de la industria y el sector energético. Concretamente, se han desarrollado partículas bimetálicas de PtRu combinando el magnetron sputtering con fuente de agregación de gas y la selección en función de su masa de los agregados junto con una técnica de "ion soft landing", que permite un control muy preciso sobre el tamaño, composición y estado electrónico de las nanopartículas depositadas. Estas NPs, además de tener un bajo contenido en Pt (mucho más caro que el Ru), según la caracterización electroquímica con voltamperometrías cíclicas en medio ácido con atmósfera de O2, presentaban una elevada estabilidad y actividad para la ORR. Además, confirman que el método de síntesis es reproducible ya que los electrodos para la ORR sintetizados en días distintos presentaban una actividad catalítica muy similar [33]. 19 En el caso de la OER, Zheng et al. desarrollaron electrodos de Ir-TaO mediante esta técnica acoplada con un sistema de selección de nanopartículas en función de su masa, obteniendo agregados de menos de 2 nm de diámetro y electrodos de bajo contenido en Ir comparado con los electrodos de IrO2 comerciales. En el caso de estas partículas de menos de 2 nm, la actividad másica es dos veces mayor en comparación con nanopartículas de IrO2 del mismo tamaño [34]. Con esta misma técnica, se sintetizaron electrodos con nanopartículas aisladas de Pd usando una fuente de agregación en fase gas y mediante RF sobre un soporte de Mg de diferentes espesores. Estos electrodos de caracterizaron electroquímicamente para la OER en medio básico y se estudió el mecanismo de las mismas mediante DFT desarrollando un método que hace posible estudiar el mecanismo de electrocatalizadores desarrollados por otros procesos de síntesis [35]. • RF magnetron sputtering Aunque la mayoría de los compuestos se pueden depositar usando el DC magnetron sputtering, hay algunos materiales que debido principalmente a la ausencia de conductividad no pueden evaporarse de esta forma. En estos casos se emplea una fuente de radiofrecuencia alterna (RF). La radiofrecuencia alterna rápidamente el voltaje en el blanco, lo que evita que se acumule carga estática en el blanco cuando se usan materiales aislantes, permitiendo una deposición estable y continua. Este tipo de magnetron sputtering también se ha usado ampliamente para la deposición de catalizadores empleados en tecnologías relacionadas con las energías renovables y la conversión directa de luz solar. Se ha empleado para obtener recubrimientos con buenas propiedades ópticas, como pueden ser las capas presentes en las celdas solares. Por ejemplo, se ha desarrollado una capa antirreflejante de ZrO2, ZnO y TiO2 (ZZTO) con una capacidad de absorción de luz solar 20 mejorada, superhidrofóbica con una eficiencia del 10.2% [36]. Enfocado hacia la degradación de tintes orgánicos se han desarrollado capas mixtas de WOx-SiO2 sobre un sustrato de TiO2, para mejorar su hidrofobicidad y fotoactividad [37]. Más innovador, es el uso del RF magnetron sputtering para fabricar fotoánodos en sistemas fotoelectrocatalíticos para producción de H2. Se formaron heteroestructuras de WO3/Fe2O3 para su uso como fotoánodo, con el fin de mejorar el rendimiento del Fe2O3, consiguiendo un aumento de 5 veces en la corriente generada [38]. También se desarrollaron estructuras de BiVO4 dopadas con Mo mediante RF, modificando la cantidad de O2 para poder observar cual era la cantidad óptima en relación con la densidad de corriente generada en un sistema PEC, llegando a producir 1.34 mA cm-2 vs 1.23 V vs RHE [39]. La cantidad de trabajos de investigación que emplean diferentes tipos de magnetron sputtering para desarrollar una amplia variedad de catalizadores para distintas aplicaciones, es una prueba de la gran versatilidad de esta técnica como método de fabricación de catalizadores. Aunque es cierto que el magnetron sputtering como método de fabricación de catalizadores necesita aún una adaptación desde la escala de laboratorio a la industrial, para que esto ocurra es necesario realizar trabajos y estudios previos que prueben el potencial de esta tecnología, colaborando así con su desarrollo y futura implementación a gran escala. Por ello en esta Tesis, se exploran tres clases de catalizadores para tres aplicaciones diferentes, sintetizando cada uno con distintos tipos de procesos de magnetron sputtering. Así de esta forma, estudiar si el magnetron sputtering es un método versátil, escalable, viable y prometedor para la futura fabricación a nivel industrial de catalizadores eficaces y de bajo coste. 21 22 Capítulo 2: Hipótesis y objetivos El principal objetivo de esta Tesis Doctoral es el estudio de la viabilidad y validación del magnetron sputtering como técnica de fabricación de catalizadores alternativo a los métodos químicos, con el fin de reducir el material catalítico, mejorar la eficiencia catalítica y optimizar el proceso de síntesis. Para ello, es fundamental la reducción del material catalítico, tanto el empleado durante el proceso de síntesis como la propia cantidad de catalizador depositado. Otro punto clave es poder llevar a cabo estrategias de mejora en el desarrollo de catalizadores, que aumenten la eficiencia, selectividad y estabilidad de los catalizadores. Y finalmente, la optimización del propio proceso sintético, reduciendo las etapas de este y la formación de subproductos, junto con una automatización del proceso de fabricación tanto a nivel laboratorio como industrial. La hipótesis que se plantea en este trabajo es que el magnetron sputtering tiene las características necesarias para poder sustituir a los métodos de síntesis químicos actuales. • Al ser una técnica automatizable y de un solo paso, en la cual no hay subproductos, permitiría la mejora e industrialización del proceso de fabricación de multitud catalizadores. • El elevado control y el propio método de deposición en sí evitan el uso excesivo o pérdida de material catalítico durante el proceso. • La modificación de parámetros de proceso, como pueden ser la densidad de potencia o la presión, permite modificar con precisión las propiedades de la capa catalítica y dota al proceso sintético de versatilidad, propiciando de esta forma un amplio margen de mejora de las capas depositadas. 23 Con el fin de comprobar la hipótesis planteada y cumpliendo con el objetivo principal de la Tesis, se marcaron tres objetivos técnicos parciales centrados en el desarrollo de foto y electrocatalizadores para purificación de aguas, producción de hidrógeno y producción de metano: 1. Mejorar la eficiencia de un fotocatalizador muy conocido y estudiado como es el TiO2, mediante la nanoestructuración y la modificación de la estructura electrónica de la capa catalítica. Para mejorar la eficiencia se optó por la nanoestructuración de la capa de TiO2 y la formación de una heterounión p-n con NiO. Sobre la capa de TiO2 se depositaron dos espesores distintos de NiO con el fin de evaluar el efecto del espesor. La evaluación de su actividad fotocatalítica y estabilidad se realizó mediante la degradación de tintes orgánicos, como el azul de metileno, presentes en el agua. Las fotoactividades de las capas se compararon entre ellas y frente a una muestra comercial de vidrio recubierto con TiO2 fotocatalítico depositado por métodos químicos. 2. Reducir la cantidad de material catalítico, mediante el desarrollo de catalizadores de baja cantidad de platino y evaluación de su actividad catalítica exsitu para la reacción de evolución de hidrógeno (HER) en medio ácido, junto con la validación y comparación de la actividad de un electrodo de Pt desarrollado por magnetron sputtering con un catalizador comercial de mayor contenido de Pt fabricado por métodos químicos tradicionales en una monocelda PEM. Para reducir el contenido en Pt empleado en la HER, sobre un sustrato de papel de carbono con una capa microporosa se depositaron tres cargas de nanopartículas de platino distintas, controlando la carga con el tiempo de deposición. El objetivo era 24 obtener electrodos con un contenido de Pt menor de 0.125 mgcm2, con un área superficial activa elevada. Estos tres electrodos se caracterizaron tanto física como electroquímicamente. A través de la caracterización electroquímica ex situ se pudo llevar a cabo una evaluación preliminar de la actividad catalítica de los electrodos desarrollados por sputtering y determinar la influencia de la carga catalítica en la eficacia de los electrodos para llevar a cabo la HER. Con el fin de validar uno de los electrodos desarrollados por sputtering se comparó su actividad catalítica tanto en celda de tres electrodos como en una monocelda PEM con un electrodo comercial con una carga de Pt de 0.3 mg cm-2. Para ello, se ensamblaron dos monoceldas iguales cambiando únicamente el electrodo del cátodo. Se llevaron a cabo medidas del rendimiento de la celda durante una semana con el propósito de comparar los resultados obtenidos en las dos celdas (relación V-j y degradación). 3. Probar la versatilidad del magnetron sputtering a través del Desarrollo de fotoánodos de BiVO4 y BiVO4/WO3 para su uso en sistemas fotoelectrocatalíticos (PEC) de reducción de CO2 a metano novedosos, empleando aguas grises como reactivo anódico, enfocando el uso del dispositivo en Marte como una tecnología ISRU (insitu resource utilization). Para probar la versatilidad del magnetron sputtering, se marcó como objetivo obtener un catalizador novedoso para un sistema de bajo TRL. Se llevó a cabo un estudio bibliográfico sobre los sistemas PEC para producción de hidrocarburos mediante reducción de CO2 para determinar cuáles eran las necesidades de esta tecnología desde el punto de vista del fotoánodo. Teniendo en cuenta estas necesidades, se depositaron capas de BiVO4 y BiVO4/WO3 sobre papel de carbono. Por un lado, se estudió la posible influencia de los parámetros de deposición y por otro la 25 influencia de la formación de una heterounión tipo II entre el BiVO4 y el WO3, además de la reducción del espesor del BiVO4 en estas capas. Para ello se evaluaron parámetros como la tasa de producción, la eficiencia faradaica (FE) y el porcentaje de ETF (energy to fuel). Las medidas se realizaron en una celda PEC de laboratorio, irradiada con luz del espectro visible, empleando aguas grises como reactivo para llevar a cabo la OER y CO2 humidificado para la obtención de metano. Estos tres objetivos parciales permitirán cumplir el objetivo principal de la Tesis, la validación del magnetron sputtering como método de síntesis de catalizadores, alineado con la reducción de material, la mejora de la eficiencia catalítica y la optimización de procesos de fabricación. De esta forma se pretende demostrar la versatilidad y validez del magnetron sputtering, al permitir desarrollar catalizadores con características y aplicaciones completamente diferentes, siendo estos catalizadores activos, eficaces y estables. 26 Capítulo 3: Desarrollo de recubrimientos fotocatalíticos nanoestructurados de TiO2 y NiO/TiO2 por magnetron sputtering para aplicaciones fotocatalíticas 3.1. Introducción Dentro de los muchos procesos catalíticos que existen, la fotocatálisis genera mucho interés, ya que como única fuente de energía emplea la luz solar, con lo cual es un proceso autónomo desde el punto de vista energético. Esto lo hace muy atractivo para aplicarlo tecnologías tan diversas como puede ser la conversión de energía, la electrolisis del agua, procesos de obtención de hidrocarburos mediante reducción de CO o CO2, o la purificación del agua o del aire. La fotocatálisis es un proceso a través del cual se lleva a cabo la reacción química usando los fotones de la luz como fuente de energía. Los materiales semiconductores, que poseen una estructura eléctrica de bandas, se pueden emplear como materiales fotocatalíticos, ya que poseen una banda de valencia (BV, donde están los electrones en estado fundamental) y una banda de conducción (BC, sin electrones) separadas por un bandgap, que es un intervalo de energía prohibido. Cuando estos materiales son irradiados por fotones con una energía mayor a ese bandgap, los electrones que se encuentran en la BV se excitan y saltan a la BC, creándose un hueco (h+) en la BV y un par e- /h+. Cuando estos pares electrónicos se encuentran en la superficie de un fotocatalizador, pueden promover reacciones de reducción (e-) y oxidación (h+) (Figura 1). Estas reacciones pueden dar lugar a la formación de radicales libres (p.e. OH.), las cuales, a su vez, intervienen en reacciones como la degradación de compuestos, la producción de hidrógeno o la síntesis de compuestos orgánicos [40–44]. 27 Figura 1. Esquema de un proceso fotocatalítico [45]. La fotocatálisis puede ser principalmente de dos clases: • Homogénea: el fotocatalizador es soluble en una fase líquida o gaseosa y acelera una reacción química mediante la absorción de luz. En este proceso, el fotocatalizador y los reactivos se encuentran en el mismo estado de agregación. • Heterogénea: en este caso el fotocatalizador y los reactivos se encuentran en fases distintas, normalmente en fase sólidolíquido y sólido gas. Los catalizadores que se van a estudiar en esta Tesis entran todos dentro de la fotocatálisis heterogénea ya que los catalizadores desarrollados son sólidos y están depositados sobre un soporte. Uno de los aspectos que más influyen en la eficacia de un fotocatalizador es la tasa de recombinación de los pares electrónicos. Cuanto más tarden estas cargas en volver a su estado energético inicial, mayor capacidad tendrá el fotocatalizador para llevar a cabo las reacciones de fotocatálisis correspondientes. Para mejorar esta tasa 34 observan cambios significativos al aumentar el espesor. La estabilidad de las capas de NiO/TiO2 se evaluó para 10 ciclos en las mismas condiciones en las cuales se realizaron las medidas de actividad fotocatalítica, confirmando la estabilidad de estas capas después del primer ciclo, en el cual se puede apreciar un descenso en el % de degradación del azul de metileno. 3.4. Conclusiones Finalmente, de este trabajo podemos extraer las siguientes conclusiones: 1. La adición de una capa de NiO al TiO2 resulta en la formación de una heterounión p-n, ya que no se observan cambios en el bandgap del recubrimiento, pero si una mejora de su actividad catalítica. El porcentaje de degradación aumenta de un 5.8% al 25.3% del TiO2 puro al NiO/TiO2 y de un 14% a un 25.3% si comparamos una muestra comercial de vidrio Pilkinton con el recubrimiento de NiO/TiO2 de 10 nm. 2. El cambio en el espesor del NiO no se traduce en cambios significativos en la fotoactividad de la capa. 3. Aunque la estabilidad de las capas de NiO/TiO2 ha sido comprobada, si es cierto que se puede observar un descenso del % de degradación el azul de metileno tras el primer ciclo. Esto puede venir derivado del efecto de la adsorción del azul de metileno. 4. Teniendo en cuenta los resultados obtenidos, se puede concluir que el objetivo inicial de este estudio, utilizar el magnetron sputtering para desarrollar y mejorar un fotocatalizador se ha conseguido, demostrando la eficacia de esta técnica para sintetizar fotocatalizadores. 35 3.5. Development of photocatalytic nanostructured TiO2 and NiO/TiO2 coatings by DC magnetron sputtering for photocatalytic applications Ceramics International 49, 19309-19317 (2023). Abstract High photocatalytic activity layers were obtained by combining a TiO2 nanostructured coating support with 10 nm and 20 nm NiO layers synthesized by direct current (DC) magnetron sputtering. In order to improve the TiO2 crystallinity several sputtering process parameters were studied. The TiO2 and NiO/TiO2 coatings were characterized through XPS, XRD, FESEM, AFM and UV transmittance measurements, in the latter case, for band gap determination. The photocatalytic activity of the coatings was tested through methylene blue degradation under UV light. The formation of a p-n heterojunction between the TiO2 and NiO layers improved the degradation rates, compared to the bare TiO2 layers; however, almost no difference can be seen when increasing the NiO sputtering power and subsequently the coating thickness to higher values than 20 nm Keywords: DC magnetron sputtering, Photocatalytic coatings, p-n heterojunctions, NiO catalyst. 3.5.1. Introduction Recently, the rise of the “fast fashion” industry has sparked profound concern, as it worsens the damaging effects of the textile industry on the environment [60]. Thus, it has become critical to find efficient and non-polluting wastewater cleaning methods that allow the removal of toxic organic dyes. 36 Photocatalysis has drawn attention for wastewater treatment, as a lightdriven, cost-effective and sustainable method that can easily degrade organic pollutants. Among photocatalytic compounds, crystalline titanium oxide [61], specifically the anatase phase, is widely known for its use as an organic pollutant oxidant in wastewater treatment [62–64], but also for hydrogen production [65–67], or to be incorporated in biological materials [68]. When a photon, with equal or higher energy than the band gap of anatase TiO2 (3.2 eV), irradiates the TiO2 conduction band (CB), electron-hole pairs are generated. These charge carriers diffuse all over the surface, reacting with surrounding compounds, forming radicals (·OH, ·O2) able to carry out oxidation reactions that mineralize organic molecules [69]. Moreover, its low toxicity, low cost, and high stability are beneficial for the aforementioned applications. Nevertheless, the photoactivity of TiO2 is hindered by the high recombination rate of the electron-hole pairs, reducing the efficiency of this material to carry out photocatalytic reactions. Several strategies can be followed to improve TiO2 photoactivity such as metal doping [70,71], the formation of p-n heterojunctions [72–74] or surface modification [75]. The formation of a p-n heterojunction with an n-type semiconductor appears to be an efficient approach to overcoming obstacles such as high recombination rates and low charge carrier transfer. When a p-n heterojunction is formed, the excited electrons migrate from the n-type semiconductor to the p-type one, which has a lower energy conduction band (CB); meanwhile, the holes move from the p-type to the n-type valence band (VB). The excitons’ movement creates a built-in electric field with a negative charge at the interface of both semiconductors, facilitating charge separation and diminishing the recombination rate, resulting in an improvement of the photocatalytic performance [76] (Figure 3). TiO2 can be combined with several semiconductors such as Fe2O3 [77], SnO2 [78], ZnO [79] or NiO [80] to achieve lower recombination rates, although recently, NiO has received more attention, due to its 37 improved charge separation and cost-effectiveness [81]. As a p-type semiconductor, NiO has remarkable optoelectronic properties, which translates into improved charge carrier separation, higher hole mobility and promoting interfacial charge transfer [82,83]. Besides, the polar surface of NiO seems to improve the adsorption of several organic compounds and dyes [80,84–87]. The combination of all these features makes this material suitable for numerous applications such as hydrogen production [88], electrochromic devices [89] or sensors [90]. Figure 3. Scheme of a NiO/TiO2 p-n heterojunction For catalyst development, wet-chemical methods such as sol-gel techniques are commonly preferred because they can lead to higher specific surface areas, a fundamental factor in achieving good photocatalytic activity [91,92]. However, these methods not only produce catalysts with low homogeneity, durability and highly toxic by-products, but they are also unsuitable for industrial upscaling, which is fundamental for large scale production of large-area thin films. Furthermore, as the catalyst is not fixed onto a substrate, the method requires a post-filtration treatment, which could be detrimental to the 38 environment [93]. Physical vapor deposition (PVD) technologies, in particular magnetron sputtering, could resolve the aforementioned drawbacks, arising as a promising alternative to chemical methods, especially in terms of industrial upscaling. Magnetron sputtering allows the manufacture of large-area catalytic thin films immobilized onto a substrate and with excellent physical properties such as high adhesion or durability. The production of photocatalytic TiO2 films by means of magnetron sputtering has been already widely explored for several applications including wastewater treatment [94–100]. Regarding the use of magnetron sputtering to develop photocatalytic TiO2 p-n heterojunctions, several combinations have been investigated, including WO3/TiO2 [56], CuO/TiO2 [57] or ZnO/TiO2 [58], however, as far as our knowledge is concerned, there are only a few examples of sputtered NiO/TiO2. Thus, a NiO/TiO2 transparent photovoltaic cell was developed through magnetron sputtering, with the TiO2 acting as a UV light detector and NiO as the visible light absorber [101]. Elsewhere, a NiO layer was sputtered by DC reactive magnetron sputtering over TiO2 for electrochemical reduction of N2 to study the “π-back donation” mechanism [102]. Given that PVD methods seem to be unusual for the fabrication of this type of heterojunctions, it is interesting to explore this technology, taking also into account its potential to upscale to an industrial level the fabrication of effective, durable and non-toxic photocatalytic films. In this work, porous, columnar thin films of TiO2 and NiO/TiO2 have been developed by reactive DC magnetron sputtering in an industrial sputtering chamber, to study the effect of NiO on their photocatalytic activity. The aim of this research is to (i) prove that magnetron sputtering is a suitable method for large-scale production of NiO/TiO2 thin films and (ii) to improve the photocatalytic performance for dye degradation of TiO2 through the formation of a p-n heterojunction. For that purpose, the study of TiO2 sputtering process parameters, such as 39 sputtering power and pressure, was performed, optimizing the TiO2 film in terms of crystallinity and morphology. Once the desired TiO2 film was obtained, NiO was deposited on top of TiO2, resulting in nominally 10 nm and 20 nm thick NiO layers. The photocatalytic activity of both TiO2 and NiO/TiO2 coatings has been tested for methylene blue (MB) degradation under UV light. 3.5.2. Materials and methods 3.5.2.1. Deposition Glass slides (2.6 x 7.6 cm2) and monocrystalline silicon (100) wafers were used as substrates. Previous to the deposition, in order to remove any residue or contamination from the samples, glass slides were cleaned with an alkaline solution, rinsed with isopropyl alcohol and dried with hot air. For the Si wafers, only the last two steps were done. To prevent the migration of Na+ ions present in glass substrates, 250 nm of SiO2 was deposited as an initial barrier layer. The SiO2 deposition was carried out by DC magnetron sputtering in reactive mode in Ar while regulating the O2 flow at 20% with a Speedflo®TM controller (Gencoa Ltd). The magnetron was driven using an Advanced Energy Pinnacle Plus in pulsed DC mode, applying a sputtering power of 2 kW and a pulse frequency of 75 kHz with a duty cycle of 70% at a pressure of 0.62 Pa. For the TiO2 coating deposition process optimization, three pressures and two sputtering power densities were studied. The sputtering parameters described below were selected for the TiO2 coating application, according to the desired crystallinity and morphology for the coatings. Titanium dioxide coatings were deposited by DC magnetron sputtering technology in an industrial vacuum chamber (80 x 60 cm2), designed and manufactured at TEKNIKER, equipped with one unbalanced 40 magnetron connected to an Advanced Energy Pinnacle DC power supply. A titanium target (55 x 12.5 cm2, 99.99% purity) was used, and the substrates were placed on a fixed substrate holder at ground potential, positioned at 5 cm in front of the target. The chamber base pressure was 2x10-4 Pa and the target was pre-sputtered with Ar for 4 min 0.95 Pa, to remove the oxide layer present on the Ti target. Following this, the deposition of TiO2 was carried out in reactive mode with a 110/40 sccm ratio of Ar/O2 gas flow, operating in fully oxidized mode due to the absence of continuous feedback control. This mode is needed because of the high instability of the TiO2 reactive deposition process in transition mode [103,104]. The sputtering power was set at 4 kW, the working pressure was 8.2 Pa and the substrate was at ground potential. The deposition rate was approximately 0.2 nm s-1 with a total thickness of around 700 nm. After the deposition process, the samples were post-annealed for 2h at 450ºC. The deposition of the NiO thin films was performed in a Kenosintec™ sputtering chamber, equipped with three confocal circular magnetron cathodes. DC powers of 25 W or 50 W were applied to a Ni target (diameter 152.4 mm, 99.99% purity) for 10 minutes with an Advanced Energy MDXTM power supply. The base pressure of the chamber was 6.5x10-3 Pa and a working pressure of 1.1 Pa was maintained during the deposition with a gas flow of 55 sccm of Ar. The distance from the samples to the target was 245 mm. The initial aim was to obtain a Ni metallic coating, but NiO was formed, probably due to oxidation after deposition or, maybe due to the presence of some oxygen inside the chamber. 3.5.2.2. Characterization Methods To assess the crystal structure of the thin films produced, X-ray diffraction (XRD) was performed on a Panalytical XpertTM system. CuKα1 radiation at 0.154 nm was used in grazing incidence mode at 3°, over a scan range from 20 to 80° 2θ. The acceleration voltage and 41 applied current were set at 40 kV and 30 mA, respectively. Morphology and thickness studies were carried out using an ULTRA Plus Carl-Zeiss field emission scanning electron microscope (FESEM) and Solver PRO NT-MDT atomic force microscope (AFM). Raman Spectroscopy was performed with a Raman Renishaw for each sample. The laser used for the acquisition of the spectra operated at a wavelength of 514 nm with a power of 10 mW. The optical characterization of the coatings deposited on glass was performed using an Ocean Optics USB4000 UVVis spectrometer. The oxidation state information of the samples was obtained by X-ray photoelectron spectroscopy (XPS), performed with a SUPRA photoelectron spectrometer (Kratos Analytical Ltd.) equipped with Mg K X-rays as the primary excitation source. The binding energy was referenced to the C 1s line at 284.8 eV for calibration, and a Gaussian function was applied for curve fitting with a Shirley background. 3.5.2.3. Photocatalytic activity measurements for methylene blue degradation The photocatalytic activity of the samples was assessed by methylene blue (MB) degradation tests. Each 2.5x 2.5 cm2 sample, cut from the original samples, was placed in a quartz cuvette containing 50 mL of methylene blue solution (from Alfa Aesar) with a concentration of 2 μmol L-1. Before testing, the samples were placed in the dark for 1h under continuous stirring to reach the adsorption/desorption equilibrium. The samples were then irradiated for 1h with UV-A sources (15 W Sankyo Denki BB lamps). The main peak of the UV source was measured at 365 nm. An Ocean Optics USB4000 UV-Vis spectrometer was used to monitor the main methylene blue absorption peak (664 nm). Finally, the pseudo-first-order rate constants kα were obtained for each sample by plotting Ln(A0/At) against time in order to compare the different photocatalytic activities [105,106]. 3.5.2.4. Stability measurements 42 The stability of both NiO/TiO2 samples was studied by repeating the degradation test of MB as described in Section 3.5.2.3. After each measurement, the samples were cleaned with distilled water and dried with compressed air and the operation was repeated with both samples nine more times. 3.5.3. Results and discussion 3.5.3.1. Characterization FESEM was used to determine the morphology of the coatings. Figures x (a, b) shows the columnar porous structure of the TiO2 coating, which is similar in appearance to the porous structure described in the Thornton Structure Zone model [18]. The measured thickness of the TiO2 coating is around 740 nm with column diameters ranging from 60 nm to 160 nm (Figure 4c, d). (a) (b) 43 Figure 4. SEM images of sputtered TiO2. For the NiO films, only the 20 nm thick film could be measured through FESEM (Figure 4c) as the 10 nm layer did not form a continuous film and could not be measured this way. Regarding the morphology of the NiO coatings, the 10 nm NiO (Figure 5a, c) layer changes subtly the surface of the columns TiO2, with the appearance of a small cluster-like structure that can be seen on top of these columns, although no changes are apparent in terms of film porosity. When the thickness is increased from nominally 10 nm to 20 nm the surface morphology changes are much more relevant (Figure 5b, d). The 20 nm NiO layer forms an (c) (c) (d) 50 Finally, the presence of Ni0 (552.6 eV) shows that some of the Ni on the coating remains metallic [109]. Figure 10. XPS spectra of Ni 2p for NiO/TiO2 coatings of (a) 10 nm and (b) 20 nm. XPS spectra of O 1s are shown in Figure 11. For TiO2 samples, the O 1s main peak can be deconvoluted in three individual peaks: two peaks at 533.2 eV and 531.7 eV that can be attributed to the presence of adsorbed oxygen species (such as O2, H2O, OH-) [110,111]. And a third peak is attributed to the Ti-O bonds from the TiO2 lattice. For NiO/TiO2 (a) (b) 51 samples, those 3 peaks can also be found, with the latter peak attributed to the Ni-O bonds from the NiO lattice. (a) (b) 52 Figure 11. XPS spectra of O1s for (a) TiO2 and (b) 10 nm and (c) 20 nm NiO/TiO2 coatings. The valence-band (VB) spectra of TiO2 and NiO/TiO2 samples are shown in Figure 12. By extrapolating the valence band spectrum using linear fitting for each sample, the valence-band maximum (VBM) values were measured. For the TiO2 sample, a VBM of 2.8 eV was measured. This value is similar to values that can be found in the literature [112]. When the NiO layer is deposited on top of the TiO2 coatings, the VBM measured shifts from 2.8 eV to 0.40 eV, which is in accordance with the formation of a p-type NiO layer [109]. (c) 53 Figure 12. VB spectra of all samples for VBM determination. In addition to the identification of the oxidation states of the elements, elemental quantification was performed on the samples. It must be noticed that a high concentration of C 1s was measured during the quantification, certainly due to impurities during the measurement. The Raman spectra shown in Figure 13 matches the reported spectra of anatase found in the literature [113,114], with specific peaks located at 144, 397, 516, and 638 cm-1. It can be also noted that the intensity of the peaks decreases with NiO layer thickness, but changes are not observed in the anatase structure [112]. 54 Figure 13. Raman spectra of all samples. To determine the optical band gap, the transmittance of the TiO2 and NiO/TiO2 films was measured between 200 nm and 850 nm wavelength range. For the band gap estimation, assuming an indirect band gap, the absorption coefficient, α, was obtained using Equation 1, where t is the thickness of the film. T≈e−αt (1) Using the Tauc plot method [115], Eg was calculated using the absorption coefficient using Equation 2, by plotting (αhν)1/2 as a function of hν and extrapolating the linear region to the abscissa. αhν=C(hν−Eg)1/n (2) where α is the absorbance coefficient, h is the Plank constant, ν is the frequency of vibration, Eg is the bandgap energy and n is the value for TiO2 indirect allowed transition, which is equal to 2 [116]. As can be seen in Figure 14, almost no change in terms of optical band gap can be observed when the NiO layer is deposited on the top of the TiO2, which suggests that a p-n heterojunction is being formed. 55 Figure 14. Optical bandgap of all samples. 3.5.3.2. Study of photocatalytic degradation of methylene blue The degradation of methylene blue (MB) was studied for the coatings on glass substrates for 1 hour, with and without the NiO layer. Initially, the adsorption-desorption equilibrium of MB for the coatings was determined by keeping the samples immersed in a MB solution with constant stirring for 1 hour without illumination. As shown in Figure 15a, using the natural logarithm of the relative absorbance (Ln[A/A0]), plotted vs. time, a pseudo-first-order rate constant (k) is obtained for the reaction [64]. Figure 15b shows the degradation rate of methylene blue, A/A0, with respect to time. All the experimental samples are compared to a commercial Pilkington Activ glass sample. Although this commercial product is produced using a different process, there is a lack of standards to compare to in this field and the Activ sample is acknowledged as providing a useful purpose in this case [117]. 56 Figure 15. (a) The degradation rate of MB for the photocatalytic coatings and (b) determination of rate constants for MB degradation reaction using a first-order kinetic model under UV light. From these results, it can be concluded that TiO2 presents photocatalytic activity under UV light with a k constant around 2x10-5 s-1 which agrees with literature reported values [105]. Also, the TiO2 (a) (b) 57 photocatalytic activity is improved with the addition of NiO coatings, with a more than 4 times improvement for the 10 nm NiO thick layer. Regarding degradation rates for the NiO/TiO2 coatings, they also are in agreement with literature values reported, the slightly lower values can be attributed to differences in purity, structure or synthesis method [118,119]. This improvement in MB degradation for NiO/TiO2 coatings could be due to the reduction of the recombination rate of TiO2 h+/epairs by the presence of NiO and the improved electronic mobility caused by the formation of a p-n heterojunction, where the electrons produced in the conduction band of the p-type semiconductor (here NiO) are moved to the conduction band of the n-type semiconductor (here TiO2). These photogenerated electrons are responsible for reducing adsorbed O2 to O2-·, which is considered a key step in MB oxidative decomposition. On the other hand, the holes generated in the valence band of the n-type semiconductor are promoted to the valence band of the p-type semiconductor, reacting with absorbed H2O or OH- , which leads to the formation of strongly oxidative OH· radicals. These radicals are thought to be the main driving force for photocatalytic oxidative reactions, reacting very fast with the MB adsorbed on the NiO surface [120–122]. The different p-n junctions created increase the charge separation efficiency of the material as confirmed by the lower band gap values achieved for pure TiO2 by UV transmittance measurements. The formation of these radicals due to the formation of a p-n heterojunction is the main reason for the improvement of the charge carriers’ separation and, consequently, photocatalytic efficiency [123,124]. As a future objective to complete this work, we intend to perform photoactivity measurements in the presence of scavengers [105,125]. Regarding the thickness of the NiO coatings, it appears that increasing thickness from 10 nm to 20 nm does not translate into a comparable improvement in photocatalytic activity. This could be because, despite 58 the observed reduction of roughness in the 20 nm layer, which is equivalent to a reduction in surface area, the 20 nm NiO layer gives better coverage of the TiO2 surface, improving the formation of new pn heterojunctions. These two contradictory effects may be the reason why there is no remarkable increase in photocatalytic activity when the thickness is increased. The influence of the thickness of NiO layers on photocatalytic activity is a subject that could be interesting to further analyse in future research. The degradation percentage and degradation rate constant of the coatings are summarized in Table 2. Table 2. Degradation percentage and degradation rate constant under UV for 1h. Sample R (%) k (s-1) Pilkington 14 4.31x10-5 TiO2 5.8 1.73x10-5 10 nm NiO/TiO2 25.3 8.22x10-5 20 nm NiO/TiO2 21.3 6.79x10-5 Finally, stability measurements were performed with NiO/TiO2 samples to confirm their photoactivity and stability over ten cycles. The obtained results are shown in Figure 16. From these measurements, several interesting conclusions can be drawn: A significant reduction between the degradation percentage of the 1st cycle and the 2nd cycle takes place, especially in the 10 nm NiO sample. The most probable reason for this reduction could be that when the 1st measurement was carried out, the adsorption of the MB hadn’t reached 59 equilibrium and was still taking place while the samples were irradiated. This effect would probably be more noticeable in the 10 nm layer, due to its high porosity. For the rest of the cycles the degradation percentage is very stable when compared to the rest of the cycles, thus we can assume that no degradation of the coatings is taking place. When the adsorption of the MB no longer takes place, the 20 nm NiO coating shows a slightly higher activity when compared with the 10 nm NiO coating. Besides the 1st cycle, the coatings show high stability for the remainder of the measurements. Figure 16. Stability test of NiO/TiO2 samples for MB degradation. 3.5.4. Conclusions 66 Figura 2. Esquema de un electrolizador PEM. Para llevar a cabo la reacción de electrólisis se necesita aplicar una energía mínima para que esta ocurra. En condiciones estándar, dado que no estamos ante una reacción espontánea, la energía libre de Gibbs (∆G0) para esta reacción es igual a 237.1 kJ/mol. Esta ∆G0 se puede expresar como potencial (E0) según la siguiente ecuación (Ec. 3): 𝐸0=∆𝐺0 𝑛𝐹 = 1.2291 𝑉 Ec. (3) Donde n es el número de electrones presentes en la reacción y F se corresponde con la constante de Faraday. Pero este valor de potencial no es el voltaje real que se necesita aplicar a un electrolizador para que tenga lugar la electrólisis, se necesita aplicar un voltaje mayor debido a las diferentes barreras energéticas a superar dentro del sistema. Este exceso de potencial que es necesario aplicar recibe el nombre de sobrepotencial y es la suma de distintos factores, que se pueden resumir en tres tipos: 67 ηtot = ηact + ηohm + ηmass Ec. (4) Figura 3. Representación de los sobrepotenciales presentes en una celda de electrólisis [150]. • Sobrepotencial de activación (ηact): este sobrepotencial está ligado a la propia cinética del electrodo para una determinada reacción y se debe a la energía de activación necesaria para llevar a cabo el proceso de transferencia electrónica. Para poder minimizar el ηact hay que utilizar el catalizador más adecuado para cada reacción. Existe un ηact anódico y ηact catódico inherentes al material del que se compongan los electrodos, y se pueden calcular usando la ecuación de Tafel. ηact,a= aa + balogj Ec. (5) ηact,c= ac + bclogj Ec.(6) Donde aa y ac son las constantes de Tafel para ánodo y cátodo, b es la pendiente y j la densidad de corriente. 68 • Sobrepotencial óhmico (ηohm): a este sobrepotencial contribuye principalmente las resistencias iónicas y eléctricas presentes en las superficies e interfases de los componentes de una celda. Se incluye la resistencia iónica de la membrana o las resistencias eléctricas entre elementos como la capa difusora de gases y la placa bipolar. Este término cumple la ley de Ohm, donde ηohm = iR, siendo el ηohm proporcional a la intensidad de corriente. • Sobrepotencial de transporte de masa (ηmass): este sobrepotencial, también llamado sobrepotencial de concentración, tiene lugar a altas densidades de corriente, cuando la producción de burbujas de gas dificulta la llegada del reactivo a la superficie del electrodo, dificultando así la reacción y la movilidad de especies en el medio. Si el objetivo es la reducción del ηact,, hay que focalizarse en encontrar los catalizadores más adecuados y mejorarlos, con el fin de minimizar la energía de activación sin afectar a la estabilidad del propio catalizador. De esta forma se consigue mejorar los parámetros cinéticos de la reacción de interés mejorando a su vez la eficiencia de la reacción. Una de las principales desventajas de estos sistemas es el tipo de catalizadores que se emplea, ya que en el ánodo se usa Ir, IrOx o IrRuOx y en el cátodo Pt/C, materiales que son escasos y caros [151–153]. Actualmente la cantidad que se emplea en los electrolizadores comerciales es bastante elevada, sobre los 3 mg cm-2. Además, la síntesis de estos catalizadores consta de varios pasos de reacción y durante el proceso se forman diversos productos secundarios lo cual aumenta aún más la cantidad necesaria de material catalítico para llegar a depositar esos 3 mg cm-2 de material. Estas vías sintéticas son complejas y difíciles de automatizar, lo cual obstaculiza aún más la posibilidad de 69 escalar a un nivel industrial la producción de catalizadores y a su vez la producción de electrolizadores PEM [154]. 4.2.1. Reducción del contenido de Pt en electrodos para la reacción de evolución de hidrógeno (HER) mediante magnetron sputtering El precio y la escasez de los catalizadores empleados en sistemas tanto de producción como de combustión de H2 supone una problemática a la hora de producir a gran escala e instaurar estas tecnologías a nivel global. Además, hay que sumarle la difícil industrialización de los procesos de fabricación de estos catalizadores. Los electrodos para pilas de combustible y electrolizadores PEM se obtienen mediante procesos muy similares. En ambos casos, se necesita obtener una tinta catalítica formada por Pt o compuestos de Ir mezclados con un ionómero y negro de carbón en polvo (carbon black), proceso sintético de varios pasos, ya que parte de la obtención de nanopartículas catalíticas. Estas tintas tienen que ser depositadas mediante pulverización sobre la membrana o la capa difusora de gases. Todos estos pasos se traducen en una pérdida de material catalítico relevante y dificultan el escalado del proceso de síntesis completo. Por ello, es fundamental encontrar una técnica de fabricación industrial que permita reducir la cantidad de catalizador empleada y sintetizar compuestos catalíticos en un solo paso y de forma precisa, evitando así el desperdicio de material catalítico durante el proceso de síntesis. En bibliografía existen ya múltiples ejemplos del uso de magnetron sputtering para sintetizar catalizadores de materiales del PGM. En pilas de combustible PEM, tecnología que emplea el H2 como fuente energética, usa como catalizadores compuestos de Pt/C. Para este sistema se ha estudiado el magnetron sputtering como técnica de fabricación de catalizadores, obteniéndose resultados que confirman su validez [155–158]. 70 En el caso de la electrólisis PEM, la mayoría de los trabajos se centran en el lado anódico, con el fin de reducir la cantidad de Ir necesaria. Entre ellos, cabe destacar los estudios realizados por Hrbek et al, basados en catalizadores con bajo contenido de Ir-Ru para la OER, donde se han desarrollado capas finas con cantidades de Ir menores de 0.1 μg cm-2 y distintas proporciones de Ir-Ru mediante magnetron sputtering. La caracterización electroquímica, tanto ex-situ como in-situ de estos catalizadores, confirma que, con mucha menos cantidad de Ir respecto a los catalizadores comerciales, ni la estabilidad ni la actividad catalítica de los electrodos se ve comprometida con la disminución de Ir y la incorporación de Ru [144,146,159,160]. Respecto al cátodo, aunque pocos, existen algunos trabajos de investigación sobre el uso de magnetron sputtering para el desarrollo de catalizadores de Pt/C. Bernsmeier et al. utilizaron el hollow cathode magnetron sputtering para obtener películas de Pt/C de elevada estabilidad y evaluaron su actividad electroquímica utilizando un RDE (electrodo de disco rotatorio) [161]. En otro trabajo muy interesante, Fedotov et al., depositaron nanopartículas de Pt directamente sobre un soporte de polvo de carbón negro, usando un recipiente con agitación para homogeneizar la cantidad de nanopartículas depositadas. La actividad catalítica de los electrodos se evaluó en un electrolizador PEM, logrando rendimientos similares a electrodos desarrollados mediante métodos químicos [162]. También, usando directamente la GDL de carbon paper como soporte, se depositaron películas delgadas densas mediante distintos tipos de magnteron sputtering (DC, DC pulsado y RF) [157,163,164]. Con el fin de evaluar y observar la formación de burbujas de H₂ dentro de un electrolizador PEM, en otro estudio se depositaron capas de Pt de distintos espesores sobre GDLs de titanio. La actividad de estos electrodos se comparó con la actividad de una CCM convencional, probando que, con una cantidad mucho menor, la actividad másica aumentaba de forma muy notable [165]. 71 Los resultados obtenidos en todos estos trabajos de investigación, tanto en ánodo como en cátodo, confirman que es posible desarrollar catalizadores para electrolisis PEM mediante magnetron sputtering, posibilitando la reducción de material, sin pérdida de estabilidad ni actividad. Además, el uso del magnetron sputtering permite simplificar enormemente el proceso de fabricación de estos electrodos, reduciéndolo a un proceso de un solo paso fácilmente escalable a nivel industrial. Una vez analizadas las dificultades asociadas a los métodos de fabricación tradicionales, se estableció como objetivo principal de esta parte de la Tesis la reducción del contenido de Pt en electrodos para la HER en electrolizadores PEM. Este método puede solventar tanto el malgasto o uso excesivo de material catalítico y a su vez facilitar el escalado de la fabricación de estos electrodos. 4.3. Resumen de resultados Este capítulo de la Tesis es el más relevante y en el cual se ha dedicado más tiempo a su consecución ya que el adquirir conocimiento y experiencia en tecnologías relacionadas con la producción de H2 es de un interés elevado desde un punto estratégico para Tekniker y para Euskadi. Esta parte de la Tesis se ha llevado a cabo dentro del proyecto H2Basque/H2Plan, que tenían como objetivo principal el desarrollo de tecnologías innovadoras para la producción de H2 verde a un coste competitivo y así poder impulsar la descarbonización de la economía. Principalmente, este proyecto se ha centrado en la mejora y optimización de componentes presentes en electrolizadores como son las placas bipolares o los electrodos. En nuestro caso, nuestro objetivo ha sido el desarrollo de electrodos de bajo contenido de Pt mediante magnetron sputtering para la producción de H2. El trabajo realizado se compone de dos etapas. Inicialmente, la obtención de tres electrodos con un contenido de Pt menor de 0.125 mg 72 cm-2, seguido de la evaluación de la actividad catalítica y durabilidad de estos para la HER en una celda de tres electrodos (caracterización ex situ) y, en una monocelda (caracterización in situ). 1. Deposición de nanopartículas de Pt mediante magnetron sputtering con fuente de agregación de gas. La deposición de las nanopartículas de Pt se llevó a cabo en un sistema de sputtering acoplado a una fuente de agregados. Este sistema facilita la formación de nanopartículas o clústers debido a las altas presiones que se puede alcanzar dentro de la cámara de agregación, del orden de 10-1 mbar. Para el desarrollo de los tres electrodos se fijaron unos parámetros de proceso previamente optimizados variando únicamente el tiempo de deposición. Las nanopartículas de Pt se depositaron directamente sobre una GDL de papel de carbono recubierto por una capa microporosa de negro de carbón, obteniéndose electrodos con cargas de Pt de 0.1, 0.062 y 0.052 mg cm-2 y con un diámetro medio de nanopartícula de 4 nm. La reducción de Pt alcanzada es de hasta 4 veces menor que un electrodo comercial análogo con una carga de 0.3 mg cm2 de Pt. 2. La evaluación de la actividad catalítica y durabilidad para la HER, dividida en dos fases: • Fase de caracterización ex situ: evaluación electroquímica ex situ en celda de tres electrodos. • Fase de caracterización in situ: evaluación electroquímica in situ en monocelda PEM. • Fase de caracterización ex situ: evaluación electroquímica ex situ en celda de tres electrodos. Con el fin de caracterizar y evaluar electroquímicamente los tres electrodos desarrollados, se llevaron a cabo medidas electroquímicas en medio ácido de voltamperometría cíclica y de barrido lineal, impedancia y cronopotenciometrías. De estas 73 técnicas cabe destacar la voltamperometría cíclica, la cual nos proporciona información sobre la pureza del material catalítico, durabilidad o la superficie electroquímicamente activa del electrodo. Otra de las técnicas más relevantes para la evaluación de la actividad catalítica catódica es la voltamperometría de barrido lineal, la cual nos permite obtener directamente el sobrepotencial de un electrodo a una densidad de corriente determinada para la reacción que nos interese evaluar, en este caso la HER. De esta evaluación ex situ, se pudieron extraer los valores de sobrepotencial catódico a -10 mA cm-2 para los tres electrodos, el cual era sumamente bajo, unos 8 mV. Además, se observó que este valor de sobrepotencial no aumentaba al disminuir la cantidad de Pt, probando que mayor cantidad de Pt no es equivalente a una mejora en la actividad catalítica. Si tenemos en cuenta el área superficial electroquímicamente activa (ECSA), estos valores de sobrepotencial se mantienen para el electrodo con 0.052 mg cm-2 de cargar de Pt, en cambio para los otros dos electrodos este valor aumenta. Esto también se ve reflejado en la actividad másica de los electrodos, siendo mayor cuanto menor es el contenido de Pt. Aparte de la actividad catalítica, otro factor determinante para la eficiencia de un electrodo es su estabilidad. En este caso, se pudo probar tanto la estabilidad a corto plazo (24 h) de los tres electrodos desarrollados como la estabilidad a largo plazo (1000 ciclos) a altas densidades de corriente para el electrodo con 0.1 mg cm-2 de carga de Pt. • Fase de caracterización in situ: evaluación electroquímica in situ en monocelda PEM. Teniendo en cuenta los resultados ex situ obtenidos tanto de actividad catalítica en la HER como de estabilidad para los 74 electrodos desarrollados por magnetron sputtering, se pasó a la siguiente fase de caracterización electroquímica. Para ello, se diseñó una monocelda (Figura 4) de forma que su configuración fuese versátil y permitiese el uso de componentes de distintos espesores. Esta monocelda, se compone de dos placas bipolares de titanio en las cuales la zona que corresponde al área activa se encaja en unos marcos de PEEK. Estos marcos tienen dos funciones principales; por un lado, distribuir el agua hacia la malla de titanio que ocupa el área activa en contacto con el electrodo, y por otro, permitir el alojamiento componentes de distinto espesor y configuración, asegurando una compresión correcta del electrodo. Junto a la placa bipolar y dentro del marco PEEK se encuentran las mallas de Ti, encargadas de distribuir reactivos y evacuar gases. A continuación, se encuentran los PTLs, en este caso de Ti platinizado, recubrimiento que evita la oxidación de Ti durante el proceso de electrólisis. Para llevar a cabo la electrólisis, se empleó la configuración membrane electrode assembly/catalyst coated electrode (MEA/CCE). La configuración MEA/CCE corresponde a una membrana comercial con catalizador solo en el lado anódico (MEA). En el lado catódico tenemos la configuración CCE, es decir, un electrodo compuesto por carbon paper con una capa microporosa de carbon black, sobre la cual se han despositado por sputtering las NPs de Pt que actuarán como catalizador. 75 Figura 4. Monocelda PEM diseñada en Tekniker. Para operar la monocelda se ha usado un banco de ensayos PEM diseñado por H2GREEM (Figura 5). Este banco de ensayos permite ajustar parámetros de operación como puede ser el caudal de agua, temperatura e intensidad aplicada. Además, puede operar de forma autónoma durante los ensayos, ya que estos se pueden programar. 82 remarkable higher mass activity is achieved [165,198]. This work proves that most of the catalyst that is located inside the porous structure is inactive which is very relevant if magnetron sputtering is the chosen method for catalyst manufacturing. Taking all this into account, since HER appears to happen mostly on the electrode surface, if the active surface area could be improved using conventional carbon GDLs with MPL (microporous layer) and nano structuring platinum into nanoparticles, excellent catalytic activity should be achieved with an ultrasmall amount of Pt. Magnetron sputtering gas aggregation (MSGA) method remains, to the best of our knowledge, unexplored for HER electrode development. The difference with conventional magnetron sputtering, is that the MSGA has higher pressure inside a small aggregation chamber, due to a high inert gas flow which forms supersaturated vapors from the sputtered material, which allows easily the formation of nanoparticles and clusters with well-defined size and shape[199–201] making MSGA the most suitable approach for nanoparticle synthesis [32,202–205] between the different magnetron sputtering techniques. In this work, a series of Pt-based electrodes with different Pt loadings of 0.68, 0.81, and 1.37 wt% (up to 4 times smaller than conventional Pt/ C GDLs), have been prepared by MSGA technique and tested as electrocatalysts for the HER in acidic electrolyte. Electrodes were prepared by depositing Pt particles onto carbon black microporous layer/carbon paper forming thin layers of Pt nanoparticles. The loading of Pt on the sputtered GDLs was determined through gravimetric thermal analysis (TGA) while compositional and morphology analyses were performed through Xray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Electrochemical analysis techniques (CV, LSV, CP) were performed in all samples to determine kinetic and electrochemical parameters for HER in an acidic electrolyte. 83 4.5.2. Materials and methods 4.5.2.1. Platinum nanoparticles deposition Pt nanoparticles were deposited onto two different substrates, namely, a piece of silicon wafer and MPL-carbon paper (Sigracet 22BB MPL, with 5% PTFE) with a coated area of 1 cm2 for the latter. Pt nanoparticles were generated using a gas aggregation system (Nano GenTrio) from Mantis Deposition Ltd. (Oxfordshire, UK). The gas aggregation chamber (Figure 6) was attached to an industrial PVD chamber (80 × 60 cm2) and connected through a 4 mm hole, which allowed nanoparticles to flow between both chambers, with a base pressure of 2 ×10-4 Pa in the whole vacuum system. The platinum target (2.54 × 3 cm2, purity of 99.999%) was placed on a mobile sputter head, fixed at 95 mm from the chamber hole. Before starting the sputtering process, the target was cleaned with an Ar pressure of 4 Pa for 2 min at 15 W. For the sputtering process, a discharge power of 25 W was applied for 15, 10, or 5 min (PtNPs_15, PtNPs_10, PtNPs_5, respectively), with an Ar flux of 30 sccm and He flux 15 sccm, which equals to a pressure of 14 Pa in the sputtering chamber. Through the sputtering of a metal target at high pressure, small nanoparticles can be obtained in the gas aggregation chamber, which is connected to a deposition chamber at a much lower pressure, where the nanoparticles formed are dragged out by an auxiliary gas (He) and the pressure difference and deposited on top of a substrate placed in the deposition chamber. 84 Figure 6. Diagram of a magnetron sputtering gas aggregation chamber. 4.5.2.2. Characterization methods The morphology and thickness of the Pt layer on the electrodes were analysed using an ULTRA Plus Carl-Zeiss field emission scanning electron microscope (FESEM) and a TECNAI G2 20 TWIN high resolution transmission electron microscope at an acceleration voltage of 200 kV (STEM). The nature and relative abundance of the surface species was analysed with a SPECS customized apparatus equipped with a nonmonochromatic X-ray source XR 50 and a hemispherical energy analyser PHOIBOS 150. X-ray MgK line (1253.6 eV) operating at 200 W/12 kV, was used as the excitation The regions of interest were scanned at increments of 0.1 eV and fixed pass energy of 20 eV. Charge effects were accounted for by setting the C 1s core-level peak at 284.6 eV. Spectra were analysed using the XPS Casa Software. The Pt loading of the commercial and experimental samples was analysed through thermogravimetric analysis (TGA) using a TA Instruments Q500 instrument equipped with an EGA furnace. The measurements were conducted in air flow (90 mL min-1) using a 10ºC min-1 heating ramp, from r. t. until 900ºC. 4.5.2.3. Electrochemical characterization 85 Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) tests were carried out in a three-electrode cell with an Autolab PGstat 302 N potentiostat along with Nova software. A 3 M Ag/AgCl electrode and a Pt wire were used as reference and counter electrodes, respectively. As a working electrode, 1 cm2 of the sputtered Pt/C samples were used. The potentials were calibrated versus reversible hydrogen electrode (RHE) according to Equation (1): E(RHE) =E(Ag/AgCl) + 0.059 pH + 0.21 (1) CV measurements (for ECSA determination and Pt surface cleaning) were collected in Ar saturated 0.5 M H2SO4 electrolyte at room temperature. First, a number of CVs were recorded between 0.05 and 1.4 V at 100 mV s-1 until a constant response was obtained. The electro chemically active surface area (ECSA) was subtracted from the charge of the Hupd region from cyclic voltammograms recorded within a potential window between 0.05 and 1.2 V at 50 mV s-1, using the linear scan voltammetry mode. The HER was evaluated by recording LSV between 0.01 and 0.3 V at 10 mV s-1 in an H2-saturated 0.5 M H2SO4 electrolyte to ensure the H+/H2 equilibrium at 0 V vs. RHE [206]. Polarization curves are reported after iR compensation. The solution resistance (Rs) has been determined by electrochemical impedance spectroscopy (EIS) at high frequencies with a 10 mV perturbation. The overpotential (η) to reach a current density of 10 mA cm-2 and the Tafel slopes obtained in the 0.015 to 0.03 V vs. RHE potential range were used to benchmark catalysts’ activity. Finally, the degradation of the sputtered electrodes was assessed by conducting chronopotentiometry (CP) measurements at 10 mA cm-2 for 24 h and CV measurements at 5 mV s 1 from 0 V to 1 V vs Ag/AgCl for 1000 cycles were performed to assess the long-term stability at high density currents. 4.5.3. Results and discussion 4.5.3.1. Characterization 86 Before moving into a more complex substrate such as MPL-carbon paper, we analysed the deposition of Pt onto the silicon wafers (Figure 7) since they are a flat substrate widely used in sputtering. FESEM images were used to determine the thickness of the coatings over a silicon wafer and surface morphology over MPL-carbon paper substrates for the three samples. The aim was to analyse if the thickness of the samples was linearly dependent on the sputtering time and, also, how they were distributed. Figure 7. FESEM image of PtNPs_15 coating over a silicon wafer. The thicknesses of the Pt layer over silicon substrates were around 50, 25 and 14 nm for the PtNPs_15, PtNPs_10, and PtNPs_5 coatings, respectively (Figure 8), confirming the linear growth with sputtering time and presented a uniform distribution. (a) (b) 87 Figure 8. Thickness measurements of (a) PtNPs_15. (b) PtNPs_10 and (c) PtNPs_5 nanoparticles films over a silicon wafer. Regarding the surface morphology of the Pt coatings, SEM images of commercial (0.3 mg cm-2 Pt/C) and experimental samples were ana lysed (Figure 9). While the commercial sample (Figure 9a) has a porous and smooth surface where Pt nanoparticles are not visible, the sputtered PtNPs_15 coating (Figure 9b) shows a clear nanoparticulated surface, as the sputtered Pt nanoparticles are deposited all over the substrate surface, so almost all the Pt is exposed to the electrolyte. (c) 88 Figure 9. FESEM images of (a) commercial Pt/C sample and (b) PtNPs_15. When the deposition time is increased, the higher Pt loading could also subtly be seen in FESEM images (Figure 9a and Figure 10a, b). The Pt nanoparticles accumulate over the porous surface when sputtering time is increased, giving a thicker and denser appearance to the surface. 89 Figure 10. SEM images of (a) PtNPs_10 and (b) PtNPs_5 sputtered electrodes. TEM images of the sputtered samples and their corresponding size distribution histogram are displayed in Figure 12 and Figure 11, respectively. The statistic histograms of the samples were constructed by measuring the diameter of no less than 100 randomly selected Pt (a) (b) (b) 90 NPs (for each sample) showing a narrow size distribution from 2 to 6 nm with an average size between ~4 and 3 nm (Figure 11). Figure 11. Pt nanoparticles size distribution of (a) PtNPs_5. (b) PtNPs_10 and (c) PtNPs_15. In general, Pt nanoparticles have a well-defined spherical shape, as shown in Figure 12a. High-angle annular dark-field scanning TEM (HAADF-STEM) images reveal that Pt nanoparticles are located on the periphery of the carbon black layer and present an uneven and nonhomogeneous thickness As observed in the images, the coverage of the surface increases with the sputtering time, resulting in uncoated areas in the PtNPs_5 sample. (a) (a) (b) (c) 91 Figure 12. STEM images of (a) Pt NPs electrode and HAAF-STEM cross section images of (b) PtNPs_5 (c) PtNPs_10 and (d) PtNPs_15. Pt loading of the samples was determined from thermogravimetric analyses. Pt loadings of 1.37 ±0.29% (0.105 mg cm-2), 0.81 ±0.18% (0.062 mg cm-2), 0.68 ±0.09% (0.052 mg cm-2) and 3.93 ±1.58 wt% (0.3 mg cm-2) were obtained for PtNPs_15, PtNPs_10, PtNPs_5 and Pt/C commercial electrodes, respectively (Table 3). Figure 13. TGA profile for the commercial Pt/C electrode 0.3 mgPt cm-2. (c) (b) (d) 98 to Pt exposure on the catalyst surface (Figure 12c). The effect of Pt loading reduction was also studied by Bernt et al. [222] in an MEA electrolyzer. They compared the performance of two electrodes, one with a 45.8 wt% Pt/C and another with 4.8 wt% Pt/C, showing almost the same V–I polarization curves and confirming that cell performance is not affected by Pt loading reduction. So, if the current used amount of Pt is unnecessary for the cell to perform perfectly, the cost of the stack can be significantly reduced through a synthetic method like MSGA. Using MSGA, not only the amount of Pt can be remarkably diminished but also, in a 5-min single-step automated process, commercial electrodes reported catalytic performance can be surpassed. (a) (b) 99 Figure 17. Linear sweep voltammetry curves of all the samples in 0.5 M H2SO4 solution at 10 mV s-1 scan rate with respect to (a) GSA, (b) ECSA (c) Mass activities at 20 mV cm-2 and (d) Tafel plot curves. The rate-determining step (rds) of the HER reaction can be obtained through the Tafel slope from the Tafel curves (η vs. log j) shown in Figure 17d. The HER in acidic solutions can follow two reaction pathways: Volmer-Tafel or Volmer Heyrovsky. As reported in the literature, polycrystalline Pt follows the Volmer-Tafel pathway, with a Tafel slope of around 30 mV dec-1 [223,224]. Tafel slopes for PtNPs_15, PtNPs_10, PtNPs_5 samples are 30.8, 31.5 and 30.6 mV dec-1, which correspond to the values usually reported for polycrystalline Pt in acid solution. Finally, the short-term durability of the sputtered samples was evaluated through chronopotentiometry method at a constant current density of 10 mA cm-2 for 24 h in Ar-saturated 0.5 M H2SO4 solution [225] showing almost no variation in potential for that period of time, (Figure 18a). As long-term durability can be an issue when operating at high current densities for a long period of time for Pt-based HER catalysts, 1000 cycles between 1 V and 0 V vs Ag/AgCl at 5 mV s-1 were (c) (d) 100 per formed in Ar atmosphere to evaluate the degradation of the PtNPs_15 electrode [211,226]. In Figure 18b is represented the 1st and the 1000th cycle of PtNPs_15 electrode showing that there is no increase in overpotential between both cycles, hence the degradation of sputtered Pt nanoparticles could be considered negligible. Usually Pt/C lack of stability is due to poor adhesion and nanoparticle agglomeration [226,227], thus the high stability reported in sputtered samples could be due to an adhesion improvement since Pt nanoparticles are ionized, arriving with higher energy to the substrate and improving adhesion [203]. (a) 101 Figure 18. (a) Chronopotentiometry measurements at 10 mA cm-2 for 24h of all sputtered samples and (b) HER polarization curves of 1st and 1000th cycle of PtNPs_15 sample at 5 mV s-1. 4.5.4. Conclusions In this work, three Pt electrodes were developed by gas aggregation magnetron sputtering to study the effect of lowering Pt loading on their catalytic activity for HER in an acidic solution. Surface composition and morphologic characterizations were done, showing that the only present element in coatings was metallic Pt, shaped into well-defined nano particles around 4 nm size and also, covering all the substrate surface except in the PtNPs_5 sample, as is shown in TEM and ECSA measurements. The Pt content reduction when compared with a commercial sample is notable, more than 4 times smaller, up to 0.052 mg cm-2. So, in terms of reducing Pt content easily and effectively, MSGA seems to be an adequate and very effective method. Regarding HER activity, when the jgeo is compared at 10 mA cm-2, the overpotentials are almost equal (8 mV) but in terms of comparing jECSA, the data shown a tendency towards higher activity when lowering the Pt content, which agrees also with mass activity values. From these (b) 102 results, two conclusions can be drawn. First, that it might not be necessary to use considerable amounts of Pt to have a remarkable HER activity, as this work proves the opposite. It seems that is more important to have active sites available, given that catalysis is a surface process, than having a huge amount of catalyst, considering that the electrode with less Pt content and with an ECSA lower than its GSA is the one with lower overpotential and higher mass activity. The other conclusion is that MSGA is a remarkable option to develop Pt electrodes since in a one-step 5-min process, a highly active electrode can be ensembled depositing the Pt nanoparticles directly onto the MLP-carbon paper substrate without any byproducts or secondary, which translates into reducing the platinum that could be wasted along the process. Furthermore, this method is easy to upscale, so large scale fabrication of these electrodes could be done effortlessly. 103 4.6. Magnetron Sputtered Low-Platinum Loading Electrode as HER Catalyst for PEM Electrolysis Coatings, 14, 868-881 (2024). Abstract The development of cost-effective components for Proton Exchange Membrane (PEM) electrolyzers plays a crucial role in the transformation of renewable energy into hydrogen. To achieve this goal, two main issues should be addressed: reducing the Platinum Group Metal (PGM) content present on the electrodes and finding a large-scale electrode manufacturing method. Magnetron sputtering could solve these hurdles since it allows the production of highly pure thin films in a single-step process and is a well-established industrial and automated technique for thin film deposition. In this work, we have developed an ultra-low 0.1 mg cm−2 Pt loading electrode using magnetron sputtering gas aggregation method (MSGA), directly depositing the Pt nanoparticles on top of the carbon substrate, followed by a complete evaluation of the electrochemical properties of the sputtered electrode. These ultra-low Pt content electrodes have been thoroughly characterized and tested in a real electrolyzer cell. They demonstrate similar efficiency to commercial electrodes with a Pt content of 0.3 mg cm-2, achieving a 67% reduction in Pt loading. Additionally, durability tests indicate that these electrodes offer greater stability compared to their commercial counterparts. Thus, magnetron sputtering has been proven as a promising technology for manufacturing optimum high-performance electrodes at an industrial scale. 4.6.1. Introduction As the concern about climate change and the urgency to mitigate its devastating consequences are increasing rapidly, it is crucial to 104 radically change our fuel-based energy system towards a more sustainable approach based on renewable energy sources that allow complete decarbonization of the energy sector. In recent years, hydrogen has arisen as the ideal energy carrier, mainly due to its high energy density per molecule; the only byproduct of H2 combustion is water, and it can be obtained directly by water electrolysis [228–231]. There are several types of electrolysis using renewable energy sources, and among these, one of the most relevant is the proton membrane exchange (PEM) electrolysis [232–235]. This is due to several factors like low-temperature operation, achievement of high current densities at low voltages, negligible membrane gas crossover, and compact system design [236]. Despite these advantages, the main issue with PEM electrolysis is the high price of its components, especially the catalysts. In the anode, for the OER (oxygen evolution reaction), the most effective catalysts are Ir or IrOx, whereas in the cathode, for the HER (hydrogen evolution reaction) is Pt, both being highly scarce and expensive elements. Currently, for PEM electrolyzers with a catalystcoated membrane configuration (CCM), the Pt loadings in the cathode are high, around 3 to 1 mg cm−2 [237,238]. In principle, decreasing the Pt loading could be achieved without compromising the electrolyzer’s efficiency. This is because the exceptionally fast kinetics of Pt for the HER allow the use of low Pt loadings, hence significantly reducing the production costs of the stack [207,239,240]. Currently, catalytic coatings are developed using wet chemical methods like wet chemical impregnation, sol-gel, or thermal decomposition. For instance, Pt nanoparticles synthesized through these methods are often mixed with an ionomer and high-surface carbon black powder to avoid nanoparticle agglomeration. The obtained catalytic ink usually has to be deposited by spraying techniques over the membrane or the gas diffusion layer (GDL) [229,241,242]. Wet chemistry methods provide a high surface area to the catalytic layer, which is crucial to obtaining a high catalytic surface. However, these techniques have multiple 105 drawbacks, such as reaction byproducts and multiple reaction steps, which are directly related to material waste and complex industrial upscaling. Magnetron sputtering appears to be a compelling manufacturing method for solving these issues. As a robust industrial single-step process for thin film deposition, magnetron sputtering not only reduces the waste of material with the absence of byproducts but also facilitates the upscaling of catalysts manufacturing for large-scale production of electrodes for PEM electrolyzers. As far as we know, the use of magnetron sputtering as a manufacturing method for platinum group metals (PGMs) PEM electrolysis catalysts is not very wide, especially for the cathode side [243–245]. Most of the existing research work is focused on the ex-situ evaluation of the HER catalyst, where there is a huge potential for magnetron sputtering for their application [239,246– 250]. To the best of our knowledge, a couple of research works have been published regarding in-situ single-cell testing. For example, Pt thin films were deposited over titanium multifunctional liquid/gas diffusion layers in order to evaluate HER activity and H2 bubbles behaviour inside a PEM single cell [251]. This research work proved that most of the catalysis seems to occur on the surface of the electrode. In another study, a stirring cup for the carbon powder was introduced inside a magnetron-sputtering laboratory chamber, and Pt nanoparticles were deposited over the carbon support using a direct current (DC) power source [252]. Several loadings of Pt were evaluated for PEM electrolysis and fuel cell, showing no performance differences with electrodes developed using chemical methods. Hence, if HER catalysis is a process that occurs only on the electrode surface with no catalytic activity differences from the traditional synthesized catalyst, magnetron sputtering could be a feasible alternative to industrially upscale PEM catalyst. 106 As far as we are concerned, there is a complete lack of research on the use of magnetron sputtering to develop Pt nanoparticles for PEMWE and critically reduce the presence of Pt in the cathode. In this work, we have developed an ultra-low 0.1 mg cm−2 Pt loading electrode using magnetron sputtering, directly depositing the Pt nanoparticles on top of a carbon microporous layer (MPL) GDL, followed by a complete evaluation of the electrochemical properties of the sputtered electrode. This method enables a full exposure of Pt nanoparticles to the reactants, maintaining a high surface area while reducing the Pt loading. The aim of this research is to prove the feasibility of magnetron sputtering as an electrode manufacturing technique, allowing Pt reduction without interfering with cell performance. The electrochemical evaluation was performed in a three-electrode cell and a single PEM cell, and the results were compared with a commercial 0.3 mg cm−2 Pt/C electrode for both electrochemical setups. Electrochemical ex-situ analysis techniques like cyclic voltammetry and linear sweep voltammetry (CV, LSV) were measured for all samples to evaluate the catalytic activity for HER in an acidic electrolyte and in-situ analysis through V-i curves and chronopotentiometry measurements (CP), was performed in a real single PEM cell to evaluate the performance and degradation of both electrodes. 4.6.2. Materials and Methods 4.6.2.1. Platinum Nanoparticles Deposition Pt nanoparticles were deposited over a microporous layer of carbon paper (Sigracet 22BB MPL, with 5% polytetrafluoroethylene (PTFE)) with a coated surface of 1 cm2 for the ex-situ and 25 cm2 for the in-situ tests. The gas aggregation system (NanoGenTrio, Mantis Deposition Ltd., Thame, UK) was attached to an industrial PVD chamber (80 × 60 cm2), coupled with a 4 mm hole, enabling particles to move from one chamber to the other. The platinum target (2.54 × 3 cm2, purity of 107 99.999%) was positioned on a mobile sputter head, with a distance of 9.5 cm from the chamber hole. The base pressure of the whole vacuum system was 2 × 10−4 Pa, and the target was cleaned before starting the deposition process for 2 min at 15 W with a 4 Pa Ar pressure. After a previous process optimization [239], an Ar flux of 30 sccm and He flux of 15 sccm, corresponding to a total pressure of 14 Pa, were introduced in the aggregation chamber. To start the sputtering process to obtain the Pt nanoparticles, a discharge power of 25 W was applied for 15 min. The sputtered electrodes were compared with a commercial 0.3 mg cm−2 Pt electrode (40%wt Pt/C, Sigracet 22BB MPL, with 5% PTFE. The MSGA method is the most suitable magnetron sputtering technique for nanoparticle deposition. In contrast with traditional magnetron sputtering, in MSGA, the sputtering takes place inside a high-pressure aggregation chamber where supersaturated vapors of the ejected material enable the formation of clusters or nanoparticles of welldefined size and shape. As the sputtered particles collide with inert gas atoms, their energy is lowered and begins to condense into small clusters. These clusters serve as nuclei, where additional particles coalesce, leading to the formation of larger nanoparticles [253–257]. 4.6.2.2. Characterization Methods The size and morphology of the nanoparticles Pt electrodes were studied using a ULTRA Plus Carl-Zeiss (Jena, Germany) field emission scanning electron microscope (FESEM) and a TECNAI G2 20 TWIN (Hillsboro, OR, USA) high-resolution transmission electron microscope at an acceleration voltage of 200 kV (STEM). A SPECS customized system with a nonmonochromatic X-ray source XR 50 and a hemispherical energy analyzer PHOIBOS 150 was used to assess the relative abundance of the surface species. Thermogravimetric analysis (TGA) was performed using a TA Instruments Q500 instrument (New Castle, DE, USA) equipped with an EGA furnace to determine the Pt loading of the experimental and commercial samples. The 114 layer region and, towards more positive potentials, the Pt-oxide formation region. Moving to more negative potential values, the reduction of Pt oxides can be observed in the reserves scan around ca. 0.8 V [259]. Figure 23. Cyclic voltammetry of PtS_GDE. To evaluate and compare the HER activity of both electrodes, a potential sweep from 0 V vs. RHE towards cathodic potential values was recorded in the Ar atmosphere. Figure 24 shows the polarization curves where the overpotentials needed to reach a current density value of −10 mA cm-2 are the reference for the evaluation of the HER activity of different catalysts [260,261]. As presented in Figure 24, the overpotential for a current density of −10 mA cm-2 of the PtCom_GDE sample is 74 mV, while the overpotential at the same current density for the PtS_GDE is 53 mV. These results are preliminary proof that Pt loading can be reduced without compromising the electrode activity towards the HER. Besides, the reduction of the overpotential, when compared to the commercial sample, could be influenced by the fact that in the sputtered electrode, all Pt nanoparticles are placed on the surface in direct contact with the reactant, while in the commercial 115 sample, the Pt active sites can be more difficult to reach since they are mixed with the MPL powder. Figure 24. LSV curves of PtS_GDE and PtCom_GDE in an Ar-saturated 0.5 M H2SO4 solution. Regarding the kinetic parameters of both electrodes, the Tafel slope for the HER reaction was extracted from the Tafel curves shown in Figure 25. At an acidic pH, the HER occurs through two reaction pathways: Volmer Heyrovsky or Volmer-Tafel. For polycrystalline Pt, the reported values are over 30–40 mV dec−1, which corresponds to the Volmer-Tafel pathway [262,263]. Tafel slopes for PtS_GDE and PtCom_GDE are 32.8 and 58.8 mV dec−1, respectively, which are in concordance with the reported values for polycrystalline Pt. In the case of the PtCom_GDE, the slope is slightly higher than for PtS_GDE, meaning that the HER is happening at a slower pace in the case of the commercial sample. As Pt nanoparticles are mixed with the carbon MPL, this could hinder the access of the reactants to the active sites and the release of gas bubbles, slowing the reaction rate [251]. 116 Figure 25. Tafel plots of PtS_GDE and PtCom_GDE. Since the results of the ex-situ tests indicated that the Pt sputtering electrodes performed better than the commercial ones, in situ tests in electrolysis cells were conducted. Measurements were carried out as described in Section 2.3 in a self-designed single cell. After the 24 h preconditioning, polarization V-i curves were performed every 24 h, applying stationary current densities of 1.5 A cm−2 between each curve. 117 Figure 26 shows the recorded V-i curves at 0 h, 24 h, 72 h, and 168 h for PtCom_GDE (Figure 26a) and PtS_GDE (Figure 26b). Figure 26. Polarization V-i curves at t 0 h, 24 h, 72 h, and 168 h of (a) PtCom_GDE and (b) PtS_GDE. When using the PtCom_GDE sample as the cathode, t = 0 h curve, the potential corresponding to a current density of 1 A cm−2 is 1.85 V. This 118 potential value increases up to 1.93 V after 168 h of operation. For the PtS_GDE, the initial t = 0 h V-i curve shows slightly higher voltages at the second half of the current density sweep compared to the rest of the polarization curves. Nevertheless, the voltage stabilizes completely after the first polarization curve. Figure 27 compares the V-i curves at 0 h and 168 h for both samples, demonstrating that, under the same conditions, the PtS_GDE electrode is more stable with time than the PtCom_GDE sample. Both start with 1.85 V at 1 A cm−2; however, at t = 168 h, the potential value is maintained nearly the same for the PtS_GDE, whereas for the PtCom_GDE electrode, the potential has increased over time. Figure 27. Comparison of polarization V-i curves at t 0 h and 168 h of PtCom_GDE and PtS_GDE. This voltage variation is further observed in stationary current measurements. To estimate the degradation rate, operation potentials at stationary current densities were measured for 2 h every 20 h over a total period of 100 h. From the slope of the curve, the degradation rate can be extracted. Figure 28 presents the V vs. t graphic of PtCom_GDE and PtS_GDE. The same tendencies as seen in V-i curves are observed, 119 hence proving the stability of the PtS_GDE. The slope extracted from the linear regression of the PtCom_GDE measurements correspond to a degradation rate of 700 μV h−1 at an operation current density of 1.5 A cm−2. Meanwhile, the degradation rate for the PtS_GDE is completely negligible, as the line is almost flat. Figure 28. Comparison of voltage evolution with time of both samples. Summarizing all the results, it is evident that reducing the amount of Pt is not hindering the cell performance. In fact, using MSGA as a deposition method could enhance cell performance not only because all of the catalytic material is on the MPL surface but also because the deposited nanoparticles could be better attached to the surface as a consequence of an adhesion improvement. One of the major issues of Pt/C catalysts is the lack of adhesion and nanoparticle agglomeration, which negatively impacts performance [264,265]. 4.6.4. Conclusions Pt-sputtered electrodes have been successfully developed by gas aggregation magnetron sputtering to study the effect of Pt loading reduction in the performance of a PEM single cell. The surface 120 morphology and composition study confirmed that the catalytic surface was formed by spherical 4 nm pure Pt nanoparticles. The achievement of a 67% reduction in Pt compared with a 0.3 mg cm−2 Pt/C commercial electrode without compromising cell performance content. Although the conditions used in this work need to be transferred and adapted to a magnetron sputtering process, this work proves that magnetron sputtering is a promising method to precisely reduce PGM loadings. Finally, from the obtained results, two main conclusions can be drawn: The amount of Pt is not directly related to an improvement in cell performance. The exposition of Pt active sites seems to be much more relevant than the Pt present in the catalytic coating. MSGA is confirmed as a practical and feasible manufacturing method for developing Pt HER electrodes. This technique successfully produces effective, stable, and effective electrodes in a step process, directly depositing pure Pt nanoparticles on the GDL. Considering the obtained results, it can be confirmed that magnetron sputtering is a feasible manufacturing method for HER low Pt loading electrodes, enabling the reduction of Pt without worsening cell stability or affecting performance. Therefore, magnetron sputtering technology could be a promising technique for the large-scale production of competitive PEMWE electrodes. 121 Capítulo 5: Desarrollo de recubrimientos de BiVO4 y BiVO4/WO3 por magnetron sputtering para su uso como fotoánodo en un sistema fotoelectrocatalítico de reducción de CO2 a metano Pendiente de publicación 5.1. Introducción Combinando los dos tipos de catálisis estudiadas en los Capítulos 1 y 2, la fotocatálisis y la electrocatálisis, nos encontramos con un proceso catalítico híbrido llamado fotoelectrocatálisis (PEC). Este tipo de sistemas parten de la misma base que un sistema fotocatalítico, en el cual tenemos un material semiconductor que al absorber luz generan huecos (h+) y electrones (e-). Estas cargas se separan y se mueven hacia la superficie del catalizador donde pueden llevar a cabo reacciones redox. Hasta este momento, el sistema PEC funciona como un sistema fotocatalítico. A partir de aquí es donde entraría la parte electrocatalítica, ya que se aplica un potencial externo, el cual ayuda a la separación de las cargas generadas, reduciendo así su tasa de recombinación. Para que esto pueda ocurrir, necesitamos tener dos electrodos (un ánodo y un cátodo) sumergidos en un electrolito líquido o en contacto con un electrolito sólido como puede ser el Nafion. A diferencia de un sistema fotocatalítico, donde las reacciones tienen lugar en el mismo catalizador, en el sistema PEC estas reacciones tienen lugar en electrodos diferentes. En resumen, el objetivo de los sistemas PEC es usar los electrones excitados generados cuando el material semiconductor absorbe la luz solar para llevar a cabo una reacción electroquímica redox y obtener así un producto determinado. 122 En función del semiconductor, sea tipo p o tipo n, van a poder actuar más eficientemente como fotoánodo o fotocátodo. Los semiconductores de tipo n, al tener exceso de electrones, suelen emplearse como fotoánodos y moverán la corriente en una dirección determinada. Mientras que los semiconductores de tipo p tienen un mayor carácter fotocátodico, generando huecos que moverán la corriente en dirección opuesta. Actualmente, existen diversas configuraciones PEC, cada una con sus ventajas y desventajas, pero en este trabajo nos centraremos exclusivamente en la configuración fotoánodo/cátodo. Esta configuración emplea semiconductores tipo n (TiO2, WO3, BiVO4…) que son muy abundantes y estables. En la Figura 1 se puede observar un ejemplo de este sistema PEC con una configuración fotoánodo/cátodo. Figura 1. Sistema genérico PEC fotoánodo/cátodo [266]. Los sistemas PEC tienen aplicaciones muy versátiles y de gran interés medioambiental ya que generalmente se pueden emplear en procesos de degradación de contaminantes orgánicos en aguas [267–270], producción de H2 verde [271–274], síntesis de hidrocarburos sintéticos o revalorización de productos orgánicos [275–279], todos ellos 123 alimentados con una mezcla de energía solar directa y fuentes de energía renovables. Pero este tipo de sistemas están aún en TRLs muy bajos, ya que no solo es una tecnología novel si no que su funcionamiento es complejo [280]. Para poder avanzar en el desarrollo de los sistemas PEC e incluso llegar a su industrialización, hay dos puntos clave en los que es necesario invertir esfuerzos: • La mejora de la eficiencia los catalizadores. • Fabricación de estos catalizadores mediante un método simple, versátil y escalable a nivel industrial. La mejora de los catalizadores se puede abordar de distintas formas, muchas de las cuales ya se han expuesto en los Capítulos 1 y 2. En este tipo de sistema PEC tenemos una parte fotocatalítica (el fotoánodo), la cual se va a componer de semiconductores de tipo n, donde son esenciales parámetros como el bandgap o la tasa de recombinación. Las estrategias de mejora de estos compuestos pasan por la nanoestructuración [281,282], el dopaje [283,284]o la formación de heterouniones de diverso tipo [285,286]. Tanto este tipo de óxidos metálicos, como las estrategias para mejorar estos compuestos, se han abordado mediante el uso de magnetron sputtering para diversos sistemas PEC [287–292]. En la parte metálica del sistema PEC (el cátodo) es donde vamos a obtener nuestro producto final, con lo cual factores como la selectividad y área superficial son claves para tener un alto rendimiento en nuestra reacción. Para ello se suele recurrir a compuestos bi o trimetálicos [293,294], ingeniería de caras cristalográficas [293,295] o la nanoestructuración [296,297]. Como veremos más adelante, esta selectividad también se puede modular en función del potencial externo que se aplique. 130 (principalmente Cu), el CO puede ser reducido aún más para formar hidrocarburos [313]. Esta cantidad de posibles productos, sumado a la competencia que existe en el cátodo con la reacción de evolución de H2, hace que sea complejo modular la selectividad de este tipo de sistemas. Por esto, la elección adecuada de catalizadores y parámetros de operación (pH, potencial) es crucial para obtener un determinado producto de forma selectiva y eficiente. Para poder evaluar la selectividad y eficiencia de un sistema PEC existen diversos parámetros [314]: Eficiencia faradaica (FE): es la ratio de cargas transferidas al producto final en relación a la carga total que pasa por el circuito. Se utiliza como una medida de la selectividad de los catalizadores. En la ecuación de arriba n es el número de moles de producto, y es el número de electrones necesarios para convertir el CO2 en ese producto, Q es la carga eléctrica calculada y F la constante de Faraday. Solar to fuel (STF): relaciona la cantidad de combustible producido con la potencia aplicada. Se considera una forma de representar la eficiencia de conversión. Donde rfuel es la cantidad de combustible producida por segundo, ∆G0 es la energía libre de Gibbs, Psolar es la densidad de potencia de la luz (AM 1.5G) y Area es el área expuesta del fotocatalizador a la luz solar. 131 En este trabajo emplearemos el término Energy to fuel (ETF), ya que la luz irradiada no comprende el total del espectro solar. 5.2.2. Materiales empleados como fotoánodo en sistemas PEC Para poder alimentar de eal cátodo, en el fotoánodo es necesario contar con semiconductores de tipo n, que tengan un badgap, con una banda de valencia de energía mayor que 1.23 V (vs RHE) para poder llevar a cabo la oxidación del agua. Estos materiales deben ser estables en medio acuoso, resistentes a la corrosión y buenos transportadores de carga. Hay que tener en cuenta que, en función del pH del medio de reacción, algunos materiales serán más estables que otros. Encontrar un solo material que cumpla estas condiciones es casi imposible, por lo tanto, el desarrollo de estrategias para la mejora de la eficiencia de los fotoánodos es fundamental para poder obtener sistemas PEC funcionales. El semiconductor tipo n por excelencia es el TiO2, siendo el material empleado como fotoánodo en el primer sistema PEC [315]. Es muy estable en un amplio rango de pH, abundante y barato, con un bandgap (3.0-3.2 eV) que es apto para llevar a cabo la reacción de OER. Pero a pesar de haber sido estudiado para su uso como fotoelectrodo, también en sistemas PEC de reducción de CO2 [290,316–320], su bandgap tan elevado limita su habilidad para absorber luz solar, ya que solo absorbe luz UV. Por esta razón, además de su alta tasa de recombinación de cargas, las fotocorrientes obtenidas por fotoánodos de TiO2 son limitadas. Otro semiconductor muy estable, abundante y que si absorbe en la parte visible del espectro es el WO3. Su bandgap es menor que el del TiO2 (2.6-2.8 eV) lo que le permite producir fotocorrientes teóricas más elevadas [321]. Además, tiene muy buena movilidad electrónica y una longitud de difusión de h+ media (150 nm) [322–325]. Una de las limitaciones del WO3 es su baja estabilidad a pH básicos, además de 132 tener un bandgap ligeramente elevado. Su estabilidad a pH más ácidos lo convierte en un material más idóneo para la OER en sistemas PEC de producción de H2 más que para reducción de CO2, aunque se ha empleado en ambos sistemas como fotoánodo [326–329]. Un semiconductor no tan común, pero con propiedades muy interesantes, es el BiVO4. Este compuesto posee tres polimorfos, de los cuales el monoclínico (m-BiVO4) es el más activo para la OER [330,331]. Su bandgap es alrededor de 2.4 eV (VB 2.4 V vs RHE), teniendo un sobrepotencial lo suficientemente alto para producir h+ que lleven a cabo la oxidación del agua al absorber luz solar, alcanzando un máximo teórico de fotocorriente 7.4 mA cm-2 (AM 1.5 G) [331–333]. A su vez, el rango de pH en el cual es relativamente estable es de 3 a 14, siendo preferible un medio acuoso neutro [334–336]. A pesar de sus ventajas, este compuesto tiene limitaciones, principalmente debido a una mala movilidad electrónica (0.02-0.044 cm2 V-1 s-1), una longitud de difusión de emuy corta (10 nm), lo que aumenta la recombinación de cargas considerablemente, y una cinética de reacción para la OER compleja [337–339]. Además, aunque es estable en un amplio rango de pH, esa estabilidad necesita mejorarse si se quieren obtener tiempos de operación óptimos. Aun así, este material es uno de los fotoánodos más prometedores tal y como se muestra en bibliografía, tanto para sistemas PEC de producción de H2 y de reducción de CO2 [340–343]. Como se ha comentado en el Capítulo 3 de esta Tesis, para mejorar las propiedades de un material fotocatalítico se pueden llevar a cabo numerosas estrategias: nanoestructuración, dopaje, uso de cocatalizadores o formación de heteroestructuras, siendo esta última en la que se centra este trabajo. 5.2.3. Formación de heterouniones de BiVO4 Una heterounión corresponde a la unión interfacial de dos materiales, por lo general semiconductores, en la cual bandas de conducción con 133 niveles de energía distintos se encuentran muy próximas. La formación de este tipo de estructuras favorece la movilidad electrónica entre ellas y disminuye la tasa de recombinación de los transportadores de carga, favoreciendo así la eficiencia fotocatalítica de los materiales que forman parte de esa heterounión. El tipo de heterouniones más comunes en sistemas fotoánodo/cátodo son las de tipo II (Figura 3). Figura 3. Esquema de una heterounión tipo II. En este tipo de heterouniones, uno de los semiconductores posee una banda de conducción con niveles de energía más negativos que el otro, favoreciendo así el flujo de edel nivel más positivo al nivel más negativo. El otro semiconductor, en cambio, tiene una banda de valencia más positiva, facilitando el flujo de h+ en el sentido opuesto a los e-. De esta forma se mejora la actividad catalítica del sistema, ya que los transportadores de carga están disponibles durante más tiempo para llevar a cabo reacciones de oxidación en el fotoánodo (en el caso de los h+) o de reducción al moverse hacia el cátodo (e-) [344,345]. El uso de heterouniones de tipo II es común a la hora de diseñar fotoelectrodos para sistemas PEC. En sistemas de reducción de CO2 134 podemos encontrar diversas heterouniones de BiVO4 con otros semiconductores como pueden ser el ZnO, el TiO2, o el WO3 [346–351]. Por ejemplo, el TiO2 es un semiconductor muy estable, abundante y fácil de sintetizar, pero su uso como fotoánodo está principalmente limitado por su alta recombinación de cargas y su bandgap, que solo le permite absorber luz UV. Con lo cual, combinarlo mediante una heterounión con un semiconductor que absorba en el visible como el BiVO4, es una excelente opción para aprovechar su gran capacidad como aceptor de electrones y su facilidad para transferir estas cargas. El BiVO4 al tener una movilidad electrónica pobre, puede ver mejorada su eficiencia mediante la formación de una heterounión con el TiO2. Aunque existe cierta ambigüedad sobre el posicionamiento de las VB y CB a la hora de formar la heterounión, varios trabajos han probado la eficacia de esta estrategia para aumentar la separación de cargas y bajar la tasa de recombinación [344]. En cambio, en el caso del WO3, la posición de sus bandas es muy favorable para la formación de una heterounión tipo II ya que las CB del WO3 y del BiVO4 son 0.42 y 0.02 V vs RHE y las VB son 3.12 y 2.42 V vs RHE respectivamente. Esto en el caso de la CB favorece la inyección de edesde del BiVO4 al WO3 y la posición de la VB también va a favorecer el flujo de h+ del WO3 al BiVO4. Además, como el WO3 tiene mayor movilidad electrónica (12 cm2 V-1 S-1), este aporta una mejora en la eficiencia a la hora separar cargas y transportar electrones, reduciendo considerablemente la recombinación [325]. Mientras que BiVO4, al tener un bandgap menor, tiene mayor capacidad de absorción de luz solar, generando más pares de cargas y su longitud de difusión de huecos también es menor (70-100 nm), siendo este más eficaz como fotocatalizador para la OER [352,353]. 135 Figura 4. Esquema de un sistema PEC con un fotoánodo de BiVO4/WO3 [354]. Estas características postulan al BiVO4/WO3 como una de las heterouniones más eficientes para su uso como fotoánodo en un sistema PEC. Esto se confirma en múltiples trabajos de investigación en donde se ha podido probar el aumento de la estabilidad y densidad de corriente en estos fotoánodos una formada la heterounión entre el WO3 y el BiVO4. 5.2.4. Fotoánodos de BiVO4 desarrollados por magnetron sputtering Aunque es cierto que la síntesis de hidrocarburos mediante reducción de CO2 en sistemas PEC está lejos de ser industrializable, si es una tecnología de elevado interés para la descarbonización de la producción de combustibles sintéticos, ya que combina el uso de luz solar con el uso de CO2 atmosférico, permitiendo la obtención de diversos combustibles sin generar emisiones contaminantes. La mejora de los fotocatalizadores es el punto clave para poder avanzar en el desarrollo de esta tecnología, ya que actualmente aún quedan por 136 resolver problemas de estabilidad y eficiencia ligados a los materiales que se emplean. Parte de la solución para reducir las limitaciones ligadas a los fotocatalizadores pasan por emplear diversos compuestos, con lo cual es fundamental contar con una técnica de fabricación práctica y versátil. De esta forma se facilitaría tanto la mejora de los propios fotocatalizadores como su proceso sintético. Al igual que la mayoría de los compuestos catalíticos, el BiVO4 y sus correspondientes heterouniones, se sintetizan mediante diversos métodos químicos húmedos combinados con técnicas de deposición como el spin-coating o el pulverizado pirolítico, siendo la descomposión de compuestos metalorgánicos (MOD) el método más empleado [355,356]. Estos métodos son tediosos, constan de varias etapas sínteticas seguidas de una etapa final de deposición sobre el sustrato, etapas que se ven multiplicadas a la hora de necesitar sintetizar compuestos diferentes. Las tecnologías de PVD, en concreto el magnetron sputtering, podría ser la opción idónea para desarrollar este tipo de recubrimientos fotocatalíticos. Mediante el sputtering es posible depositar capas de diferentes compuestos en un solo paso, depositados en el sustrato final que se puede incorporar de forma directa al sistema PEC. Es un proceso a alto vacío, lo que asegura que las impurezas sean inexistentes o mínimas. Además, nos permite controlar parámetros de proceso con elevada precisión lo que se traduce en un control preciso de las capas desarrolladas. Todas estas características hacen del magnetron sputtering un método muy versátil, adaptable a la deposición de una gran variedad de compuestos y estructuras. Dado que la propia reducción de CO2 en sistemas PEC es un campo que está muy poco desarrollado, la bibliografía existente sobre desarrollo de fotoánodos mediante magnetron sputtering no es muy extensa. Hay varios ejemplos de deposición de BiVO4 mediante sputtering para 137 aplicaciones fotocatalíticas. Se llevó a cabo la deposición mediante sputtering por RF de BiVO4 para comprobar el efecto del post-annealing en la actividad fotocatalítica del BiVO4 para la degradación de la Rodamina 6G, en el cual se comprobó que los defectos presentes en la capa depositada a t.a. favorece la degradación de este compuesto orgánico [357]. En otro estudio, se depositó BiVO4 mediante DC pulsado reactivo, a partir de targets de Bi y V, aplicando distinta densidad de potencia a cada blanco metálico, ya que estos dos componentes no tienen la misma ratio de pulverización. Se evaluó la actividad fotocatalítica en función del espesor y del pH, para la degradación de la Rodamina B, donde la capa con un espesor de 700 nm y menor tamaño de cristal resultó en la capa con mayor capacidad de degradación [358]. Ya enfocado en emplear el BiVO4 como fotoánodo en un sistema PEC de electrólisis de H2O, se desarrolló un fotóanodo de BiVO4 mediante magnetron sputtering usando una fuente RF y partiendo directamente de un blanco de BiVO4. A la hora de estudiar su fotoactividad para la OER, la densidad de corriente resultó ser muy baja (0.03 mA cm-2), lo que solventaron depositando distintos óxidos metálicos (ZnO, MoO3 y Fe2O3) para formar heterouniones tipo II. De estos tres óxidos metálicos, el mejor resultado se obtuvo con el MoO3, llegando a una densidad de corriente de 0.22 mA cm-2, aumentado el número de portadores de carga y bajando su resistencia al contacto eléctrico [359]. Más recientemente, se preparó mediante sputtering DC reactivo, partiendo de blancos de Bi y V a los que se les aplico distintas densidades de potencia, un fotoánodo de BiVO4 con un espesor de 345 nm. Este fotoánodo se irradió con luz solar (AM 1.5 G) para llevar a cabo una voltametría lineal, en la cual para un potencial de 1.23 V vs RHE la densidad de corriente obtenida alcanzó los 2.1 mA cm-2. Para mejorar su actividad fotocatalítica, se depositó una capa de NiFeOx, llegando a aumentar el doble su densidad de corriente al mismo potencial [360]. 138 Respecto a la fabricación de heterouniones de BiVO4/TiO2 o BiVO4/WO3, desarrollados por magnetron sputtering los ejemplos son más escasos. Se ha hecho algún trabajo en sistemas PEC para electrólisis de agua en el que o el TiO2 [361] o el WO3 [362,363] eran desarrollados por sputtering formando parte de una heterounion de BiVO4 y otros compuestos. En el caso del BiVO4/WO3, se ha desarrollado un fotoánodo mediante sputtering DC reactivo, depositados en dos procesos diferentes ya que la capa de WO3 fue calcinada después de su deposición. En la caracterización fotoelectroquímica se pudo observar que la formación de la heterounión entre en WO3 y el BiVO4 aumentaba la fotocorriente generada al irradiar las muestras, probando así el efecto beneficioso de esta unión [364]. También se sintetizó un fotoánodo de ITO/WO3-NRs/BiVO4 completamente por sputtering, en el mismo proceso, sin romper vacío. Los nanotubos de WO3 fueron depositados mediante sputtering de ángulo rasante (GLAD, glancing angle deposition) y el BiVO4 directamente por RF. Al llevar iluminar la muestra con luz solas Am 1.5 G se alcanzaron corrientes de 3.2 mA cm-2 a 1.23 V vs RHE. [365]. En mismo proceso y mediante magnetron sputtering DC para generar capas de ITO/BiVO4 y ITO/WO3/BiVO4 se estudió la influencia del tipo de iluminación (frontera o trasera) en la densidad de corriente fotoanódica, con el fin de vislumbrar como afectaba a la movilidad de cargas. Las capas de BiVO4 se obtuvieron a partir de un blanco de Bi2O3 y V, y comprobaron la influencia de un exceso de Bi o de V en el rendimiento del fotoánodo [366]. Aunque el magnetron sputtering no sea un método ampliamente utilizado para este tipo de aplicaciones, si hay indicios de que este método de deposición es prometedor para poder desarrollar fotoelectrodos de forma automatizada y en un solo proceso. 5.2.5. Producción de metano en sistemas PEC como tecnología ISRU (Proyecto HISRU) 139 Dentro del proyecto HISRU surgió la oportunidad de desarrollar un sistema fotoelectrocatalítico en el cual se llevaría a cabo la producción de metano aprovechando tanto el CO2 de la atmósfera marciana (98%) como las aguas grises que se generan dentro de las estaciones espaciales. Básicamente, el objetivo principal era el desarrollo de un sistema PEC de reducción de CO2 acoplado con el uso y tratamiento previo de las aguas grises. El interés de este proyecto está por un lado en el aprovechamiento de los recursos in situ (ISRU) en el espacio, tanto del CO2 como de las aguas grises y por otro en el desarrollo de un sistema muy novedoso de producción fotoelectroquímica de metano. En este proyecto, se decidió proponer el desarrollo de fotoánodos por magnetron sputtering, ya que era una oportunidad de probar la versatilidad del método, tanto por el tipo de fotoánodos como por la tecnología en sí. Además, al realizar una búsqueda de trabajos publicados, no se encontró ningún estudio que evaluase la viabilidad del magnetron sputtering como técnica de fabricación de fotoelectrodos en sistemas PEC de reducción de CO2. Con lo cual, teniendo en cuenta que esta Tesis Doctoral tiene como línea de investigación general el desarrollo de catalizadores por sputtering, el hecho de desarrollar fotocatalizadores tan innovadores para un sistema tan poco desarrollado es de gran de interés para este trabajo de investigación. Por lo tanto, se decidió fabricar fotoánodos de BiVO4, BiVO4/TiO2 y BiVO4/WO3, caracterizar sus propiedades fisicoquímicas y hacer una evaluación de su comportamiento en un sistema PEC. Ceramics International 49 (2023) 19309–19317 19311 and thickness studies were carried out using an ULTRA Plus Carl-Zeiss field emission scanning electron microscope (FESEM) and Solver PRO NT-MDT atomic force microscope (AFM). Raman Spectroscopy was performed with a Raman Renishaw for each sample. The laser used for the acquisition of the spectra operated at a wavelength of 514 nm with a power of 10 mW. The optical characterization of the coatings deposited on glass was performed using an Ocean Optics USB4000 UV–Vis spectrometer. The oxidation state information of the samples was obtained by X-ray photoelectron spectroscopy (XPS), performed with a SUPRA photoelectron spectrometer (Kratos Analytical Ltd.) equipped with Mg K X-rays as the primary excitation source. The binding energy was referenced to the C 1s line at 284.8 eV for calibration, and a Gaussian function was applied for curve fitting with a Shirley background. 2.3. Photocatalytic activity measurements for methylene blue degradation The photocatalytic activity of the samples was assessed by methylene blue (MB) degradation tests. Each 2,5 ×2,5 cm 2 sample, cut from the original samples, was placed in a quartz cuvette containing 50 mL of methylene blue solution (from Alfa Aesar) with a concentration of 2 μ molL −1 . Before testing, the samples were placed in the dark for 1 h under continuous stirring to reach the adsorption/desorption equilibrium. The samples were then irradiated for 1 h with UV-A sources (15 W Sankyo Denki BB lamps). The main peak of the UV source was measured at 365 nm. An Ocean Optics USB4000 UV–Vis spectrometer was used to monitor the main methylene blue absorption peak (664 nm). Finally, the pseudo-first-order rate constants k α were obtained for each sample by plotting Ln(A 0 /A t ) against time in order to compare the different photocatalytic activities [49,50]. 2.4. Stability measurements The stability of both NiO/TiO 2 samples was studied by repeating the degradation test of MB as described in Section 2.3. After each measurement, the samples were cleaned with distilled water and dried with compressed air and the operation was repeated with both samples nine more times. 3. Results and discussion 3.1. Characterization FESEM was used to determine the morphology of the coatings. Fig. 2 (a, b) shows the columnar porous structure of the TiO 2 coating, which is similar in appearance to the porous structure described in the Thornton Structure Zone model [51]. The measured thickness of the TiO 2 coating is around 740 nm with column diameters ranging from 60 nm to 160 nm (Figure S3). For the NiO films, only the 20 nm thick film could be Fig. 2. SEM images of (a, b) TiO 2 (c, d) 10 nm NiO/TiO 2 and (e, f) 20 nm NiO/TiO 2 coatings over a silicon substrate. A. Villamayor et al. Ceramics International 49 (2023) 19309–19317 19312 measured through FESEM (Figure S2) as the 10 nm layer did not form a continuous film and could not be measured this way. Regarding the morphology of the NiO coatings, the 10 nm NiO (Fig. 2 c, d) layer changes subtly the surface of the columns TiO 2 , with the appearance of a small cluster-like structure that can be seen on top of these columns, although no changes are apparent in terms of film porosity. When the thickness is increased from nominally 10 nm–20 nm the surface morphology changes are much more relevant (Fig. 2 e, f). The 20 nm NiO layer forms an almost continuous film over the TiO 2 columns with a cauliflower-like morphology, reducing its porosity remarkably. The reason for film densification could be the change of the sputtering power from 25 W to 50 W, because not only is the deposition rate increased at higher powers, but the incident sputtered particles arrive at the film surface with higher energy. The surface and thickness of the NiO films were further analysed by atomic force microscopy (AFM). The thickness of both NiO films was obtained by step height measurements in different regions of the coatings. The average thicknesses of the samples were 23,8 ±3,6 nm and 8,7 ±1 nm (Figure S1), respectively. Through AFM analysis, the roughness of the NiO films was also investigated. Average roughness (Ra) values, which correspond to the arithmetic average of the deviations of a height profile compared to a mean line and root-mean-square roughness (Rq), which is the quadratic average of a height profile compared to a mean line were the parameters selected for roughness evaluation. 3D-AFM images of the 10 nm and 20 nm NiO samples, deposited at a sputtering power of 25 W and 50 W, respectively, are depicted in Fig. 3a and b. When sputtering power was increased a smoother surface was formed, lowering the roughness of the 20 nm NiO film compared to the 10 nm one, as indicated by the Ra and Rq values of 7,4 nm and 9,4 nm for the 20 nm film, compared with the 10 nm layer Ra and Rq values of 10,2 nm and 12,7 nm (Table S1). These values can be used as a comparison between the NiO films, but not as a quantitative measure due to the discontinuity of the thinner NiO layer and the porous nature of the underlying TiO 2 surface. XRD measurements were performed for the identification of the crystalline structure of the TiO 2 and NiO/TiO 2 films (Fig. 4.). The TiO 2 coatings were analysed after annealing in air for 2 h at 450 ◦C, showing only anatase peaks at 2θ of 25.3◦(101), 36.9◦(103), 37.9◦(004), 38,5◦ (112), 48.1◦(200), 53.9◦(105), 55.1◦(211), 62,7◦(204), 68.9◦(116), 70.3◦(220), 75,1◦(215), 76◦(301). All those peaks were identified with the JCPDS card 96-720-6076. Regarding the NiO/TiO 2 coatings (Fig. 4) no characteristic peaks related to NiO species were observed in the X-ray traces, most likely because the coatings were too thin. Additionally, it can be seen that the anatase TiO 2 structure did not measurably change with the addition of the NiO coatings. The surface elemental composition and electronic state of the elements of the coatings were determined by X-ray photoelectron spectroscopy (XPS). The overall XPS spectra are presented in Fig. 5a, showing the presence of Ti 2p, O 1s, C 1s and Ni 2p peaks. Carbon peaks can be explained by the presence of adventitious carbon on the surface. A high-resolution XPS spectrum for Ti 2p is shown in Fig. 5b. Two main peaks can be attributed to Ti 2p 1/2 and Ti 2p 3/2 , which are in good agreement with literature values for the Ti 4+ valence state present in TiO 2 [52]. As can be observed, the intensity of the signal decreases for TiO 2 –Ni due to the presence of a NiO coating. Through XPS, the formation of NiO instead of Ni was confirmed. As discussed earlier, the Ni was sputtered in metallic mode but instead of Ni metallic films, NiO films were obtained. A high-resolution XPS spectrum for the Ni 2p region is shown in Fig. 6. This spectrum can be divided into two parts: the first one on the left is composed of two peaks located at 879.4 eV and 872.8 eV, which can be attributed to the satellite peak and the Ni 2p 1/2 peak, respectively. The second part of the spectra exhibits a satellite peak located at 860.5 eV and the Ni 2p 3/2 peak located around 856 eV. Both of Ni 2p peaks can be deconvoluted in several peaks, confirming the presence of several oxidation states for Ni in the coating: Ni 3+ (with peaks located at 874.5 eV and 855.5 eV), Ni 2+ (with peaks located at 873.8 eV and 854.2 eV) and N 0 (with peaks located at 872.3 and 852.6 eV). The latter one is composed of several peaks, showing different oxidation states for Ni. The peaks attributed to the presence of the Ni +3 oxidation state can be correlated to the high amount of oxygen adsorbed (531.2 eV), revealing the presence of NiOOH [53] on the surface of the sample. The peaks attributed to the presence of Ni +2 show the formation of the NiO layer. Finally, the presence of Ni 0 (552.6 eV) shows that some of the Ni on the coating remains metallic [54]. XPS spectra of O 1s are shown in Fig. 7. For TiO 2 samples, the O 1s main peak can be deconvoluted in three Fig. 3. 3D-AFM images of (a) 10 nm NiO/TiO 2 and (b) 20 nm NiO/ TiO 2 coatings. Fig. 4. XRD spectra of anatase TiO 2 and NiO/TiO 2 coatings. A. Villamayor et al. Ceramics International 49 (2023) 19309–19317 19313 individual peaks: two peaks at 533.2 eV and 531.7 eV that can be attributed to the presence of adsorbed oxygen species (such as O 2 , H 2 O, OH − ) [55,56]. And a third peak is attributed to the Ti–O bonds from the TiO 2 lattice. For NiO/TiO 2 samples, those 3 peaks can also be found, with the latter peak attributed to the Ni–O bonds from the NiO lattice. The valence-band (VB) spectra of TiO 2 and NiO/TiO 2 samples are shown in Fig. 8. By extrapolating the valence band spectrum using linear fitting for each sample, the valence-band maximum (VBM) values were measured. For the TiO 2 sample, a VBM of 2.8 eV was measured. This value is similar to values that can be found in the literature [57]. When the NiO layer is deposited on top of the TiO 2 coatings, the VBM measured shifts from 2.8 eV to 0.40 eV, which is in accordance with the formation of a p-type NiO layer [54]. In addition to the identification of the oxidation states of the elements, elemental quantification was performed on the samples. It must be noticed that a high concentration of C 1s was measured during the quantification, certainly due to impurities during the measurement (Tables S2 and S3). The Raman spectra shown in Fig. 9 matches the reported spectra of anatase found in the literature [58,59], with specific peaks located at 144, 397, 516, and 638 cm −1 . It can be also noted that the intensity of the peaks decreases with NiO layer thickness but changes are not observed in the anatase structure [57]. To determine the optical band gap, the transmittance of the TiO 2 and NiO/TiO 2 films was measured between 200 nm and 850 nm wavelength range. For the band gap estimation, assuming an indirect band gap, the absorption coefficient, α , was obtained using Equation (1), where t is the thickness of the film. T≈e− α t(1) Using the Tauc plot method [60], Eg was calculated using the absorption coefficient using Equation (2), by plotting ( α h ν ) 1/2 as a function of h ν and extrapolating the linear region to the abscissa. α h ν =C(h ν −Eg)1/n(2) where α is the absorbance coefficient, h is the Planck constant, ν is the frequency of vibration, Eg is the band gap energy and n is the value for TiO 2 indirect allowed transition, which is equal to 2 [61]. As can be seen in Fig. 10, almost no change in terms of optical band gap can be observed when the NiO layer is deposited on the top of the TiO 2 , which suggests Fig. 5. XPS spectra of (a) all samples and (b)Ti 2p peaks for TiO 2 . Fig. 6. XPS spectra of Ni 2p for NiO/TiO 2 coatings of (a) 10 nm and (b) 20 nm. A. Villamayor et al. Ceramics International 49 (2023) 19309–19317 19314 that a p-n heterojunction is being formed. 3.2. Study of photocatalytic degradation of methylene blue The degradation of methylene blue (MB) was studied for the coatings on glass substrates for 1 h, with and without the NiO layer. Initially, the adsorption-desorption equilibrium of MB for the coatings was determined by keeping the samples immersed in a MB solution with constant stirring for 1 h without illumination. As shown in Fig. 11a, using the natural logarithm of the relative absorbance (Ln[A/A 0 ]), plotted vs. time, a pseudo-first-order rate constant (k) is obtained for the reaction [5]. Fig. 11b shows the degradation rate of methylene blue, A/A 0, with respect to time. All the experimental samples are compared to a commercial Pilkington Activ glass sample. Although this commercial product is produced using a different process, there is a lack of standards to compare to in this field and the Activ sample is acknowledged as providing a useful purpose in this case [62]. From these results, it can be concluded that TiO 2 presents photocatalytic activity under UV light with a k constant around 2 ×10 −5 s −1 Fig. 7. XPS spectra of O1s for (a) TiO 2 and (b) 10 nm and (c) 20 nm NiO/TiO 2 coatings. Fig. 8. VB spectra of all samples for VBM determination. Fig. 9. Raman spectra of all samples. A. Villamayor et al. Ceramics International 49 (2023) 19309–19317 19315 which agrees with literature reported values [49]. Also, the TiO 2 photocatalytic activity is improved with the addition of NiO coatings, with a more than 4 times improvement for the 10 nm NiO thick layer. Regarding degradation rates for the NiO/TiO 2 coatings, they also are in agreement with literature values reported, the slightly lower values can be attributed to differences in purity, structure or synthesis method [63, 64]. This improvement in MB degradation for NiO/TiO 2 coatings could be due to the reduction of the recombination rate of TiO 2 h + /e − pairs by the presence of NiO and the improved electronic mobility caused by the formation of a p-n heterojunction, where the electrons produced in the conduction band of the p-type semiconductor (here NiO) are moved to the conduction band of the n-type semiconductor (here TiO 2 ). These photogenerated electrons are responsible for reducing adsorbed O 2 to O 2 −⋅ , which is considered a key step in MB oxidative decomposition. On the other hand, the holes generated in the valence band of the n-type semiconductor are promoted to the valence band of the p-type semiconductor, reacting with absorbed H 2 O or OH − , which leads to the formation of strongly oxidative OH•radicals. These radicals are thought to be the main driving force for photocatalytic oxidative reactions, reacting very fast with the MB adsorbed on the NiO surface [65–67]. The different p-n junctions created increase the charge separation efficiency of the material as confirmed by the lower band gap values achieved for pure TiO 2 by UV transmittance measurements. The formation of these radicals due to the formation of a p-n heterojunction is the main reason for the improvement of the charge carriers’ separation and, consequently, photocatalytic efficiency [68,69]. As a future objective to complete this work, we intend to perform photoactivity measurements in the presence of scavengers [49,70]. Regarding the thickness of the NiO coatings, it appears that increasing thickness from 10 nm to 20 nm does not translate into a comparable improvement in photocatalytic activity. This could be because, despite the observed reduction of roughness in the 20 nm layer, which is equivalent to a reduction in surface area, the 20 nm NiO layer gives better coverage of the TiO 2 surface, improving the formation of new p-n heterojunctions. These two contradictory effects may be the reason why there is no remarkable increase in photocatalytic activity when the thickness is increased. The influence of the thickness of NiO layers on photocatalytic activity is a subject that could be interesting to further analyse in future research. The degradation percentage and degradation rate constant of the coatings are summarized in Table 2. Finally, stability measurements were performed with NiO/TiO 2 samples to confirm their photoactivity and stability over ten cycles. The obtained results are shown in Fig. 12. From these measurements, several interesting conclusions can be drawn: (i) A significant reduction between the degradation percentage of the 1st cycle and the 2nd cycle takes place, especially in the 10 nm NiO sample. The most probable reason for this reduction could be that when the 1st measurement was carried out, the adsorption of the MB hadn’t reached equilibrium and was still taking place while the samples were irradiated. This effect would probably be more noticeable in the 10 nm layer, due to its high porosity. For the rest of the cycles the degradation percentage is very stable when compared to the rest of the cycles, thus we can assume that no degradation of the coatings is taking place. (ii) When the adsorption of the MB no longer takes place, the 20 nm NiO coating shows a slightly higher activity when compared with the 10 nm NiO coating. Fig. 10. Optical bandgap of all samples. Fig. 11. (a) The degradation rate of MB for the photocatalytic coatings and (b) determination of rate constants for MB degradation reaction using a first-order kinetic model under UV light. Table 2 Degradation percentage and degradation rate constant under UV for 1 h. Sample R (%) k (s −1 ) Pilkington 14 4.31 ×10 −5 TiO 2 5.8 1.73 ×10 −5 10 nm NiO/TiO 2 25.3 8.22 ×10 −5 20 nm NiO/TiO 2 21.3 6.79 ×10 −5 A. Villamayor et al. Ceramics International 49 (2023) 19309–19317 19316 (iii) Besides the 1st cycle, the coatings show high stability for the remainder of the measurements. 4. Conclusions TiO 2 and NiO/TiO 2 coatings were prepared by DC and pulsed DC reactive magnetron sputtering and, after annealing, their photocatalytic activity was tested for methylene blue degradation. In this work, it was concluded that the incorporation of 10 nm and 20 nm NiO coatings on top of the TiO 2 , modified the surface morphology without significantly changing the bandgap or crystalline structure of the coatings, but clearly resulted in a remarkable improvement of the MB degradation rate due to the formation of a p-n heterojunction that favoured charge carrier separation and electronic mobility. Although both NiO coatings improved the performance of the TiO 2 film, the thicker coating does not have a much bigger effect on activity probably due to the countering effect of reducing surface area and increasing p-n heterojunctions. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This research has received funding from the KMM-VIN Research Fellowship 2021 program, granted to Antía Villamayor (https://www. kmm-vin.eu/fellowships/). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.ceramint.2023.03.058. References [1] K. Niinimamp, G. Peters, H. Dahlbo, P. Perry, T. Rissanen, A. Gwilt, environ. price of fast fash., (n.d.). https://doi.org/10.1038/s43017-020-0039-9. [2] T. Luttrell, S. Halpegamage, J. Tao, A. Kramer, E. Sutter, M. Batzill, Why is anatase a better photocatalyst than rutile? - model studies on epitaxial TiO 2 films, Sci. Rep. 4 (2015) 1–8, https://doi.org/10.1038/srep04043. [3] E. Pelizzetti, C. Minero, Mechanism of the photo-oxidative degradation of organic pollutants over TiO 2 particles, Electrochim. Acta 38 (1993) 47–55, https://doi.org/ 10.1016/0013-4686(93)80009-O. [4] J.M. Herrmann, J. Disdier, P. Pichat, S. Malato, J. Blanco, TiO 2 -based solar photocatalytic detoxification of water containing organic pollutants. Case studies of 2,4-dichlorophenoxyaceticacid (2,4-D) and of benzofuran, Appl. Catal. B Environ. 17 (1998) 15–23, https://doi.org/10.1016/S0926-3373(97)00098-2. [5] A. Houas, H. Lachheb, M. Ksibi, E. Elaloui, C. Guillard, J.M. Herrmann, Photocatalytic degradation pathway of methylene blue in water, Appl. Catal. B Environ. 31 (2001) 145–157, https://doi.org/10.1016/S0926-3373(00)00276-9. [6] A. Naldoni, M. Altomare, G. Zoppellaro, N. Liu, ˇ S. Kment, R. Zboˇ ril, P. Schmuki, Photocatalysis with reduced TiO 2 : from Black TiO 2 to cocatalyst-free hydrogen production, ACS Catal. 9 (2019) 345–364, https://doi.org/10.1021/ acscatal.8b04068. [7] X. Kong, Z. Peng, R. Jiang, P. Jia, J. Feng, P. Yang, Q. Chi, W. Ye, F. Xu, P. Gao, Nanolayered heterostructures of N-doped TiO 2 and N-doped carbon for hydrogen evolution, ACS Appl. Nano Mater. 3 (2020) 1373–1381, https://doi.org/10.1021/ ACSANM.9B02217/SUPPL_FILE/AN9B02217_SI_001.PDF. [8] D. Jia, X. Li, Q. Chi, J. Low, P. Deng, W. Wu, Y. Wang, K. Zhu, W. Li, M. Xu, X. Xu, G. Jia, W. Ye, P. Gao, Y. Xiong, Direct Electron Transfer from Upconversion Graphene Quantum Dots to TiO 2 Enabling Infrared Light-Driven Overall Water Splitting (2022), https://doi.org/10.34133/2022/9781453. [9] A.E. Wiącek, A. Gozdecka, M. Jurak, Physicochemical characteristics of chitosanTiO 2 biomaterial. 1. Stability and swelling properties, Ind. Eng. Chem. Res. 57 (2018) 1859–1870, https://doi.org/10.1021/acs.iecr.7b04257. [10] J.M. Janusz, J.A. Berson, Heterogeneous photocatalytic synthesis of methane from acetic acid—new kolbe reaction pathway, J. Am. Chem. Soc. 100 (1978) 2239–2240, https://doi.org/10.1021/JA00475A049/ASSET/JA00475A049.FP. PNG_V03. [11] E. Santos, A.C. Catto, A.F. Peterline, W. Avansi, Transition metal (Nb and W) doped TiO 2 nanostructures: the role of metal doping in their photocatalytic activity and ozone gas-sensing performance, Appl. Surf. Sci. 579 (2022), 152146, https://doi. org/10.1016/J.APSUSC.2021.152146. [12] N. Sriharan, T.S. Senthil, P. Soundarrajan, M. Kang, Surface modification of TiO 2 nanorods with Mg doping for efficient photoelectrodes in dye sensitized solar cells, Appl. Surf. Sci. 585 (2022), 152719, https://doi.org/10.1016/J. APSUSC.2022.152719. [13] T. Gao, T.-T. Li, X. Liao, J.-H. Lin, B.-C. Shiu, C.-W. Lou, Construction of Cu 2 O/TiO 2 heterojunction photoelectrodes for photoelectrochemical determination of glucose, J. Mater. Res. Technol. 21 (2022) 798–809, https://doi.org/10.1016/J. JMRT.2022.09.043. [14] F. Pinto, A. Wilson, B. Moss, A. Kafizas, Systematic exploration of WO 3 /TiO 2 heterojunction phase space for applications in photoelectrochemical water splitting, J. Phys. Chem. C 126 (2022) 871–884, https://doi.org/10.1021/ACS. JPCC.1C08403/ASSET/IMAGES/LARGE/JP1C08403_0009.JPEG. [15] Q. Chi, G. Zhu, D. Jia, W. Ye, Y. Wang, J. Wang, T. Tao, F. Xu, G. Jia, W. Li, P. Gao, Built-in electric field for photocatalytic overall water splitting through a TiO 2 / BiOBr P–N heterojunction, Nanoscale 13 (2021) 4496–4504, https://doi.org/ 10.1039/D0NR08928A. [16] H. Zhang, J. Cai, Y. Wang, M. Wu, M. Meng, Y. Tian, X. Li, J. Zhang, L. Zheng, Z. Jiang, J. Gong, Insights into the effects of surface/bulk defects on photocatalytic hydrogen evolution over TiO 2 with exposed {001} facets, Appl. Catal. B Environ. 220 (2018) 126–136, https://doi.org/10.1016/J.APCATB.2017.08.046. [17] S.A. Rawool, M.R. Pai, A.M. Banerjee, A. Arya, R.S. Ningthoujam, R. Tewari, R. Rao, B. Chalke, P. Ayyub, A.K. Tripathi, S.R. Bharadwaj, Pn Heterojunctions in NiO:TiO 2 composites with type-II band alignment assisting sunlight driven photocatalytic H 2 generation, Appl. Catal. B Environ. 221 (2018) 443–458, https:// doi.org/10.1016/J.APCATB.2017.09.004. [18] L. Wen, B. Liu, X. Zhao, K. Nakata, T. Murakami, A. Fujishima, Synthesis, characterization, and photocatalysis of Fe-doped TiO 2 : a combined experimental and theoretical study, Int. J. Photoenergy 2012 (2012), https://doi.org/10.1155/ 2012/368750. [19] H. Mour˜ ao, W. Junior, C.R.-M.C. and Physics, undefined, Hydrothermal Synthesis of Ti Oxide Nanostructures and TiO 2 : SnO 2 Heterostructures Applied to the Photodegradation of Rhodamine B, Elsevier, 2012 (n.d.), https://www.sciencedi rect.com/science/article/pii/S025405841200466X. (Accessed 5 January 2023). [20] J. Wang, Z. Jiang, L. Zhang, P. Kang, Y. Xie, Y.L.-U., undefined, Sonocatalytic Degradation of Some Dyestuffs and Comparison of Catalytic Activities of NanoSized TiO 2 , Nano-Sized ZnO and Composite TiO 2 /ZnO Powders, Elsevier, 2009 (n. d.), https://www.sciencedirect.com/science/article/pii/S1350417708001326. (Accessed 5 January 2023). [21] Q. Zhu, N. Liu, N. Zhang, Y. Song, M.S. Stanislaus, C. Zhao, Y. Yang, Efficient photocatalytic removal of RhB, MO and MB dyes by optimized Ni/NiO/TiO 2 composite thin films under solar light irradiation, J. Environ. Chem. Eng. 6 (2018) 2724–2732, https://doi.org/10.1016/j.jece.2018.04.017. [22] P. Brault, A. Caillard, A.L. Thomann, J. Mathias, C. Charles, R.W. Boswell, S. Escribano, J. Durand, T. Sauvage, Plasma sputtering deposition of platinum into porous fuel cell electrodes, J. Phys. D Appl. Phys. 37 (2004) 3419–3423, https:// doi.org/10.1088/0022-3727/37/24/010. [23] W. Liu, S.C.-A.M. Research, undefined, Visible-light Activity Evaluation of Pn Junction Photocatalyst NiO/TiO 2 Prepared by Sol-Gel Method, Trans Tech Publ, 2011 (n.d.), https://www.scientific.net/AMR.152-153.441. (Accessed 5 January 2023). Fig. 12. Stability test of NiO/TiO 2 samples for MB degradation. A. Villamayor et al. Ceramics International 49 (2023) 19309–19317 19317 [24] Y. Chen, J. Crittenden, S.H.-… science & technology, undefined 2005, Preparation of a novel TiO 2 -based p−n junction nanotube photocatalyst, ACS Publ 39 (2005) 1201–1208, https://doi.org/10.1021/es049252g. [25] Y. Zheng, B. Zhu, H. Chen, W. You, C. Jiang, J. Yu, Hierarchical flower-like nickel (II) oxide microspheres with high adsorption capacity of Congo red in water, J. Colloid Interface Sci. 504 (2017) 688–696, https://doi.org/10.1016/J. JCIS.2017.06.014. [26] F. Motahari, M.R. Mozdianfard, M. Salavati-Niasari, Synthesis and adsorption studies of NiO nanoparticles in the presence of H 2 acacen ligand, for removing Rhodamine B in wastewater treatment, Process Saf. Environ. Protect. 93 (2015) 282–292, https://doi.org/10.1016/J.PSEP.2014.06.006. [27] T. Sreethawong, S. Ngamsinlapasathian, S. Yoshikawa, Surfactant-aided sol–gel synthesis of mesoporous-assembled TiO 2 –NiO mixed oxide nanocrystals and their photocatalytic azo dye degradation activity, Chem. Eng. J. 192 (2012) 292–300, https://doi.org/10.1016/J.CEJ.2012.04.006. [28] L. Chu, M. Li, P. Cui, Y. Jiang, Z. Wan, S. Dou, The study of NiO/TiO 2 photocatalytic activity for degradation of methylene orange, Energy Environ. Focus. 3 (2014) 371–374, https://doi.org/10.1166/EEF.2014.1125. [29] C. Yu, M. Li, D. Yang, K. Pan, F. Yang, Y. Xu, L. Yuan, Y. Qu, W. Zhou, NiO nanoparticles dotted TiO 2 nanosheets assembled nanotubes P-N heterojunctions for efficient interface charge separation and photocatalytic hydrogen evolution, Appl. Surf. Sci. 568 (2021), 150981, https://doi.org/10.1016/J.APSUSC.2021.150981. [30] J.H. Yu, H. Yang, R.H. Jung, J.W. Lee, J.H. Boo, Hierarchical NiO/TiO 2 composite structures for enhanced electrochromic durability, Thin Solid Films 664 (2018) 1–5, https://doi.org/10.1016/J.TSF.2018.07.049. [31] U.T. Nakate, R. Ahmad, P. Patil, Y.T. Yu, Y.B. Hahn, Ultra thin NiO nanosheets for high performance hydrogen gas sensor device, Appl. Surf. Sci. 506 (2020), 144971, https://doi.org/10.1016/J.APSUSC.2019.144971. [32] I. Barton, V. Matejec, J. Matousek, Photocatalytic activity of nanostructured TiO 2 coating on glass slides and optical fibers for methylene blue or methyl orange decomposition under different light excitation, J. Photochem. Photobiol. Chem. 317 (2016) 72–80, https://doi.org/10.1016/j.jphotochem.2015.11.009. [33] A. Najafidoust, S. Allahyari, N. Rahemi, M. Tasbihi, Uniform coating of TiO 2 nanoparticles using biotemplates for photocatalytic wastewater treatment, Ceram. Int. 46 (2020) 4707–4719, https://doi.org/10.1016/j.ceramint.2019.10.202. [34] M. Yamagishi, S. Kuriki, P.K. Song, Y. Shigesato, Thin film TiO 2 photocatalyst deposited by reactive magnetron sputtering, Thin Solid Films 442 (2003) 227–231, https://doi.org/10.1016/S0040-6090(03)00987-8. [35] J. Singh, S.A. Khan, J. Shah, R.K. Kotnala, S. Mohapatra, Nanostructured TiO 2 thin films prepared by RF magnetron sputtering for photocatalytic applications, Appl, Surf. Sci. 422 (2017) 953–961, https://doi.org/10.1016/J.APSUSC.2017.06.068. [36] G. ˇ Zerjav, G. Scandura, C. Garlisi, G. Palmisano, A. Pintar, Sputtered vs. sol-gel TiO 2 -doped films: characterization and assessment of aqueous bisphenol A oxidation under UV and visible light radiation, Catal. Today 357 (2020) 380–391, https://doi.org/10.1016/J.CATTOD.2019.09.027. [37] B. Hao, J. Guo, L. Zhang, H. Ma, Cr-doped TiO 2 /CuO photocatalytic nanofilms prepared by magnetron sputtering for wastewater treatment, Ceram. Int. 48 (2022) 7106–7116, https://doi.org/10.1016/J.CERAMINT.2021.11.270. [38] K.R. Reyes-Gil, D.B. Robinson, WO 3 -enhanced TiO 2 nanotube photoanodes for solar water splitting with simultaneous wastewater treatment, ACS Appl. Mater. Interfaces 5 (2013) 12400–12410, https://doi.org/10.1021/AM403369P/SUPPL_ FILE/AM403369P_SI_001.PDF. [39] M. Ratova, R. Klaysri, P. Praserthdam, P.J. Kelly, Pulsed DC magnetron sputtering deposition of crystalline photocatalytic titania coatings at elevated process pressures, Mater. Sci. Semicond. Process. 71 (2017) 188–196, https://doi.org/ 10.1016/j.mssp.2017.07.028. [40] S. Takeda, S. Suzuki, H. Odaka, H. Hosono, Photocatalytic TiO 2 thin film deposited onto glass by DC magnetron sputtering, Thin Solid Films 392 (2001) 338–344, https://doi.org/10.1016/S0040-6090(01)01054-9. [41] P. Pansila, N. Witit-Anun, S. Chaiyakun, Influence of sputtering power on structure and photocatalyst properties of DC magnetron sputtered TiO 2 thin film, Procedia Eng. 32 (2012) 862–867, https://doi.org/10.1016/j.proeng.2012.02.024. [42] L. Zhang, J. Guo, B. Hao, H. Ma, WO 3 /TiO 2 heterojunction photocatalyst prepared by reactive magnetron sputtering for Rhodamine B dye degradation, Opt. Mater. 133 (2022), 113035, https://doi.org/10.1016/J.OPTMAT.2022.113035. [43] B. Hao, J. Guo, L. Zhang, H. Ma, Magnetron sputtered TiO 2 /CuO heterojunction thin films for efficient photocatalysis of Rhodamine B, J. Alloys Compd. 903 (2022), 163851, https://doi.org/10.1016/J.JALLCOM.2022.163851. [44] D. Ramírez-Ortega, D. Guerrero-Araque, P. Acevedo-Pe˜ na, E. Reguera, H. A. Calderon, R. Zanella, Enhancing the photocatalytic hydrogen production of the ZnO–TiO 2 heterojunction by supporting nanoscale Au islands, Int. J. Hydrogen Energy 46 (2021) 34333–34343, https://doi.org/10.1016/J. IJHYDENE.2021.08.019. [45] T.T. Nguyen, M. Patel, S. Kim, R.A. Mir, J. Yi, V.A. Dao, J. Kim, Transparent photovoltaic cells and self-powered photodetectors by TiO 2 /NiO heterojunction, J. Power Sources 481 (2021), 228865, https://doi.org/10.1016/J. JPOWSOUR.2020.228865. [46] Y. Tian, B. Chang, G. Wang, L. Li, L. Gong, B. Wang, R. Yuan, W. Zhou, Magnetron sputtering tuned “ π back-donation” sites over metal oxides for enhanced electrocatalytic nitrogen reduction, J. Mater. Chem. A. 10 (2022) 2800–2806, https://doi.org/10.1039/D1TA10273G. [47] H. Tomaszewski, H. Poelman, D. Depla, D. Poelman, R. De Gryse, L. Fiermans, M. F. Reyniers, G. Heynderickx, G.B. Marin, TiO 2 films prepared by DC magnetron sputtering from ceramic targets, Vacuum 68 (2002) 31–38, https://doi.org/ 10.1016/S0042-207X(02)00279-8. [48] S. Schiller, U. Heisig, K. Steinfelder, J. Strümpfel, Reactive D.C. sputtering with the magnetron-plasmatron for tantalum pentoxide and titanium dioxide films, Thin Solid Films 63 (1979) 369–375, https://doi.org/10.1016/0040-6090(79)90042-7. [49] M. Grao, M. Ratova, C.C. Amorim, R.B.P. Marcelino, P. Kelly, Crystalline TiO 2 supported on stainless steel mesh deposited in a one step process via pulsed DC magnetron sputtering for wastewater treatment applications, J. Mater. Res. Technol. 9 (2020) 5761–5773, https://doi.org/10.1016/j.jmrt.2020.03.101. [50] M. Grao, J. Redfern, P.J. Kelly, M. Ratova, Magnetron co-sputtered Bi 12 TiO 20 / Bi 4 Ti 3 O 12 composite – an efficient photocatalytic material with photoinduced oxygen vacancies for water treatment application, Appl. Surf. Sci. 552 (2021), 149486, https://doi.org/10.1016/J.APSUSC.2021.149486. [51] J.A. Thornton, Influence of apparatus geometry and deposition conditions on the structure and topography of thick sputtered coatings, Undefined 11 (1974) 666–670, https://doi.org/10.1116/1.1312732. [52] P. Stefanov, M. Shipochka, P. Stefchev, Z. Raicheva, V. Lazarova, L. Spassov, XPS characterization of TiO 2 layers deposited on quartz plates, J. Phys. Conf. Ser. 100 (2008), https://doi.org/10.1088/1742-6596/100/1/012039. [53] M.C. Biesinger, B.P. Payne, L.W.M. Lau, A. Gerson, R.S.C. Smart, X-ray photoelectron spectroscopic chemical state quantification of mixed nickel metal, oxide and hydroxide systems, Surf. Interface Anal. 41 (2009) 324–332, https://doi. org/10.1002/SIA.3026. [54] R. Karsthof, H. von Wenckstern, J. Zú˜ niga-P´ erez, C. Deparis, M. Grundmann, Nickel oxide–based heterostructures with large band offsets, Phys. Status Solidi Basic Res. 257 (2020), https://doi.org/10.1002/PSSB.201900639. [55] B. Zhou, X. Jiang, Z. Liu, R. Shen, A.V. Rogachev, Preparation and characterization of TiO 2 thin film by thermal oxidation of sputtered Ti film, Undefined 16 (2013) 513–519, https://doi.org/10.1016/J.MSSP.2012.05.001. [56] B. Bharti, S. Kumar, H.-N. Lee, R. Kumar, Formation of Oxygen Vacancies and Ti 3+ State in TiO 2 Thin Film and Enhanced Optical Properties by Air Plasma Treatment OPEN, Nat. Publ. Gr., 2016, https://doi.org/10.1038/srep32355. [57] Z.H. Ibupoto, M.A. Abbasi, X. Liu, M.S. Alsalhi, M. Willander, The Synthesis of NiO/TiO 2 Heterostructures and Their Valence Band Offset Determination (2014), https://doi.org/10.1155/2014/928658. [58] O. Frank, M. Zukalova, B. Laskova, J. Kürti, J. Koltai, L. Kavan, Raman spectra of titanium dioxide (anatase, rutile) with identified oxygen isotopes (16, 17, 18), Phys. Chem. Chem. Phys. 14 (2012) 14567–14572, https://doi.org/10.1039/ C2CP42763J. [59] S. Challagulla, K. Tarafder, R. Ganesan, S. Roy, Structure sensitive photocatalytic reduction of nitroarenes over TiO 2 , Sci. Rep. 7 (2017), https://doi.org/10.1038/ S41598-017-08599-2. [60] J. Tauc, R. Grigorovici, A. Vancu, Optical properties and electronic structure of amorphous germanium, Phys. Status Solidi 15 (1966) 627–637, https://doi.org/ 10.1002/PSSB.19660150224. [61] R.B.P. Marcelino, C.C. Amorim, M. Ratova, B. Delfour-Peyrethon, P. Kelly, Novel and versatile TiO 2 thin films on PET for photocatalytic removal of contaminants of emerging concern from water, Chem. Eng. J. 370 (2019) 1251–1261, https://doi. org/10.1016/J.CEJ.2019.03.284. [62] M. Ratova, P.J. Kelly, G.T. West, X. Xia, Y. Gao, materials Deposition of Visible Light Active Photocatalytic Bismuth Molybdate Thin Films by Reactive Magnetron Sputtering, (n.d.). https://doi.org/10.3390/ma9020067. [63] Z. Wu, Y. Wang, L. Sun, Y. Mao, M. Wang, C. Lin, An ultrasound-assisted deposition of NiO nanoparticles on TiO 2 nanotube arrays for enhanced photocatalytic activity, J. Mater. Chem. A. 2 (2014) 8223–8229, https://doi.org/10.1039/C4TA00850B. [64] J.Z. Chen, T.H. Chen, L.W. Lai, P.Y. Li, H.W. Liu, Y.Y. Hong, D.S. Liu, Preparation and characterization of surface photocatalytic activity with NiO/TiO 2 nanocomposite structure, Materials 8 (2015) 4273–4286, https://doi.org/ 10.3390/MA8074273, 8 (2015) 4273–4286. [65] X. Wan, M. Yuan, S.L. Tie, S. Lan, Effects of catalyst characters on the photocatalytic activity and process of NiO nanoparticles in the degradation of methylene blue, Appl. Surf. Sci. 277 (2013) 40–46, https://doi.org/10.1016/J. APSUSC.2013.03.126. [66] R. Zuo, G. Du, W. Zhang, L. Liu, Y. Liu, L. Mei, Z. Li, Photocatalytic degradation of methylene blue using TiO 2 impregnated diatomite, Adv. Mater. Sci. Eng. 2014 (2014), https://doi.org/10.1155/2014/170148. [67] M. Shaban, A.M. Ahmed, N. Shehata, M.A. Betiha, A.M. Rabie, Ni-doped and Ni/Cr co-doped TiO 2 nanotubes for enhancement of photocatalytic degradation of methylene blue, J. Colloid Interface Sci. 555 (2019) 31–41, https://doi.org/ 10.1016/J.JCIS.2019.07.070. [68] M. Faisal, F.A. Harraz, A.A. Ismail, A.M. El-Toni, S.A. Al-Sayari, A. Al-Hajry, M. S. Al-Assiri, Novel mesoporous NiO/TiO 2 nanocomposites with enhanced photocatalytic activity under visible light illumination, Ceram. Int. 44 (2018) 7047–7056, https://doi.org/10.1016/J.CERAMINT.2018.01.140. [69] J. Yu, W. Wang, B. Cheng, Synthesis and enhanced photocatalytic activity of a hierarchical porous flowerlike p–n junction NiO/TiO 2 photocatalyst, Chem. Asian J. 5 (2010) 2499–2506, https://doi.org/10.1002/ASIA.201000550. [70] I. Khan, K. Saeed, I. Zekker, B. Zhang, A.H. Hendi, A. Ahmad, S. Ahmad, N. Zada, H. Ahmad, L.A. Shah, T. Shah, I. Khan, Review on methylene blue: its properties, uses, toxicity and photodegradation, Water 14 (2022) 242, https://doi.org/ 10.3390/W14020242, 14 (2022) 242. A. Villamayor et al. International Journal of Hydrogen Energy 64 (2024) 50–57 Available online 25 March 2024 0360-3199/© 2024 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Highly active ultralow loading Pt electrodes for hydrogen evolution reaction developed by magnetron sputtering Antía Villamayor a , * , Dmitry Galyamin b , Laura V. Barrio c , Eva G. Berasategui a , Sergio Rojas b a TEKNIKER, Basque Research and Technology Alliance (BRTA), Parke Teknologikoa, Calle I˜ naki Goenaga 5, 20600, Eibar, Gipuzkoa, Spain b Grupo de Energía y Química Sostenibles, Instituto de Cat´ alisis y Petroleoquímica, CSIC, Calle Marie Curie 2, 28049, Madrid, Spain c Dept. of Chemical and Environmental Engineering, Faculty of Engineering, University of the Basque Country (UPV/EHU), Plaza Ingeniero Torres Quevedo, 1, 48013, Bilbao, Spain ARTICLE INFO Handling Editor: Dr. E.A. Veziroglu Keywords: Magnetron sputtering Ultralow Pt loading HER catalyst Proton exchange membrane water electrolyzer (PEMWE) ABSTRACT Developing cost-effective components for PEM electrolyzers is key to transforming renewable energy into H 2 . Platinum is a scarce and expensive material but also, is the most active known catalyst for hydrogen evolution reaction in PEM electrolysis; thus it is of utmost importance to focus on developing electrocatalysts that utilize the minimum amount of platinum without losses in cell performance to make PEM electrolysis affordable. In addition, and with the aim of scaling up electrode production, it is imperative to develop a robust automated industrial manufacturing process to avoid the waste of catalyst during electrode preparation. In this work, magnetron sputtering gas aggregation method is studied for developing three low loading Pt electrodes (0.105, 0.062, and 0.052 mg cm −2 ), achieving remarkable overpotentials as low as 8 mV at 10 mA cm −2 for HER, showing similar activity than commercial Pt electrodes with more than 4 times less Pt (compared to 0.3 mgcm −2 Pt commercial GDL). 1. Introduction The rising global temperature is undoubtedly one of the most concerning issues today. Therefore, it is crucial to urgently replace the current fuel-based energy system with a sustainable system that solely relies on environmentally friendly and renewable energy sources. Recently, hydrogen has drawn much attention as an energy carrier due to its high-density energy per molecule [1–3]. If this production of hydrogen is going to be powered by renewable sources, one of the most promising methods is proton membrane exchange (PEM) water electrolysis, due to its high current density and low temperature operation, high voltage efficiency and compact system design [4–6]. The main drawback is the use of scarce and expensive materials, such as Pt in the cathode and IrO x in the anode. Pt is the most active and durable catalyst for the hydrogen evolution reaction (HER) and currently high amounts of Pt (between 1 and 3 mg cm −2 ) are used in PEM electrolyzers with membrane electrode assembly (MEA) configuration [4,7]. Several strategies have been investigated to reduce Pt loading or, even replace it with other catalysts to avoid the use of platinum group metals (PGMs) [8–12]. In principle, due to the very high kinetics for the HER of Pt catalysts, it will be possible to decrease the Pt used in state-of-the-art electrodes without significant activity penalties [13–15], hence allowing a significant reduction in the stack production costs. This can be achieved by optimizing Pt deposition process on the electrode to conform the catalyst coated membrane (CCM). The most common method for preparing Pt-based electrodes for the cathode of PEMWE electrolyzers is by mixing Pt nanoparticles with carbon black powder, primarily using wet chemical impregnation to obtain high surface area catalysts [16–19]. Afterward, the Pt/C catalytic ink is deposited over the polymeric membrane, the porous transport layer (PTL) or the gas diffusion layer (GDL). Unfortunately, chemical synthetic paths have multiple drawbacks, such as complex reaction pathways, byproducts generation and intricate industrial upscaling [20, 21]. Physical vapor deposition (PVD) techniques such as magnetron sputtering, can allow to overcome these issues as highly pure thin films are obtained in a one-step process without generating any byproducts. Moreover, magnetron sputtering is a robust industrial method for thin film deposition appearing as an attractive alternative to manufacture PEM catalytic material at an industrial scale. Magnetron sputtering has been studied for catalyst development in PEM fuel cells with remarkable results [22–27]. Regarding PEM electrolyzers, some work has been done for anode catalysts development * Corresponding author. E-mail address: [email protected] (A. Villamayor). Contents lists available at ScienceDirect International Journal of Hydrogen Energy journal homepage: www.elsevier.com/locate/he https://doi.org/10.1016/j.ijhydene.2024.03.226 Received 30 August 2023; Received in revised form 13 March 2024; Accepted 17 March 2024 International Journal of Hydrogen Energy 64 (2024) 50–57 51 [28–34], suggesting that this technique could be a suitable method to obtain efficient catalysts for PEM electrolysis. However, to the best of our knowledge, the use of magnetron sputtering for the preparation of electrodes for PEM electrolysis has been scarcely studied to date. Some of the relevant research work that exists on this topic using different types of magnetron sputtering, are summarized below. Interesting research was done by Bernsmeier et al. using hollow cathode gas flow sputtering to obtain highly stable Pt/C catalyst films and the electrochemical activity was measured using a glassy carbon electrode [35]. Fedotov et al. developed Pt/C electrodes using a laboratory scale DC sputtering system equipped with a stirring cup for the carbon powder support, directly depositing the Pt nanoparticles over a powdered support. The catalytic activity of the electrodes was tested in a PEMWE, achieving similar performances when compared with electrodes developed through chemical methods [36]. More recently, research on the properties of dispersed Pt deposited by DC magnetron sputtering over several metallic substrates was conducted in both acidic and alkaline media. The results showed that the Pt–W electrodes had a catalytic activity towards HER comparable to bulk Pt with an overpotential of 47 mV at −10 mA/cm 2 [37]. Moving towards the more suitable approach of using directly the carbon GDL as a support, some works have been reported where Pt and Pt/C dense films were deposited directly over carbon GDL substrates, using three different traditional sputtering sources (DC, DC with bias voltage and RF) [38–41]. Another remarkable research work was done to study HER and H 2 bubbles dynamics inside a PEM electrolyzer, using different thicknesses of Pt films, sputtered over titanium multifunctional liquid/gas diffusion layers, instead of microporous carbonaceous supports, and were compared to conventional catalyst coated membranes (CCMs), showing that with a much smaller amount than conventional CCMs a remarkable higher mass activity is achieved [42,43]. This work proves that most of the catalyst that is located inside the porous structure is inactive which is very relevant if magnetron sputtering is the chosen method for catalyst manufacturing. Taking all this into account, since HER appears to happen mostly on the electrode surface, if the active surface area could be improved using conventional carbon GDLs with MPL (microporous layer) and nanostructuring platinum into nanoparticles, excellent catalytic activity should be achieved with an ultrasmall amount of Pt. Magnetron sputtering gas aggregation (MSGA) method remains, to the best of our knowledge, unexplored for HER electrode development. The difference with conventional magnetron sputtering, is that the MSGA has higher pressure inside a small aggregation chamber, due to a high inert gas flow which forms supersaturated vapors from the sputtered material, which allows easily the formation of nanoparticles and clusters with well-defined size and shape [44–46] making MSGA the most suitable approach for nanoparticle synthesis [47–50] between the different magnetron sputtering techniques. In this work, a series of Pt-based electrodes with different Pt loadings of 0.68, 0.81, and 1.37 wt% (up to 4 times smaller than conventional Pt/ C GDLs), have been prepared by MSGA technique and tested as electrocatalysts for the HER in acidic electrolyte. Electrodes were prepared by depositing Pt particles onto carbon black microporous layer/carbon paper forming thin layers of Pt nanoparticles. The loading of Pt on the sputtered GDLs was determined through gravimetric thermal analysis (TGA) while compositional and morphology analyses were performed through X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Electrochemical analysis techniques (CV, LSV, CP) were performed in all samples to determine kinetic and electrochemical parameters for HER in an acidic electrolyte. 2. Materials and methods 2.1. Platinum nanoparticles deposition Pt nanoparticles were deposited onto two different substrates, namely, a piece of silicon wafer and MPL-carbon paper (Sigracet 22BB MPL, with 5% PTFE) with a coated area of 1 cm 2 for the latter. Pt nanoparticles were generated using a gas aggregation system (NanoGenTrio) from Mantis Deposition Ltd. (Oxfordshire, UK). The gas aggregation chamber (Fig. 1) was attached to an industrial PVD chamber (80 ×60 cm 2 ) and connected through a 4 mm hole, which allowed nanoparticles to flow between both chambers, with a base pressure of 2 ×10 −4 Pa in the whole vacuum system. The platinum target (2.54 ×3 cm 2 , purity of 99.999%) was placed on a mobile sputter head, fixed at 95 mm from the chamber hole. Before starting the sputtering process, the target was cleaned with an Ar pressure of 4 Pa for 2 min at 15 W. For the sputtering process, a discharge power of 25 W was applied for 15, 10, or 5 min (PtNPs_15, PtNPs_10, PtNPs_5, respectively), with an Ar flux of 30 sccm and He flux 15 sccm, which equals to a pressure of 14 Pa in the sputtering chamber. Through the sputtering of a metal target at high pressure, small nanoparticles can be obtained in the gas aggregation chamber, which is connected to a deposition chamber at a much lower pressure, where the nanoparticles formed are dragged out by an auxiliary gas (He) and the pressure difference and deposited on top of a substrate placed in the deposition chamber. 2.2. Characterization methods The morphology and thickness of the Pt layer on the electrodes were analysed using an ULTRA Plus Carl-Zeiss field emission scanning electron microscope (FESEM) and a TECNAI G2 20 TWIN high resolution transmission electron microscope at an acceleration voltage of 200 kV (STEM). The nature and relative abundance of the surface species was analysed with a SPECS customized apparatus equipped with a nonmonochromatic X-ray source XR 50 and a hemispherical energy analyser PHOIBOS 150. X-ray MgK line (1253.6 eV) operating at 200 W/12 kV, was used as the excitation The regions of interest were scanned at increments of 0.1 eV and fixed pass energy of 20 eV. Charge effects were accounted for by setting the C 1s core-level peak at 284.6 eV. Spectra were analysed using the XPS Casa Software. The Pt loading of the commercial and experimental samples was analysed through thermogravimetric analysis (TGA) using a TA Instruments Q500 instrument equipped with an EGA furnace. The measurements were conducted in air flow (90 mL min −1 ) using a 10 ◦C min −1 heating ramp, from r. t. until 900 ◦C. 2.3. Electrochemical characterization Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) tests were carried out in a three-electrode cell with an Autolab PGstat 302 N potentiostat along with Nova software. A 3 M Ag/AgCl electrode and a Pt wire were used as reference and counter electrodes, respectively. As a working electrode, 1 cm 2 of the sputtered Pt/C samples were used. The potentials were calibrated versus reversible hydrogen electrode (RHE) according to Equation (1): E (RHE) =E (Ag/AgCl) +0.059 pH +0.21 (1) Fig. 1. Diagram of a magnetron sputtering gas aggregation chamber. A. Villamayor et al. International Journal of Hydrogen Energy 64 (2024) 50–57 52 CV measurements (for ECSA determination and Pt surface cleaning) were collected in Ar saturated 0.5 M H 2 SO 4 electrolyte at room temperature. First, a number of CVs were recorded between 0.05 and 1.4 V at 100 mV s −1 until a constant response was obtained. The electrochemically active surface area (ECSA) was subtracted from the charge of the H upd region from cyclic voltammograms recorded within a potential window between 0.05 and 1.2 V at 50 mV s −1 , using the linear scan voltammetry mode. The HER was evaluated by recording LSV between 0.01 and −0.3 V at 10 mV s −1 in an H 2 -saturated 0.5 M H 2 SO 4 electrolyte to ensure the H + /H 2 equilibrium at 0 V vs. RHE [51]. Polarization curves are reported after iR compensation. The solution resistance (R s ) has been determined by electrochemical impedance spectroscopy (EIS) at high frequencies with a 10 mV perturbation. The overpotential ( η ) to reach a current density of 10 mA cm −2 and the Tafel slopes obtained in the −0.015 to −0.03 V vs. RHE potential range were used to benchmark catalysts’ activity. Finally, the degradation of the sputtered electrodes was assessed by conducting chronopotentiometry (CP) measurements at 10 mA cm −2 for 24 h and CV measurements at 5 mV s −1 from 0 V to −1 V vs Ag/AgCl for 1000 cycles were performed to assess the long-term stability at high density currents. 3. Results and discussion 3.1. Characterization Before moving into a more complex substrate such as MPL-carbon paper, we analysed the deposition of Pt onto the silicon wafers (Fig. 2) since they are a flat substrate widely used in sputtering. FESEM images were used to determine the thickness of the coatings over a silicon wafer and surface morphology over MPL-carbon paper substrates for the three samples. The aim was to analyse if the thickness of the samples was linearly dependent on the sputtering time and, also, how they were distributed. The thicknesses of the Pt layer over silicon substrates were around 50, 25 and 14 nm for the PtNPs_15, PtNPs_10, and PtNPs_5 coatings, respectively (Fig. 2 and Fig. S1), confirming the linear growth with sputtering time and presented a uniform distribution. Regarding the surface morphology of the Pt coatings, SEM images of commercial (0.3 mg cm −2 Pt/C) and experimental samples were analysed (Fig. 3). While the commercial sample (Fig. 3a) has a porous and smooth surface where Pt nanoparticles are not visible, the sputtered PtNPs_15 coating (Fig. 3b) shows a clear nanoparticulated surface, as the sputtered Pt nanoparticles are deposited all over the substrate surface, so almost all the Pt is exposed to the electrolyte. When the deposition time is increased, the higher Pt loading could also subtly be seen in FESEM images (Fig. S2). The Pt nanoparticles accumulate over the porous surface when sputtering time is increased, giving a thicker and denser appearance to the surface. TEM images of the sputtered samples and their corresponding size distribution histogram are displayed in Fig. 4 and Fig. S3, respectively. The statistic histograms of the samples were constructed by measuring the diameter of no less than 100 randomly selected Pt NPs (for each sample) showing a narrow size distribution from 2 to 6 nm with an average size between ~4 and 3 nm (Fig. S3). In general, Pt nanoparticles have a well-defined spherical shape, as shown in Fig. 4a. High-angle annular dark-field scanning TEM (HAADF-STEM) images reveal that Pt nanoparticles are located on the periphery of the carbon black layer and present an uneven and non-homogeneous thickness As observed in the images, the coverage of the surface increases with the sputtering time, resulting in uncoated areas in the PtNPs_5 sample. Pt loading of the samples was determined from thermogravimetric analyses. Pt loadings of 1.37 ±0.29% (0.105 mg cm −2 ), 0.81 ±0.18% (0.062 mg cm −2 ), 0.68 ±0.09% (0.052 mg cm −2 ) and 3.93 ±1.58 wt% (0.3 mg cm −2 ) were obtained for PtNPs_15, PtNPs_10, PtNPs_5 and Pt/C commercial electrodes, respectively (Table S1). The initial weight loss of approximately 17% between 500 and 590 ◦C corresponds to the PTFE present in the carbon paper substrate, followed by the degradation of the carbon species in the range of 590–750 ◦C (around 81%). At higher temperatures, Pt is the only component remaining stable; hence the Pt loading was obtained from the mass at the highest analysis temperature. Compared to the commercial samples, the reduction of Pt loading in the sputtered electrodes is remarkable, up to more than 4 times smaller. It should be noted that the deposition time and Pt loading is not linear, probably due to that several factors like the temperature of the substrate/target, pressure, or particle collision energy have an influence in the sputter rate. The temperature of the target or the collision energy are parameters that are going to vary within the first minutes of the sputtering process, so at such small deposition times or Pt loadings, the effect of these parameters over the sputtering rate are more remarkable [52–54]. Finally, the nature of the Pt species on the electrodes was analysed by XPS. As shown in Fig. 5, the Pt 4f region shows the characteristic 4f 7/2 and 4f 5/2 components, with an asymmetric Pt 4f 7/2 peak at ca. 71.5 eV characteristic of metallic Pt species, confirming the purity of the coatings. Fig. 2. FESEM image of PtNPs_15 coating over a silicon wafer. Fig. 3. FESEM images of (a) commercial Pt/C sample and (b) PtNPs_15. A. Villamayor et al. Coatings 2024,14, 868 2 of 13 because the exceptionally fast kinetics of Pt for the HER allow the use of low Pt loadings, hence significantly reducing the production costs of the stack [ 12 – 14 ]. Currently, catalytic coatings are developed using wet chemical methods like wet chemical impregnation, solgel, or thermal decomposition. For instance, Pt nanoparticles synthesized through these methods are often mixed with an ionomer and high-surface carbon black powder to avoid nanoparticle agglomeration. The obtained catalytic ink usually has to be deposited by spraying techniques over the membrane or the gas diffusion layer (GDL) [ 2 , 15 , 16 ]. Wet chemistry methods provide a high surface area to the catalytic layer, which is crucial to obtaining a high catalytic surface. However, these techniques have multiple drawbacks, such as reaction byproducts and multiple reaction steps, which are directly related to material waste and complex industrial upscaling. Magnetron sputtering appears to be a compelling manufacturing method for solving these issues. As a robust industrial single-step process for thin film deposition, magnetron sputtering not only reduces the waste of material with the absence of byproducts but also facilitates the upscaling of catalysts manufacturing for large-scale production of electrodes for PEM electrolyzers. As far as we know, the use of magnetron sputtering as a manufacturing method for platinum group metals (PGMs) PEM electrolysis catalysts is not very wide, especially for the cathode side [ 17 – 19 ]. Most of the existing research work is focused on the ex-situ evaluation of the HER catalyst, where the a huge potential of magnetron sputtering for their application [ 12 , 20 – 24 ]. To the best of our knowledge, a couple of research works have been published regarding in-situ single-cell testing. For example, Pt thin films were deposited over titanium multifunctional liquid/gas diffusion layers in order to evaluate HER activity and H 2 bubbles behavior inside a PEM single-cell [ 25 ]. This research work proved that most of the catalysis seems to occur on the surface of the electrode. In another study, a stirring cup for the carbon powder was introduced inside a magnetron-sputtering laboratory chamber, and Pt nanoparticles were deposited over the carbon support using a direct current (DC) power source [ 26 ]. Several loadings of Pt were evaluated for PEM electrolysis and fuel cell, showing no performance differences with electrodes developed using chemical methods. Hence, if HER catalysis is a process that occurs only on the electrode surface with no catalytic activity differences from the traditional synthesized catalyst, magnetron sputtering could be a feasible alternative to industrially upscale PEM catalyst. As far as we are concerned, there is a complete lack of research on the use of magnetron sputtering to develop Pt nanoparticles for PEMWE and critically reduce the presence of Pt in the cathode. In this work, we have developed an ultra-low 0.1 mg cm −2 Pt loading electrode using magnetron sputtering, directly depositing the Pt nanoparticles on top of a carbon microporous layer (MPL) GDL, followed by a complete evaluation of the electrochemical properties of the sputtered electrode. This method enables a full exposure of Pt nanoparticles to the reactants, maintaining a high surface area while reducing the Pt loading. The aim of this research is to prove the feasibility of magnetron sputtering as an electrode manufacturing technique, allowing Pt reduction without interfering with cell performance. The electrochemical evaluation was performed in a three-electrode cell and a single PEM cell, and the results were compared with a commercial 0.3 mg cm −2 Pt/C electrode for both electrochemical setups. Electrochemical ex-situ analysis techniques like cyclic voltammetry and linear sweep voltammetry (CV, LSV) were measured for all samples to evaluate the catalytic activity for HER in an acidic electrolyte and in-situ analysis through V-icurves and chronopotentiometry measurements (CP), was performed in a real single PEM cell to evaluate the performance and degradation of both electrodes. 2. Materials and Methods 2.1. Platinum Nanoparticles Deposition Pt nanoparticles were deposited over a microporous layer of carbon paper (Sigracet 22BB MPL, with 5% polytetrafluoroethylene (PTFE)) with a coated surface of 1 cm 2 for the ex-situ and 25 cm 2 for the in-situ tests. The gas aggregation system (NanoGenTrio, Mantis Deposition Ltd., Thame, UK) was attached to an industrial PVD chamber (80 × 60 cm 2 ), Coatings 2024,14, 868 3 of 13 coupled with a 4 mm hole, enabling particles to move from one chamber to the other. The platinum target (2.54 × 3 cm 2 , purity of 99.999%) was positioned on a mobile sputter head, with a distance of 9.5 cm from the chamber hole. The base pressure of the whole vacuum system was 2 × 10 −4 Pa, and the target was cleaned before starting the deposition process for 2 min at 15 W with a 4 Pa Ar pressure. After a previous process optimization [ 12 ], an Ar flux of 30 sccm and He flux of 15 sccm, corresponding to a total pressure of 14 Pa, were introduced in the aggregation chamber. To start the sputtering process to obtain the Pt nanoparticles, a discharge power of 25 W was applied for 15 min. The sputtered electrodes were compared with a commercial 0.3 mg cm −2 Pt electrode (40%wt Pt/C, Sigracet 22BB MPL, with 5% PTFE. The MSGA method is the most suitable magnetron sputtering technique for nanoparticle deposition. In contrast with traditional magnetron sputtering, in MSGA, the sputtering takes place inside a high-pressure aggregation chamber where supersaturated vapors of the ejected material enable the formation of clusters or nanoparticles of well-defined size and shape. As the sputtered particles collide with inert gas atoms, their energy is lowered and begins to condense into small clusters. These clusters serve as nuclei, where additional particles coalesce, leading to the formation of larger nanoparticles [27–31]. 2.2. Characterization Methods The size and morphology of the nanoparticles Pt electrodes were studied using a ULTRA Plus Carl-Zeiss (Jena, Germany) field emission scanning electron microscope (FESEM) and a TECNAI G2 20 TWIN (Hillsboro, OR, USA) high-resolution transmission electron microscope at an acceleration voltage of 200 kV (STEM). A SPECS customized system with a nonmonochromatic X-ray source XR 50 and a hemispherical energy analyzer PHOIBOS 150 (SPECS Surface Nano Analysis GmbH, Berlin, Germany) was used to assess the relative abundance of the surface species. Thermogravimetric analysis (TGA) was performed using a TA Instruments Q500 instrument (New Castle, DE, USA) equipped with an EGA furnace to determine the Pt loading of the experimental and commercial samples. The measurements were conducted from r. t. until 900 ◦ C using a 10 ◦ C min −1 heating ramp in air flow (90 mL min−1). 2.3. Electrochemical Characterization For the ex-situ test in a three-electrode cell, an Autolab PGstat 302 N (Tokyo, Japan) potentiostat was used to perform cyclic voltammetry (CV) and linear sweep voltammetry (LSV) measurements recorded with Nova software (https://www.metrohm.com/es_es/ service/software-center/nova.html). The three-electrode cell setup had an Ag/AgCl (3 M KCl) electrode and a Pt wire as reference and counter electrodes, respectively. The working electrodes were the sputtered and commercial electrodes with an exposed area of 1 cm 2 . CV measurements for Pt surface cleaning and activation were recorded between 0.05 and 1.23 V at 50 mV s −1 until a constant signal was obtained. The data was collected in 0.5 M H 2 SO 4 electrolyte at room temperature in an Ar atmosphere. LSV measurements were recorded between 0.01 and − 0.3 V at 5 mV s −1 in an Ar-saturated 0.5 M H 2 SO 4 electrolyte for HER evaluation. To determine the iR drop [ 32 ], the solution resistance (Rs) was obtained by electrochemical impedance spectroscopy (EIS) at a frequency interval of 0.1 Hz–100 kHz with a 10 mV perturbation. The overpotential ( η ) at a current density of 10 mA cm −2 and the Tafel slopes obtained in the corresponding cathodic overpotential range were extracted from the corrected iR polarization curves in order to benchmark catalysts’ activity. A PEM test bench from H2GREEM was used for in-situ characterization in a single cell designed in Tekniker (Figure 1). The MEA was composed of commercial catalyst-coated Nafion 115 membranes (127 µ m of thickness) only coated with 3 mg cm −2 loading of IrRuO x on the anode side (Fuel Cell Store, Bryan, TX, USA, EE.UU.). For the cathode, a commercial GDL with 0.3 mg cm −2 Pt electrode (40%wt Pt/C, Sigracet 22BB MPL, with 5% PTFE, from Fuel Cell Store) was established as the reference and with the same MPL-GDL. Also, a 0.1 mg cm −2 Pt sputtered electrode was measured. Both anode and cathode were Coatings 2024,14, 868 4 of 13 equipped with an ultrathin platinized titanium (grade 2) porous transport layer (PTL) with a nominal thickness of 250 µ m, a 30% porosity, and 5–10 µ m of pore diameter (Fuel Cell Store, EE.UU.). As flow channels, two PEEK frames and two titanium meshes of 1 mm were used to distribute and evacuate the water and formed gases. To stabilize the cell and activate the catalysts, a current density of 0.12 A cm −2 was applied for 24 h. Measurements were performed at 60 ◦ C with a deionized H 2 O flow rate of 0.2 L min −1 . Polarization curves were measured every 24 h for a week, ranging from 0.12 to 2 A cm −2 . Between these measurements, the cell was maintained at a current density of 1.5 A cm −2 to assess the stability of the two Pt GDLs. Coatings 2024, 14, x FOR PEER REVIEW 4 of 13 corresponding cathodic overpotential range were extracted from the corrected iR polarization curves in order to benchmark catalysts’ activity. A PEM test bench from H2GREEM was used for in-situ characterization in a single cell designed in Tekniker (Figure 1). The MEA was composed of commercial catalystcoated Nafion 115 membranes (127 µm of thickness) only coated with 3 mg cm−2 loading of IrRuOx on the anode side (Fuel Cell Store, Bryan, TX, USA, EE.UU.). For the cathode, a commercial GDL with 0.3 mg cm−2 Pt electrode (40%wt Pt/C, Sigracet 22BB MPL, with 5% PTFE, from Fuel Cell Store) was established as the reference and with the same MPL-GDL. Also, a 0.1 mg cm−2 Pt sputtered electrode was measured. Both anode and cathode were equipped with an ultrathin platinized titanium (grade 2) porous transport layer (PTL) with a nominal thickness of 250 µm, a 30% porosity, and 5–10 µm of pore diameter (Fuel Cell Store, EE.UU.). As flow channels, two PEEK frames and two titanium meshes of 1 mm were used to distribute and evacuate the water and formed gases. To stabilize the cell and activate the catalysts, a current density of 0.12 A cm−2 was applied for 24 h. Measurements were performed at 60 °C with a deionized H2O flow rate of 0.2 L min−1. Polarization curves were measured every 24 h for a week, ranging from 0.12 to 2 A cm−2. Between these measurements, the cell was maintained at a current density of 1.5 A cm−2 to assess the stability of the two Pt GDLs. Figure 1. Scheme of the PEM single-cell components. 3. Results and Discussion 3.1. Characterization XPS measurements were performed to determine the chemical composition and oxidation state of the Pt nanoparticles on the sputtered electrode (Figure S1), confirming the presence of only metallic Pt species, thus the purity of the coatings. For the observation and comparison of the surface morphology of the sputtered and commercial electrodes, images of the surface were obtained by SEM (Figure 2). The sputtered Pt electrode (PtS_GDE) clearly shows a rough surface formed by nanoparticles (Figure 2a), while the commercial sample (PtCom_GDE) presented a smoother surface as Pt nanoparticles are mixed with the carbon black powder (Figure 2b). In the sputtered sample, all of the Pt nanoparticles are on the electrode surface, potentially increasing Pt exposure to the electrolyte. Figure 1. Scheme of the PEM single-cell components. 3. Results and Discussion 3.1. Characterization XPS measurements were performed to determine the chemical composition and oxidation state of the Pt nanoparticles on the sputtered electrode (Figure S1), confirming the presence of only metallic Pt species, thus the purity of the coatings. For the observation and comparison of the surface morphology of the sputtered and commercial electrodes, images of the surface were obtained by SEM (Figure 2). The sputtered Pt electrode (PtS_GDE) clearly shows a rough surface formed by nanoparticles (Figure 2a), while the commercial sample (PtCom_GDE) presented a smoother surface as Pt nanoparticles are mixed with the carbon black powder (Figure 2b). In the sputtered sample, all of the Pt nanoparticles are on the electrode surface, potentially increasing Pt exposure to the electrolyte. The size and nanoparticle distribution of the electrodes were analyzed using TEM images. The diameter of 100 nanoparticles was measured from the images to elaborate the PtS_GDE nanoparticle size statistic size histogram. An average size of around 4 nm was obtained with a narrow size distribution from 2 to 6 nm (Figure 3b) for the PtS_GDE sample, and most parts of the Pt nanoparticles showed a well-defined spherical shape (Figure 3a). High-angle annular dark-field scanning TEM (HAADF-STEM) images reveal that Pt nanoparticles on the PtCom_GDE sample are scattered inside the carbon black powder (Figure 3d), while Pt nanoparticles on PtS_GDE are deposited only on the surface of the carbon black MPL. In the case of the PtS_GDE, as a consequence of the high surface area of the MPL, the thickness is non-homogeneous (Figure 3c). Finally, the Pt content of both samples was obtained from thermogravimetric analyses. The initial weight loss (approximately 17% between 500 and 590 ◦ C) is related to the presence of PTFE in the carbonaceous substrate. In the range of 590–750 ◦C (around 81%), the rest of the carbon species are calcinated, and the remaining mass corresponds to Pt, the only compound that is stable at high temperatures. Hence, Pt loading was obtained from the mass at the highest analysis temperature. Two measurements were performed for each electrode, and the results show that PtCom_GDE has a Pt loading of 3.93 ± 1.58 wt% (0.3 mg cm −2 ) while PtS_GDE contains 1.37 ± 0.29% (0.105 mg cm −2 ) (Figure 3e). This is a remarkable reduction in the Pt loading of 67% with respect to the 0.3 mg cm −2 Pt/C commercial electrode. Coatings 2024,14, 868 5 of 13 Coatings 2024, 14, x FOR PEER REVIEW 5 of 13 Figure 2. SEM images of (a) PtS_GDE and (b) PtCom_GDE. The size and nanoparticle distribution of the electrodes were analyzed using TEM images. The diameter of 100 nanoparticles was measured from the images to elaborate the PtS_GDE nanoparticle size statistic size histogram. An average size of around 4 nm was obtained with a narrow size distribution from 2 to 6 nm (Figure 3b) for the PtS_GDE sample, and most parts of the Pt nanoparticles showed a well-defined spherical shape (Figure 3a). High-angle annular dark-field scanning TEM (HAADF-STEM) images reveal that Pt nanoparticles on the PtCom_GDE sample are scattered inside the carbon black powder (Figure 3d), while Pt nanoparticles on PtS_GDE are deposited only on the surface of the carbon black MPL. In the case of the PtS_GDE, as a consequence of the high surface area of the MPL, the thickness is non-homogeneous (Figure 3c). (a) (b) Figure 2. SEM images of (a) PtS_GDE and (b) PtCom_GDE. 3.2. Electrochemical Characterization The cyclic voltammogram of the sputtered sample after stabilization in the Ar-saturated 0.5 M H 2 SO 4 electrolyte is depicted in Figure 4. The observed peaks correspond with the reported for polycrystalline Pt with an H upd region between 0.05 and 0.4 V, followed by the double layer region and, towards more positive potentials, the Pt-oxide formation region. Moving to more negative potential values, the reduction of Pt oxides can be observed in the reserves scan around ca. 0.8 V [33]. Coatings 2024,14, 868 6 of 13 Coatings 2024, 14, x FOR PEER REVIEW 6 of 13 Figure 3. STEM images of (a) PtS_GDE, (b) size distribution histogram and cross-section colored HAAF-STEM images of (c) PtS_GDE and (d) PtCom_GDE and (e) TGA measurements of PtS_GDE. Finally, the Pt content of both samples was obtained from thermogravimetric analyses. The initial weight loss (approximately 17% between 500 and 590 °C) is related to the presence of PTFE in the carbonaceous substrate. In the range of 590–750 °C (around 81%), the rest of the carbon species are calcinated, and the remaining mass corresponds to Pt, the only compound that is stable at high temperatures. Hence, Pt loading was obtained from the mass at the highest analysis temperature. Two measurements were performed for each electrode, and the results show that PtCom_GDE has a Pt loading of 3.93 ± 1.58 wt% (0.3 mg cm −2 ) while PtS_GDE contains 1.37 ± 0.29% (0.105 mg cm −2 ) (Figure 3e). This (a) (b) (c) (d) Figure 3. STEM images of (a) PtS_GDE, (b) size distribution histogram and cross-section colored HAAF-STEM images of (c) PtS_GDE and (d) PtCom_GDE and (e) TGA measurements of PtS_GDE. Coatings 2024,14, 868 7 of 13 Coatings 2024, 14, x FOR PEER REVIEW 7 of 13 is a remarkable reduction in the Pt loading of 67% with respect to the 0.3 mg cm−2 Pt/C commercial electrode. 3.2. Electrochemical Characterization The cyclic voltammogram of the sputtered sample after stabilization in the Ar-saturated 0.5 M H2SO4 electrolyte is depicted in Figure 4. The observed peaks correspond with the reported for polycrystalline Pt with an Hupd region between 0.05 and 0.4 V, followed by the double layer region and, towards more positive potentials, the Pt-oxide formation region. Moving to more negative potential values, the reduction of Pt oxides can be observed in the reserves scan around ca. 0.8 V [33]. Figure 4. Cyclic voltammetry of PtS_GDE. To evaluate and compare the HER activity of both electrodes, a potential sweep from 0 V vs. RHE towards cathodic potential values was recorded in the Ar atmosphere. Figure 5 shows the polarization curves where the overpotentials needed to reach a current density value of −10 mA cm2 are the reference for the evaluation of the HER activity of different catalysts [34,35]. As presented in Figure 6, the overpotential for a current density of −10 mA cm2 of the PtCom_GDE sample is 74 mV, while the overpotential at the same current density for the PtS_GDE is 53 mV. These results are preliminary proof that Pt loading can be reduced without compromising the electrode activity towards the HER. Besides, the reduction of the overpotential, when compared to the commercial sample, could be influenced by the fact that in the sputtered electrode, all Pt nanoparticles are placed on the surface in direct contact with the reactant, while in the commercial sample, the Pt active sites can be more difficult to reach since they are mixed with the MPL powder. Figure 4. Cyclic voltammetry of PtS_GDE. To evaluate and compare the HER activity of both electrodes, a potential sweep from 0 V vs. RHE towards cathodic potential values was recorded in the Ar atmosphere. Figure 5 shows the polarization curves where the overpotentials needed to reach a current density value of − 10 mA cm 2 are the reference for the evaluation of the HER activity of different catalysts [ 34 , 35 ]. As presented in Figure 6, the overpotential for a current density of −10 mA cm2 of the PtCom_GDE sample is 74 mV, while the overpotential at the same current density for the PtS_GDE is 53 mV. These results are preliminary proof that Pt loading can be reduced without compromising the electrode activity towards the HER. Besides, the reduction of the overpotential, when compared to the commercial sample, could be influenced by the fact that in the sputtered electrode, all Pt nanoparticles are placed on the surface in direct contact with the reactant, while in the commercial sample, the Pt active sites can be more difficult to reach since they are mixed with the MPL powder. Coatings 2024, 14, x FOR PEER REVIEW 8 of 13 Figure 5. LSV curves of PtS_GDE and PtCom_GDE in an Ar-saturated 0.5 M H2SO4 solution. Regarding the kinetic parameters of both electrodes, the Tafel slope for the HER reaction was extracted from the Tafel curves shown in Figure 6. At an acidic pH, the HER occurs through two reaction pathways: Volmer Heyrovsky or Volmer-Tafel. For polycrystalline Pt, the reported values are over 30–40 mV dec−1, which corresponds to the VolmerTafel pathway [36,37]. Tafel slopes for PtS_GDE and PtCom_GDE are 32.8 and 58.8 mV dec−1, respectively, which are in concordance with the reported values for polycrystalline Pt. In the case of the PtCom_GDE, the slope is slightly higher than for PtS_GDE, meaning that the HER is happening at a slower pace in the case of the commercial sample. As Pt nanoparticles are mixed with the carbon MPL, this could hinder the access of the reactants to the active sites and the release of gas bubbles, slowing the reaction rate [25]. Figure 6. Tafel plots of PtS_GDE and PtCom_GDE. Since the results of the ex-situ tests indicated that the Pt sputtering electrodes performed better than the commercial ones, in situ tests in electrolysis cells were conducted. Measurements were carried out as described in Section 2.3 in a self-designed single cell. After the 24 h preconditioning, polarization V-i curves were performed every 24 h, applying stationary current densities of 1.5 A cm−2 between each curve. Figure 7 shows the Figure 5. LSV curves of PtS_GDE and PtCom_GDE in an Ar-saturated 0.5 M H2SO4solution. Regarding the kinetic parameters of both electrodes, the Tafel slope for the HER reaction was extracted from the Tafel curves shown in Figure 6. At an acidic pH, the HER occurs through two reaction pathways: Volmer Heyrovsky or Volmer-Tafel. For Coatings 2024,14, 868 8 of 13 polycrystalline Pt, the reported values are over 30–40 mV dec −1 , which corresponds to the Volmer-Tafel pathway [ 36 , 37 ]. Tafel slopes for PtS_GDE and PtCom_GDE are 32.8 and 58.8 mV dec −1 , respectively, which are in concordance with the reported values for polycrystalline Pt. In the case of the PtCom_GDE, the slope is slightly higher than for PtS_GDE, meaning that the HER is happening at a slower pace in the case of the commercial sample. As Pt nanoparticles are mixed with the carbon MPL, this could hinder the access of the reactants to the active sites and the release of gas bubbles, slowing the reaction rate [ 25 ]. Coatings 2024, 14, x FOR PEER REVIEW 8 of 13 Figure 5. LSV curves of PtS_GDE and PtCom_GDE in an Ar-saturated 0.5 M H2SO4 solution. Regarding the kinetic parameters of both electrodes, the Tafel slope for the HER reaction was extracted from the Tafel curves shown in Figure 6. At an acidic pH, the HER occurs through two reaction pathways: Volmer Heyrovsky or Volmer-Tafel. For polycrystalline Pt, the reported values are over 30–40 mV dec−1, which corresponds to the VolmerTafel pathway [36,37]. Tafel slopes for PtS_GDE and PtCom_GDE are 32.8 and 58.8 mV dec−1, respectively, which are in concordance with the reported values for polycrystalline Pt. In the case of the PtCom_GDE, the slope is slightly higher than for PtS_GDE, meaning that the HER is happening at a slower pace in the case of the commercial sample. As Pt nanoparticles are mixed with the carbon MPL, this could hinder the access of the reactants to the active sites and the release of gas bubbles, slowing the reaction rate [25]. Figure 6. Tafel plots of PtS_GDE and PtCom_GDE. Since the results of the ex-situ tests indicated that the Pt sputtering electrodes performed better than the commercial ones, in situ tests in electrolysis cells were conducted. Measurements were carried out as described in Section 2.3 in a self-designed single cell. After the 24 h preconditioning, polarization V-i curves were performed every 24 h, applying stationary current densities of 1.5 A cm−2 between each curve. Figure 7 shows the Figure 6. Tafel plots of PtS_GDE and PtCom_GDE. Since the results of the ex-situ tests indicated that the Pt sputtering electrodes performed better than the commercial ones, in situ tests in electrolysis cells were conducted. Measurements were carried out as described in Section 2.3 in a self-designed single cell. After the 24 h preconditioning, polarization V-i curves were performed every 24 h, applying stationary current densities of 1.5 A cm −2 between each curve. Figure 7shows the recorded V-i curves at 0 h, 24 h, 72 h, and 168 h for PtCom_GDE (Figure 7a) and PtS_GDE (Figure 7b). When using the PtCom_GDE sample as the cathode, t = 0 h curve, the potential corresponding to a current density of 1 A cm −2 is 1.85 V. This potential value increases up to 1.93 V after 168 h of operation. For the PtS_GDE, the initial t = 0 h V-i curve shows slightly higher voltages at the second half of the current density sweep compared to the rest of the polarization curves. Nevertheless, the voltage stabilizes completely after the first polarization curve. Figure 8compares the V-i curves at 0 h and 168 h for both samples, demonstrating that, under the same conditions, the PtS_GDE electrode is more stable with time than the PtCom_GDE sample. Both start with 1.85 V at 1 A cm −2 ; however, at t = 168 h , the potential value is maintained nearly the same for the PtS_GDE, whereas for the PtCom_GDE electrode, the potential has increased over time. This voltage variation is further observed in stationary current measurements. To estimate the degradation rate, operation potentials at stationary current densities were measured for 2 h every 20 h over a total period of 100 h. From the slope of the curve, the degradation rate can be extracted. Figure 9. presents the Vvs. tgraphic of PtCom_GDE and PtS_GDE. The same tendencies as seen in V-i curves are observed, hence proving the stability of the PtS_GDE. The slope extracted from the linear regression of the PtCom_GDE measurements corresponds to a degradation rate of 700 µ V h −1 at an operation current density of 1.5 A cm −2 . Meanwhile, the degradation rate for the PtS_GDE is completely negligible, as the line is almost flat. Coatings 2024,14, 868 9 of 13 Coatings 2024, 14, x FOR PEER REVIEW 9 of 13 recorded V-i curves at 0 h, 24 h, 72 h, and 168 h for PtCom_GDE (Figure 7a) and PtS_GDE (Figure 7b). Figure 7. Polarization V-i curves at t 0 h, 24 h, 72 h, and 168 h of (a) PtCom_GDE and (b) PtS_GDE. When using the PtCom_GDE sample as the cathode, t = 0 h curve, the potential corresponding to a current density of 1 A cm−2 is 1.85 V. This potential value increases up to 1.93 V after 168 h of operation. For the PtS_GDE, the initial t = 0 h V-i curve shows slightly higher voltages at the second half of the current density sweep compared to the rest of the polarization curves. Nevertheless, the voltage stabilizes completely after the first polarization curve. Figure 8 compares the V-i curves at 0 h and 168 h for both samples, demonstrating that, under the same conditions, the PtS_GDE electrode is more stable with time than the PtCom_GDE sample. Both start with 1.85 V at 1 A cm−2; however, at t = 168 h, the potential value is maintained nearly the same for the PtS_GDE, whereas for the PtCom_GDE electrode, the potential has increased over time. Figure 7. Polarization V-i curves at t 0 h, 24 h, 72 h, and 168 h of (a) PtCom_GDE and (b) PtS_GDE. Summarizing all the results, it is evident that reducing the amount of Pt is not hindering the cell performance. In fact, using MSGA as a deposition method could enhance cell performance not only because all of the catalytic material is on the MPL surface but also because the deposited nanoparticles could be better attached to the surface as a consequence of an adhesion improvement. One of the major issues of Pt/C catalysts is the lack of adhesion and nanoparticle agglomeration, which negatively impacts performance [38,39]. Coatings 2024,14, 868 10 of 13 Coatings 2024, 14, x FOR PEER REVIEW 10 of 13 Figure 8. Comparison of polarization V-i curves at t 0 h and 168 h of PtCom_GDE and PtS_GDE. This voltage variation is further observed in stationary current measurements. To estimate the degradation rate, operation potentials at stationary current densities were measured for 2 h every 20 h over a total period of 100 h. From the slope of the curve, the degradation rate can be extracted. Figure 9. presents the V vs. t graphic of PtCom_GDE and PtS_GDE. The same tendencies as seen in V-i curves are observed, hence proving the stability of the PtS_GDE. The slope extracted from the linear regression of the PtCom_GDE measurements corresponds to a degradation rate of 700 µV h−1 at an operation current density of 1.5 A cm−2. Meanwhile, the degradation rate for the PtS_GDE is completely negligible, as the line is almost flat. Figure 9. Comparison of voltage evolution with time of both samples. Figure 8. Comparison of polarization V-i curves at t 0 h and 168 h of PtCom_GDE and PtS_GDE. Coatings 2024, 14, x FOR PEER REVIEW 10 of 13 Figure 8. Comparison of polarization V-i curves at t 0 h and 168 h of PtCom_GDE and PtS_GDE. This voltage variation is further observed in stationary current measurements. To estimate the degradation rate, operation potentials at stationary current densities were measured for 2 h every 20 h over a total period of 100 h. From the slope of the curve, the degradation rate can be extracted. Figure 9. presents the V vs. t graphic of PtCom_GDE and PtS_GDE. The same tendencies as seen in V-i curves are observed, hence proving the stability of the PtS_GDE. The slope extracted from the linear regression of the PtCom_GDE measurements corresponds to a degradation rate of 700 µV h−1 at an operation current density of 1.5 A cm−2. Meanwhile, the degradation rate for the PtS_GDE is completely negligible, as the line is almost flat. Figure 9. Comparison of voltage evolution with time of both samples. Figure 9. Comparison of voltage evolution with time of both samples. 4. Conclusions Pt-sputtered electrodes have been successfully developed by gas aggregation magnetron sputtering to study the effect of Pt loading reduction in the performance of a PEM single cell. The surface morphology and composition study confirmed that the catalytic surface was formed by spherical 4 nm pure Pt nanoparticles. The achievement of a 67% reduction in Pt compared with a 0.3 mg cm −2 Pt/C commercial electrode without compromising cell performance content. Although the conditions used in this work need to be transferred and adapted to a magnetron sputtering process, this work proves that magnetron sputtering is a promising method to precisely reduce PGM loadings. Finally, from the obtained results, two main conclusions can be drawn: Coatings 2024,14, 868 11 of 13 • The amount of Pt is not directly related to an improvement in cell performance. The exposition of Pt active sites seems to be much more relevant than the Pt present in the catalytic coating. • MSGA is confirmed as a practical and feasible manufacturing method for developing Pt HER electrodes. This technique successfully produces effective, stable, and effective electrodes in a step process, directly depositing pure Pt nanoparticles on the GDL. Considering the obtained results, it can be confirmed that magnetron sputtering is a feasible manufacturing method for HER low Pt loading electrodes, enabling the reduction of Pt without worsening cell stability or affecting performance. Therefore, magnetron sputtering technology could be a promising technique for the large-scale production of competitive PEMWE electrodes. Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/coatings14070868/s1, Figure S1. XPS Spectrum of the Pt 4f core-level region of PtS_GDE Pt species. Author Contributions: Conceptualization, A.V., E.G.-B., L.V.B. and S.R.; methodology, A.V., S.R. and A.A.; validation, A.V. and A.A.; formal analysis, A.V.; investigation, A.V.; data curation, A.A. and A.V.; writing—original draft preparation, A.V.; writing—review and editing, A.V., E.G.-B. and L.V.B.; visualization, A.V.; supervision, E.G.-B. and L.V.B.; project administration, E.G.-B.; funding acquisition, E.G.-B. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the H2PLAN project funded by MCIN and Basque Government with funding from European Union NextGenerationEU (PRTR-C17.I1). Spanish grants PID2020-116712RB-C21 and TED2021-131033B–I00 from MCIN/AEI/10.13039/501100011033 are also acknowledged. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Data are contained within the research article. Acknowledgments: The author would like to acknowledge the contribution of Ion Alberdi to the design and assembly of the PEM single cell. Conflicts of Interest: The authors declare no conflicts of interest. References 1. Baykara, S.Z. Hydrogen as fuel: A critical technology? Int. J. Hydrogen Energy 2005,30, 545–553. [CrossRef] 2. Wang, Y.; Pang, Y.; Xu, H.; Martinez, A.; Chen, K.S. PEM Fuel cell and electrolysis cell technologies and hydrogen infrastructure development—A review. Energy Environ. Sci. 2022,15, 2288–2328. [CrossRef] 3. Wang, T.; Cao, X.; Jiao, L. PEM water electrolysis for hydrogen production: Fundamentals, advances, and prospects. Carbon Neutrality 2022,1, 1–19. [CrossRef] 4. Ayers, K.E.; Renner, J.N.; Danilovic, N.; Wang, J.X.; Zhang, Y.; Maric, R.; Yu, H. Pathways to ultra-low platinum group metal catalyst loading in proton exchange membrane electrolyzers. Catal. Today 2016,262, 121–132. [CrossRef] 5. El-Shafie, M. Hydrogen production by water electrolysis technologies: A review. Results Eng. 2023,20, 101426. [CrossRef] 6. Kamaroddin, M.F.A.; Sabli, N.; Abdullah, T.A.T.; Siajam, S.I.; Abdullah, L.C.; Jalil, A.A.; Ahmad, A. Membrane-Based Electrolysis for Hydrogen Production: A Review. Membranes 2021,11, 810. [CrossRef] 7. Hancke, R.; Holm, T.; Ulleberg, Ø. The case for high-pressure PEM water electrolysis. Energy Convers. Manag. 2022,261, 115642. [CrossRef] 8. Grigoriev, S.A.; Fateev, V.N.; Bessarabov, D.G.; Millet, P. Current status, research trends, and challenges in water electrolysis science and technology. Int. J. Hydrogen Energy 2020,45, 26036–26058. [CrossRef] 9. Carmo, M.; Fritz, D.L.; Mergel, J.; Stolten, D. A comprehensive review on PEM water electrolysis. Int. J. Hydrogen Energy 2013,38, 4901–4934. [CrossRef] 10. Hassan, N.S.; Jalil, A.A.; Rajendran, S.; Khusnun, N.F.; Bahari, M.B.; Johari, A.; Kamaruddin, M.J.; Ismail, M. Recent review and evaluation of green hydrogen production via water electrolysis for a sustainable and clean energy society. Int. J. Hydrogen Energy 2024,52, 420–441. [CrossRef] 11. Tang, H.; Peng, Z.; Tian, R.; Ye, L.; Zhang, J.; Rao, M.; Li, G. Platinum-Group Metals: Demand, Supply, Applications and Their Recycling from Spent Automotive Catalysts. J. Environ. Chem. Eng. 2023,11, 110237. [CrossRef]