A Review on Preparation of Palladium Oxide Films

Abstract

Fabrication aspects of PdO thin films and coatings are reviewed here. The work provides and organizes the up-to-date information on the methods to obtain the films. In recent years, the interest in Pd oxide for different applications has increased. Since Pd can be converted into PdO, it is instructive to pay attention to the preparation of the pure and the alloyed Pd films, heterostructures, and nanoparticles synthesized on different substrates. The development of PdO films is presented from the early reports on coatings’ formation by oxidation of Pd foils and wires to present technologies. Modern synthesis/growth routes are gathered into chemical and physical categories. Chemical methods include hydrothermal, electrochemical, electroless deposition, and coating methods, such as impregnation, precipitation, screen printing, ink jet printing, spin or dip coating, chemical vapor deposition (CVD), and atomic layer deposition (ALD), while the physical ones include sputtering and cathodic arc deposition, laser ablation, ion or electron beam-induced deposition, evaporation, and supersonic cluster beam deposition. Analysis of publications indicates that many as-deposited Pd or Pd-oxide films are granular, with a high variety of morphologies and properties targeting very different applications, and they are grown on different substrates. We note that a comparative assessment of the challenges and quality among different films for a specific application is generally missing and, in some cases, it is difficult to make a distinction between a film and a randomly oriented, powder-like (granular), thin compact material. Textured or epitaxial films of Pd or PdO are rare and, if orientation is observed, in most cases, it is obtained accidentally. Some practical details and challenges of Pd oxidation toward PdO and some specific issues concerning application of films are also presented.

1. General Considerations on the Growth of PdO Films and Early Research as the Background of the Present Study

1.1. General Considerations on the Growth of PdO Films

According to the International Union of Pure and Applied Chemistry (IUPAC) [1], “A film is defined as a generic term referring to condensed matter restricted in one dimension. The term thin film is associated to a film whose thickness is of the order of a characteristic scale (i.e., a parameter which characterizes the density profile of a given physical quantity, which can have a static—equilibrium—or dynamic character that should be specified, e.g., terms out of plane and in plane refer to characteristic lengths normal or parallel to film surface, respectively) or smaller. Since a film may ‘look’ operationally thin or thick, according to the procedure applied, it is recommended that the measurement procedure employed be specified (e.g., ellipsometrically thin film, optically thin film, or, e.g., thick compared to the electron mean free path, thin compared to the optical wavelength, etc.)”.

In this review on preparation techniques of palladium oxide films, the presentation will stick to understanding that films are 2D objects and they are obtained on a substrate, unless otherwise mentioned, as for the cases of free-standing films. Thickness below about 10–30 nm, i.e., containing up to ~100 Pd(II) oxide unit cells (tetragonal with lattice constants a = 0.3035 nm and c = 0.5323 nm [2]), is often associated in the literature with the term ‘ultrathin’ films. Typical thin films are usually of few hundreds of nm, while thick films exceed 1 µm in thickness.

It is emphasized that many of the palladium oxide films are granular (i.e., composed of individual, well-defined particles with simple or complex morphologies). Often, a clear delimitation between the methods and procedures to obtain particles in a powder (on a substrate) and in a thin film is not possible. In this work, mostly the technologies demonstrated to produce a conventional relatively continuous film, usually with macroscopic surfaces useful for device fabrication, are briefly introduced. These films are Pd(II) monoxide films, although palladium shows different oxidation states. The common ones are 0, +1, +2, and +4 [3]. The most important oxide forms are the monoxide PdO, dioxide PdO₂, and trioxide PdO₃, as well as a hydrated sesquioxide Pd₂O₃ × H₂O [4]. In the organic synthesis methods, Pd(0) and Pd(II) are usually considered [5].

Oxidation states +3 and +6 have been reported [6,7,8] in intermediate compounds proposed to occur in the palladium-catalyzed cross-coupling reactions. Authors of [9] theoretically studied the stability and activity of Pd₄O_x (x = 1–7) films on Pd metal within a CO oxidation reaction model. Other articles also present nonstoichiometric palladium oxides PdO_1−x [4,10]. Finally, there are films in the literature referred to as Pd/PdO that assume a Pd core region covered by a Pd-oxide surface layer. These types of particles and thin films (including materials marked with PdO/Pd), containing both metal and oxide species, are of interest for various applications, especially where oxidation–reduction processes play an important role. Since there are no generally accepted terms, on many occasions, materials described as Pd or PdO are in fact a mixture of the metal and oxide [11].

1.2. Early Research as the Background of the Present Study

Three groups of methods for film preparation are available:

(1)

Preparation of a Pd metal coating on a substrate, followed by oxidation.

(2)

Preparation of a PdO film by thermal decomposition of salts.

(3)

Preparation of the PdO oxide films by reactive methods.

The proposed classification is arbitrary, and often technological routes share the same processing steps and raw materials. Considering this, in this presentation, this classification is not strictly followed.

An early approach was to use a Pd metal foil or wire as a substrate, subject to surface oxidation. The functional product would be a PdO film coated on the Pd metal. This configuration was used as a pH sensor at high temperatures in geothermal brines, corrosive liquids in contact with nuclear waste containers, pressurized steam/water lines in power generation, and pressurized chemical reactors [12]. These electrodes are also suitable for measurements of pH in biological environments, such as blood and extracellular fluids [13,14], or for neurophysiological measurements [15]. It is well known that PdO is biocompatible, hence, it can also be used in wearable devices.

A high-temperature pH electrode should provide a direct voltage response to a pH value, negligible ionic and redox interfaces, minimum hysteresis, and drift. Only the oxide should be in contact with the solution. These requirements are difficult to fulfill in terms of palladium oxide thin films’ quality and thickness. An additional needed condition is miniaturization and low cost [13,16].

Three methods were attempted for fabrication of the Pd oxide film on Pd metal foils or wires:

(i)

Betteridge and Rhys [17] demonstrated that, when the Pd metal is heated to 700 °C in air, a PdO film forms on its surface. The electrode was not stable, and the utility of the thermal oxidation method for these purposes was reconsidered.

(ii)

Grubb and King [14] used chemical oxidation of the metal surface. The Pd metal surface, after cleaning in aqua regia for 20 s and roughening, was dipped into a 50% aqueous solution of NaOH and dried for 10 min in a stream of dry nitrogen. The coated Pd wire was heated in a flow of oxygen at 800 °C for 20 min. The portion covered with NaOH resulted in formation of a black PdO layer. The oxide layer was uniform, nonporous, and its thickness was of about 6 µm. Kinoshita et al. [18] expanded this research and studied different parameters, such as oxidation temperature in the range of 600 and 870 °C. They found that the best results were obtained for the electrodes processed at 750 °C after having been immersed in a 50% NaOH solution.

Thermal oxidation and chemical oxidation, as presented in points (i) and (ii), need relatively high processing temperatures. This impedes the preparation of oxide films and devices on low-temperature substrates, such as flexible ones (plastics, paper, and textiles), although Pd films have been coated by other methods on flexible, elastomeric, or optical fiber substrates [19,20,21,22]. Another inconvenience is that the reproducibility and stability of the oxide electrodes produced by the two addressed routes have been found to be highly sensitive to processing conditions, electrodes’ design, and sizes. Other important issues are stoichiometry control of PdO vs. oxidation temperature and decomposition of PdO. Decomposition of PdO was established both theoretically and experimentally to occur above 870–900 °C [23]. Some practical aspects regarding oxidation are addressed in the next paragraphs, while details in terms of time and spatial development of the metal–oxide interface will not be approached in detail.

Considering the presented information, the search for new methods to obtain PdO films on Pd or on other substrates has continued.

(iii)

An alternative method was proposed, namely, the electrochemical oxidation. Kinoshita et al. [18] oxidized the metal electrodes by treating them in 0.2 M NaOH for 5 min at 2.7 V and for 17 h at 0.74 V. Electrodes lost their near-Nernstian response to pH after a few acid–base cycles. A longer lifetime of less than 6 days was determined for an electrode fabricated by Liu et al. [13]. The authors used a molten salt electrolyte of NaNO₃/LiCl for oxidation, with a current density of 20 mA·mm⁻² at 5.9–6.2 V for 90 s. In [24], the authors applied electrolysis waveforms containing various levels of the ac and dc components for a processing time up to 24 h. The pH sensor’s response depends on the processing conditions. The authors also observed surface morphology changes.

Recent review articles summarize the work on synthesis of powders and coatings of Pd [3,25,26,27]. Powders and granular films of Pd or Pd-based materials are intensively investigated and used as catalysts. Rare are cases when they are oxidized to obtain palladium oxide as a product, although the methods from points (i)–(iii) are expected to also be successful in these cases. A straightforward example is thermal oxidation (in air at 300 °C for 2h) of PdCo nanoparticles and formation of a PdO-doped Co₂O₃ nano-powder [28,29].

Oxidation can be a general route to obtain Pd-based oxide, oxide compounds, or oxide composites from a large class of Pd-based alloys. Although, as already pointed out, oxidation of many reported Pd products was not performed, in the next paragraphs in this section, some details regarding the types of alloys and the methods used for their synthesis are introduced. Whenever available, the information on oxidation is also provided.

Palladium is shown to combine with other metals. Bimetallic or ternary alloys in the nano-powder form were synthesized: Pd-Pt, Pd-Au, Pd-Ag, Pd-Co, Pd-Co-Au, Pd-Cu, Pd-Ni, Pd-Rh, PdBi, Pt-Pd-Bi ([25,27] and references therein, [11,30,31,32,33]), Pd-Fe [34], and Pd-Ti [3]. Heterostructures and supported catalysts can enhance the chemical performance due to cooperative and synergetic interactions of the components [35]. In these cases, one takes advantage at the interface between Pd and the support on the charge transfer, strain effects, nanoscale surface stabilization, preferential faceting, prevention of the catalyst deactivation and regeneration, control of selective catalytic oxidation reactions, and improved gas-poisoning resistance. Supports of Pd tested in the literature are carbon black, activated carbon, carbon spheres, carbon nanofibers, single-wall (SWCNT) and multiwall (MWCNT) carbon nanotubes, graphene (G) and graphene oxide (GO), nanodiamond [36,37,38,39,40,41,42,43,44,45], Ti, TiO₂, TiO₂ nanotubes, Fe₂O₃, MnO₂, V₂O₃, ZnO ([25] and references therein, [46]), Fe₃O₄, Fe₃O₃-C core-shell microsphere, Fe₃O₄-polyaniline and Fe₃O₄-NiSiO₃ composites ([47] and references therein), CeO₂ [48,49], CuO [26,50], sulfur-doped WO₃ [51,52], Cr₂O₃ [53], MgO [54], La₂O₃ [55], kaolin [56], CuMn₂O₄ [57], MnSiO_x [58], FeCrAl fibers [59], NiO [60], Co oxide [61], In-Sn-Zn-O-sputtered thin films [62], and wool [63]. Catalysts of Pd nanoparticles in which GO particles are encapsulated are also reported [64]. We observed that many of the supports are oxides stable at oxidation temperatures of metallic Pd and, therefore, they can be useful, depending on applications, as substrates for thin-film preparation of PdO-based materials. Other supports for Pd metal or alloy clusters are micelles, dendrimers, and mesoporous materials (e.g., polycarbonate, carbon tube, alumina, porous Ni, SiO₂, and zeolite) [65,66]. Often, these supporting materials play the role of hard nanostructured templates, and the skeleton of the synthesized Pd objects follows the porous architecture of the support [3]. Materials of Pd/PdO clusters in glasses were also fabricated [67]. Apart from the use of templates, the crystal chemistry of Pd as a cubic face-centered (fcc) metal is itself highly favorable to formation of different geometrical shapes/morphologies [68] of the particles, both in the powders and in the coatings. The resulting morphology is controlled through manipulation of the thermodynamic aspects, involving high reduction rates, or of the anisotropy kinetics of the crystal growth. Size and especially morphology have a strong impact on the catalytic properties of nanocrystals. Fabrication and catalytic activity of Pd or Pd-based powders composed of particles with well-defined morphology, uniformity, and size is currently a field of significant interest both from academia and industry. This is a rich domain that is developing fast and in which many excellent articles were published in recent years. Despite this tremendous effort, as already mentioned, research and development rarely proceed toward synthesis of Pd oxide products from the as-fabricated Pd or Pd alloys.

Instead of alloys, other precursors or raw materials can be used, and the technological routes usually employ two processing steps. In the first one, a precursor film is obtained and, in the second one, the precursor is decomposed, reacted, or oxidized to obtain the Pd oxide. An example of a modern route is the growth of a nanoparticulate PdO thin film by the Langmuir–Blodgett (LB) technique, followed by thermal decomposition of the multilayer precursor film of octadecylamine (ODA)–chloropalladate complex [69]. Many other modern synthetic routes usually have several processing steps, one of which is mandatory of a chemical type: a granular metallic Pd⁽⁰⁾, alloy, or organometallic film/coating is obtained in the first step, and it is further used as a precursor for preparation of the Pd oxide film final product by decomposition/reaction/oxidation. Nevertheless, the final oxide film can also be prepared by physical methods and by one-step routes, reactive or not. In the next sections, the review will consider two groups of technologies, namely, chemical and physical. This classification is not strict and should be taken as arbitrary.

Depending on the application, it is of interest to also take into account the reduction of PdO into Pd. In [70], Pd nanoparticles were decorated on PdO hollow-shell supports through a one-step treatment with NaBH₄. In this process, PdO is partially reduced to Pd. The H₂ gas detection limit of this material is lower than 1 ppm, and the proposed strategy can be employed to extend the lifetime of the H₂ gas sensors. Another approach is precipitation on the support particles. In [71], PdCl₂ was precipitated on ZnO nanoparticles and decomposed at 500 °C to Pd. Decoration of different support particles (WO₃, ZnO-Eu, ZnO, β-Bi₂O₃, In₂O₃, SnO₂, and V₂O₅) with functional Pd [71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89] or PdO [88,90,91,92,93,94,95], mainly for improved gas-sensing applications, is gaining much attention.

2. Chemical Routes of PdO and Pd Precursor Films’ Growth

In the category of chemical preparation methods of PdO or Pd precursor films (Figure 1), hydrothermal, electrochemical, electroless deposition, and coating methods can be included, such as the already mentioned Langmuir–Blodgett method, impregnation, precipitation, screen printing, ink jet printing, spin or dip coating, chemical vapor deposition (CVD), including its popular variant atomic layer deposition (ALD), and green biosynthesis. Chemical routes may involve in situ oxidation or reduction processes (e.g., in reactive synthesis routes or salts’ decomposition routes). As mentioned in Section 1.2, additional processing steps for reduction or oxidation (such as thermal, chemical, and electrochemical treatments) are sometimes applied on the precursor films already deposited by physical or chemical routes.

2.1. Hydrothermal Method

The hydrothermal method was employed in [96] to obtain PdCd nanostructures on a Ti plate. The Ti substrate was introduced in the Teflon autoclave containing 10 mL of 5 mM PdCl₂, 5 mM Cd(NO₃)₂·4H₂O, and 1 M ammonium formate as a reducing agent. The Pd/Cd ratio was varied by the amount of Cd(NO₃)₂·4H₂O precursor. The processing temperature was 180 °C for 2 h. The coated Ti plates were further heated in a furnace with a controlled Ar-flow atmosphere at 250 °C for 2 h. The PdCd film had a dendritic microstructure.

In other examples ([25] and references therein) of hydrothermal synthesis of Pd or Pd-based bimetallic nanomaterials, a popular source of Pd is H₂PdCl₄, and formaldehyde (CH₂O) and ethylene glycol (EG) are used as reducing agents. Composite PdO/CoNi₂S₄ particles were obtained via the hydrothermal method in [97].

The hydrothermal method is considered a single-step green technology that can use nontoxic solvents, and it minimizes waste. On the other hand, the productivity of this method is relatively low, and synthesis of large amounts of nano-powders or of large film surfaces is challenging. Via hydrothermal synthesis, different controlled shapes, such as nanospheres, nano-cubes, nano-tetrapods, nanoporous bulks, nanowires/dendrites, and core-shell particles on different substrates, are obtained [25].

A facile sol-gel hydrothermal method to obtain PdO particles is presented in [98]. After dissolving PdCl₂ in C₂H₅OH (0.1 Molar solution), 1 vol.% HCl is added to enhance the solubility of PdCl₂. Stirring for 30 min results in a transparent brown solution. The solution is aged for 24 h, loaded into a Teflon hydrothermal capsule, and heated at 200 °C. The as-obtained brown powder is washed with D.I. water and annealed at 600 °C. The final powder is fine and black and is used in the electrophoretic deposition to obtain a PdO film covered with ITO on a glass substrate for H₂ gas detection at room temperature (Figure 2).

Wang et al. [99], via the hydrothermal method at 200 °C, obtained a ring-like PdO-NiO architecture, with a lamellar structure of 3.5 μm in diameter composed of nanosheets with a thickness of ~15 nm, for CO gas detection, showing a relatively low operating temperature (180 °C) and a rapid response/recovery (2–3 s when exposed to 50 ppm CO).

2.2. Electrochemical Deposition

In the group of electrochemical deposition methods, in an electrochemical cell with two or three electrodes (working electrode, counter electrode, and reference electrode), the electrolyte is conductive, and it is also the source of Pd. A current or potential is applied on the electrochemical cell, and ions of Pd^(+II) in solution will deposit in the metal state (Pd⁽⁰⁾) at the surface of electrodes. Depending on the current or potential features, different versions of electrochemical depositions were defined: cyclic voltammetry (CV), square-wave voltammetry (SWV), chronoamperometry (CA), chronopotentiometry (CP), chrono-coulometry (CC), and pulsed electrodeposition ([25] and references therein).

The electrochemical deposition is considered a simple, affordable, and relatively fast method. Popular Pd raw materials are PdCl₂ and H₂PdCl₄, but other materials can be used, e.g., K₂PdCl₆ [100] or Pd(NH₃)₄Cl₂ [101]. Different morphologies are obtained in the electrochemical deposition. Nanoparticles with diameters of 5–10 nm and nanorods or nanowires are the most common. Pd nano-horns or nanoparticles of PdAu tube-shaped heterostructures were electrochemically deposited in [100,102], while an AgPd coating (on an oxidized/activated stainless-steel substrate), tested as a beneficial catalyst for low-temperature fuel cells’ application, was obtained in [103]. Fabrication of nanostructured bimetallic Pd-Fe thin films and their electrodechlorination activity are reported in [104]. The electrochemical deposition allows the use of template substrates, e.g., anodized aluminum oxide (AAO) or porous anodic alumina (PAA) [25]. After the directional template-assisted growth into the parallel pores, the substrate is removed, and the product is a sort of film consisting of an array of Pd nanowires [101]. By using an AAO template and subsequent etching of PdCu nanorods, Pd nano-springs were obtained in [102]. An open-pore Pd foam was electrodeposited, and a Pd-CeO₂ catalyst for environment processes was reported in [105]. It was observed that most works report deposition of metal particles and films, rather than of the Pd oxide. Additional information can be found in [25].

In [98], films of PdO nanoparticles were deposited on ITO-coated glass substrates (5 mm × 10 mm) with electrophoretic deposition (EPD). The anode is the ITO substrate, and the cathode is a Pt rod. The distance between electrodes was set to 1 cm, and the electrodes were placed in a 20 mL glass vessel. A PdO powder (0.1 g) obtained by the sol-gel hydrothermal method (see the previous paragraph) was dispersed in 10 mL of ethanol. In this electrolyte solution, as a dispersant and to enhance conductivity and mobility of the ions, magnesium nitrate (Mg(NO₃)₂·6H₂O, 100 μL) was added. The deposition potential applied for 2 min was between 10 and 35 V. The authors indicated that the quality of the film depends on the deposition voltage and time. The optimum voltage would be 30 V. Films were used for H₂ gas detection at room temperature by using an amperometric gas-sensing setup (Figure 2).

It is important to emphasize that PdO oxide films can be used not only for reduction gases’ detection, but also for oxidizing ones. An example of the results obtained with a chemoresistive (conductomeric) sensor sensitive to ozone is presented in Figure 3 [106].

The somehow related method of electrospinning is also a straightforward fabrication coating technique, and it was applied in [107] to obtain a decorated Pd nanolayer upon aligned polyurethane nanofibers for H₂ gas sensors. Among the coatings obtained by electrospinning of PdO, Au, and CdO on SnO₂ nanofibers, the highest sensitivity to toluene was determined for PdO-SnO₂ [108].

2.3. Electroless Deposition

The group of electroless deposition techniques has two main variants, namely, the displacement (galvanic replacement) and autocatalytic deposition [25]. These methods take advantage of a photo-reductant, a stabilizer, or a chemical reducing agent. Microemulsion and photochemical synthesis also belong to these types of methods, which are without an external current source but share a common ground with electrochemical routes.

In the displacement deposition, 25 µL of 10 mM H₂PdCl₂ was placed on Fe thin films for 30 min [104]. Reactions between PdCl₄²⁻/Pd and Fe²⁺/Fe developed with standard redox potentials of 0.95 and −0.44 V, respectively. The final product was washed with ultrapure water and dried. The less noble metal acted as a reducing agent, and the reaction proceeded at its surface. This produced certain limitations. Other limitations were imposed by the required favorable difference in the reduction potentials of the two metals.

The autocatalytic deposition uses citric acid, ascorbic acid, ethylene glycol, sodium borohydride, and others as reducing agents. The impact of the reducing agent is very strong and can modify growth kinetics with changes in the particles’ morphology and in the surface properties ([25] and references therein). In general, the reducing agents are added into aqueous solutions of PdCl₂, NaPdCl₄, or H₂PdCl₄. The method is suitable to obtain bimetallic Pd-based nanomaterials. It also has the advantage that the processes proceed at the surface and in the bulk of the metals. The polyol synthesis is a derivate method. Metal-containing compounds are obtained through the use of ethylene glycol, playing the role of both the solvent and the reducing agent. Without templates, Wang et al. [109,110] synthesized nanowires with uniform diameters of 2 nm via this technique. Synthesis proceeded from a solution of (CF₃COO)₂Pd dissolved in diethylene glycol (DEG). When instead of (CF₃COO)₂Pd, the authors used Na₂PdCl₄, Pd polyhedrons were obtained.

Microemulsion techniques use water in oil to control the particle size. Surfactant-stabilized reverse micelles are the microreactors. Typical surfactants are dioctyl sodium sulfosuccinate (AOT) and cetyl trimethyl ammonium bromide (CTAB). In some cases, heptane is used as an oil phase, while toluene is a co-surfactant. Application of surface-confining agents, such as CO that are preferentially binding on the Pd(111) surface, produced nanosheets of Pd with a thickness of 1.8 nm [111]. Microemulsion is shown to be suitable for the synthesis of Pd-Co-Au/C tri-metallic catalysts [112]. Yuasa et al. [113] investigated preparation of composite particles of nano-sized PdO loaded on SnO₂ nanoparticles by precipitating Pd(OH)₂ and Sn(OH)₄ inside a reverse micelle. CO gas-sensing properties of the particles were highly sensitive and were maximized for a small amount of PdO (0.1 mol %), at which PdO agglomeration is avoided. This method allows loading of small amounts of PdO when compared to the conventional impregnation method.

Nanoparticles of PdAu were obtained on TiO₂ via photochemical synthesis. Tian et. al. [114] immersed a plate with TiO₂ nanotubes into a water solution of 0.04 M Pd(NO₃)₂ and AuCl₃. In the solution, 5 mL (50 vol.%) of methanol was also introduced. After 30 min of exposure to UV radiation, in the solution, PdAu particles with the size of ~20 nm formed.

2.4. Coating Techniques

The group of coating techniques, such as Langmuir–Blodgett, impregnation, precipitation, screen printing, ink jet printing, and dip and spin coating, due to their simplicity and low cost, are often employed to obtain Pd-based precursor powders or films [10,94,115,116,117,118]. The idea of these synthesis methods is to synthesize particles or films, usually of palladium chloride, nitrate, or hydroxide, which are subsequently thermally decomposed.

Jin et al. [119] used the conventional impregnation method by dipping a honeycomb cordierite substrate coated with Ce_xZr_y oxides (synthesized by co-precipitation with urea of a Ce, Zr nitrate solution, and calcination at 550 °C in air) into the aqueous solution of palladium nitrate, followed by drying and calcination in air at 500 °C for 2 h, to obtain a brown-colored catalyst complex with Pd (metal loading on the monolith of about 0.3 wt.%). This material was investigated as an alternative catalytic technology to oxygen removal of coal mine methane. To optimize CO₂ methanation pathways, coatings of Pd/P-CeO₂-Al₂O₃ were prepared via multistep impregnation in [120].

In [121], decorated flower-like ZnO structures with PdO obtained by precipitation in a methanol solution with dispersed ZnO particles and with dissolved PdCl₂ (60 mg ZnO, 3.3 mL methanol, and 6 mg PdCl₂, shaken for 30 min), followed by calcination at 350 °C for 1 h (heating rate of 5 °C/min), showed better sensing activity of toluene and ethanol than of ZnO (Figure 4).

Use of PdCl₂ as an initial reagent is economically convenient since it is produced at an industrial scale and it is relatively cheap, but traces of residual chloride anions can show a strong negative influence on the sensitivity of PdO gas sensors. For this type of device, it is recommended to convert PdCl₂ (e.g., through a reaction with NaOH or AgNO₃ [10]) or to use the palladium hydroxide or nitrate directly as initial reagents. The typical media to dissolve the initial reagents is water. Balamurugan et al. [118] dissolved Pd(NO₃)₂·H₂O in weak triprotic citric acid monohydrate (C₆H₈O₇·H₂O). Powders or films are thermally decomposed in air or in dry oxygen to obtain PdO. In [10], PdO was obtained by decomposition of Pd nitrate at 400 and 600 °C in dry oxygen. In [115], samples of PdNO₃ impregnated on La_0.6Sr_0.4Co_0.2Fe_0.8O_3−x for cathodes in solid oxide fuel cells working at intermediate temperatures were calcinated at 800 °C for 2 h, in air. Nanohybrids of PdO-WO₃ were obtained by annealing at 400 °C for 24 h in air to decompose PdCl₂ into PdO on WO₃ nanorods [117].

Other initial reagents than those already mentioned can be used. A proposed precursor material for spin coating is the β-ketoiminato palladium(II) complex [Pd(OAc)L] (L=PhC(O)CHC(Me)NCH₂CH₂NHCH₂CH₂NH(Me)C=CHC(O)Ph) [122]. It decomposes at 200–500 °C and forms Pd in argon and PdO in oxygen atmospheres. Kabcum et al. [117] impregnated 0.25–2.0 wt.% Pd onto WO₃ powder using palladium acetylacetonate (Pd(C₅H₇O₂)₂ dissolved in ethanol. After drying at 80 °C for 24 h, a heat treatment was conducted at 300 °C for 2 h. The powder (60 mg) was mixed (30 min) with a homogeneous binder solution (0.28 mL) of ethyl cellulose dissolved (at 80 °C for 12 h) in α-terpineol solvent to obtain a paste. The paste was spin-coated at 3000 rpm for 30 s on Al₂O₃ substrates. Annealing was performed at 150 °C for 1 h and then at 450 °C for 2 h (heating rate of 1 °C/min) for binder removal. The advantage of using organometallic reagents as a precursor is that it is convenient for decreasing the decomposition temperature.

Thick films in the SnO₂–CeO₂–PdO_x system were obtained by sol-gel screen printing on alumina substrates in [123]. A solution of 20 wt.% of SnCl₄ in water was mixed with solutions of ceric ammonium nitrate and palladium chloride. Ammonium hydroxide solution (25% in water) was added dropwise to the mixture to obtain a pH of 9.0, followed by stirring for 20 min. The resulting solution was aged for one day at room temperature. The volatiles were removed under reduced pressure at 70 °C to obtain powders. Powders were used to obtain a paste that was screen-printed on ultrasonically cleaned alumina substrates with Au electrodes and with a heater on the backside. The printed films were heated at 700 °C for 2 h. The as-fabricated sensor had enhanced sensitivity for CO gas detection.

In [116], the precursor solution of a Pd-organoamine was spin-coated (500 rpm for 10 s, followed by 3000 rpm for 60 s, using an acceleration of 800 rpm/s in both stages) on glass substrates and annealed at 200 and 250 °C (4 min, 24 h, and 48 h). This approach allows ink jet printing, as demonstrated in [124]. Amine-stabilized Pd clusters were obtained after decomposition of the precursor. Subsequent thermolysis of the clusters under air yielded smooth films. Under nitrogen atmosphere, the films consisted of large micron aggregates, while in low vacuum, Pd/PdO bilayer films were obtained (Figure 5).

2.5. Chemical Vapor Deposition

A review on the growth via chemical vapor deposition (CVD) of iridium, platinum, rhodium, and palladium thin films is presented in [125]. The work indicated that for the deposition of Pd thin films, with the exception of PdCl₂, the precursors were of organic type: metal β-diketonates (Pd(acac)₂, Pd(hfac)₂), allyl complexes (Pd(allyl)₂, Pd(Me allyl)₂, CpPd(allyl), Pd(allyl)(acac), Pd(allyl)(hfac), Pd(Me allyl)(hfac), Pd(Me allyl)(acac), alkyl complexes (PdMe₂(PMe₃)₂, PdMe₂(PEt₃)₂, and PdMe₂(tmeda)), metal carboxylate (Pd(acetate)₂), and ethylene complexes (Pd(C₂H₄)₂), where acac = acetylacetonate = C₅H₇O₂, hfac = hexafluoroacetylacetonate = C₅HO₂F₆, allyl = η³-C₃H₅, Cp = cyclopentadienyl = η⁵-C₅H₅, tmeda = tetramethylethylenediamine = C₆H₁₂N₂H₂, acetate = C₂H₃O₂, and vinyl = C₂H₄. Other organic precursors proposed and reported in recent years will be mentioned in the next paragraphs. The Pd thin films were also grown via CVD in [126,127,128,129]. The PdCu alloy films on Si or SiO₂ substrates were deposited by CVD from a single-source precursor (β-diketonate complex) at 250–400 °C [129].

Among the CVD group techniques, the aerosol-assisted chemical vapor deposition (AACVD) is considered a promising variant, being simple and easy [130]. The precursor material was the commercially available palladium acetylacetonate, Pd(acac)₂ (Pd(C₅H₇O₂)₂, 100 mg, 0.328 mM), dissolved in toluene (10 mL). The aerosol mist from Pd(acac)₂ solution was generated by a piezoelectric ultrasonic humidifier. The aerosol was injected into the reactor tube with the help of a nitrogen (N₂) gas flow (150 cm³/min). No hydrogen gas or any other reducing agents were used. Deposition for 30 min was performed on glass substrates heated to 475 °C.

Via AACVD, synthesis of palladium-nanoparticle-decorated WO₃-nanoneedles was accomplished in a single step [131,132]. Two source materials were tested: palladium acetylacetonate and ammonium hexachloropalladate. The first one led to better quality: the response of the PdO-decorated WO₃ toward hydrogen gas sensing was about 680–750 times higher than that of bare WO₃ nanoneedles. The operation temperature was at relatively safe and low values of 50–150 °C, and the addition of PdO minimized the effect of humidity on the sensor’s response.

The AACVD was successfully demonstrated for fabrication of Pd-decorated WO₃ H₂ gas sensors on a flexible kapton substrate [133].

Pd nanoparticles were deposited by AACVD on an array of WO₃ nanorods [134]. The source material was (NH₄)₂PdCl₄ in a methanol solvent, the deposition temperature was 350 °C, and the N₂ gas flow rate was 300 sccm. Oxidation to PdO was performed in air at 500 °C for 2 h.

In [135], films of Pd (20–120 nm) were prepared on glass substrates via plasma-enhanced metalorganic chemical vapor deposition (PEMOCVD). The source material was Pd(acac)₂ (vapor pressure 0.1 hPa at 433 K) and the carrier gas was Ar (0.71 cm³ min⁻¹). The total pressure in the deposition chamber was 4.5 Pa. The glow discharge power was 40 W. The sublimation temperature of Pd(acac)₂ was found to influence the films’ composition relative to the Pd and C ratio in the as-deposited films. The optimum sublimation temperature was determined to be 413 K. Films were thermally decomposed at 623 K for 900 s, and they were composed of Pd nanoparticles in the range of 5–10 nm, which is considered suitable for catalytic activity.

In the CVD group methods, precursors are evaporated or sublimated and often show problems of thermal instability, resulting in poor quality and reproducibility of the as-deposited Pd film. To avoid formation of oxides is challenging, and the typically slow deposition rates in CVD are not convenient for high productivity and commercial fabrication. Some of the most popular gas-phase precursors are η³-allyl and β-diketonato palladium complexes [136,137,138]. Among those of the first type, [Pd(η³-C₃H₇)₂] [139] and [Pd(η³-2-Me-C₃H₄)₂] [140] show thermal instability during evaporation and storage. They are also highly sensitive to air and moisture [137]. Another problem is that without the aid of a reactive gas [140], the deposited Pd films contain a low amount of carbon impurity (1 wt.%). The precursors of the second type, bis (β-diketonato) palladium (II) complexes, such as [Pd(acac)₂] (acac = acetylacetonate), are thermally stable, but their volatility is low, and they require higher decomposition temperatures [138]. A combination of the allyl and β-diketonate groups within one complex was proposed in [141] to overcome the deficiencies of each type. Other synthesized CVD precursors are β-ketoiminate [142], β-ketoiminato palladium (II) complexes, such as [Pd(CF₃C(O)CHC(CF₃)NⁿBu)₂] [143,144], [Pd(CH₃C(NH)CHC(O)CH₃)₂] [145,146], [Pd(CH₃(O)CCHCN(CH₃)(CH₂)₃)₂)] [147], and [Pd(2-NC₅H₄)NHNCH-1-(O)-4-OMe-Ph)(PPh₃)] [148], and palladium(II) allyl (β-ketoiminato) complexes [149], e.g., [Pd(ƞ³-2-Me-C₃H₄)(MeC(O)CHCMeNPh)]. The oxygen content in some of these precursors is minimized, leading to a lower oxidation of the substrate during formation of the metal film [145,146]. The complexes have a higher stability to air and moisture [150]. The search for new MOCVD source materials [151] continues, and it is one of the key priorities for the development of CVD technology.

Via atomic layer deposition (ALD), a variant of CVD, Pd thin films were successfully grown in [144,152,153,154,155,156,157]. In ALD, precursors are used that react with the surface of the substrate in a sequential way, one at a time. The chemisorbed metalorganic compound (e.g., β-ketoiminate [154] or Pd^(II)(hfac)₂ precursors [152]) decomposes under the influence of the second precursor, and the organic ligand is removed to form a metal film. The second precursor can be an organic (e.g., glyoxylic acid [152]) or a gas, such as oxygen [144]. In general, ALD is recognized as a technique that allows better control of the thickness and density. A recent review [157] on ALD of Pd pointed out the enhanced possibilities of this technique. Precise control of Pd deposited on challenging surfaces with different precursor chemistries and by using customizable processing conditions can generate low-dimensional nanostructures (single atoms, nanoclusters or nanoparticles, core-shell particles, and ultrathin films). The disadvantages of ALD are the slow deposition rate (in the range of nm/min) and that it allows growth of films with rather small surface areas.

2.6. Green Biosynthesis

The need for clean technologies led to a search of green biosynthesis routes ([5] and references therein). Natural, non-toxic, and eco-friendly precursor substances can be used to obtain usually metal particles. Examples of such substances are Aspalathus linearis (Rooibos) and Annona squamosa peel extracts, banana peel extract, Cinnamon zeylanicum bark, Curcuma longa tuber, broth of Cinnamon camphora leaf, Gum acacia, Solanum trilobatum leaf, Soybean leaf, Zanthan gum, or, more recently, Crocus sativus [158], coconut coir [159], Trigonella foenum-graecum (fenugreek) seed [159], Rosmarinus officinalis L. (rosemary) [160], and Zaleya decandra [161]. Among the most bioactive substances, aspalathin, nothofagin, and aspalalinin are considered. However, the choice of the plant and processing method and conditions are critical regarding the quality and quantity of the extracted compound. Aspalathus linearis leaves (0.03 g) were added to 300 mL of deionized water at room temperature for ~30 min [5]. The orange extract (pH~5) was filtered three times to remove solid residuals and was slightly fermented. In the as-prepared extract, PdCl₂ (3 g) was added and fully dissolved under stirring at room temperature. After 30 min of reaction time necessary to complete the oxidation–reduction reaction, the color of the solution changed. By drying in air at 90 °C for 1 h, an amorphous product was obtained, which was further annealed for 2 h at 100 °C or at a higher temperature. The product annealed at 100 °C was composed of non-agglomerated crystalline nanoparticles with different morphologies and with an average mean size diameter of 12.7 nm. Via TEM, particles were ascribed to fcc cubic Pd. At higher annealing temperatures (200–600 °C), the product was oxidized. At intermediate temperatures, X-ray diffraction indicated co-existence of Pd and tetragonal PdO, while at 600 °C, only PdO was detected. Time and/or temperature to obtain PdO were considered [5,162] lower than for other processing routes, such as heating of PdCl₂ in a NaNO₃ melt, thermal decomposition of Pd(NO₃)₂·2H₂O, oxidation of Pd in air or in oxygen, or direct carbon arc vaporization of Pd.

In [37], citric acid was used as a mild and environmentally friendly stabilizing and reducing agent to obtain nanohybrids of graphene oxide decorated with Pd. Graphene oxide was treated with aliphatic amines 1-octadecylamine and 1,8-diaminooctane, followed by addition of a palladium chloride in a liquid medium and citric acid.

3. Physical Routes of Pd and PdO Thin Films’ Growth

Through physical technological methods, either the Pd metal precursor or the Pd oxide films can be prepared in a single processing step. Among the physical methods are sputtering and cathodic arc deposition, laser ablation, ion or electron beam-induced deposition, evaporation, and supersonic cluster beam deposition (Figure 1).

3.1. Sputtering and Cathodic Arc Deposition

Sputtering a target of a metal through the bombardment with energetic ions of an inert gas (e.g., argon or helium) produces collision of these energetic ions with the target and ejects target metal atoms into space. These metal atoms are deposited on a substrate material, forming a metallic film. Sputtering is recognized as a fast-rate vacuum coating technique that allows creation of strongly adhesive coatings on substrates with complex geometries. Magnetron sputtering is popular for deposition of metals and oxides. The energetic ionic species and electrons are confined in a magnetically enhanced glow discharge. Confinement produces a relatively high plasma density and promotes acceleration of the ions, which will vaporize the source material (target). The ion or molecular ion-sputtered species from the target, i.e., Pd in this case, fly and deposit on the substrate. If oxygen, as a reactive gas, is introduced in the deposition chamber, one can obtain in situ Pd oxide films. In practice, to avoid oxygen contamination of the sputtering chamber, thermal oxidation of the as-prepared Pd films is preferred to direct reactive sputtering. Simultaneous co-sputtering from different targets is applied to obtain bi- or multi-component metal alloys. This is performed by tuning the power applied on each sputtering gun/target. The drawback of sputtering is considered its limitation in deposition of films into deep structures. The literature presents different solutions of directional sputtering approaches, such as collimated sputter deposition, elevated temperature/reflow deposition, long-throw sputter deposition, and ionized magnetron sputtering [163,164].

Thin films of PdO on (111) Si-wafers and alumina circuit-board substrates were deposited by magnetron sputtering from a Pd target in [12]. The deposition atmosphere was O₂:Ar = 1:1 for a total pressure of 0.35 Pa (starting base pressure was (2–4) × 10⁻⁴ Pa). The as-grown film consisted of Pd oxide, according to XRD, but an unidentified impurity peak was detected at 2θ = 56°. The lattice parameters of PdO increased, and the XRD line of the impurity disappeared during post-annealing at 420 °C in air.

Yoon at et al. [165] deposited, for hydrogen gas sensors, Pd thin films (50 nm) by magnetron sputtering in Ar gas at a pressure of 1.2 × 10⁻² Torr (initial vacuum pressure was 1.2 × 10⁻⁶ Torr) and for an input power of 1000 W.

Via dc magnetron sputtering, Pd films were obtained in [166] on oxidized Si (SiO₂ thickness was 100 nm) and for a source power of 60 W. The distance between the target and substrate was 10 cm, and an Ar atmosphere was used (initial vacuum was 10⁻⁶ Torr). When the deposition pressure was increased from 22 to 122 mTorr, the average grain size in the films changed from 10 to 30 nm.

Hao et al., via dc magnetron sputtering, obtained porous films of Pd on porous anodic aluminum oxide (AAO) templates for room-temperature chemoresistive hydrogen sensors [167] (Figure 6). After degassing the AAO substrate in a vacuum (1 × 10⁻³ Pa) by heating in the sputtering chamber at 100 °C, Pd films were deposited at ~50 °C in an Ar (99.999%) atmosphere (made from a vacuum of 4 × 10⁻⁴ Pa) at a pressure of 0.7 Pa. As-prepared Pd films (45 nm) were annealed for 30 min in the deposition chamber at 150, 200, 300, and 400 °C. It was shown that annealing decreased the response time (time delay to reach 90% of resistance), and the drift of the baseline for the annealed sensor was not observed.

Magnetron sputtering was also used in [168] to grow Pd thin films on Ti–6Al–7Nb plates. After a heat treatment in air at 450, 550, and 650 °C for 1 h, their mechanical, corrosive, and wettability properties and in vitro bioactivity were studied. These materials are considered for implant applications.

Other examples of PdO-sputtered films were reported in [169,170,171,172]. Reactive sputtering was employed to obtain PdO nanoflake thin films for CO gas sensing at low temperatures in [170,173]. Layers of Pd were incorporated by rf sputtering into an electrode-functional layer of a thin-film-based nickel–yttria-stabilized zirconia (Ni–YSZ) solid oxide cell (TF-SOC) [174]. Selective doping (Gd or Tb) of Pd films on (001) MgO for hydrogen gas sensing was performed by dc magnetron sputtering at 550 °C, and they showed the (002) preferred orientation [175]. PdO thin films [176] with nano-sized cracks (Figure 7), obtained in a vacuum atmosphere (base pressure of 4 × 10⁻⁸ Torr with 5, 10, 15, 20, 25, and 30% oxygen) by reactive sputtering at room temperature on oxidized (100) Si, demonstrated an ultrahigh sensitivity (~4.5 × 10³%) and a fast response time at room temperature to H₂ gas in nitrogen.

A variant of sputtering, namely, the plasma-assisted sputtering, was performed in [177] for an ionization current of 3 mA under 1.0 × 10⁻¹ Torr. The thickness of the Pd nanostructures deposited on (001) GaN was in the range of 1–30 nm. Samples were annealed for 450 s at 650, 700, and 750 °C in a vacuum of 1.0 × 10⁻⁴ Torr, and evolution of the morphology was studied.

Thin films in the Pd-Zr-Y-O and Pd-Pt-O systems were grown via reactive magnetron sputtering and the pulsed filtered cathodic arc deposition technique in [178] and [179], respectively.

3.2. Electron Beam Coating and Electron and Ion Beam-Induced Deposition

For the growth of Pd films, electron and ion beam-induced deposition (EBID and IBID) were demonstrated. These methods were successful in fabrication of 3D structures/patterns [180]. In these vacuum-based techniques, a precursor material (usually of a metal-organic type) is introduced (injected) in the deposition chamber, and it is absorbed on a substrate. A focused electron or ion beam is used for processing the precursor and through its decomposition to obtain the film. Via IBID, and using Pd-bis(hexafluoroacetylaceacetonate) as the precursor film, a film composed of nanowires of Pd was obtained in [181]. EBID was applied in [182] on a spin-coated precursor film of Pd-hexadecathiolate on a Si substrate. These methods share similar features with two-step chemical methods that use a precursor Pd-based compound deposited in the first step and decomposed in the second one. A typical conventional electron beam coating to obtain a Pd film (with a thickness of 12.3 nm) was applied in [183]. The film was deposited on glass substrates. The base pressure before deposition was 3 × 10⁻⁵ mbar. The substrate temperature was 32 °C, cathode voltage was 8.5 kV, and coating rate was 12 A·s⁻¹. Annealing was performed in air at 350, 450, 550, and 650 °C for 1 h. According to XRD, the films consisted of Pd phase. In [184], the electron beam gun was used to evaporate a Pd ingot target and deposit Pd thin films (17–100 nm) on a 2-inch Si wafer substrate with a sputtered buffer layer of SiO₂ of 400 nm in thickness. The substrate was positioned at about 35 cm above the target. The film deposition was performed at 8 × 10⁻⁵ Pa in a vacuum chamber. XRD indicated that the Pd film was (111) oriented. The sheet resistance of the Pd film (50 nm) annealed at 100, 150, and 200 °C for 3 h was not affected up to 150 °C. In [185], Pd films were deposited by e-beam evaporation of palladium pellets (99.95%). The substrate was thick fused silica (1 mm) and, during deposition in the vacuum (base pressure of 10⁻⁴ Pa), it was heated at 90 °C. The growth rate of the film was set at 0.5 nm/s. The Pd films with thicknesses of 7.2 (S1), 12.6 (S2), 17.5 (S3), and 27.3 (S4) nm were oxidized by annealing at 500 °C in a vacuum at 150 Pa. The annealing time was selected to be 1 h for each 10 nm of the film’s thickness, ensuring the full oxidation of the films. The oxidized films were exposed for 10 min to H₂ atmosphere (5 vol.% in nitrogen) at room temperature. This procedure reduced PdO to r-Pd, and the thickness of the samples decreased without restoring the initial one, while porosity increased (density decreased by 1.68–2.6 times). Films were used in room-temperature optical H₂ gas sensors (Figure 8). A higher film thickness increased the response and recovery times of the as-prepared Pd films, and it was almost constant for r-Pd-treated films. Palladium films were deposited by electron beam evaporation on Si/SiO₂/Pt substrates, and after annealing in oxygen at 500 °C for one hour to form the palladium oxide (PdO), films were tested for hydrogen ion (pH) sensing [186]. The authors found that the sensitivity of the device (extended-gate field-effect transistor (EGFET)) was 42.36 mV/pH.

3.3. Pulsed Laser Deposition

In the pulsed laser deposition (PLD), a pulsed laser is focused on a Pd or PdO [187,188,189,190,191] bulk target. Through ablation, the material from the target evaporates/sublimates or converts to a plasma, and it is directed on a substrate to form a film. Laser type (pulsed or continuous), wavelength, and its fluence, gas pressure in the deposition chamber, target–substrate distance, and substrate temperature are among the parameters to control the process and the film quality. Ablation of the Pd bulk targets can also be realized in a liquid instead of a gas environment, and the product is usually a nano-powder. The type of liquid was shown to influence not only the size and morphology, but also the oxidation and hydrogenation level of the resulting Pd particles [192]. Via this approach, ultra-small Pd nanoparticles were deposited on CdS nanorods for the photocatalytic hydrogen production [193]. The morphology and porosity of the bilayer films of Pd/TiO₂ deposited by PLD (Nd-YAG laser, emission wavelength of 355 nm, 5 ns pulse duration, repetition rate of 10 Hz, energy per pulse of 71 mJ, energy density on the target of 25 J/cm², target–substrate distance of 40 mm, and substrate temperature is room temperature) were modified by changing the oxygen and argon pressures. The films were deposited on the quartz substrate of surface acoustic wave (SAW) room-temperature H₂ gas sensors [188]. The quality of Pd deposits versus the ablation process and Al₂O₃ substrate characteristics were reported in [187]. Pd was deposited by PLD on wood for Ni electroless plating in [194]. Different helium background gas pressures (50, 350, and 800 mTorr) during Pd thin films’ PLD deposition influence the crystallite size and the tendency of their orientation in the <111> direction [189]. Pd thin films of 0.4–4 monolayers were deposited on (001) Cu at room temperature by PLD. The Pd–Cu interface formation is characterized by an alloying–dealloying mechanism. Up to two monolayers, Pd atoms are incorporated into the Cu substrate for less than half-filled layers but expelled if the Pd coverage is close to a complete layer. In this latter case, the top layer is composed of Pd [195]. PLD was used for doping the surface of polycrystalline thin-film SnO₂ or SnO₂(Cu) gas sensors for the enhancement of the sensitivity to 1% H₂ in N₂ (at 200–380 °C), with two orders of magnitude [196]. Pd-doped SnO₂ thin films for gas sensors were grown on (100) Si substrates using PLD at room temperature in [197]. The microstructure of the films depends on the O₂ background pressure (10–100 Pa) and on the target–substrate distance. When the target–substrate distance was close to the plume length, textured thin films were obtained, and the authors explained this phenomenon considering an adiabatic expansion. Nanostructured Pd films obtained by PLD (KrF excimer laser, 248 nm, pulse time of 20 ns, repetition rate of 100 Hz, background pressure in the deposition chamber of 10⁻⁵ mbar, laser fluence of 5 ± 0.1 J/cm² or 250 MW/cm², the laser beam was focused at a 45° angle on the target with a spot area of 1 mm², and Pd metal target with purity of 99.9%) have very different electrochemical characteristics from those of common coarse-grained films [198]. Stability was good under hydrogen charge/discharge cycling, and this was related to the lack of an abrupt α-to-β phase transition. Upon oxidation of Pd films with controlled thickness and grain size deposited by PLD, spontaneous formation of uniformly distributed arrays of conical PdO “tips” useful for field emission applications was reported in [191,199].

Chemoresistive acetone sensors of ZnO were decorated with Pd particles by PLD [200]. Sensors operated in the temperature range 159–200 °C, and they showed an enhancement of the response factor between 2 and 7 when compared with pure ZnO sensors. The detection limit was 26 ppm at 200 °C.

3.4. Evaporation (Sublimation) of Pd

Evaporation (sublimation) of Pd (e.g., a metal foil resistively heated by a V-shaped tungsten filament) and deposition is a simple and popular technique [106,201,202,203,204,205]. Evaporation was realized in a high-vacuum chamber (e.g., at 6.65 × 10⁻⁴ Pa in [84] or 18–57 × 10⁴ Pa in [87]). These films were deposited on (1120) α-Al₂O₃ [86], polished polycrystalline Al₂O₃ [87], SiO₂/(100)Si [83,84], Pyrex glass [82], (100)Si [84], optical quality quartz [106,202], and KCl with a buffer layer of amorphous carbon [106,205]. The palladium vapor pressure, P, in the deposition chamber was set according to Equation (1) [206]:

log₁₀P(Pa) = −20,150T⁻¹ − 0.419log₁₀T − 0.302 × 10⁻³T + 13.670

(1)

In [164], the palladium vapor pressure was:

log₁₀P(Pa) = 8.749 − 18,655T⁻¹

(2)

in the temperature range 927–1427 °C. When the rate of evaporation in the vacuum was compared among the platinum-group metals (Pt, Ir, Os, Pd, Rh, and Ru), the values at lower temperatures were the highest for Pd. The heat of vaporization was the lowest (H_292K = 89.2 ± 0.8 kcal/mol). The presented information promoted Pd as the most convenient element among the platinum-group metals for thin-film deposition through physical evaporation: it is considered that Pd has an anomalously high metal vapor pressure. Substrate temperature during Pd thin films’ deposition was selected to be at room temperature [106,202], but in some experiments, a low temperature was used, e.g., 77 K in [201]. The physical evaporation allows deposition of Pd films with different thicknesses. Attempts to grow Pd ultrathin films are presented in [202], and films with thicknesses of 1, 5, 10, and 80 nm are compared. The study pointed out the challenging limitations for ultrathin films regarding their tendency for formation of an island-type microstructure featuring a high possibility of discontinuity, accompanied by the lack of a percolation path for electrical conduction. It also pointed out the strong substrate influence, implying occurrence of a high crack density and stoichiometry changes due to the substrate–film interdiffusion. In [106,205], PdO ultrathin and thin films with thickness of 5–40 nm were obtained. The thermal annealing of the Pd precursor films was performed under dry oxygen. Films with low thickness of 5–15 nm were heated for 1 h, while thicker ones for 2 h. Annealing temperatures were 237, 297, 397, 497, 597, and 797 °C. According to X-ray diffraction measurements, oxidation of the Pd precursor film with partial or full formation of PdO took place for annealing temperatures of 297–797 °C. Heras et al. [201] found that for Pd films deposited on cold substrates (glass), during subsequent annealing by heating from −196 °C to 400 °C, they changed their morphology and absorbed oxygen without formation of a bulk PdO oxide. A fiber morphology and texture with the axis (111) normal to the substrate developed, and the process was stronger for temperatures closer to the higher values of the indicated interval. The penetration of oxygen into the bulk depends on the surface defect density, and significant oxygen incorporation into polycrystalline palladium without formation of PdO was also reported in [207,208] at 247 °C.

3.5. Supersonic Cluster-Beam Deposition

In the supersonic cluster-beam deposition [209], a Pd target rod is subject to a He plasma ignited by a pulsed electric discharge between the Pd rod (cathode) and anode, producing the ablation of the target. The ablated Pd atoms thermalize inside the cavity via collision with the inert gas and condense into clusters. The mixture of the cluster and He is expanded in a vacuum chamber through a nozzle and forms a supersonic beam. Substrates intercept the supersonic beam in a second chamber separated from the expansion chamber by an electro-skimmer. The deposition rate is about 1 mm/min, and the kinetic energy is small enough (0.5 eV order) to avoid cluster fragmentation.

In [210], deposition was performed on glass substrates. The authors adopted two strategies for oxidation of Pd clusters. In the two-step ex situ route, Pd was deposited with He carrier gas, followed by oxidation annealing in air at 200 and 400 °C. In the one-step in situ route, the carrier gas was a mixture of 80 at.% He and 20 at.% O₂. Films were nanostructured with a particle size below 20 nm, with the smallest particles below 5 nm. They were porous and very soft and could be easily detached from the substrate [211].

4. Some Practical Aspects of Pd Films’ Oxidation Toward PdO and Specific Issues Concerning Applications of the Films

The above-presented information indicates that oxidation of Pd is a complex process, and some details are not completely understood. The next paragraphs will briefly screen some practical aspects, but details of the oxidation mechanisms/models will be neglected.

When a Pd foil is thermally oxidized in ambient conditions at 650 °C, a bulk PdO layer develops, and after 1.52 min [212] of oxidation, it is detected by XRD. The thickness of the PdO layer increases with time, and the oxidation process is governed by a parabolic law directly related to temperature and oxidation time parameters [212]. The PdO film on Pd presents a preferential growth corresponding to the (002) PdO direction, since the Pd foil substrate is polycrystalline, with grains showing a preferred (002)-normal direction. The lattice constants of the Pd oxide film are a = 0.3035 nm and c = 0.5323 nm, as expected for the bulk PdO [2]. The authors noted that the oxidation rate is high as a result of lower surface energy activation for the (002) direction than for the (111) one. A model for oxidation of Pd with formation of PdO based on motion and creation of cation vacancies was proposed, and within it the necessary energy (E_a) for movement or migration of cationic vacancies depends on the thickness of the oxide layer. A lower E_a of 0.034528 eV is found for an ultrathin film, compared to E_a of 0.115883 eV for a thick one. This explains the faster diffusion process in a thin oxide layer than in a thick one and the already mentioned parabolic dependence. The activation energy for oxygen diffusion in Pd bulk has been estimated to be in the range of 84–98 kJ·mol⁻¹ [201], while the free energy of formation of PdO (at 727–875 °C) is [202]:

ΔG⁰_T(PdO) = −27460 + 23.9T ± 280 (cal/mol)

(3)

It is worthy to note that the solid oxide formation results in a significant increase in weight. Other features of interest are:

(i)

Kinetics of the Pd oxidation (in the mbar pressure range) depends on the crystal surface: oxidation of Pd(110) proceeds at ~100 K lower temperatures than Pd(111) [213].

(ii)

The PdO layer formed as a skin on Pd is green when relatively thin [23]. When thicker (annealed in air at 800 °C), it turns gray.

(iii)

Depending on the heating rate and annealing temperature, the PdO layer decomposes in air at 825–900 °C. A decomposition temperature of 870 °C was estimated for annealing under 1 atm of oxygen [23,214,215,216].

Samoylov et al. [214] found that the thermal stability of the PdO film in an oxygen atmosphere enhanced from 810 ± 5 to 860 ± 5 °C when the thickness of the initial Pd layer increased from 10 to 95 nm, respectively. The authors also noted that the resulting Pd film, unlike the continuous Pd film used as a precursor to obtain the oxygen-decomposed PdO film, was composed of isolated hemispherical Pd nanocrystalline particles with well-defined faceting and preferential (111) or (100) alignment.

In the ultrahigh vacuum, PdO was reported to be unstable at temperatures above 147 °C [215].

Okamoto and Aso [203] claimed decomposition of PdO at room temperature under H₂ atmosphere from the rapid increase in electrical conductivity.

The results of more recent first principles studies indicated that the thermodynamic stability of PdO depends on the crystal surfaces [216]. These aspects significantly impact the catalytic efficiency. Although for methane oxidation, Pd is one of the best catalysts, the relation between the presence of oxides and the catalytic activity is complex. In methane oxidation over a Pd(100) single crystal, it was found [217] that the metallic surface was active, but it was unstable under reaction conditions. Growth of the oxide promoted low activity, but for a thicker oxide, the activity increased, and above a certain level, it decreased, even if the thickness of the oxide continued to expand. In these processes, depending on the thickness, the exposed surface of the oxide changed, being (101) or (100), with the second one showing low activity due to lack of coordinatively unsaturated (CUS) sites (i.e., Pd atoms on the surface) where the reactants can adsorb. The first layer was of (101) PdO. For very thick films, the (100) oxide was stable, and it was no longer coupled to the substrate, so the surface became inactive. For a few layers of the oxide, a ligand effect [218] with the Pd layer below occurred, and it contributed high activity for the intermediate thickness of the oxide.

To overcome the instability of the catalytic activity, Pd alloys were obtained and tested, e.g., PdAu in [11]. Composites, such as conductive polymers (polyaniline)/PdO (PANI/PdO), are another promising solution. In [219], a CH₄ gas sensor of the quartz crystal microbalance (QMS—a mass sensitive device) type was obtained by a layer-by-layer self-assembly method that is based on electrostatic force and in which an in situ chemical oxidation polymerization approach in the presence of PdO nanoparticles at room temperature is applied. The method uses dip coating on hydrophilic glass substrates. Substrates were treated sequentially and ultrasonically with CHCl₃, C₂H₅OH, and deionized H₂O, each for 20 min, and dried in nitrogen gas. Treated substrates were coated by immersing in 1.0 wt.% PDDA (poly(diallyldimethylammonium chloride), molecular weight 200,000–350,000, polycation) aqueous solution for 5 min and washing with deionized water. Washing prevents cross-contamination with the second coating solution of 0.2 wt.% PSS (poly(sodium-pstyrenesulfonate), molecular weight 70,000, polyanion), with the opposite charge to PDDA. The PSS polyanion layer provides the charges for the adsorption of the first-layer polycation in the formation of a polyaniline/PdO thin film. The film PANI/PdO was deposited on a glass/PDDA/PSS substrate. Dip coating of the substrate was performed in a solution prepared as follows: (a) Aniline monomer was added into the 1.0 M camphor sulfonic acid (CSA) solution with PdO nanoparticles at room temperature. The ratio of aniline to PdO was 5 to 1. (b) The mixture was reacted with ammonium persulfate for 20 min. The progress in this research direction can have high importance and impact because the greenhouse effect of methane (CH₄) gas is about 30–80 times higher than for carbon dioxide (CO₂) [220]. Methane is also a colorless and odorless gas, and the key component of the natural gas. It is flammable and explosive when its concentration is 5–14% in air.

(iv)

At rather high annealing temperatures (~1000 °C) of Pd, a slight weight increase was detected because of higher oxygen solubility at higher temperatures [221,222].

The questions to be addressed in the future studies will also have to deal with possibilities and novel techniques of fast films’ characterization, e.g., in situ local mapping of the degree of oxidation on the surface and in the volume of the film. In [11,172], the authors used advanced characterization methods, such as grazing-incidence X-ray diffraction (GIXRD) and photoemission spectroscopy, respectively, to observe the details of oxidation. Evolution during oxidation of Pd particles (2.3 nm) was investigated with in situ time-resolved X-ray absorption spectroscopy (XAS), supported by theoretical simulations [223]. Obtaining coatings with organized surfaces, e.g., with ordered and periodic patterns of hillocks of Pd thin films, was studied by XRD and AFM in [224]. The authors observed the variation in compressive stress and the total hillock surface area versus the thickness of the Pd film. They introduced formation of a PdO layer at the interface of Pd as one relaxation mechanism and found that this mechanism was the dominant relaxation process in the films with a thickness higher than 40 nm. Hence, to generate spontaneously ordered hillocks over the Pd surface, a thickness of over 40 nm is required since it is expected to lead to a constant value of the total hillock surface area. The aspects of oxygenation and morphology formation will help in designing the films and finding their optimum deposition route to achieve the desired and controlled functional characteristics that are needed for the targeted application. Ultimately, it will enhance the performance (response time, recovery time, response percent, sensitivity, selectivity, and reliability [225]) of the sensor and will provide its reproducible and stable operation. Notably, the effect of the preparation process parameters on the responses of the gas sensors of SnO₂ loaded with Pt, Pd, and Au was investigated for prediction purposes with artificial neural networks in [226].

Since H₂ gas is highly flammable and explosive when its concentration is 4–75% in air and it cannot be detected by human senses, in the emerging clean economy of hydrogen, detection of this gas in a very sensitive, highly stable, and selective manner and in a wide range of temperatures will be crucial. This is challenging, and the current typical conductometric and electrochemical (amperometric and potentiometric) sensors show limitations.

Review articles [225,227] indicated eight types of commercial hydrogen gas sensors based on electrical conductivity, thermal conductivity, acoustic, mechanical, optical, electrochemical, catalytic, and work functions (Figure 9). Metal oxide sensors of conductometric-type work optimally only at high temperatures, in low relative humidity, and their sensitivity for hydrogen is quite low. On the other hand, Pd is a highly electroactive element, but undesirable processes occur, and we mention hydrogen embrittlement due to formation of the palladium hydride phase (accompanied by lattice expansion and resistance increase [167]) and of the irreversible contamination with the other gases from the atmosphere. These processes decrease the sensing efficiency of a sensor made of a pure metal. A sensor made of PdO will not be poisoned due to its higher working temperature. Hence, these sensors are thought of as candidates with a higher potential than those made of Pd but, ultimately, this will also depend on the application needs. In addition, PdO electrochemical H₂ gas sensors show high sensitivity and selectivity and might be a convenient solution for sensing at room temperature [228,229].

The present strategies for improvement of gas sensors are to use composite and/or alloyed films and to modify/control their morphology, e.g., particle size, shape, distribution, porosity, nano-structuring, and defects (e.g., cracks) [20,230,231,232,233,234,235,236,237,238,239,240,241,242], and all these features depend on preparation and post-processing technology. A low detection limit of 300 ppb, a fast response time of 3 s toward 500 ppb, excellent selectivity at 200 °C, long-term stability, and outstanding tolerance to humidity were determined for a H₂ gas sensor built with Fe₂O₃-core/Pd/PdO-shell nanoparticles [243]. Excellent sensing properties were discussed within the frame of the Pd/PdO ‘spillover effect’, where the electron transfer and charge accumulation related to adsorbed oxygen are promoted by formation of PdO/Fe₂O₃ p-n heterojunctions.

The indicated strategies also promote new practical opportunities. For example, indium oxide-palladium (In₂O₃-Pd) sensors have a good sensitivity toward nitric oxide; hence, it can be used for detection of nitroaromatic explosives [244]. In another example, binary metal oxide solid solution (Ir_(1−x)M_xO_y, where M = Pd, Rh, and Ru) thin films obtained by dc reactive magnetron sputtering (each of 2 Ir targets were supplied 25 W power at a frequency of 144 kHz, while for the Pd target, power was 50 W, and pressure in the chamber was ~4 Pa, with 20% oxygen partial pressure) on 316 stainless-steel or SiO₂/Si substrates were assessed, targeting implantable neural interfacing applications [245]. Undesirable nanoflake growth of IrO_x is suppressed by introduction of Pd (Figure 10). The presence of Pd also activates higher oxidation states, including a +5 oxidation state. Both effects are believed to be the reason for better performance of the binary metal oxide films when compared to single metal oxides. Salagare et al. [246], in a selective three-electrode electrochemical sensor, modified a graphite electrode with PdO-rGO composite flakes and demonstrated a high degree of consistency of electrode sensitivity measurements of nitrite at room temperature. In a recent article [247], palladium nanoparticles were decorated on TiO₂ nanotubes (TNT) by gamma-ray irradiation. They degraded nitrogen monoxide (NO) in the photocatalysis process, achieving 53.30% NO removal efficiency, which is 1.6 times greater than that of pure TNTs. A layered nanoelectrode based on GO and rGO decorated with n/p nanoparticles of palladium oxide and cadmium sulfide has shown high efficiency for supercapacitor applications [248]. Nanoparticles of PdO (n-type) and SrO (p-type) on GO and rGO were tested [249] for decomposition of organic pollutants, such as mixed dyes of Rhodamine B and methylene blue (RhB/MB), insecticides, such as imidacloprid, and the removal of heavy metals, such as chromium ions. Results were considered promising for water cleaning.

5. Conclusions

This review presented different deposition routes of PdO and Pd films and heterostructures. The largest contribution in the literature is on Pd films’ fabrication, without a further investigation toward PdO formation as a product. This situation is due to the dominant interest from the industry in catalytic processes where Pd rather than PdO has a key role. However, the development of different applications, such as gas sensors, especially of ozone, NO_x, H₂, CH₄, etc., has become important and promoted a higher motivation, expanded attention, and research on fabrication of PdO films. Gas sensors are used in different industries, such as medical, aerospace, petrochemical, mining, and automotive.

In the literature, there are very few papers to compare films for a specific application obtained by different technological routes. An isolated example is in [250], in which the electrochemically grown Pd nanoparticle films were observed to show better hydrogen sensing responses than the sputtered Pd thin films.

Some of the reviewed deposition routes are highly complex or sophisticated, require expensive and specific raw materials, and are using or generating harmful, undesirable substances. This situation leads to fabrication cost increases and environmental and integration problems. Recent approaches tried looking at green synthesis or biosynthesis routes, and some progress was achieved in recent years ([5] and references therein, [37,251]), taking advantage of non-toxic, clean, eco-friendly, natural substances. Although these developments are promising, in general, green synthesis produced particles rather than films.

Films of Pd or PdO are usually composed of randomly oriented crystallites, although in some reports, the authors observed a certain degree of texture [175,184,252,253], which was accidentally obtained in most cases. As we presented above, the activity of different crystal planes of the palladium oxide is very different. Therefore, preparation and exploration of textured or epitaxial, single-crystal or polycrystalline palladium-oxide-based films might provide interesting new opportunities for applications and may afford a better understanding of the complex catalytic/sensing processes. Despite the huge effort of the international scientific community, revealing the details of these processes requires further research—the physical–chemical mechanisms governing the interaction with the medium or taking place inside the material containing Pd are insufficiently understood. In this regard, there is an urgent need for new, fast, in situ, and reliable investigation techniques that are effective at different scales.

The number of devices reported and fabricated on PdO films is significantly lower than those on Pd, although, as addressed above, there are advantages in using the oxide. Gas and pH sensors are the main ones to use PdO films. These sensors are expanding their domains of application due to rapid developments linked to wellbeing devices and gadgets, clean and safe environments, and sustainable economies. Therefore, these devices based on PdO films can contribute to improvements of the quality of life. Palladium oxide has p-type conductivity and, due to this, it can successfully compete with other oxides, such as SnO₂, ZnO, or In₂O₃, especially in detection of oxidizing gases [214].