Volatile Iridium and Platinum MOCVD Precursors: Chemistry, Thermal Properties, Materials and Prospects for Their Application in Medicine

Abstract

Interest in iridium and platinum has been steadily encouraged due to such unique properties as exceptional chemical inertia and corrosion resistance, high biological compatibility, and mechanical strength, which are the basis for their application in medical practice. Metal-organic chemical vapor deposition (MOCVD) is a promising method to fabricate Ir and Pt nanomaterials, multilayers, and heterostructures. Its advantages include precise control of the material composition and microstructure in deposition processes at relatively low temperatures onto non-planar substrates. The development of MOCVD processes is inextricably linked with the development of the chemistry of volatile precursors, viz., specially designed coordination and organometallic compounds. This review describes the synthesis methods of various iridium and platinum precursors, their thermal properties, and examples of the use of MOCVD, including formation of films for medical application and bimetallics. Although metal acetylacetonates are currently the most widely used precursors, the recently developed heteroligand Ir(I) and Pt(IV) complexes appear to be more promising in both synthetic and thermochemical aspects. Their main advantage is their ability to control thermal properties by modifying several types of ligands, making them tunable to deposit films onto different types of materials and to select a combination of compatible compounds for obtaining the bimetallic materials.

1. Introduction

Interest in platinum group metals has not waned for a long time due to their unique properties and the possibility of application of their films as protective, anticorrosive, buffer, optical, sensor, or electrocatalytic layers in various fields, such as aerospace engineering, catalysis, hydrogen energy, fuel cells, and microelectronics [1,2,3,4,5]. In recent years, the number of publications devoted to the medical application of platinum group metals, including platinum and iridium, has increased. These metals are distinguished by exceptional chemical inertness and corrosion resistance in biological environments, high biological compatibility, and mechanical strength. Traditionally, platinum- and iridium-containing film materials have been used in medicine to cover the contact poles of electrodes for cardio- and neurostimulation and diagnostics in order to improve their electrochemical characteristics [6,7,8]. In recent works [9,10,11,12,13], the prospect of using platinum and iridium in composite coatings and film heterostructures to modify the surface of permanent implants, which are in demand in oncological, orthopedic, and reconstructive surgery, has also been shown. In particular, new bioactive materials consisting of platinum metal coatings with micron dots of silver or gold demonstrate an enhanced antibacterial effect compared to the traditionally used silver [12,13]. A winning feature of such heterostructures is also their ability to configure time-limited activity involving low amounts of Ag or Au, which prevents cytotoxicity and undesirable tissue reactions. Moreover, platinum and iridium coatings can improve the biocompatibility and osseointegration of implants [9,10,11]. The principle of action of such composites is currently under discussion, and apparently includes not only the electrochemical, but also the synergistic aspect [13]. The dependence of the observed effects on the combination of noble metals and the composition and morphology of coatings has been noted in a number of studies [9,10,11,12,13].

The full potential of the application of functional noble metal coatings in medical practice can only be realized if the problem of deposition of the layers with a given composition and structure on materials of different natures and non-planar geometry can be successfully solved. For this reason, the development of universal and precision methods for the deposition of platinum and iridium coatings is one of the most important steps. Metal-organic chemical vapor deposition (MOCVD) fully meets the specified criteria, since it allows obtaining coatings and nanoparticles, multilayers and heterostructures with controlled characteristics (composition, morphology, particle size, etc.) on objects of complex shapes and does not impose special requirements upon the coated material, except for its stability. In addition, the MOCVD principle allows avoiding the difficulties associated with the refractoriness of platinum and iridium, since the use of volatile compounds (precursors) of these metals makes it possible to obtain the required materials at relatively low temperatures (200–600 °C).

The MOCVD process consists of several main steps (Figure 1): heating of the volatile compound source to transfer this precursor into the gas phase, transport of the precursor vapors to the substrate, activated precursor decomposition on the substrate surface to form the target material (coating or nanoparticles), and gaseous by-products that are removed from the reactor chamber. The decomposition reaction could be activated by high temperatures (classical approach), plasma, irradiation, etc. Regarding the thermodynamic and kinetic aspects, the deposition process in MOCVD reactors is a complicated non-equilibrium dynamic system that includes a sequence of equilibrium or partially non-equilibrium chemical reactions of the precursor decomposition. Therefore, the thermochemical properties of precursors play a key role in the reactor construction and experimental parameters necessary for the preparation of the materials with the required properties. The quantitative data on thermal characteristics of volatile compounds are essential to precisely control the deposition process. Specifically, the information on the phase transitions (melting, sublimation, vaporization) and mass transfer gives the opportunity to manage the amount of substance supplied into the reaction zone. The characteristics of precursor vapor thermal destruction processes allows for determining the effective deposition temperature range and the effect of the introduction of various reagent gases or activation forces. Knowledge of thermal behavior is the perfect tool for tailoring the precursor structure to specific technological tasks on deposition of the target materials onto the substrates of various natures. For medicine, this approach means the ability to effectively coat a wide range of non-planar materials, including metals, polymers, carbon composites, etc., which differ dramatically in the thermochemical characteristics such as thermal stability, oxidation, or embrittlement resistance.

Thus, the development of MOCVD processes is inextricably linked with the development of the chemistry of iridium and platinum precursors, viz., specially designed coordination and organometallic compounds. Another important issue is the expansion of the collection of compounds for the effective selection of compatible combinations of a precursor in order to obtain bimetallic systems based on platinum and iridium, which are of particular interest due to the possibility of a synergistic effect.

The growing interest in the chemistry of volatile platinum and iridium compounds is confirmed by the regular appearance of thematic reviews [14,15,16,17,18]. While the latest data on new iridium precursors were collected in 2015 [18], the latest review of platinum precursors was published in 2008 [15], and further reviews have only covered certain aspects of their application [17]. It is also worth mentioning that the recent reviews covered various classes of precursors, including specific and general synthesis concepts, and then focused mainly on applications in MOCVD processes, while the thermal properties of the compounds have received much less attention.

In this regard, this review presents the main classes of volatile precursors of platinum and iridium, including compounds synthesized in recent years, in order to provide a basis for the further development of MOCVD processes, primarily in the aspect of medical applications. Thus, the thermal properties of compounds will constitute the mainstream of this review, since it is this knowledge that contributes to the most efficient use of the compounds as volatile MOCVD precursors. This information will be supplemented with examples of specific deposition conditions and characteristics of the resulting materials. In addition, the role of the precursor properties in the preparation of bimetallic coatings will be specially emphasized. Finally, the latest information on the success of MOCVD of platinum and iridium for coatings used in medicine will be presented, and the future scope of their application will be outlined.

2. Volatile Precursors

The main requirements for effective MOCVD precursors are substance purity (≥98.0 mass.%), relatively high vapor pressure (volatility), stability (both in condensed and vapor phases during volatilization), the presence of a “window” between the temperatures of volatilization and vapor decomposition, and a high degree of conversion into the target material. To minimize contamination of the formed material, the decomposition of the precursor should occur through the formation of only gaseous by-products. In the case of noble metals, another important aspect is the possibility of synthesizing a precursor with a high yield from available reagents.

As noted above, iridium and platinum have a high chemical inertness, which is one of the reasons for the limited range of volatile precursors of these metals used in MOCVD experiments. Various classes of volatile compounds of iridium and platinum will be considered separately below. The structure and abbreviations of ligands are shown in

Table 1. Structure and abbreviation of the ligands of various classes employed to volatile iridium and/or platinum MOCVD precursors ^a.

NEUTRAL LIGANDS
hexadiens					norbornadiens				cyclooctadiens
Structure	R1		R2	Abbreviation	Structure	R	Abbreviation		Structure		R	Abbreviation
	H		H	hd		H	nbd				H	cod
	H		Me	hd-Me		Me	cod-Me				Et	cod-Et
	Me		Me	hd-Me2		Et	nbd-Et				Bu	cod-Bu
						iPr	nbd-iPr				iBu	cod-iBu
ANIONIC LIGANDS
β-diketonates					cyclopentadienyls
Structure	R1		R2	Abbreviation	Structure	R1	R2		R3	R4	R5	Abbreviation
	Me			acac		H	H		H	H	H	Cp
	tBu			thd		Me	H		H	H	H	CpMe
	tBu/CF3		C(OMe)Me2	zis/zif		Et	H		H	H	H	CpEt
	tBu		CF3	ptac		allyl	H		H	H	H	Cp-allyl
	Me		CF3	tfac		nPr	H		H	H	H	Cp-propenyl
	CF3			hfac		iPr	H		H	H	H	Cp-ipropenyl
	CF3		Ph	btfac		nBu	H		H	H	H	CpBu
	Ph			dbac		Me	H		Me	Me	H	CpMe3
	Me		Ph	bac		Me	Me		Me	Me	H	CpMe4
	CF3		C4H4S	ttfac		Me	Me		Me	Me	Me	Cp*
β-iminoketonates					β-iminoalcoholates				ω-Alkenyls
Structure	R1	R2	R3	Abbreviation	Structure	R	Abbreviation		Structure			Abbreviation
	CF3		Et	Eti-hfac		Me	Mei-hfda					C5H9
			nPr	nPri-hfac
	CF3	Me	Me	Mei-tfac		Et	Eti-hfda		Structure		n	Abbreviation
	CF3	H	CH2CF3	TFB-TFEA							0	C6H11
	Me		H	i-acac		nPr	nPri-hfda				1	C7H13
			Me	Mei-acac							2	C8H15
α-aminoalcoholates				β-heteroarylketonate		β-alkenols			dithio/diselenoimidodiphosphinato
Structure	R	Abbreviation		Structure	Abbreviation	Structure	R	Abbreviation	Structure		Hal	Abbreviation
	Me	amN(Me)2			ThTFP		Me	alk(Me)			S	S/S-idpp
	CF3	amakN(Me)2					CF3	alk(CF3)			Se	Se/Se-idpp

^a group symbols: Me = CH₃ (methyl), ^tBu = C(CH₃)₃ (tert-butyl), Ph = C₆H₅ (phenyl), Et = C₂H₅ (ethyl), ⁿPr = CH₂C₂H₅ (n-propyl), ⁱPr = CH(CH₃)₂ (iso-propyl), ⁿBu = CH₂CH₂C₂H₅ (n-butyl), ⁱBu = (C₂H₅)CH(CH₃) (iso-butyl).

.

2.1. Volatile Iridium Precursors

As a rule, metal-organic and coordination compounds of iridium in the oxidation state +3 and +1 are used in the deposition of iridium coatings.

2.1.1. Iridium(III) Precursors

Iridium(III) compounds that can be used as precursors in MOCVD processes are restricted to the classes of β-diketonates and allyls [14,18]. At the same time, β-diketonate chelates are synthetically more accessible than tris-allyl iridium; however, their synthesis with high yields still present a significant challenge.

The traditional approach to the synthesis of iridium(III) β-diketonates consists in the interaction of iridium(III) chloride with the corresponding β-diketone or its salt in water-alcohol solutions with controlled pH [19,20]. However, the yields of the products do not exceed 15%–20%, since the reaction is complicated by the formation of monomeric and dimeric complexes of iridium(III) with γC-bound β-diketones [19], as well as heteroligand complexes of polymer structure formed as a result of competition between acetylacetonate and acido-ligands [21]. Isakova et al. [21] developed a universal method of the preparation of Ir(L)₃ β-diketonates with high yields by the reaction of β-diketone with iridium(III) aquafluorocomplexes formed when a K₃[IrF₆] solution was heated in hydrofluoric acid. The method was based on the lability of fluoro- and aqua-ligands, which made it possible to substitute them with β-diketonate anions with high yields (up to 90%) and for a wide range of β-diketonates (L = acac, thd, tfac, hfac, ptac). All iridium(III) β-diketonates were shown to be mononuclear molecular complexes formed by chelating coordination of all three ligands. In the case of an asymmetric tfac-ligand, the formation of cis/trans isomers differing in thermal properties was confirmed; however, the crystal structure was determined only for trans-Ir(tfac)₃ [22]. The structure of only one complex of this class, viz., Ir(acac)₃, was also determined by the single-crystal XRD method [23]. Crystallographic data confirmed the bidentate coordination of β-diketonate ligands and the octahedral coordination environment of the metal in the molecule (Figure 2).

Thermal properties of a series of iridium(III) β-diketonates were described in [21]. Within the investigated series, Ir(acac)₃ has the highest melting point (269–270 °C). A decrease in the melting point is observed when both CF₃ and ^tBu substituents are introduced into the β-diketonate ligand, so that the most fusible complex is Ir(ptac)₃ (83–84 °C). The vapor pressure increases with the introduction of these terminal groups, and the influence of the CF₃ group is more pronounced (Figure 3). The general range of variation of vapor pressure values from the least volatile Ir(acac)₃ complex to the most volatile Ir(hfac)₃ exceeds two orders of magnitude. The equations of temperature dependences of saturated vapor pressure and thermodynamic parameters of sublimation and vaporization processes are given in

Table 2. Equations of the temperature dependences of saturated vapor pressure over the solid (sublimation, subl.) and liquid (vaporization, vap.) Ir(III) β-diketonates. The sublimation and vaporization enthalpies (ΔH_T*) and entropies (ΔS⁰_T*) are referred to the average temperature (T*) of the interval measured (ΔT), p₀ = 760 Torr = 1 atm = 10⁵ Pa. The ligand abbreviations correspond to Table 1.

Ref.	Complex	Process	ΔT, K	ln(p/p0) = A − B/(T/K)		ΔHT*, kJ·mol−1	ΔS0T*, J·(mol·K)−1
				A	B
[24]	Ir(acac)3	Subl.	468–518	15.65	10,397	86 ± 2	130 ± 4
[21]	trans-Ir(tfac)3	Subl.	423–453	26.30	14,013	116 ± 3	218 ± 17
		Vap.	458–488	16.29	9462	79 ± 3	136 ± 6
	cis-Ir(tfac)3	Subl.	414–426	20.62	11,389	95 ± 8	171 ± 18
		Vap.	426–441	15.15	9045	75 ± 3	126 ± 6
	Ir(hfac)3	Subl.	358–396	23.29	10,688	88 ± 6	193 ± 9
		Vap.	401–443	13.79	6973	56 ± 1	115 ± 2
	Ir(thd)3	Vap.	418–522	14.55	8927	74 ± 2	116 ± 4
	Ir(ptac)3	Vap.	413–518	13.88	8259	68 ± 1	116 ± 2

. It is interesting to note that the cis-isomer of Ir(tfac)₃ is characterized by the lower values of enthalpy and entropy of sublimation and the lower melting point compared to the trans-isomer (153–155 °C vs. 186–187 °C).

It was shown by the method of in situ mass spectrometry that the decomposition of Ir(acac)₃ vapors in vacuum and in the presence of hydrogen on a heated surface began at 410 °C with the release of fragments of the acetylacetonate ligand into the gaseous phase. In the presence of oxygen, the initial decomposition temperature reduces to 220 °C, and the final decomposition products are water vapor and carbon dioxide [25]. Recently, a mechanism of Ir(tfac)₃ vapor under electron impact at 160 °C has been proposed [20].

2.1.2. Iridium(I) Precursors

The chemistry of iridium(I) precursors is more diverse, since these complexes are multi-ligand, which makes it possible to expand the series of compounds not only by varying the structure of ligands of the same type, but also by combining various moieties of the molecule. These complexes are much more readily available synthetically than their Ir(III) counterparts. Alkenes or carbonyls act as neutral ligands in volatile iridium(I) complexes, while ligands containing such donor atoms as carbon, oxygen, nitrogen, and sulfur act as the anionic part.

Below we will consider the main classes of compounds proposed for MOCVD applications and discuss the effective synthesis methods and present their structure, and summarize the information of their thermal properties. It is worth mentioning that the synthesis of Ir(I) compounds is carried out in an inert atmosphere, since their solutions are sensitive to air components. However, as solids, most of these complexes are stable for storage in the air for months. Although information about crystal structures is sketchy, the data obtained by other methods clearly show that all Ir(I) precursors are mononuclear molecular complexes, with the exception of the carboxylate and acetate derivatives.

Complexes with (O)- and (O^O)-coordinated ligands are represented by the following three classes:

• alcoholate [Ir(cod)(OMe)]₂ (OMe = methylate ion) [26,27,28],
• carboxylate [Ir(cod)(OAc)]₂ (OAc = acetate ion) [29,30],
• β-diketonates [Ir(cod)(L)], L = acac [31,32,33], tfac [31,33], hfac [31,33,34], thd [32,33,34], btfac, ptac [33], zis [35] and [Ir(CO)₂(L)], L = acac [36,37], hfac [37,38], dbac, bac, ttfac [37], and [Ir(C₂H₄)₂(acac)] [39].

In the synthesis of complexes with cyclooctadiene [Ir(cod)(L)], regardless of the type of L ligand, [Ir(cod)Cl]₂ is usually used as the initial iridium compound. The alcoholate complex [Ir(cod)(OMe)]₂ is obtained by the reaction with a stoichiometric amount of potassium hydroxide in methanol with a yield of 85% [27]. For the synthesis of the [Ir(cod)(OAc)]₂ acetate complex, a twofold excess of alkali metal acetate was used, and the yield was more than 80% [22]. For the synthesis of β-diketonate complexes, a similar approach is used, but an excess of the β-diketone salt is not required. Yields exceed 85%. In the case of non-fluorinated ligands (L = thd, acac), it is preferable to generate the β-diketone salt in situ. In the case of fluorinated ligands L = hfac, tfac, ptac, btfac, it is convenient to use pre-synthesized β-diketonate M^I(L), and the variation of the alkali metal cation M^I does not affect the product yield [33]. It was shown using ligands L = acac, tfac, hfac as examples, that HL β-diketones themselves could substitute methylate ions in the [Ir(cod)(OMe)]₂ complex with the formation of [Ir(cod)(L)]; however, the yield of the reaction was noticeably lower (about 70%) [31].

The complexes [Ir(cod)(L)], in their turn, are effective starting reagents for further preparation of the corresponding carbonyl derivatives [Ir(CO)₂(L)] via a bubbling of carbon monoxide. Indeed, it was shown for L = acac [36] and hfac [38] that this approach led to fairly high yields (>75%). At the same time, the yield of the target products was only 14%–21% in the reverse order of stages of anionic and neutral ligand introduction, i.e., when the β-diketone HL interacted with Na₂[Ir₂(CO)₄Cl_4.6] in the presence of an excess of sodium carbonate [37]. The only described complex with an acyclic alkene, [Ir(C₂H₄)₂(acac)], was obtained using the [Ir(coe)Cl]₂ complex (coe = cyclooctaene) as an iridium source with the yield of 45% via ethylene bubbling followed by the reaction with in situ generated K(acac) [39].

Crystal structures have been determined only for a number of β-diketonate derivatives, some of which have been studied recently: [Ir(cod)(L)], L = acac [40], hfac [34], tfac, thd, btfac, ptac [33]; [Ir(CO)₂(L)], L = acac [36], hfac [38]; [Ir(C₂H₄)₂(acac)] [39]. In these molecules, the iridium atom is located in a typical distorted square environment: the coordination node IrO₂C’₂ (C’ is the center of the C=C bond) is realized for complexes with alkenes or IrO₂C₂ is for carbonyl derivatives (Figure 4). In the case of the [Ir(cod)(OAc)]₂ acetate complex, the binuclear structure was proposed based on the measurement of molecular mass by the method of osmometry [30].

Complexes with (O^N)-coordinated ligands are represented by the following four classes:

• β-heteroarylketonate [Ir(cod)(ThTFP)] [41];
• α-aminoalcoholates [Ir(cod)(amakN(Me)₂)] [42] and [Ir(CO)₂(amakN(Me)₂)] [43];
• β-iminoalcoholates [Ir(cod)(L)], L = Mei-hfda and ⁿPri-hfda [42], and [Ir(CO)₂(L)], L = ⁿPri-hfda, Eti-hfda, [43];
• β-iminoketonates [Ir(cod)(L)], L = Eti-hfac [42], Mei-tfac [44], Mei-acac, i-acac [45], and [Ir(CO)₂(L)], L = Eti-hfac, ⁿPri-hfac [43], TFB-TFEA [46].

In general, approaches to the synthesis of Ir(I) complexes with (O^N)-coordinated ligands are similar to those described above for β-diketonate derivatives. In particular, fluorinated compounds with cyclooctadiene [Ir(cod)(L)] (L = ThTFP [41], Eti-hfac, ⁿPri-hfda, and amakN(Me)₂ [42], Mei-tfac [44]) were obtained by the reaction of [Ir(cod)Cl]₂ with a sodium salt of the corresponding ligand. The yields were 75%–95%. For the synthesis of fluorine-free complexes (L = Mei-acac, i-acac), the corresponding salts were generated in the reaction mixture in situ [45].

Carbonyl complexes [Ir(CO)₂(L)] were synthesized by passing carbon monoxide gas into a solution of the corresponding cyclooctadiene derivatives [Ir(cod)(L)] [43,46]. However, in contrast to the methods of preparation of β-diketonate complexes, the [Ir(cod)(L)] complexes themselves were not isolated, but generated in situ through the reaction of stoichiometric quantities of [Ir(cod)Cl]₂, NaH, and HL with the high (70–80%) yields of all target complexes, with the exception of L = ⁿPri-hfac (50%).

Although most of the presented compounds decompose in a polar solvent in air, and carbonyl complexes decompose faster, solutions of two imine complexes with cyclooctadiene [Ir(cod)(L)] with L = ⁿPri-hfac and ⁿPri-hfda are stable [42].

Crystal structures were determined only for single representatives of complexes with different anionic ligands: [Ir(cod)(L)], L = Mei-tfac [44], L = Mei-hfda [42] and ThTFP [41] and [Ir(CO)₂(L)], L = TFB-TFEA [46] and amakN(Me)₂ [43]. Similar to β-diketonate complexes, anionic ligands here also perform a chelating function, implementing the coordination nodes IrONC’₂ (C’ is the center of C=C bond) or IrONC₂ (Figure 5). The absence of conjugation in the chelate cycle of an anionic ligand (e.g., L = amakN(Me)₂ [43]) leads to a noticeable spatial distortion of the molecule, i.e., deviation from planarity.

Complexes with cyclopentadienyl ligands are a fairly large class of compounds not only due to the possibility of variation of substituents in the aromatic ring, but also due to a large number of neutral ligands that can complement the coordination environment of iridium, such as alkenes, alkynes, dienes, carbonyls, phosphines, amines, imines, and so on. Among them, only compounds with a limited number of neutral ligands have been proposed for MOCVD applications:

• complexes with cyclooctadiene [Ir(cod)Cp^X], Cp^X = Cp [47], Cp^Me [47,48], Cp^Me3 [47], Cp^Me4 [49], Cp* [50], Cp^Et [50], Cp-allyl [51], Cp-propenyl [51], Cp-ⁱpropenyl [52]);
• complexes with cyclohexadiene [Ir(chd)Cp^X] (Cp^X = Cp [53], Cp^Me [54], Cp^Et [55]);
• complexes with ethylene [Ir(C₂H₄)₂Cp^X] (Cp^X = Cp [56], Cp^Et [56,57]);
• complexes with carbon monooxide [Ir(CO)₂Cp*] [58].

The synthesis of cyclopentadienyl complexes of iridium (I), as in other cases, is carried out in an inert atmosphere, which is caused not only by the possibility of oxidation of iridium, but also by the complexity of working with MCp^X salts (M is an alkali metal), since they decompose in air and are often pyrophoric.

A technique described by Angelici in 1991 [47], in which the reagents are [Ir(cod)Cl]₂ and the substituted cyclopentadiene is salt, is mainly used to obtain the complexes [Ir(cod)Cp^X]. The yields in the reactions vary from 30% to 95% depending on the substituents in the Cp ligand.

The highest yields (90%–95%) are obtained in the synthesis of complexes with a monosubstituted Cp^X (X = Me, Et) ligand [54,55,56], although they decrease with an increasing substituent length (80% at X = allyl) [51].

Complexes with polysubstituted cyclopentadienyl ligands are obtained with the lower yields of 45%–85% [47]. In addition, the yield value strongly depends on the cation in the MCp^X salt and the selected solvent. For example, the yield of [Ir(cod)Cp^Me4] complex is 45% when using LiCp^Me4 in tetrahydrofuran (THF) [47], whereas the product yields vary from 15% to 54% when using NaCp^Me4 or KCp^Me4 as a ligand source and benzene, THF or diethyl ether as a solvent [49]. Alternative methods of the preparation of [Ir(cod)Cp^X] complexes are also described. The first one is the direct interaction of cyclopentadiene with the complex [Ir(cod)(OH)]₂ [59]. This method eliminates the stage of synthesis of cyclopentadiene salts and further formation of insoluble chlorides in the reaction mixture, which results in a cleaner product. Despite the attractiveness of the method, as alkyl substituents are added to the cyclopentadienyl ring, the acidity of the C–H proton decreases, which leads to a significant decrease in the yield in this reaction. The second method is based on the reduction of dimeric cyclopentadienyl complexes of iridium(III) in the presence of cyclooctadiene. The interaction of [Ir(Cp*)(µ-H)(µ-Cl)(Cl)₂] with an excess of cyclooctadiene in the presence of sodium carbonate makes it possible to isolate [Ir(cod)Cp*] with the yield of 73% [60]. Complexes with cyclohexadiene can also be obtained using this technique.

It should be noted that the neutral cod-ligand in cyclopentadienyl complexes [Ir(cod)Cp^X] is more strongly bound to the central atom than in [Ir(cod)(L)] complexes with β-diketonate ligands and their derivatives. This does not allow to obtain [Ir(CO)₂Cp^X] complexes directly by the substitution of the cod-ligand in the reaction with CO, as described above. Therefore, alternative synthetic strategies are developed. For example, the complex [Ir(CO)₂Cp*] was obtained by the reduction from the cyclopentadienyl complex of iridium(III), [IrCp*Cl₂]₂, using the iron carbonyl complex Fe₃(CO)₁₂ [58].

Despite the fact that methods for the synthesis of cyclopentadienyl complexes of iridium(I) have been known since the mid-20th century, only five complexes with cyclooctadiene and one with cyclohexadiene have been structurally characterized: [Ir(cod)Cp^Me] [48], [Ir(cod)Cp^Me4] [49], [Ir(cod)Cp^Me5] [50], [Ir(cod)(Cp-propenyl)] [51], [Ir(cod)(Cp-ⁱpropenyl)] [52], [Ir(chd)Cp^Me] [54]. In all complexes, the cyclopentadienyl ligand is coordinated by the central atom according to the η⁵-type, while cyclooctadiene and cyclohexadiene are coordinated according to the η⁴-type (Figure 6).

Thermal properties of iridium(I) precursors

Information about thermal properties of the considered iridium compounds is limited and is mainly represented by melting characteristics, thermogravimetry results, and sublimation parameters [32,33,34,41,42,43,46,58]. Moreover, among a wide range of synthesized cyclopentadienyl complexes, only precursors containing Cp^X = Cp^Me, Cp^Et, Cp^Me4, and Cp* [48,49,50,57] were studied in the thermochemical aspect. The thermochemical data for dimeric acetate [Ir(cod)(OAc)]₂ and alcoholic [Ir(cod)(OMe)]₂ complexes, with the exception of the melting points, are completely absent [27,29].

A feature of [Ir(cod)Cp^X] complexes is an extremely low melting point manifesting when one alkyl substituent is introduced into the cyclopentadienyl ligand. For example, complexes with Cp^Me and Cp^Et melt at 38–40 and 14 °C, respectively. Thus, these compounds are easy to use in MOCVD processes as liquid precursors, which provide convenient dosage and a constant surface area during vaporization.

Other volatile iridium(I) complexes have noticeably higher melting points. An increase in the number of methyl groups in the Cp ligand leads to an increase in the melting temperature, namely, up to 124 °C in the case of four methyl groups (Cp^X = C^Me4) and up to 170 °C in the case of five methyl groups (Cp^X = Cp*). In the series of β-diketonate complexes [Ir(cod)(L)], the introduction of a CF₃ group in L leads to a decrease in the melting temperature, while the presence of a ^tBu or Ph substituent results in its increase [33]. Interestingly, the replacement of the methyl group in the ^tBu substituent with a methoxide group (L = zis vs. thd) leads here to such a decrease in the melting point that the properties of this non-fluorinated complex approach those of fluorinated analogues, which was not observed for all previously studied volatile β-diketonate precursors [35]. Thus, in the series of considered compounds, the melting point varies in the range of 105–170 °C, where the complexes [Ir(cod)(tfac)] and [Ir(cod)(zis)] have the lowest melting points, while [Ir(cod)(thd)] has the highest one.

The tendency for a decrease of the melting point when fluorine atoms are introduced into the ligand is also characteristic of β-diketonate complexes [Ir(CO)₂(L)] [36,37,38] and β-iminoketonate derivatives [Ir(cod)(L)] [42,44,45]. For the other classes of (O^N)-coordinated ligands, only fluorinated complexes were synthesized.

It was shown that in the case of β-iminoketonate complexes, the introduction of the Me group at the nitrogen atom also led to a decrease in the melting point, apparently due to the absence of intermolecular N–H...O interactions [45]. In general, in the series of β-iminoketonate complexes [Ir(cod)(L)] (L = Eti-hfac < Mei-tfac < Mei-acac < i-acac), the melting point varies in the range of 111–167 °C. Presumably, the change in the melting point during the transition from [Ir(cod)(L)] to [Ir(CO)₂(L)] does not have a definite character even within one class of L ligands. Specifically, this parameter can decrease (L = hfac: 116 °C vs. 92 °C, L = Cp*: 170 °C vs. 110 °C), change slightly (L = acac: 152 °C vs. 160 °C), or increase (L = tfac: 109 °C vs. 148 °C). However, iridium(I) carbonyl complexes are always noticeably more volatile than their cyclooctadiene-containing analogues. This result was obtained by thermogravimetry (TG) with a wide variation of anionic ligands L (β-diketonates, β-iminoketonates, β-iminoalcoholates, α-aminoalcoholates, cyclopentadienyls). A similar study of a series of cyclopentadienyl complexes [Ir(Q)Cp^Et] demonstrates the effect of changing the neutral Q ligand within the class of alkenes [57]. In this case, the volatility regularly increased with a decrease in the molecular weight of the ligand: Q = cod < chd << C₂H₄.

The effect of the anionic ligand L on volatility has also been mainly investigated by the TG method. The most complete picture is observed for the complexes [Ir(cod)(L)] with β-diketonate derivatives investigated in a series of our recent papers [33,35,44,45]. The presence of fluorinated substituents in the β-diketonate ligand L leads to an increase in the volatility of the complex, while the presence of the Ph group leads to a decrease in the volatility, and in the latter case, the vaporization process is accompanied by significant decomposition. The introduction of bulk ^tBu substituents also reduces the volatility of the compound, which is a feature of complexes of this class in comparison with monoligand β-diketonates. The replacement of the Me group in the ^tBu substituent with OMe results in the opposite effect, and this effect of increasing volatility (L = zis vs. thd) is also opposite to that observed for monoligand planar-square complexes. Replacement of β-diketonate ligand (L = acac) with the corresponding β-iminoketonate derivative (L = i-acac, Mei-acac) leads to a decrease in volatility.

It should be noted that vaporization of all the studied complexes [Ir(cod)(L)] with (O^N)-donor ligands is accompanied by decomposition [42,44,45] and for β-heteroarylketonate [Ir(cod)(ThTFP)] the mass loss curve corresponds only to decomposition with a clear cleavage of the anionic ligand at the first stage [41]. According to [42] and [43], it can be assumed that β-iminoketonates are the most volatile among [Ir(cod)(L)] and [Ir(CO)₂(L)] complexes with the studied classes of (O^N)-coordinated ligands, whereas β-iminoalcoholates are the least volatile. However, very different masses and substituents in the ligands of this series (L = Eti-hfac, amakN(Me)₂, ⁿPri-hfda) do not allow us to make a final conclusion.

In a series of cyclopentadienyl complexes [Ir(cod)Cp^X], an increase in the size of a single substituent or the number of substituents leads to a slight decrease in volatility: Cp^X = Cp^Me > Cp^Et ≥ Cp^Me4 > Ir(cod)Cp* [50]. However, the temperature difference is very small, so quantitative studies are required to better identify the considered effect.

The volatilization processes were studied quantitatively for a limited number of compounds, namely, for the following five complexes: [Ir(cod)(L)], L = acac [45,58], Cp^Me [58,61], zis [35], i-acac, Mei-acac [45], and two complexes [Ir(CO)₂(L)], L = acac, Cp* [58]. The data obtained are summarized in Figure 7 and

Table 3. Equations of the temperature dependences of saturated vapor pressure over the solid (sublimation, subl.) and liquid (vaporization, vap.) Ir(III) β-diketonates. The sublimation and vaporization enthalpies (ΔH_T*) and entropies (ΔS⁰_T*) are referred to the average temperature (T*) of the interval measured (ΔT), p₀ = 760 Torr = 1 atm = 10⁵ Pa. The ligand abbreviations correspond to Table 1.

Ref.	Complex	Process	ΔT, K	ln(p/p0) = A − B/(T,K)		ΔHT*, kJ·mol−1	ΔS0T*, J·(mol·K)−1
				A	B
[45]	[Ir(cod)(acac)]	Subl.	363–423	23.03	12,817	106.6 ± 0.7	191 ± 2
[35]	[Ir(cod)(zis)]	Subl.	353–376	28.04	14,886	124 ± 4	233 ± 10
		Vap.	381–403	18.95	11,477	96 ± 5	158 ± 13
[45]	[Ir(cod)(i-acac)]	Subl.	383–431	25.31	14,312	119 ± 2	210 ± 5
	[Ir(cod)(Mei-acac)]	Subl.	383–420	22.56	13,398	111 ± 3	188 ± 7
[58]	[Ir(cod)CpMe]	Subl.	304–310	33.43	14,990	124.6 ± 5.0	279 ± 16
		Vap.	310–330	19.39	10,592	88.1 ± 1.3	161.2 ± 4.2
	[Ir(CO)2Cp*]	Subl.	297–332	27.91	12,641	105.0 ± 3.4	232 ± 11
	[Ir(CO)2(acac)]	Subl.	306–333	22.40	11,293	94 ± 2	186 ± 8

. All Ir(I) complexes are more volatile than the traditional non-fluorinated Ir(III) MOCVD precursor, Ir(acac)₃. The order of their volatility is consistent with thermogravimetry data. In the series of [Ir(cod)(L)] complexes with β-diketonate derivatives, the change in the vapor pressure values when changing the terminal substituents (L = acac vs. zis) and the coordination environment of the metal (L = acac vs. i-acac, Mei-acac) is quite small. It is interesting to note the unusual order of the minor decrease (L = acac > i-acac > Mei-acac) in volatility being partially opposite to that for the planar-square complexes [M(L)₂] with the same ligands [45]. Replacing the cod ligand with carbonyl ligands, i.e., switching from [Ir(cod)(L)] to [Ir(CO)₂)(L)], leads to a significant increase in the vapor pressure of complexes with both β-diketonate and cyclopentadienyl ligands L. In this case, the complex [Ir(CO)₂Cp*] is the most volatile among the Ir(I) compounds under consideration; however, synthetic difficulties (see above) hinder its practical application.

As for the stability during volatilization, sublimation is congruent for almost all the studied complexes, except for [Ir(cod)(Mei-acac)], for which partial decomposition is observed. After melting, the thermal stability of the compounds usually decreases. This is most pronounced for [Ir(CO)₂(L)] complexes, and least pronounced for [Ir(cod)Cp^X] complexes. For example, the composition of the gas phase of [Ir(cod)Cp^Me] was shown by mass spectrometry to be stable over time during vaporization at 60 °C, and only its holding at 120 °C results in the appearance of signal corresponding to [Ir(cod)Cp^Me]₂ and a decrease in the intensity of signal corresponding to the molecular ion [61].

Recently, we studied the thermal stability of vapors of a representative series of β-diketonate complexes [Ir(cod)(L)] (L = hfac, tfac, ptac, acac, thd, btfac) on a heated surface [33]. It turned out that the vapor of complex [Ir(cod)(thd)] was the most stable, which appeared to be due to the presence of two bulk tert-butyl substituents. Indeed, a relatively low degree of vapor decomposition even at maximum temperatures (475–500 °C) was also observed for complexes with one bulk substituent in the ligand, i.e., tert-butyl (L = ptac) or phenyl (L = btfac).

The [Ir(cod)(hfac)] complex is characterized by the lowest vapor stability, which indicates the possibility of deposition of coatings at deposition temperatures lower than those for the other compounds from this series. This property together with its highest volatility makes this compound a particularly interesting precursor. More detailed information about the thermal behavior of vapors on a heated surface is provided for two Ir(I) volatile complexes: [Ir(cod)Cp^Me] [62] and [Ir(CO)₂(acac)] [25,63].

Using in situ mass-spectrometry [Ir(cod)Cp^Me] vapors have been shown to be stable up to 280 °C in the absence of a gas-reagent, but the rate of thermolysis was greatly increased at 500 °C (the intensity of ions corresponding to HCp^Me and cod increases sharply) [62]. When oxygen is introduced into the reaction zone, vapor decomposition begins at a temperature of ~200 °C; the main gaseous products of thermolysis are CO, CO₂, and H₂O. In this case, the molecular ion peak [Ir(cod)Cp^Me]⁺ and the peaks of its defragmentation products under electron impact practically disappear already at temperatures above 230 °C, which indicates complete decomposition of the precursor [62].

For the [Ir(CO)₂(acac)] complex, experimental results obtained by in situ mass spectrometry [25] were recently supplemented by quantum chemical modeling, which allowed us to present a more complete picture of the process [63]. Calculations have shown that the decomposition of the molecule in the gas phase begins with the cleavage of one of the Ir–O bonds. When the molecule is adsorbed on the substrate, a structural rearrangement occurs, which consists of a noticeable increase in the lengths of the Ir–O and Ir–C bonds, and the carbonyl and methyl groups are located out of the complex plane. Thus, the decomposition of the precursor on a heated surface proceeds by a different mechanism than in the gaseous phase, i.e., with the cleavage of Ir–C bonds at the first stage during a removal of CO molecule. The process is schematically presented in Figure 8.

Threshold temperature of the beginning of [Ir(CO)₂(acac)] vapor decomposition in vacuum is 200 ± 10 °C. It is shown that the introduction of gases-reagents (oxygen or hydrogen) does not significantly affect the process temperatures. Moreover, the main decomposition products (CO, Hacac) in vacuum and in the presence of hydrogen are the same. Peaks corresponding to the appearance of water and CO₂ were registered in the oxidizing atmosphere.

2.1.3. Application of Iridium Precursors in MOCVD Processes

Nowadays, the most commonly used iridium precursor remains Ir(acac)₃. The main regularities of MOCVD processes with the use of Ir(acac)₃ and the characteristics of the resulted coatings were described in detail in our previous review [18] and supplemented in the recent paper [64] focused on the coating deposition in the presence of hydrogen. In general, this precursor is successfully used to produce both metallic iridium (in the presence of hydrogen or oxygen) and its oxide (in an oxidizing atmosphere).

At the same time, this compound has some notable disadvantages, such as high thermal stability of vapors in hydrogen atmosphere, which requires high deposition temperatures (>460 °C) and relatively low volatility, as a result of which the source temperature is usually set above 200 °C to create the necessary concentration of precursor vapors, and, finally, a rather intricate synthesis. Comparing Ir(I) volatile precursors with Ir(acac)₃ based on these parameters (see above), we can conclude that this type of compound looks more attractive for MOCVD processes. In addition to the advantages associated with these aspects, the variety of combinations of neutral and anionic ligands provides the ability of “fine-tuning” of the thermochemical properties of the precursor, based on the characteristics of the target material and the object to be coated. For example, if we want to deposit a coating on a substrate with low thermal stability, we can choose a compound with a low thermal stability of vapors or use ligands that are sensitive to additional activation of the decomposition process.

Despite this potential, the use of Ir(I) precursors in MOCVD processes is still limited. The representative examples are summarized in Table 4 in order to show the range of possibilities for varying deposition conditions, composition, growth rate, and some other characteristics of the resulting coatings depending on the specific precursor. This information will be briefly discussed below.

The first MOCVD experiments using Ir(I) complexes were conducted in 1967, when high-purity iridium coatings (90–99 mass.% Ir) were deposited from [Ir(cod)(OMe)]₂ and [Ir(cod)(acac)] in the presence of hydrogen at 550 and 600 °C, respectively [26]. Then the row of precursors used for the deposition of pure metal films (about 1 mass.% C and O) in a reducing atmosphere was expanded. It was shown that when using [Ir(cod)(hfac)] the coating formed already at 350 °C, and the growth rate was quite high (15 nm/min) [34]. At the same time, the formation of coatings from [Ir(cod)(thd)] at this temperature was possible only by additional passing of hydrogen through isopropanol [32]. These results are consistent with our data on the thermal stability of the vapors of these complexes: high stability in the case of [Ir(cod)(thd)] and a relatively low one for [Ir(cod)(hfac)] [33]. Using [Ir(cod)Cp^Me] it is possible to obtain coatings with a high growth rate (up to 25 nm/min) at 300–400 °C [65], and the precursor is removed from the reactor without decomposition at the lower temperatures, which also correlates perfectly with in-situ mass spectrometry data [62]. Finally, [Ir(CO)₂(acac)] vapors were shown by mass-spectrometry to be the least stable in the presence of hydrogen, and the possibility of deposition of coatings from this precursor at the lowest temperature to date (240–300 °C) was confirmed using pulsed MOCVD [25].

As for the use of (O^N)-coordinated precursors in a reducing atmosphere, it has been shown using a fluorinated β-iminoalcoholate complex [Ir(cod)(ⁿPri-hfda)] as an example that the growth rate was relatively low (1.3 nm/min) even at a deposition temperature of 500 °C [42]. The latter temperature was higher than that used for deposition of coatings from a fluorinated β-diketonate precursor [Ir(cod)(hfac)]. At the same time, the obtained samples had a high degree of contamination (41 mass.% C). Replacing the gas-reagent with oxygen results in a decrease in both the amount of impurities (2 mass.% O) and the deposition temperature (375 °C), as well as a noticeable increase in the growth rate. The formation of metal coatings under similar conditions was also demonstrated for other (O^N)-coordinated cyclooctadiene complexes [42].

The use of oxygen as a gas-reagent for the production of metallic iridium has proven itself well in the case of liquid cyclopentadienyl complexes. When using the [Ir(cod)Cp^Me] precursor, pure coatings (<1 at.% C and O) are formed already at 270 °C [29,66]; however, the growth rate is low. The application of the special system of “liquid delivery” of the precursor using a tetrahydrofuran (THF) solution allows attainment of exceptionally high growth rates (up to 70 nm/min at 350 °C) [67]. However, the purity of the coatings noticeably reduces (12.5 at.% C and 4 at.% O). Orientation of crystallites in films deposited from [Ir(cod)Cp^Me] strongly depends on the substrate material and deposition temperature [62,68]. An interesting example is the preparation of highly oriented <111> films on amorphous native silicon oxide at 700 °C: the ratio of intensities of the (111) and (200) reflections is two orders of magnitude higher than in the non-textured sample [68]. It is important that despite the high deposition temperature, these coatings are also pure metallic iridium, whereas when using [Ir(cod)Cp^Et] with an increase in the deposition temperature (300–500 °C), the probability of formation of iridium oxide increases [69,70,71,72,73]. Another factor affecting the phase composition of films formed from [Ir(cod)Cp^Et] is the oxygen concentration. With a significant increase in the O₂/[Ir(cod)Cp^Et] ratio, the content of the oxide phase increases [70]; however a balance that allows obtaining metal coatings with a high growth rate (up to 5 nm/min) can be found [72,73].

Table 4. Deposition conditions and some characteristics of the coatings obtained by the MOCVD method from volatile iridium(I) complexes. The ligand abbreviations correspond to Table 1 ^a.

Ref.	Precursor	Reagent Gas	Source/Deposition Temperature, °C	Substrate	Thickness, nm	Growth Rate, nm/min	Composition (XRD)	Purity and Other Details
[26]	[Ir(cod)(OMe)]2	H2	–/550	–	–	–	Ir	5–10 mass.% impurities
	[Ir(cod)(acac)]	H2	–/600	–	–	–		1 mass.% C and O
[29]	[Ir(cod)(OAc)]2	no	130/250	Si (100)	40	13.3	Ir	1 mass.% C and O
[34]	[Ir(cod)(hfac)]	H2	60–65/250–350	SiO2/Si, Pt/Si, Cu/Si	–	15	Ir	0.1–0.3 μm agglomerates, 1 mass.% C
[32]	[Ir(cod)(thd)]	H2 + iPrOH	130/350	Quartz glass	25	4.1	Ir	1 mass.% C and O, islands
		H2	130/350–550		–	<0.7	Ir	–
[25]	[Ir(CO)2(acac)]	H2 or O2	–/240–300	Si (100)	–	–	Ir	Crystallite size 4–20 nm
[42]	[Ir(cod)(Eti-hfac)]	O2	–/400	Si (100)	278	7.0	Ir	2 mass.% O
	[Ir(cod)(nPri-hfda)]	H2	–/500	Si (100)	322	1.3		30–90 nm agglomerates, 41 mass.% C
		O2	–/375	Si (100)	320	3.8		1 mass.% O
	[Ir(cod)(amakN(Me)2)]	O2	–/350	Si (100)	264	4.4		2 mass.% O
[43]	[Ir(CO)2(nPri-hfda)]	O2	50/400	Si (100)	290	4.8	Ir	60–90 nm agglomerates, 2 mass.% impurities
			75/425		400	1.7	IrO2	Polycrystalline
			75/425	LiTaO3 (012)	2000	8.3		tilted needles: 15–25 nm in diameter, 1.5 μm in length
			75/425	LiNbO3 (100)	3000	12.5		vertical needles
			95/425	LiNbO3 (100)	430	7.1		vertical pillars
[41]	[Ir(cod)(ThTFP)]	no	110–130/700–800	Si (100), SiO2, Al2O3	30–140	0.5–0.6	Ir	Uniform, crystallites < 30 nm
[65]	[Ir(cod)CpMe]	H2	100/300–400	W	700–1500	11.7–25	Ir	Untextured constituted of small grains
[29]		O2	80/270	Si (100)	95	0.42		<1 mass.% C and O
[67]			C = 0.05M(THF) 200/300–450	Si(100), SiO2/Si, TiN/SiO2/Si, Pt/Ti/Si, IrOx/poly-Si/SiO2/Si	420	70		0.9 mass.% C, 0.4 mass.% O
[73]	[Ir(cod)CpEt]	O2	150–180/300–500	Si, SiO2/Si, TiO2/SiO2/Si	1.2–300	0.02–5	Ir	0.4 mass.% O
[57]	[Ir(C2H4)2CpEt]	O2	40/400	SiO2/Si	–	–	Ir	Low O2 flow (0–2 sccm)
							IrO2	Higher O2 flow (10–160 sccm)
[55]	[Ir(chd)CpEt]	O2	70/250, 350	SiO2	65, 120	1.6, 30	Ir	Uniform crystallites, roughness—1.2 nm

^a XRD = powder X-Ray diffraction, ⁱPrOH = propanol-2, C = concentration, THF = tetrahydrofurane.

Table 4. Deposition conditions and some characteristics of the coatings obtained by the MOCVD method from volatile iridium(I) complexes. The ligand abbreviations correspond to Table 1 ^a.

Ref.	Precursor	Reagent Gas	Source/Deposition Temperature, °C	Substrate	Thickness, nm	Growth Rate, nm/min	Composition (XRD)	Purity and Other Details
[26]	[Ir(cod)(OMe)]2	H2	–/550	–	–	–	Ir	5–10 mass.% impurities
	[Ir(cod)(acac)]	H2	–/600	–	–	–		1 mass.% C and O
[29]	[Ir(cod)(OAc)]2	no	130/250	Si (100)	40	13.3	Ir	1 mass.% C and O
[34]	[Ir(cod)(hfac)]	H2	60–65/250–350	SiO2/Si, Pt/Si, Cu/Si	–	15	Ir	0.1–0.3 μm agglomerates, 1 mass.% C
[32]	[Ir(cod)(thd)]	H2 + iPrOH	130/350	Quartz glass	25	4.1	Ir	1 mass.% C and O, islands
		H2	130/350–550		–	<0.7	Ir	–
[25]	[Ir(CO)2(acac)]	H2 or O2	–/240–300	Si (100)	–	–	Ir	Crystallite size 4–20 nm
[42]	[Ir(cod)(Eti-hfac)]	O2	–/400	Si (100)	278	7.0	Ir	2 mass.% O
	[Ir(cod)(nPri-hfda)]	H2	–/500	Si (100)	322	1.3		30–90 nm agglomerates, 41 mass.% C
		O2	–/375	Si (100)	320	3.8		1 mass.% O
	[Ir(cod)(amakN(Me)2)]	O2	–/350	Si (100)	264	4.4		2 mass.% O
[43]	[Ir(CO)2(nPri-hfda)]	O2	50/400	Si (100)	290	4.8	Ir	60–90 nm agglomerates, 2 mass.% impurities
			75/425		400	1.7	IrO2	Polycrystalline
			75/425	LiTaO3 (012)	2000	8.3		tilted needles: 15–25 nm in diameter, 1.5 μm in length
			75/425	LiNbO3 (100)	3000	12.5		vertical needles
			95/425	LiNbO3 (100)	430	7.1		vertical pillars
[41]	[Ir(cod)(ThTFP)]	no	110–130/700–800	Si (100), SiO2, Al2O3	30–140	0.5–0.6	Ir	Uniform, crystallites < 30 nm
[65]	[Ir(cod)CpMe]	H2	100/300–400	W	700–1500	11.7–25	Ir	Untextured constituted of small grains
[29]		O2	80/270	Si (100)	95	0.42		<1 mass.% C and O
[67]			C = 0.05M(THF) 200/300–450	Si(100), SiO2/Si, TiN/SiO2/Si, Pt/Ti/Si, IrOx/poly-Si/SiO2/Si	420	70		0.9 mass.% C, 0.4 mass.% O
[73]	[Ir(cod)CpEt]	O2	150–180/300–500	Si, SiO2/Si, TiO2/SiO2/Si	1.2–300	0.02–5	Ir	0.4 mass.% O
[57]	[Ir(C2H4)2CpEt]	O2	40/400	SiO2/Si	–	–	Ir	Low O2 flow (0–2 sccm)
							IrO2	Higher O2 flow (10–160 sccm)
[55]	[Ir(chd)CpEt]	O2	70/250, 350	SiO2	65, 120	1.6, 30	Ir	Uniform crystallites, roughness—1.2 nm

^a XRD = powder X-Ray diffraction, ⁱPrOH = propanol-2, C = concentration, THF = tetrahydrofurane.

The possibility of deposition of metallic iridium from carbonyl precursors in an oxidizing atmosphere was also shown. For example, films containing iridium in a single form (Ir⁰, according to X-ray photoelectron spectroscopy, XPS) were obtained from [Ir(CO)₂(acac)] [25] and [Ir(CO)₂(ⁿPri-hfda)] [43] at 280 and 400 °C, respectively.

It was shown for (O^N)-coordinated precursor [Ir(CO)₂(ⁿPri-hfda)] that for the simultaneous increase in the total pressure in the reactor, the source and deposition temperatures led to the formation of IrO₂ films [43]. Both the temperature parameters and the substrate nature had a significant effect on the shape and orientation of particles, as well as on the growth rate of these layers. In particular, replacing the Si(100) substrate with LiTaO₃ (012) or LiNbO₃ (100) led to the formation of obliquely or vertically oriented needles, respectively, and to a significant increase in the growth rate. With the same LiNbO₃ (100) substrate, an increase in the source temperature led to the formation of vertical columns instead of needles, accompanied by a decrease in the growth rate.

The possibility of preparation of iridium oxide films has also been tested for a series of such cyclopentadienyl complexes as [Ir(cod)Cp^Et], [Ir(chd)Cp^Et] and [Ir(C₂H₄)₂Cp^Et] [55,57]. It was shown that only the use of [Ir(C₂H₄)₂Cp^Et] precursor made it possible to obtain pure oxide films with an increase in the amount of oxygen. In the case of [Ir(cod)Cp^Et], the iridium oxide phase did not form even at the maximal oxygen concentrations at temperatures of 250–400 °C, and in the case of [Ir(chd)Cp^Et] only mixed Ir–IrO₂ layers could be obtained.

Experiments on the deposition of coatings in the absence of a gas-reagent were carried out using precursors [Ir(cod)(ThTFP)] and [Ir(cod)(OAc)]₂. In the case of the first precursor, metal layers formed at high temperatures (700–800 °C) [41]. An increase in the deposition time led to an increase in the film crystallinity and thickness, and the substrate material did not affect the decomposition of the precursor. The purity of the obtained samples was not discussed. An unexpected result was obtained when using [Ir(cod)(OAc)]₂: the possibility of the formation of pure metal films at a very low temperature (250 °C) was shown [29]. The growth rate was two orders of magnitude higher than that observed in the case of [Ir(cod)(ThTFP)]. However, this compound was not further used as a precursor.

In general, the use of Ir(I) precursors of different classes, as well as the variation of ligands within the same class, allowed us to obtain metal and oxide coatings in a wide range of deposition temperatures, and in some cases, to control the orientation of crystallites and the morphology of the formed layers.

2.2. Volatile Platinum Precursors

Platinum is able to coordinate various types of C donor ligands, including olefins, carbonyl, dienes, vinyls, etc., and form strong bonds with methyl groups [15,74,75], which causes a variety of organoplatinum compounds that can be promising precursors for MOCVD.

In contrast to organoplatinum compounds, the number of platinum complexes with organic O, N, S, and Se-donor ligands (Table 1) is significantly limited due to the high chemical inertness of platinum in exchange reactions.

2.2.1. Platinum(0) Precursors

There are only a few examples of platinum(0) precursors, namely, Pt(PF₃)₄ and Pt(C₂H₄)₃ [15]. Both compounds are obtained by heating anhydrous PtCl₂ in the presence of reducing agents and gases (C₂H₄, PF₃) at the pressure above 100 MPa [76]. Pt(PF₃)₄ and Pt(C₂H₄)₃ are liquids (Pt(PF₃)₄ (boiling temperature is 87 °C) and decompose to metal with the elimination of ligands in air. Due to low storage stability and synthetic unavailability, these compounds are currently almost not used in MOCVD.

2.2.2. Platinum(II) Precursors

As a rule, MOCVD platinum(II) precursors are represented by the following two classes: organoplatinum compounds and platinum complexes with organic O,N,S,Se-donor ligands.

Organoplatinum compounds, in turn, are represented by three classes:

• alkylplatinum derivatives with monodentate ligands: Me(R)PtQ_2, where R = Me, CO, HC = CH₂, ^tBuC≡C, ɳ³-C₃H₅; Q = HCN, R’CN [74,75];
• dimethylplatinum derivatives with bidentate ligands: Me₂Pt(Q), where Q = tmeda (N,N,N′,N′-tetramethylethylenediamine) [15], hd and its derivatives [77], nbd and its derivatives [78], cod and its derivatives [79,80] (Table 1);
• platinum homoleptic ω-alkenyl derivatives: Pt(X)₂, where X = C₆H₁₁, C₇H₁₃ and C₈H₁₅ [81,82] (Table 1).

Alkylplatinum derivatives with monodentate ligands are among the first compounds synthesized and tested in MOCVD processes. Kumar et al. [74,75] were among the first who synthesized and studied the series of Me(R)PtQ₂ compounds (R = Me, CO, HC = CH₂, ^tBuC≡C, ɳ³-C₃H₅, Q = HCN, R’CN) by spectral methods in 1989. Compounds were obtained with yields up to 60% under inert conditions by exchange reactions using cis-[PtMe(Cl)(SMe₂)] as an initial reagent. Most of Me(R)PtQ₂ is a white or light-yellow powder unstable in air. Due to the complexity of synthesis and the high content of carbon impurities (up to 40 at.%) in the resulting films [74,75], Me(R)PtQ₂ precursors have not found a real application in MOCVD processes.

The first example of a dimethylplatinum derivative with a bidentate ligand, namely Me₂Pt(cod), was presented as a MOCVD precursor in the same paper by Kumar et al. [38] This direction turned out to be more promising, and the chemistry of these derivatives has been intensively developing over the past decade. To date, a wide range of Me₂Pt(Q) compounds, where Q is some olefins, namely, hd, nbd, cod, and their derivatives [77,78,79], has been synthesized. In addition, an example of the Me₂Pt(Q) compound with Q = tmeda (N,N,N′,N′-tetramethylethylenediamine) is also known [15]. The developed approaches to the synthesis of Me₂Pt(Q) compounds are summarized in Figure 9. Note that the yields of Me₂Pt(Q) complexes do not exceed 60% using approaches based on interactions of Pt(Q)Cl₂ or Pt(Q)I₂ intermediates with alkyl sources (MeLi, Cu₂LiMe, MeMgX) (Figure 9). On the contrary, the approach based on a single step interaction of Pt(acac)₂ with AlMe₃ in the presence of Q allows us to achieve yields of the Me₂Pt(Q) compounds of up to 85% [77,78].

According to single-crystal XRD data, the platinum atom in the molecules of Me₂Pt(Q) compounds is located in a tetrahedral environment with PtC₂C’₂ (C’ is the center of C=C bond) coordination core (Figure 10), and Pt–C_Me bond lengths is approximately 0.1 Å shorter than Pt–C_C=C ones. This trend also persists for Me₂Pt(cod) [80], while the replacement of Q from nbd to cod is accompanied by a shortening of the Pt–C_Me and Pt–C_C=C distances.

The platinum homoleptic ω-alkenyl derivatives Pt(X)₂, X = C₆H₁₁, C₇H₁₃ and C₈H₁₅ (Table 1) are synthesized by interaction of (cod)Pt(Hal)₂ with the corresponding LiX salts or Grignard reagents (XMgBr) in an inert atmosphere [81,82]. When using Grignard reagents, the target products Pt(C₆H₁₁)₂, Pt(C₇H₁₃)₂, and Pt(C₈H₁₅)₂ are obtained in a mixture with by-products such as Pt(cod)(C₇H₁₃)₂, Pt(cod)(Br)(C₆H₁₁). The use of LiX reagents leads to an increase in the yields of Pt(X)₂ compounds to 60%.

Compounds Pt(C₆H₁₁)₂, Pt(C₇H₁₃)₂ and Pt(C₈H₁₅)₂ adopt solid-state structures with (idealized) C2 symmetries, in which the two α-carbon atoms are mutually cis (Figure 11). Significant influence of steric and packing effects on the conformations of Pt(C₆H₁₁)₂, Pt(C₇H₁₃)₂ and Pt(C₈H₁₅)₂ suggest that the olefin−Pt interactions are relatively weak. Indeed, Pt(C₆H₁₁)₂ and Pt(C₈H₁₅)₂ compounds are dynamic in solution, i.e., C=C bonds reversibly decomplex at the rates that are fast on the NMR time scale, while Pt(C₇H₁₃)₂ shows the slowest decomplexation rates and the greatest thermal stability.

Platinum complexes with organic O,N,S,Se-donor ligands are represented by the following five classes: β-diketonates [83,84,85,86], β-iminoketonates [87,88], β-alkenols [89], α-aminoalcoholates [90], and dithio/diselenoimidodiphosphinates [91] (Table 1).

Although the first mention of the successful isolation of a volatile coordination compound of platinum, namely, Pt(acac)₂, was published in the first half of the 20th century, the search for synthesis methods aimed at increasing the yields of platinum β-diketonate complexes is still relevant. In fact, this first synthetic approach that consisted in to the reaction of K₂PtCl₄ with H(acac) in an alkaline medium gives a yield of 33%. The most successful strategy, apparently, is the synthesis through the formation of a labile aqua-ion Pt[(H₂O)₄]²⁺, which is then introduced into the reaction with potassium β-diketonate. This way allows obtaining the target complexes with yields up to 95% [83,84,92]. Since the formation of such an ion is possible only in an acidic medium, this approach is not applicable for obtaining platinum(II) complexes with other organic ligands (Table 1). The platinum complexes with β-iminoketonate, β-alkenol, α-aminoalcoholate, and dithio/diselenoimidodiphosphinate ligands (Table 1) are synthesized by substitution reactions with appropriate ligands in neutral or alkaline media using K₂PtCl₄ or cis-(Py)₂PtI₂ (Py = pyridine) as initial reagents, and the yields do not exceed 50%.

Regardless of the class of platinum(II) complexes, all the above complexes have monomeric molecular structures, in which platinum ions are in a plane-square geometry (Figure 12). In the case of asymmetric ligands, the formation of cis,trans-isomeric platinum(II) complexes is observed [83,87,88,93].

Thermal properties of platinum(II) precursors

Despite the variety of synthesized organoplatinum compounds, the data on their thermal properties are almost absent and presented only by melting characteristics and thermogravimetry results [75,78,79]. The melting points of Me₂Pt(Q) compounds can vary over a wide range depending on the Q ligand: from liquids (Q = cod derivatives) [79] and low-melting compounds (31–49 °C, Q = hd derivatives) [77] to solids (80–110 °C, Q = nbd derivatives) [78]. The melting points of platinum homoleptic ω alkenyl derivatives Pt(X)₂ increase with the growth of the carbon chain in the X ligand: from liquid (Pt(C₆H₁₁)₂, freezing point −15 °C) to solid Pt(C₇H₁₃)₂ and Pt(C₈H₁₅)₂ compounds with melting points 20 °C and 46–47 °C, respectively [81,82].

The majority of Me(R)PtQ₂ and Me₂Pt(Q) compounds decompose when heated. The Me(R)PtQ₂ and Me(R)PtQ₂ compounds with Q = RCN, nbd and their derivatives possess the lowest thermal stability, while Me₂Pt(cod) and related compounds are partially decomposes when heated. According to [75], the decomposition of Me(R)Pt(R’CN)₂ compounds (R = Me, C = CH, ɳ³-C₃H₅, R’ = H, Me) in the condensed phase is accompanied by the elimination of the R’CN ligand and its subsequent oligomerization. The analysis of gaseous products (methane, ethane, hexadiene, and butadiene) indicates elimination and further recombination of the alkyl radical R (Me, HC = CH₂, ɳ³-C₃H₅). Decomposition Pt(C₅H₉)₂ in aromatic solvents revealed the formation of a transient three-coordinate platinum hydride [81] that makes PtNPs deposited from Pt(C₅H₉)₂ a promising catalyst for the hydrogenation of chloronitrobenzene.

Thermal properties of volatile platinum complexes with organic O,N,S,Se-donor ligands, viz., Pt(β-diketonate)₂, Pt(β-iminoketonato)₂, Pt(β-alkenols)₂, Pt(amN(Me)₂)₂, are studied mainly by the thermogravimetry method [83,84,85,86,87,88,89,90].

Most of all platinum (II) complexes of the classes considered above have a melting point above 100 °C. In contrast to organoplatinum compounds, platinum(II) complexes usually pass into the gaseous phase almost quantitatively under the conditions of TG experiments (inert atmosphere), with the exception of Pt(acac)₂ and Pt(alk(Me))₂ complexes.

A tendency to increase the stability during vaporization with the introduction of a fluorinated substituent into the ligand was shown in the series of Pt(β-diketonate)₂ and Pt(β-alkenols)₂. A comparison of a series of platinum complexes with ligands of different classes but containing the same terminal substituent (Me) shows that the volatility of platinum complexes decreases in the order Pt(amN(Me)₂)₂ > Pt(acac)₂ > Pt(alk(Me))₂. The lower volatility of the Pt(alk(Me))₂ complex can be clearly explained by the presence of an aromatic fragment in the β-alkenol ligands. To study the effect of the ligand on the volatility of platinum(II) complexes in detail, quantitative data on the volatility of the complexes are necessary.

Until now, the volatilization processes have been quantitatively studied using tensimetic data (p-T dependences of saturated vapor pressure) only for two organoplatinum compounds, one platinum(II) β-iminoketonate and several platinum(II) β-diketonate complexes. This information is summarized in

Table 5. Equations of the temperature dependences of saturated vapor pressure over the solid (sublimation, subl.) and liquid (vaporization, vap.) Pt(II) compounds. The sublimation and vaporization enthalpies (ΔH_T*) and entropies (ΔS⁰_T*) are referred to as the average temperature (T*) of the interval measured (ΔT), p₀ = 760 Torr = 1 atm = 10⁵ Pa. The ligand abbreviations correspond to Table 1.

Ref.	Compounds	Process	ΔT, K	ln(p/p0) = A − B/(T,K)		ΔHT*, kJ·mol−1	ΔS0T*, J·(mol·K)−1
				A	B
[94]	Me2Pt(cod)	Subl.	343–353	17.2	5441	45.2	142.9
[82]	Pt(C7H13)2	Subl.	–	14.2	5520	44.1	118.2
[94]	Pt(hfac)2	Subl.	323–358	29.30	9320	77 ± 2	243.5
[95]		Subl.	377–419	21.9	10,520	83.6 ± 0.4	183
		Vap.	419–439	12.9	6670	52.8 ± 1.5	107.2
[85]	cis-Pt(tfac)2	Subl.	412–461	23.8	12,900	106.6 ± 0.4	196.46 ± 4.59
	trans-Pt(tfac)2	Subl.	437–506	23.14	13,210	109.93 ± 2.92	192.28 ± 5.43
[95]	Pt(thd)2	Vap.	447–467	16.2	9280	77.1 ± 0.7	134.3 ± 1.6
[85]	Pt(acac)2	Subl.	393–453	23.40	14,136	117.52 ± 1.38	117.52 ± 1.38
		Subl.	332–399	22.15	13,391	111.33 ± 0.81	184.18 ± 2.20
	Pt(i-tfac)2	Subl.	393–453	23.62	13,429	111.54 ± 4.17	196.20 ± 9.82

and Figure 13. The partial vapor pressure of Me₂Pt(MeCN)₂ measured only at 298 K was 3.3 × 10⁻⁶ Torr [75].

In general, the vapor pressure of organoplatinum compounds is several orders of magnitude higher than that of platinum complexes. The volatility of platinum complexes with organic ligands depends on the combination of terminal groups in the ligand and to a lesser extent on the combination of donor atoms in the coordination environment of platinum. Like iridium(III) β-diketonates, the vapor pressure of platinum(II) β-diketonates increases by more than 2–3 orders of magnitude when CF₃ or ^tBu groups are introduced into ligands [85]. For example, at 353 K, the vapor pressures of Pt(hfac)₂ and Pt(acac)₂ are 10⁻¹ and 10⁻⁴ Torr, respectively. In contrast, Pt(tfac)₂ and Pt(i-tfac)₂ complexes, which differ only in the combination of donor atoms in ligands ((O^O) vs. (O^NH)), have a very close volatility (at 393 K): 2 × 10⁻² vs. 1.5 × 10⁻² Torr, respectively [85].

The information on the decomposition of vapors of platinum(II) compounds is highly sporadic. According to [96], the pyrolysis of Me₂Pt(cod) vapors in a high vacuum (5 × 10⁻⁸ Torr) is accompanied by the loss of the CH₃∙ radical, which reacts actively with an organic substrate and recombines with PtMe. The introduction of a radical trap minimizes vapor recombination and promotes an increase in the growth rate of the nanomaterial. In the presence of hydrogen, Me₂Pt(cod) vapor decomposition is accompanied by the release of methane and cod ligands [15]. In an oxidative atmosphere, oxygen dissociates into adsorbed Me₂Pt(cod-R) precursors and binds to platinum, releasing methane. On the second step, the organic ligands are decomposed and platinum species are created [79]. Decomposition of adsorbed platinum compounds with ω-alkenyl ligands is accompanied by the α-elimination and decomplexation of the olefin groups from Pt, resulting in the formation of highly reactive platinum species that serve as nucleation initiators [82]. The ligand structure affects the nature of gaseous byproducts. In particular, the main gaseous product of vapor decomposition of Pt(C₇H₁₃)₂ is the hydrogenated ligand [82], while both the hydrogenated ligand and its dehydrogenated fragments (pentandiene isomers) are observed at decomposition of Pt(C₅H₉)₂ vapors [97]. The rapid desorption of most of byproducts in the presence of a reactive gas leads to the growth of active Pt surface-demonstrating autocatalytic behavior.

Unlike organoplatinum compounds, platinum forms strong bonds with ligands in the coordination compounds. For example, the binding energy of Pt–O_acac (180 kJ·mol⁻¹) is comparable with that of C–C bonds (140 kJ·mol⁻¹) [15]. This means that the decomposition of adsorbed vapors of Pt(acac)₂ occur without ligand decoordination, which causes a large amount of carbon in the films [15]. A similar conclusion has been formulated for Pt(hfac)₂ using the mass spectrometry data since the highly intense [(CO)Pt(hfac)]⁺ fragment was registered [94]. Moreover, the fluorosubstitution in the ligand of platinum β-diketonates leads to an increase of the vapor thermal stability. Specifically, the full decomposition of Pt(hfac)₂ vapors to metallic platinum and gaseous carbon oxides takes place only at 527 °C as it has been shown using temperature-programmed reaction spectroscopy. Note that the temperature of decomposition of Pt(hfac)₂ vapors can be reduced by using copper substrates, which is due to the formation of Cu-hfac fragments during the adsorption of Pt(hfac)₂ [98].

2.2.3. Platinum(IV) Precursors

The volatile platinum(IV) precursors are represented by derivatives of trimethylplatinum with cyclopentadienyl Me₃Pt(Cp^X) (Cp^X = Cp, Cp^Me, Cp^Et) or β-diketonate ligands Me₃Pt(L)(H₂O) (L = hfac, tfac, ptac, btfac), Me₃Pt(L)(Py) (Py = pyridine, L corresponds to Table 1).

The Me₃Pt(Cp^X) complexes with 50% yields are synthesized by exchange reactions using [Me₃PtI]₄ as an initial reagent [99]. An alternative “one-pot” technique based on the interaction of K₂PtCl₆ with CH₂Br₂ and MeLi with the subsequent addition of NaCp^X allows to obtain Me₃Pt(Cp^X) complexes with yields of 60% [15].

Two synthetic approaches to obtain Me₃Pt(L)A (A = H₂O, Py) complexes with β-diketonate ligands are known. The earlier one is similar to those described in [99] and based on the reaction of [Me₃PtI]₄ with potassium β-diketonates [100,101]. A shift in the equilibrium towards the target Me₃Pt(L)A product is achieved by AgF introducing, which leads to the formation of an AgI precipitate. Thus, the use of AgF makes it possible to obtain Me₃Pt(L)A complexes with yields up to 85%, whereas without this reagent, the yields do not exceed 50%. This approach was used mainly for the synthesis of Me₃Pt(L)A complexes with fluorinated β-diketonate ligands. An alternative technique based on the interaction of [Me₃PtI]₄ with lead β-diketonates with the formation of a PbI₂ precipitate is universal and allows obtaining Me₃Pt(L)Py complexes with yields up to 90% [102].

All Pt(IV) compounds under consideration are monomeric molecular complexes. According to single-crystal XRD data, platinum atoms in Me₃Pt(Cp^X) molecules lie above a point approximately located in the centers of the five-membered cyclopentadienyl rings [103] (Figure 14a). Since a β-diketonate ligand is only bidentate, a free coordination position remains in the “Me₃Pt(β-diketonato)” fragment. A donor molecule, namely water or pyridine, complements this position. In such adducts, the platinum cation is located in a distorted octahedral environment (Figure 14b,c).

Thermal properties of platinum(IV) precursors

Thermal properties of Me₃Pt(Cp^X) and Me₃Pt(L)A (A = H₂O, Py) complexes are studied in detail in Refs. [15,100,101,102] Cyclopentadienyl complexes Me₃Pt(Cp^X) are low-melting, and the introduction of an alkyl substituent X and an increase in its carbon chain is accompanied by a decrease in melting temperatures. For example, complexes with Cp^Me and Cp^Et melt at 30 °C and −78 °C, respectively [15]. The pyridine adducts of Me₃Pt(L)Py are characterized by the lower melting points (30–97 °C) compared to their Me₃Pt(L)H₂O analogues (>100 °C).

In contrast to Me₂Pt(Q) compounds, the most of Me₃Pt(Cp^X) and Me₃Pt(L)Py complexes pass into the gas phase almost quantitatively under the conditions of TG experiments (inert atmosphere). Only fluorine-free Me₃Pt(L)Py compounds partially decompose with a mass loss of 20%–30% when heated. The appearance of the steps corresponding to pyridine loss on TG curves of Me₃Pt(zif)Py and Me₃Pt(ttfa)Py complexes indicates that the processes of vaporization and decomposition of these complexes proceed in parallel. The thermal stability of Me₃Pt(L)A complexes with the same β-diketonate ligands deteriorates when the neutral Py ligand is replaced with H₂O [100,101].

The quantitative data on volatilization processes of several cyclopentadienyl complexes of the series Me₃Pt(Cp^X) and Me₃Pt(L)Py complexes and one Me₃Pt(hfac)(H₂O) complex that has the highest thermal stability are presented in

Table 6. Equations of the temperature dependences of saturated vapor pressure over the solid (sublimation, subl.) and liquid (vaporization, vap.) Pt(II) compounds. The sublimation and vaporization enthalpies (ΔH_T*) and entropies (ΔS⁰_T*) are referred to the average temperature (T*) of the interval measured (ΔT), p₀ = 760 Torr = 1 atm = 10⁵ Pa. The ligand abbreviations correspond to Table 1.

Ref.	Compounds	Process	ΔT, K	ln(p/p0) = A − B/(T,K)		ΔHT*, kJ·mol−1	ΔS0T*, J·(mol·K)−1
				A	B
[15]	Me3Pt(Cp)	Subl.	–	25.9	8590	71.4	215
	Me3Pt(CpMe)	Subl.	–	26.1	8690	72.21	217
[95]	Me3Pt(hfac)H2O	Subl.	338–365	32.78	14,103	117.2 ± 0.5	272.4 ± 1.5
[100]	Me3Pt(hfac)Py	Vap.	343–393	11.23	6787	56.4 ± 0.5	93.3 ± 1.5
	Me3Pt(tfac)Py	Vap.	348–393	18.51	10,060	83.6 ± 0.3	153.8 ± 0.8
	Me3Pt(thd)Py	Vap.	366–403	12.72	8147	67.7 ± 0.7	105.7 ± 1.8
[102]	Me3Pt(ptac)Py	Subl.	309–343	34.17	15,631	129.9 ± 2.5	284.2 ± 7.7
	Me3Pt(zis)Py	Subl.	326–353	32.79	15,896	132.1 ± 1.8	272.5 ± 5.3
[104]	Me3Pt(acac)Py	Subl. *	–	12407	23.46	103.1 ± 0.7	195.0 ± 1.3
		Vap.	390–414	9699	16.47	80.6 ± 0.5	136.9 ± 1.2

* Calculated from differential scanning calorimetry data.

. In general, the vapor pressure of complexes differs by three orders of magnitude in the transition from Me₃Pt(Cp^X) to Me₃Pt(L)A, with cyclopentadienyl complexes being the most volatile (Figure 15). For example, the vapor pressures for Me₃Pt(Cp), Me₃Pt(hfac)H₂O, Me₃Pt(hfac)Py, and Me₃Pt(acac)Py, measured at the same temperature of 323 K, are 5 × 10⁻¹, 1.5 × 10⁻², 6.5 × 10⁻², and 10⁻⁴ Torr, respectively.

The introduction of Me substituent into the cyclopentadienyl ligand leads to an approximately threefold increase in the vapor pressure of Me₃Pt(Cp^X) complexes: 0.1 and 0.044 Torr at 300 K for Me₃Pt(Cp^Me) and Me₃Pt(Cp), respectively. The saturated vapor pressure (in Torr) of Me₃Pt(L)Py complexes measured at 323 K decreases in the following order: 6.5 × 10⁻² (L = hfac) > 2.5 × 10⁻³ (L = tfac) > 2 × 10⁻³ (L = ptac) > 8 × 10⁻⁴ (L = thd) > 5 × 10⁻⁴ (L = acac) > 2 × 10⁻⁴ (L = zis). The obtained order is in a good correlation with TG data for Me₃Pt(L)Py complexes. Thus, the introduction of each CF₃ group into the ligand (L = hfac vs. tfac vs. acac) leads to the increase of the vapor pressure by about 0.5–1 order, while replacing one ^tBu group with (CH₃)₂OCH₃ (L = thd vs. zis) reduces the volatility by four times. Comparison of the p(T) data for Me₃Pt(L)Py with the corresponding platinum(II) β-diketonates, Pt(L)₂, indicates an increase in volatility by approximately two orders of magnitude, with an increase in the coordination number of Pt. For example, the vapor pressure of Me₃Pt(acac)Py at 373 K is 4 × 10⁻² Torr, while in the case of Pt(acac)₂ this value is equal to 5 × 10⁻⁴ Torr.

The study of thermal decomposition of platinum(IV) precursor vapors was carried out by several scientific groups [15,100,102,105,106] with the main focus on cyclopentadienyl complexes.

In the presence of hydrogen, the decomposition of Me₃Pt(Cp^X) vapors via an oxidative addition on the platinum followed by a reductive elimination of cyclopentane and methane [15]. According to IR data, decomposition of absorbed {Me₃Pt(Cp^Me)} complex molecules on Si substrate in the oxygen presence occurs through a formation of {Me₂Pt(Cp^Me)} adsorbed species that further oxidize forming {Pt} species and releasing of CO₂ and CH₄ gases [105]. According to [106], the most kinetically favorable elimination pathway of Me₃Pt(Cp^Me) vapors on the hydroxylated graphene surfaces also includes the formation of {Me₂Pt(Cp^Me)} adsorbed species when methyl interacts with hydrogen from the surface OH-groups with the release of methane.

Decomposition of vapors of Me₃Pt(L)Py (L = acac, ptac, zis, zif) in the presence of hydrogen on SiO₂/Si substrates is also accompanied by the release of methane (Figure 16). However, the most stable fragment in the gas phase recorded in all mass spectra of Me₃Pt(L)Py complexes is [Me₃Pt]⁺ [100,102]. This may indicate an increase in the strength of Pt–C_Me bonds in Me₃Pt(L)Py complexes in comparison with their cyclopentadienyl Me₃Pt(Cp^X) analogues. Indeed, according to single-crystal XRD data, the Pt–C_Me distances shorten by about 0.1 Å when moving from Me₃Pt(C_P^X) to Me₃Pt(L)A.

The main gaseous by-products of decomposition of Me₃Pt(L)Py vapors in an oxidative atmosphere are CO, CO₂, H₂O; the peak corresponding to [Me₃Pt]⁺ fragments are also detected in the mass spectra. In a recent work [102], the effect of terminal groups in β-diketonate ligands on the binding strength of platinum with neutral ligands in the series of Me₃Pt(L)Py complexes (L = hfac, tfac, ttfac, ptac, acac, thd, zis) has been examined. It has been shown that the introduction of each CF₃ group into L increased the Pt–N bond energy by ~2.5 kJ/mol.

2.2.4. Application of Platinum Precursors in MOCVD Processes

Since 2000, MOCVD processes of fabrication of Pt nanomaterials have been actively developing mainly in the direction of deposition of smooth layers for microelectronics and Pt nanoparticles for catalysis. Due to the peculiarities of the synthesis and thermal behavior of volatile platinum compounds, there are only a few numbers of platinum compounds used in MOCVD processes. The most widely used Pt precursors are Me₂Pt(Q), Pt(acac)₂, and Me₃Pt(Cp^Me), while the examples of application of other compounds, e.g., Pt(C₅H₉)₂, Pt(C₇H₁₃)₂, Pt(amN(Me)₂)₂, are sporadic (

Table 7. Deposition conditions and some characteristics of Pt nanoparticles (PtNPs) and films obtained by MOCVD from different classes of platinum precursors. The ligand abbreviations correspond to Table 1 ^a.

Ref.	Precursor	Reagent Gas	Source/Deposition Temperature, °C	Substrate	Thickness, nm	Growth Rate, nm/min	Composition (XRD)	Purity and Other Details
[94]	Me2Pt(cod)	H2	–/90–120	SiO2 + TEOS	–	–	Pt	FB-MOCVD, PtNPs with average d = 1–3 nm Pt0, 2.3–3.8 mass.%; C, 4 mass.%
[79]	Me2Pt(cod-Et)	O2	100/380	Aerosil®200 SiO2	–	–		CVS/MOCVD, PtNPs with average d = 2–2.2 nm Pt0, 4–5.7 mass.%; C, 4 mass.% PtO, up to 8.6 mass.%; PtO2 up to 0.7 mass.%
	Me2Pt(cod-Bu)		100/380
	Me2Pt(cod-iBu)		110/380
[77]	Me2Pt(hd)	H2	40/200–300	SiO2 + TEOS	100	–		Smoothed films (Ra~1 nm), 25 μΩ·cm or less
	Me2Pt(hd-Me)
	Me2Pt(hd-Me2)
[78]	Me2Pt(nbd)	N2 + O2	C = 0.1 M(tolulene)/300	Si/SiO2	60	2		LD-MOCVD, 1 mass.% C and O
[81]	Pt(C5H9)2	no	25/500	mesoporous silica	–	–		PtNPs with average d = 4.6 ± 1.1nm, Pt0, 2.27–4.95 mass.%,
[82]	Pt(C7H13)2	O2	75–80/250–330	Al2O3 SiO2, Si, Au	12 (PACVD)	–		PA-MOCVD (Ra~1.7–2.2 nm) or MOCVD 47–56% Pt, 38–40% C, and 5–13% O
					300 (MOCVD)
[90]	Pt(amN(Me)2)2	H2	80/400–500	TaN/Ta/Si	–	0.35		Smoothed films (Ra~1 nm), 13.6 μΩ·cm 9 mass.% C and 1 mass.% O
[91]	Pt(S/S-idpp)2 Pt(Se/Se-idpp)2	no	150/200–600	glass	250–300	–	PtS2 PtSe2	Large particles of PtS2 with d = 2–3 μm PtSe2 films with a worm-like morphology
[108]	Pt(acac)2	O2	180–220/260–380	Ti	200–1700	–	Pt	Electrode life up to 698 hrs
[109]		H2O	140/280–380	glass rods, TiO2	10–200	0.1–1.1	Pt or Pt-TiO2	100 nm Pt film behaves like a bulk Pt electrode
[110]		O2	C = 0.04 M(THF) 200/350	Si—trench structures	200–600	1.8	Pt	LD-MOCVD, Ra = 5 nm 27.6 μΩ·cm, 2 mass.% C
[111]		O2, H2O	140/260–300(O2) vs. 280–400(H2O)	Soda glass	–	0.7–2.4(O2) vs. 0.1–1.1(H2O)		110 vs. 91 kJ/mol activation energy
[112]		N2–O2, N2–O2-H2O, N2–H2	160–170/420	quartz CaF2	50–80 280–310	15		53–16 mass.% C
[117]		O2	130/300–440	sappfire	–	1.3–21.2	Pt-SiO2	11–18 mass.% C
[110]		O2	C = 0.06 M(THF) 70/350	Si—trench structures	12–140	2	Pt	(111)-oriented films, Ra = 15 nm 23.6 μΩ·cm, 6 mass.% C
[113]	Me3Pt(CpMe)	O2	35/200–400	Si	20–600	–		(111)-oriented films, >15 μΩ·cm
[118]		O2	35/150	TiO2, ZrO2	–	–		FB-MOCVD, PtNPs with average d = 2.1 nm Pt0, 3.4–5.7 mass.%; C, 4 mass.%
[119]		O2	48/350–400	TiO2/SiO2/Si, SiO2/Si	60–100	0.5		LD-MOCVD, Ra = 3.1–4.5 nm(TiO2/SiO2/Si), Ra = 4.9–7.6 nm(SiO2/Si), 8–10 mass.% C
[115]	Me3Pt(CpEt)	O2	C = 0.05 M(THF) 130/300–450	Si—trench structures, LiCoO2	–	–		LD-MOCVD, (111)-oriented films, Ra = 5.4 nm, 11.8–25 μΩ·cm
[100]	Me3Pt(acac)Py	H2/O2	85/280–310	Si(100), Ti electrodes	110–270 800–850	1.1(H2) 3.6 (O2)		15 mass.% C and 1 mass.% O (H2) 5 mass.% C and 1 mass.% O (O2)
[116]	Me3Pt(hfac)Py	H2	50/200–270	Si(100), Ti electrodes	–	0.7–3.2		UV-MOCVD, fractal Pt films were deposited at a high hydrogen concentration

^a XRD = powder X-Ray diffraction, C = concentration, THF = tetrahydrofurane, FB-MOCVD = fluidizes bed-MOCVD, LD-MOCVD = liquid delivery-MOCVD, PA-MOCVD = plasma-assisted-MOCVD, UV-MOCVD = ultraviolet stimulation-MOCVD, d = diameter, R_a = average roughness, TEOS = tetraethoxysilane.

).

A modern trend in the field of platinum MOCVD is the use of new modifications of the method aimed at increasing the efficiency of both the mass transfer of precursors to the reaction zone (fluidizes bed-MOCVD, FB-MOCVD, liquid delivery-MOCVD, LD-MOCVD) and the deposition process itself by means of additional activation (plasma-assisted-MOCVD, PA-MOCVD, ultraviolet stimulation-MOCVD, UV-MOCVD). The FB-MOCVD has been successfully used for one-step preparation of Pt-supported catalysis with an average size of Pt 1–5 nm and a narrow size distribution. In addition, utilization of FB-MOCVD installations allows the use of Me₂Pt(Q) precursors with a low thermal stability when heated. The LD-MOCVD was adapted to produce smooth Pt films on Si structures with trenches. PA-MOCVD is suitable for the preparation of smooth Pt films, since the use of plasma helps to reduce carbon impurities and achieve the lowest resistivity of the films [82].

Nowadays, platinum(II) acetylacetonate, Pt(acac)₂, is the most demanded MOCVD precursor used for the deposition of platinum nanomaterials for various purposes, including catalysis [107], microelectrodes [108,109], and microelectronic [110] applications. Its relevance is related to its ability to change the morphology of Pt films deposited from Pt(acac)₂ in a wide range depending on MOCVD parameters. In particular, according to the zone growth model, Pt films can be deposited with porous cauliflower (zone I, T_h < 0.3), columnar (zone II, 0.3 < T_h < 0.5), or equiaxed (zone III, T_h > 0.5) structures (T_h = T_deposition/m.p.(Pt), m.p.(Pt) = 2045 K) can be deposited (T_deposition—temperature of deposition process, m.p.—melting point) [111]. The features of Pt films growth depending on the deposition temperature, the type of gas reagent, and the substrate were studied in detail in several papers [111,112,113]. According to [112], the morphology of Pt films was mainly influenced by the nature of the growth surface rather than by the reaction atmosphere. The addition of H₂O promotes homogeneous growth of Pt films, while the addition of oxygen increases the growth rates of Pt films, which is accompanied by an increase in the level of impurities. In the low-pressure MOCVD process, depending on the deposition conditions, the impurity level of Pt samples deposited from Pt(acac)₂ can reach 50 at.% due to the high strength of Pt-O bonds in Pt(acac)₂. The possibility of deposition of Pt films with a 95% step of coverage coating in Si trench structures from Pt(acac)₂ using a LD-MOCVD method was reported in [110]. The Pt(acac)₂ precursor is also of interest for the deposition of PtO_x layers by ALD (atomic layer deposition) [114].

Note that in contrast to Pt(acac)₂ the use of Me₃Pt(Cp^Me) and its analogue Me₃Pt(Cp^Et) in LD-MOCVD processes leads to the growth of layers with a moderate coverage step of 35%–51% in Si trench structures [110,115]. This may be due to the fact that during LD-MOCVD, the growth of Pt layers from Me₃Pt(Cp^Me) and Me₃Pt(Cp^Et) precursors is limited by mass transfer. According to [82], Me₃Pt(Cp^Me) exhibits long nucleation delays in CVD, which leads to the formation of smoothed layers. This feature determines its demand for ALD applications [114]. Therefore, the precursors of this type are typically applied to produce smooth films with low resistance values, as well as Pt nanoparticles.

In the last decade, interest in the deposition of Pt layers from Me₃Pt(L)Py in oxidative and reducing atmospheres has increased markedly [100,116]. Comparison of data on the deposition of Pt coatings in hydrogen from Me₃Pt(L)Py (L = acac, hfac) indicates an increase in the stability of precursor vapors when fluorinated groups are introduced into the ligand, which makes it possible to achieve relatively high growth rates of Pt layers up to 3.2 nm/min when using Me₃Pt(L)Py. At the same time, activation of the gas phase (UV-MOCVD) allows to minimize the level of fluorine impurities, and to obtain Pt layers already at 200 °C. It was shown using Me₃Pt(acac)Py as an example that the replacement of the reagent gas from hydrogen to oxygen while maintaining the deposition temperature led to a significant decrease in the amount of carbon impurities and an approximately three-fold increase in the growth rate. It should be noted that for both Me₃Pt(L)Py precursors, the deposition temperatures sufficient for the formation of high-quality films are comparable, and in many cases even lower than those used to obtain platinum nanomaterials from more commonly used precursors. This fact, as well as the possibility of varying the thermal properties of Me₃Pt(L)A complexes over a wide range by directed modification of several types of ligands, encourages further development of MOCVD processes using precursors of this class.

3. Precursors for the Preparation of Bimetallic Platinum- and Iridium-Containing Materials by MOCVD

Bimetallic materials are of particular interest due to the mutual influence of components, which provides improved characteristics compared to monometallic analogues (synergetic effect). In MOCVD processes, bimetallic systems can be obtained using a single heterometallic precursor or a combination of individual precursors of each component. The advantage of the first approach is that the metals are already mixed at the molecular level. However, a given ratio of metals limits the possibility of varying the composition of the resulting material. The second approach does not have this problem; however, the success of its implementation depends entirely on the chosen combination of precursors. Such precursors must be compatible in terms of thermochemical properties, which primarily means that there is a common temperature range for the decomposition of compound vapors and the absence of ligand exchange reactions that complicate the process control.

In general, studies dealing with the preparation of bimetallic materials containing platinum and iridium by the MOCVD method are sporadic, which is mainly due to the precursor factor. At the same time, this direction is more developed for platinum than for iridium. In particular, no iridium compounds, which can be used as a heterometallic precursor in gas-phase processes, have yet been described. For platinum, such a complex was obtained using ruthenium cyclopentadienyl, modified so as to be able to coordinate the second metal atom: [CpRu(η⁵-C₅H₃CH₂NMe₂)Pt(hfac)] [120]. This precursor was successfully tested for the deposition of bimetallic PtRu films in the temperature range up to 400 °C in an oxygen atmosphere; however, synthetic difficulties limit the use of the precursors of such types.

For this reason, co-deposition from combinations of individual precursors is a priority approach to the formation of Pt–M and Ir–M systems by the MOCVD method. The literature analysis shows that homoligand metal acetylacetonates are mainly used for this purpose. In particular, PtRu and PtCo nanoparticles [107], as well as thin films containing PtNi, Pt₃Ni, PtCo, Pt₃Co and PtCoNi phases [121], were obtained from such combinations.

However, a simple combination of Pt(acac)₂ and Ir(acac)₃ does not look optimal for the preparation of Pt–Ir metal coatings that are most interesting for medical applications. In fact, in the only example of the usage of this precursor combination, the deposition of bimetallic nanoparticles was carried out at 400 °C without an additional gas-reagent and without control of the purity and composition of the resulting product [122]. The introduction of a gas-reagent changes the picture dramatically. In the presence of hydrogen, [Pt(acac)₂] vapors are less thermally stable than Ir(acac)₃, and this difference exceeds 200 °C (see Section 2). In the presence of oxygen, a region of possible co-deposition exists; however, the process is complicated by the tendency to form IrO₂ from Ir(acac)₃. For this reason, for the preparation of metallic Pt_xIr_1–x (x = 0.5–0.9) coatings in an oxidizing environment, we proposed an alternative combination of the precursors, viz., [Pt(acac)₂] and [Ir(cod)(acac)], that seemed to be more convincing [123].

Typically, co-deposition is performed using independent heated sources in a MOCVD setup for each precursor. However, a single-source delivery system for the vapor transportation regardless of the number of precursors used is much more preferable from a technological point of view, since it allows for reducing the number of MOCVD parameters to be controlled and simplifying the reactor design. The single-source delivery system implies the usage of precursor mixtures. In this case, the processes of volatilization of precursors become crucial, since there is always the possibility of mutual influence of thermal properties on each other in the mixture, especially if the volatility of the precursors differs significantly [124]. The methods of thermogravimetry (TG) and XRD were used to study mixtures of [Ir(cod)(acac)] and [Pt(acac)₂] precursors. The precursor mixtures were prepared in two ways: (i) by mechanical mixing of the necessary substance samples (mech) and (ii) by dissolving the substance samples in chloroform and evaporating the solution in an inert medium (solv). The second method was used to detect the effect of co-crystallization. To identify the dynamics of changes in physical and chemical properties, the ratio of [Pt(acac)₂] and [Ir(cod)(acac)] components in the mixtures varied as 9:1 and 1:1 (ratios similar to those in the obtained films) and 1:9. Comparison of the diffraction patterns of all the investigated mixtures with different ratios of components (Figure 17) shows that, regardless of the method of preparation, they are a superposition of phases of two individual compounds. This fact is clearly expressed for the mixtures prepared by mechanical mixing (mech). In the case of using a solvent while maintaining the additivity of the peaks on the diffraction patters, a reproducible peak at 2θ = 12.2° is observed (see Figure 17, dotted), the appearance of which is apparently due to partial decomposition of the iridium complex.

TG curves (all experiments were performed under identical conditions) describing the processes of volatilization of individual compounds [Ir(cod)(acac)] and [Pt(acac)₂] have a single-stage appearance; the temperature difference corresponding to 50% mass loss of precursors is 30 °C (Figure 18). The platinum precursor sublimes completely (<0.3 mass.% residue), while the iridium complex passes into the gaseous phase with partial decomposition (3.4 mass.% residue).

TG curves of the mixtures with the ratio of components 9:1 and 1:1 (both mech and solv) have a two-stage form, in which the first stage completely reproduces the volatilization of the iridium compound, while the second stage corresponds to that of the platinum precursor. However, at the second stage, the transition to the gas phase is accelerated, and the mixture completely passes into the gas phase at temperatures lower by 10 °C (9:1) and 20 °C (1:1). Herewith, the mixture enriched with the iridium component (1:9) completely reproduces the TG curve of the individual compound [Ir(cod)(acac)]. These data may indicate an increase in the volatility of the platinum component with an increase in the content of the iridium complex in the mixture. The use of the solution method (solv) for the mixture preparation does not affect the speed of the mixture volatilization process; however, it significantly reduces the thermal stability of the system.

Summing up the study of mixtures, one could conclude that the preparation method does not result in forming any solid solutions and, consequently, affect volatilization processes of the binary system. However, it significantly changes the mixture stability: a mutual effect of reducing the thermal stability of substances in the mixture is observed. When the mixtures are heated, a synergistic effect is observed, which is manifested in an increase in the rate of volatilization of the less volatile component [Pt(acac)₂] with an increase in the content of the more volatile [Ir(cod)(acac)]. The effect is insignificant for the mixture [Pt(acac)₂]-[Ir(cod)(acac)] 9:1 and pronounced for the mixture of the inverse proportion of 1:9. The obtained results are very important and should be taken into account when choosing the temperature conditions of a single-channel source of MOCVD setup when depositing multicomponent layers, as well as when doping the main composition of the coating. The use of [Pt(acac)₂]-[Ir(cod)(acac)] mixture in the ratio of 9:1 and 1:1 most likely leads to coatings with a gradient distribution of metal elements in depth. The use of the [Pt(acac)₂]-[Ir(cod)(acac)] mixture in the ratio of 1:9 may be quite justified, because the volatilization process appears to take place with a preserved ratio of components.

4. MOCVD of Platinum and Iridium for Medical Application

When discussing the use of a MOCVD method for deposition of specific functional film structures, it is necessary to take into account both the characteristics of the material to be coated and the features that the target coating should have. A specific requirement for the coating material that is used for medical purposes is a developed topography, i.e., a large effective surface area. In fact, high-roughness surfaces have been proven to be the most effective for promoting osseointegration of medical implants [126,127]. As for the electrodes for cardio and neurological devices, an increase in the area of the electrochemically active surface gives both a higher capacitance and a lower polarization, which are required to optimize pacing and sensing parameters, respectively [7,8].

The prospects of the application of the MOCVD technique to fabricate high-surface-area noble metal coatings of various compositions on the pole tips of medical electrodes have been demonstrated in a series of our recent works [100,123,128,129,130]. In particular, metallic iridium coatings were produced using a conventional precursor Ir(acac)₃ in a reducing atmosphere [128]. These samples were further employed as templates for the electrochemical preparation of activated iridium oxide films (AIROFs). The resulting AIROFs exhibited charge storage capacity values higher than those of AIROFs obtained from electron-beam evaporated iridium films and were comparable to commercial electrodes. Metallic platinum coatings with an enhanced surface area and, consequently, high capacitance characteristics and low impedance were obtained on endocardial and diagnostic electrodes made of titanium and stainless steel, respectively, using Pt(II) β-diketonate, Pt(acac)₂ [129,130], and the related Pt(IV) complex Me₃Pt(acac)Py [100]. It was shown that the roughness of coatings formed in an oxidizing atmosphere was higher than that of the coatings deposited in reducing conditions [100]. Bimetallic platinum-enriched coatings Pt_xIr_(1−x) in a wide range of metal ratios (x = 0.5–0.9) were successfully obtained onto the endocardial electrodes using a combination of Pt(acac)₂ and Ir(cod)(acac) precursors and oxygen as a gas-reagent [123]. The surface roughness was found to raise with an increase in the iridium content. The activated Pt_xIr_(1−x) coatings formed by MOCVD were comparable or better than those obtained by electrodeposition and magnetron sputtering.

One of the current trends in the development of medical devices and implants is the transition from titanium and metal alloys to new materials based on carbon fibers, carbon composites, and polymers. These materials are characterized by a relatively low thermal stability and/or sensitivity to oxygen. Recent work suggests strategies for the successful deposition of noble metals also on such carriers [116,130]. In particular, deposition, a special sub-layer in a reducing atmosphere, can be effective for protection against oxygen. We have shown that the use of an adhesive Pd layer protects the stainless steel substrate during further deposition of platinum from Pt(acac)₂ in an oxidizing atmosphere without reducing the capacitance characteristics of the resulting platinum layers [130]. Another approach is to develop low-temperature deposition processes under moderate conditions by selecting special precursors and/or using additional activation. In particular, it has recently been shown that Pt layers with a high active surface area could be obtained in a reducing atmosphere already at 200–250 °C using Me₃Pt(hfac)Py under vacuum ultraviolet irradiation [117].

With regard to permanent medical products, i.e., those that imply a long-time stay in the human body, such as implants reconstructing body parts and pacemakers, the biocompatibility of the introduced material becomes a particularly serious aspect. The application of film materials from the considered noble metals can be useful to improve this characteristic. In fact, Pt, Ir, IrO_x, Pt–Ir nanomaterials obtained by different methods have been proven to be perfectly biocompatible with various cell colonies [131,132,133]. This phenomenon is traditionally associated with the inertia and non-cytotoxicity of these materials. Recent reports [131,132,133] also emphasize the positive impact of noble metal surfaces and nanoparticles on the processes of growth, integration, and differentiation of some human cells, including the effects induced by specific binding. For example, the great affinity of IrO_x to proteoglycans stimulates a direct growth of neural cells for network formation [131]. Although the noble metals have not yet been deposited on the surface of medical implants by MOCVD, the capability of this method to obtain pure nanomaterials (see Section 2.1.3 and Section 2.2.4) guarantees the absence of any problems with the biocompatibility of such coatings. On the other hand, a number of recent results in the field of implantation materials, obtained by various research groups, also indicate that a MOCVD method is promising for the preparation of biologically compatible coatings with improved characteristics [134,135,136,137,138].

Finally, it should be noted that for medical permanent implants, it is extremely important to impart antibacterial properties to the surface in order to prevent infectious complications [139]. In this aspect, in our opinion, there are two promising areas of application for the MOCVD method. On the one hand, the introduction of iridium and platinum improves bactericidal properties of silver and gold traditionally used for this purpose. MOCVD provides a convenient single process chain for the preparation of such mixed film materials with a developed surface, because both gold and silver are also metals, which can be deposited by this method [140,141], and a common deposition temperature can be achieved by selecting combinations of volatile precursors. On the other hand, mono- and bimetallic platinum nanoparticles, PtNPs [142], AgPtNPs [133], and AuPtNPs [143], are now considered as alternative antibacterial agents with improved biocompatibility. Here, MOCVD allows us to obtain the nanoparticles without impurities that are typical for solution methods. This can also help to better understand the principle of the antibacterial action of PtNPs, which remains unproven [144]. The active development of the MOCVD method for the formation of PtNPs demonstrated in Section 2.2.4 creates the basis for successful research in this direction.

5. Conclusions

The interest in platinum group metals, in particular iridium and platinum, has not disappeared for many years due to such unique properties as exceptional chemical inertia, high biological compatibility, mechanical strength, and corrosion resistance, which are the basis for their application in medical practice. MOCVD is a promising method to fabricate Ir and Pt nanoparticles and film materials, multilayers and heterostructures. Its advantages include precise control of the composition and microstructures of the formed materials in deposition processes at relatively low temperatures on a wide range of the substrates including non-planar ones. The development of MOCVD processes is inextricably linked with the development of the chemistry of volatile precursors, viz., specially designed coordination and organometallic compounds.

This review describes the main classes of volatile precursors of iridium and platinum iridium in all the oxidation state, namely, Ir(I), Ir(III), Pt(0), Pt(II), and Pt(IV). The synthesis methods are briefly discussed and the most efficient approaches are presented. The detailed data on thermal properties of the compounds are collected here and are useful for choosing the temperature regimes of MOCVD processes. The effects of a ligand structure and a metal coordination environment on the precursor thermal behavior are demonstrated. A number of examples of the usage of various volatile precursors in MOCVD including the formation of Pt, Ir and IrO₂ films, Pt nanoparticles, and bimetallic nanomaterials are presented.

A comparative analysis of the precursor chemistry allows us to draw the following conclusion. Although metal acetylacetonates, Ir(acac)₃ and Pt(acac)₂ are currently the most widely used precursors, Ir(I) and Pt(IV) complexes seem to be more promising due to a relative simplicity of preparation and their higher volatility, while the films from these precursors can be deposited at comparable or lower temperatures. The main advantage of these precursors is their ability to control thermochemical properties by modifying several types of ligands. This makes them tunable both to deposit films onto different substrate materials having specific requirements to deposition conditions and to select a combination of compatible compounds for obtaining the bimetallic materials.

Finally, the prospects for using the MOCVD approach to obtain medical Ir- and Pt-based nanomaterials are discussed through the prism of recent works and current trends in this area.