Diarylplatinum(II) scaffolds for kinetic and mechanistic studies on the formation of platinacycles via an oxidative addition/reductive elimination/oxidative addition sequence

Oxidative addition and reductive elimination reactions are fundamental steps in processes related to synthetic chemistry involving organometallic compounds. In these reactions a metal in two available oxidation states (generally differing in two units) is needed, platinum centres being a very good example. The relative inertness of diamagnetic Pt and Pt organometallic species (having respectively d square-planar or d octahedral arrangements), enables an easy monitoring of time-resolved reactivity, including its posterior kinetic analysis. Specifically, imine ligands containing C-X bonds have been observed to oxidatively add to {Pt(Aryl)2} moieties, which sequentially undergo C-C reductive elimination and C-H bond activation on the new ligand formed. These new species have been found to contain mostly seven-membered metallacycles, despite the obvious thermodynamic preference for five-membered cycles, which are found only in some rather specific instances. The kinetic preference of the complexes obtained has been studied from a kinetico-mechanistic perspective, that included obtaining thermal and pressure derived activation parameters, and a dramatic influence on the spectator halido ligands and the substituents on the aryl groups has been established. To complete this kinetic and mechanism (kinetico-mechanistic) study, theoretical calculations have also been conducted to model the data collected and propose both the elementary steps and the factors determining the specificity of the full process.


Introduction
Kinetics experiments in solution lead to the collection of kinetic data that can be analyzed in terms of rate laws, which together with activation parameters acquired from temperature and pressure variable experiments, may lead to mechanistic proposals. The value of such analyses and proposals is compromised unless due diligence regarding reactant and solvent purity is followed, and furthermore reproducibility of measurements is assured, and primary data and derived secondary data and parameters are subject to appropriate error analysis.
In the last several decades the throughput of kinetics experiments has changed dramatically by instrument improvement, automation and validated data analysis software. Developments in chemical synthesis have yielded a wider range of subtle reactant variations leading to wider data sets. Consequently, provided the caveats within the experimental approach are followed, reliable new reaction mechanisms frequently ensue.
In order to understand further the detail of reaction mechanism pathways, the experimentalists have, over the last two to three decades, at their disposal, the ability to take advantage of the application of density functional theory (DFT) computation methodology. This methodology represents a significant and appealing addition to the overall repertoire of approaches by investigators for a detailed explanation of the time course, energetics and structural aspects of their chemical reaction systems. Successes in using this computational approach are widely reported. Indeed it will be shown in this contribution that a combination of kinetics and mechanism studies and modeling by DFT computations has led to a comprehensive understanding of the formation of platinacycles through various stages starting from diarylplatinum(II) scaffolds.
Oxidative addition and reductive elimination are fundamental reactions in organometallic chemistry in both stoichiometric and catalytic processes. The inertness of platinum compounds and the availability of different oxidation states make them suitable for mechanistic studies of these reactions. In particular, although cyclopalladated complexes are much used in catalytic processes, their equivalent platinum compounds have been often used as model compounds for fundamental reactions due to their high stability and inertness. The syntheses of cyclometallated compounds traditionally involves intramolecular C-H bond activation, but a great deal of interest is also focused in intramolecular C-X bond activation since, in these cases, the initial oxidative addition can be followed by a subsequent reductive elimination producing new σ bonds such as C-C, C-O, C-H or C-X. Among the large plethora of cycloplatinated compounds available, those containing aryl ligands might give rise to oxidative addition/reductive elimination process involving C-X and C-H bond activation as well as C aryl -C aryl reductive elimination. Therefore arylplatinum compounds provide a useful platform for a thorough study of these fundamental processes.
The relative inert character of such complexes allows the monitoring of their timedependent reactivity in an easy and reproducible way. Furthermore, the possibility of detecting and isolating reaction intermediates precisely increases with the increase of the inert character on the reactivity involved. Unfortunately the possible actuation of dead-end processes also increases along the same line, and a comprehensive monitoring off all the processes involved in the reactivity studied is desirable. In this respect, the use of experimental mechanistic procedures has been suffering a decrease in use owing to the preference by some for computational methods, which usually come into play when the situation becomes too complicated to monitor experimentally, or when the number of experiments needed is excessive. However, we have to keep in mind that computational methods have to be employed wisely; although calculations may corroborate or even guide experiments, the final conclusions have always to be ascertained by proper real experimental situations.
Platinum chemistry is a terrain that has been explored thoroughly by computational means. A tremendous number of publications regarding homogeneous catalysis, surface science or medicinal applications are reported each year. Bond activation by platinum complexes, and its implication in homogeneous catalysts, has been explored thoroughly.
In particular cycloplatination reactions are a convenient computational area of study.
Association and dissociation steps, along with well-known oxidative addition and reductive elimination processes between closed-shell diamagnetic Pt II /Pt IV species, are involved in the reactivity. Furthermore, relativistic effects for platinum, which should be the most remarkable issue in computational studies of these reactions, can be effectively tackled nowadays by most DFT methods. Either the use of scalar relativistic calculations, or basis sets with pseudopotentials (that include the corresponding corrections for the heavier atoms), represent a standard solution to the issue. This chapter describes a series of experimental, kinetics, mechanistic and computational studies on the formation of five-and seven-membered metallacycles obtained from diarylplatinum(II) precursors with N-donor ligands. One aim of the material presented is to illustrate the enormous advantage of synergic collaborations between specialized research groups that adds great value to the studies conducted. However, by doing so, some aspects, which are normally taken for granted, have to be discussed and reformulated to account for the overall set of observations and results.

Compounds
Although the complexes [PtMe 2 (NN)], where NN is a diimine ligand, such as 2,2′bipyridine or 1,10-phenanthroline, are among the most reactive transition-metal complexes in intermolecular oxidative addition of alkyl halides, aryl halides fail to react with these Pt II complexes. In this respect, the first reported oxidative additions of directed aryl halides to Pt II have been achieved using ligands of formula RCH=NCH 2 CH 2 NMe 2 . [1][2][3] In these, R is an ortho-halogenoaryl group, and the ligand coordinates via both nitrogen atoms to a dimethylplatinum(II) center producing a [PtMe 2 (RCH=NCH 2 CH 2 NMe 2 )] species that ultimately leads to intramolecular oxidative addition of the aryl-halogen bond. This strategy was successful in producing C-X (X = Br, Cl, F) bond activation leading to Pt IV compounds of type [PtMe 2 X(C 5 CH 4 CH=NCH 2 CH 2 NMe 2 )]. In the same studies C-H bond activation was also achieved; in this case a final Me-H reductive elimination producing methane and Pt II compounds of type [PtMe(C 5 CH 4 CH=NCH 2 CH 2 NMe 2 )] takes place (Scheme 1, top).
Further work in this area involved the use of ligands with the same type of organometallic moieties, but with a single N-donor directing group, with a general formula 2-XC 6 H 4 CH=NCH 2 -2′-X′C 6 H 4 . [4][5][6] In this case, the reaction of these ligands with [Pt 2 Me 4 (µ-SMe 2 ) 2 ] may produce different types of five-membered platinacycles, either containing (endocycles) or not (exocycles) the imine functionality. As for the amino-imino ligands indicated above, intramolecular oxidative addition of C-X bonds (X = Br, Cl, F) produced cyclometallated Pt IV compounds, while activation of C-H bonds with loss of methane gave the corresponding cyclometallated Pt II compounds.
Interestingly, formation of endo-platinacycles (Scheme 1, bottom) is favored in all cases for these ligands. Exo-platinacycles have only been found when imines 2,4,6- single N-donor atom proved that not only initial C-Br bond activation but also C-Cl bond activation may lead to such seven-membered platinacycles as indicated in Scheme 2. 11,12 It is interesting to note that these compounds are obtained in good yields not only when the ortho position of the imine is blocked with a fluorine substituent (Y = F) but also when Y = H, where a C-H bond activation leading to more stable five-membered Pt II metallacycles is also possible.  Interestingly, the reaction of cis-[Pt(C 6 F 5 ) 2 (SEt 2 ) 2 ] with imine 2-BrC 6 H 4 CH=NCH 2 (4-ClC 6 H 4 ) (Scheme 3) produced exclusively a five-membered metallacycle with an exo C aryl -C aryl bond. In this case, formation of a seven-membered platinacycle which requires C-F bond activation is not favored and indeed is not observed. 19 F NMR monitoring allows the detection of all proposed intermediates, 15 i.e. initial formation of a Pt IV compound arising from the activation of the C-Br bond of the imine, C aryl -C aryl reductive elimination and final C aryl -H bond activation leading to a five-membered cyclometallated Pt II compound with concurrent elimination of pentafluorobenzene. involve intramolecular C-X (X = Br or Cl) bond activation followed by C aryl -C aryl reductive elimination to produce non-cyclometallated intermediates that could be detected when the reactions were monitored by 1 H or 19 F NMR spectroscopy. 13,15 However, none of these intermediates could be isolated and characterized crystallographically. Nevertheless, ligands containing two nitrogen donor atoms such as  Figure 2).

SCHEME 3 HERE
Moreover antitumor properties have been studied for some of them. 14 systems, these reactions have been thoroughly studied. 18 Careful selection of the reaction conditions extracted from the kinetic studies indicated in the following section allowed the detection and characterization, including by X-ray crystallography, of two isomers of the non-cyclometallated Pt II compound arising from C aryl -C aryl reductive elimination, thus supporting the existence of the reaction intermediates assumed for the reactivity with [C,N] systems.
In this respect and in an attempt to obtain a seven-membered [C,N,N′]-Pt II cyclometallated compound for X = Br, the reaction of [Pt 2 (4-MeC 6 H 4 ) 4 (µ-SEt 2 ) 2 ] with imine 2-Br,6-FC 6 H 3 CH=NCH 2 CH 2 NMe 2 has also been studied 11 and, effectively, the presence of an inert C-F bond at the ortho position of the aryl ring of the imine ligand, prevents the C-H activation at the imine and thus, the reaction is driven towards the formation of a seven-membered platinacycle containing an internal biaryl linkage (Scheme 5). Again, for this system two isomers of the non-cyclometallated compound containing a biaryl ligand were isolated and crystallographically characterized.

Kinetic studies
Reactions involving platinum organometallic complexes with simple cis-{Pt II R 2 } units ( R = Me, Ph) have been proved to be extremely well-behaved for the directed oxidative addition reaction of C-X or C-H bonds indicated in the previous section (Scheme 1). [4][5][6]9,20,21 This has allowed us to study from a kinetico-mechanistic perspective the reaction with the ligands indicated in Scheme 6 having monofunctional (imine) or bifunctional (amino-imine) directing units. 22 SCHEME 6 HERE Scheme 6. Ligands utilised in the preliminary studies of oxidative addition reactions on cis-{Pt II R 2 } units.
Independently of the directing groups used, the process has been found to be occurring via the initial formation of an unsaturated tri-coordinated species that reacts in a concerted oxidative addition fashion to produce a penta-coordinated intermediate. 23,24 The rapid coordination of the sixth (dangling or available in the medium) ligand produces the final compound (Scheme 1). Figure 3a collects the isokinetic plot generated with all the data available, indicating that the process occurs indeed via the same concerted tri-centered mechanism. 22 Even the available data for systems lacking hydrogen bonding characteristics show a very good and extended enough ΔV ‡ /ΔS ‡ correlation 25 (Figure 3b). and diverse behavior. Although, in some cases well-behaved two consecutive kinetic steps were observed, leading initially to complexes of type A and finally to species of type 7C, 13 in some other cases the kinetic profiles showed up to four recognizable timeresolved steps. 11,18 Parallel NMR time-resolved monitoring was thus needed in order to ascertain what are the processes observed. Furthermore, the nature of the final platinum species formed was also found to be extraordinarily dependent on a plethora of variables.

FIGURES 3a AND 3b HERE
The general reaction scheme that could be generalized from the data collected is already shown in Scheme 2, where the first step, i), corresponds to the process indicated in Scheme 1 for X ≠ H (formation of compounds of type A). 22 The steps following this first oxidative addition reaction, ii) and iii), correspond to a reductive elimination coupling (formation of compound of type B), 15 and a new oxidative addition process of the biaryl ligand formed (formation of compound of type C). 11,13,18 With these data at hand, and given the good time-profile knowledge acquired for the formation of complexes of type A, isolation of such a Pt IV five-membered metallacycle has been possible in most of the cases. Consequently the ii)+iii) set of reactions (producing compounds of type B and C) could be studied independently. Interestingly, the careful time-resolved isolation/characterization of the species involved in the process lead to the obtention of rather diverse reaction intermediates and products, which corresponds to the more detailed reaction sequence indicated in Scheme 5. As a whole, the species

Spontaneous processes with monodentate N (imine) ligands
The serendipitous seminal kinetico-mechanistic work of this reactivity series has been successfully concentration-tuned to evaluate in a separate manner, reductive elimination coupling and final cycloplatination reactions (Scheme 2). 7 As a follow up from these studies the kinetico-mechanistic monitoring of the reaction of the cis-{Pt II (C 6 F 5 ) 2 } unit with ligand 6, where R = H and R′ = 4-Cl, was also pursued (Scheme 6). 15 The process shows a neat two step rate-limiting sequence perfectly associated with reactions ii) + iii) in Scheme 2. That is, reaction Scheme 9 (with L = SMe 2 , Aryl = C 6 F 5 , X = Br) applies to this process which modifies the reactivity indicated above. In this case the final compound corresponds to a five-membered platinacycle with a dangling C 6 F 5 unit of type 5C. Interestingly, the measured kinetics indicates that process ii) is very fast and the initial oxidative addition of ligand 6 (with R = H and R' = 4-Cl) on cis-[Pt II (C 6 F 5 ) 2 (SEt 2 ) 2 ] (step i) in Scheme 2) cannot be resolved; i.e. the compound of type A is not observed, only species B is initially present as an intermediate. Table 1 collects the relevant kinetic and activation parameters collected for the resolved reaction in Scheme 9 (B " 5C).

SCHEME 9 HERE
Scheme 9. Modification of the reactivity shown in Scheme 8 for the L = SEt 2 , Aryl = C 6 F 5 , X = Br system producing a five-membered final platinacycle.
Further kinetico-mechanistic studies have also been carried out on complexes with a systematic variation of the Pt II -attached aryl ligands in the starting cis-{Pt II (Aryl) 2 } moiety. Furthermore, an ortho substituent on the initial metallated ligand has been introduced (ligand 6 in Scheme 6, with R = 2-F and R′ = H). 13 Figure 4 shows clear indication of the diverse trends observed for these systems.

Spontaneous processes with chelate N (amino) -N (imine) ligands
In As indicated in Table 1, for the L, N (imine) systems isolation of the final compound as a compounds of type B that only produce final cyclometallated complexes of type 7C as found for the systems indicated before. Nevertheless, for the ortho unsubstituted complexes of type A the formation of the more thermodynamically stable complexes of type 5C is also observed in some cases. This fact indicates that for these chelate complexes the formation of seven-membered metallacycles of type 7C can be associated, at least in part, with the presence of a C-Y bond that is too strong to produce such a cyclometalation reaction.
The summary of this chemistry is indicated in Scheme 5 from the previous section.
Scheme 12 collects the full series of species of type 5C and 7C encountered as the final compounds produced from the spontaneous reductive elimination/oxidative addition sequence occurring on complexes of type A. The kinetic study of the process occurring on compounds of type A proved to be much more complex than expected, with a very diverse behavior, 11,18 in line with that observed for the complexes in the last entries of Table 1. Figure 5 shows the Eyring plot of all the time-resolved reaction steps observed for the spontaneous reactivity of the bromo complex of type A with no chloro or fluoro substituents on the initial metallated ligand.

SCHEME 12 HERE
Scheme 12. Series of species of type 5C and 7C encountered as the final compounds produced from the spontaneous reductive elimination/oxidative addition sequence occurring on the complexes of type A indicated.
FIGURE 5a AND 5b HERE From the plot it is evident that at least three steps can be resolved by using carefully tuned conditions. Furthermore, careful choice of time/temperature conditions allowed for the isolation of some of the species appearing during the process. As indicated in Scheme 11 reductively coupled intermediates have been isolated and fully characterized ( Figure 6).  Table 2 collects the relevant associated data. Nevertheless for some of the systems studied the isomerization reaction has not been observed, thus indicating that the process is fast under the conditions studied, as already observed for some of the systems indicated in Table 1. As a whole, when the processes on the chelate N (amino) -N (imine) and monodentate N (imine) -L systems are compared, both appear to be extremely complex and depending on the large number of variables involved. This is not surprising given the number of reaction steps involved in the full reactivity, which makes any claim such as faster than or slower than meaningless. Furthermore, the characterization of the intermediate species during the full process also brings in some news facts that have to be considered with respect to the isolation of only the lessreactive species of the full set.

FIGURE 6 HERE
Nevertheless, the most striking feature of the reactivity indicated in the previous pages relates to the five-or seven-membered platinacyclic nature of the final complex obtained. As seen in Table 2    and coordination chemistry studies. [33][34][35][36][37] The synergy between experiment and theory has also evolved considerably, allowing the development of both fields and producing some outcomes that would be hardly achievable using any of the single techniques. In this respect, the use of DFT calculations began by supporting results in a post-experimental manner. i.e. determining reaction mechanisms and selectivity when the rational chemical intuition and mechanistic experimental techniques have been exhausted. More recently, however, theory computations have begun to be used in parallel to experiments, or, even prior to them in some cases. By doing so, they have led, in some cases, to the discovery of new reactions and chemical systems. [38][39][40] We have employed DFT theoretical calculations to ascertain some relevant aspects of the cycloplatination reactions explored from a kinetico-mechanistic perspective indicated in the previous sections. The aim has been to clarify some of the relative chemical stabilities, isomerization processes and kinetics observed. 11 Although the studies related to the Pt II complexes having N (amino) -N (imine) ligands have been reported, some new calculations related to monodentate N (imine) ligand systems have also been carried out and are reported here for the first time. The computational methodology used is equivalent to that utilized before, 11 and frequency calculations were carried out to confirm stationary points and transition states. Additional single point calculations on the optimized geometries were also employed to obtain improved solvated free energy values with larger basis sets at the corresponding reaction temperatures; unless otherwise stated all the free energy values reported correspond to those obtained with these larger sets. As for the concentration/time kinetic models, they have been built using Copasi software 41 using the deterministic (LSODA) method with relative and absolute tolerance values of 10 -6 and 10 -12 , respectively.

Five-versus seven-membered platinacycle stability
As mentioned in the previous sections, the final five-membered platinacycle compounds, 5C, are obviously expected to be more thermodynamically stable than the alternative seven-membered, 7C, analogues. This fact has been confirmed in a facile manner using DFT calculations. The possible 5C and 7C products have been computed for different starting Pt II complexes, A (even including some that have not been experimentally studied); the free energies of reaction (ΔG R ) thus derived are shown in Table 3. Clearly the five-membered platinacycles (5C) are the more stable in all cases, in agreement with what is expected for this kind of motif. Consequently, the obtention of the larger seven-membered platinacycles, from the isolated Bcis intermediates indicated before, has to be due to kinetic preferences (see below). Furthermore, the calculated free energies for compounds 5C and 7C also confirm the expected larger stabilization of the N (amino) -N (imine) chelated Pt II species; in most cases these are found 5-20 kJ mol -1 more stable than their corresponding monodentate N (imine) analogues. Entries 9 and 10 do not include the calculated value for the corresponding 5C complex given its non-feasibility due to the blocking indicated in Schemes 2, 5, and 12.

cis versus trans stability for B-type intermediates
As stated in the previous sections, the Btrans ⇄ Bcis interconversion plays an important role in the reactions studied, 11,18 in some cases the process even becomes the ratedetermining step of the general cycloplatination reaction. 13 The   In practice this isomerization process has been comprehensively computed only for compounds of type A with chelating N (amino) -N (imine) ligands with Aryl = 4-MeC 6 H 4 and X = Br or Cl (complexes on entries 7 and 8 on Table 3   (solid) and 8 (dashed) in Table 3.  (Table 4). Interestingly for the monodentate N (imine) ligands (L = SMe 2 ) this tendency is less pronounced.

Complexes with monodentate N (imine) ligands
The spontaneous reaction of three type A monodentate N (imine) complex systems (L = SMe 2 and X = Br, i.e. entries 1, 3 and 5 from Table 3 Other conformers with a different ligand arrangement were also explored, but the calculated energies were found to be higher than the ones shown in Scheme 15. The final platinacycles of type 5C or 7C are finally obtained by the reductive elimination and release of the corresponding aryl by-products. Thus, the formation of the five-or seven-membered products should be dominated by the relative energy requirements of the oxidative addition (OATS-CH x ) and reductive elimination (RETS-CH x ) transition states. Experimentally, only the 7C type of platinacycles is observed for the bromido complexes on entries 1, 3 and 5 in Table 3. Clearly the transition states along the route leading to 7C (II) should be lower than those involved in the formation of 5C (I).
Indeed this is what is found when the relative free energies are computed (Table 5). Table 5. Computed relative free energies (in kJ mol -1 and in toluene solution at 70 °C) for the reaction leading to Pt II compounds of type 5C and 7C from complexes of type A with monodentate N (imine) ligands (entries 1, 3, and 5 of Table 3).

TABLE 5 HERE
As shown in Table 5, all the computed free energy profiles are quite similar, despite the diverse identity of the R group on the aryl ring. Nevertheless, the fact that the ratedetermining step completely changes depending on the pathway followed by the reaction is rather interesting. For pathway I, that leading to the five-membered platinacycle product 5C, the highest energy demanding stage corresponds to the final C-H reductive elimination producing the PhR moiety (RETS-CH 5 ). Contrarily, for pathway II, that leading to the seven-membered platinacycle product 7C, the highest energy value corresponds to the initial C-H bond activation on the distal phenyl ring of the imine ligand (OATS-CH 7 ). These differences can be easily rationalized by the careful observation of the structures of the transition states involved in each transformation. Figure 8 shows, as an example, the four transition states involved for the A compound on entry 1 of Table 3 Table 6 indicate that the energy requirements to get to the transition state for the studied compounds are: OATS-CH 5 < RETS-CH 7 < OATS-CH 7 < RETS-CH 5 , in perfect agreement with the computed free energy values in Table 5. (right, pathway II) from compounds of type Bcis originated form compound of type A in entry 1 of Table   3 (distances in Å, non-relevant H atoms have been omitted for clarity).

FIGURE 8 HERE
The computed relative free energy values collected in Table 5 have also been used to build a qualitative kinetic profile simulation model to evaluate the alternative formation of the final five-and seven-membered platinacycles, 5C and 7C, starting from the corresponding Bcis parent intermediates of entries 1, 3 and 5 in Table  of Table 3, respectively. Table 6. Smallest H-Pt-(Aryl) torsion angles (in degrees) and computed free energies (in kJ mol -1 ) for the transition states indicated in Table 5 for the C-H activation and reductive elimination reactions studied.

Complexes with chelating N (amino) -N (imine) ligands
The reactions of four different Pt II complexes bearing chelate N (amino) -N (imine) ligands, i.e. entries 7-10 in Table 3, have also been computationally studied in the same manner as that indicated above for the systems containing the monodentate N (imine) ligands. The experimental reactivity of the compounds of type A shown in entries 7 and 8 of this which correspond to a five-membered platinacycle compound, type 5C, for X = Br, but to a seven-membered analogue, type 7C, for X = Cl. For the reactions occurring on the compounds of type A of the last two entries (9 and 10), only compounds of type 7C are produced. Obviously this is a result of the hydrogen by fluoride substitution on the proximal aryl ring of the N (amino) -N (imine) ligand; nevertheless, their reactivity was also theoretically calculated for comparison with the experimental data available.
As stated above for the monodentate systems, two possible pathways (leading to 5C or 7C) are possible for the systems on entries 7 and 8 of For the bromido complex on entry 7 of Table 3, the energy requirements for pathway I, producing the thermodynamically preferred five-membered platinacycle 5C, should be lower in energy than those leading to 7C. Conversely, for the analogous chlorido system (entry 8 of Table 3), the energy requirements should favor pathway II, being the sevenmembered platinacycle of type 7C the one generated experimentally. Indeed, the computed relative free energies for these two systems confirm these trends, as shown in Table 7.
It may be noted that the dissociation of the chelating dimethylamino group requires a rather high input of energy (90.1 and 97.9 kJ mol -1 for systems in entries 7 and 8 of  ligands (entries 7 and 8 of Table 3).

TABLE 7 HERE
The computed free energies shown in Table 7 fully agree with the reactivity observed with respect to the nature and size of the final platinacycles. For the bromido complex in entry 7 of Table 3, the highest energy demands for pathways I and II are predicted to be 141.5 (RETS-CH 5 ) and 146.4 (OATS-CH 7 ) kJ mol -1 , respectively; this fact indicates that the preferred product should be the five-membered platinacycle of type 5C. This behavior is totally contrary to that for the equivalent chlorido compound (entry 8 of Table 3), for which the highest barriers of pathways I and II are 155.6 (RETS-CH 5 ) and 149.8 (OATS-CH 7 ) kJ mol -1 , respectively. In this case the formation of the seven-membered ring, type 7C, should be more favorable, as experimentally observed.
It should be indicated that, both for X = Br and Cl (entries 7 and 8 of  Table 3. Nevertheless, as in the previous section, a correlation exists between the torsion angle values and the energy demands of the corresponding transition state (Table 7). Angles closer to 90 º are associated with the lower energy transition states, while smaller angles relate to an increase of the energy requirements. However, for the system of entry 7 in the previous tables, the reductive elimination transition state RETS-CH 5 (on the way to a type 5C species) appears to be quite low in energy when associated with the torsion angle value (Table 8). Even so, a careful analysis reveals that the relatively small torsion angle is in fact larger than those found for the analogous transition states of the reactivity of the systems of entries 1, 3, 5, and 8 from  Table 8. Smallest H-Pt-(Aryl) torsion angles (in degrees) and computed free energies (in kJ mol -1 ) for the transition states indicated in Table 7 for the C-H activation and reductive elimination reactions studied.  Table 3 is much faster than that of the system of entry 8, can be immediately associated with the calculated energy demands above (Table 8). 11 The computed free energies for the reactions, despite differences in reaction conditions and calculations, indicate that for the bromido complex (entry 7) the energy requirement is 8 kJ mol -1 less than for the chloride analogue (entry 8). Furthermore, for the systems from entries 7 and 8 of Table 3, the reactivity shown in Scheme 16 has been found to take place after a kinetically and spectroscopically detected Btrans ⇄ Bcis rearrangement (see Kinetic studies section). Consequently, the highest barrier found in the preferred reaction pathway for each compound in Scheme 16 has to be lower than that for the isomerization reaction. Indeed, this is observed for the computed data for both systems; even for some other systems the isomerization reaction has been found to be ratedetermining. 13 For the bromido compound (entry 7 of Table 3 As a whole, the data collected in Table 7 allow the building of a qualitative kinetic simulation model for evaluating the formation, over time of the final five-and sevenmembered platinacycle compounds of types 5C and 7C from the starting Bcis complexes of entries 7 and 8 in Table 3 ( Figure 10). A very satisfactory qualitative agreement is found, even though the minor product is present in significant amounts in both complexes: for the bromido compound in entry 7 the 5C:7C ratio is 87:13, while for the chlorido complex in entry 8 the ratio diminishes to 73:24. These results indicate that, unfortunately, the computed free energy differences between pathways I and II are not as important as they should be from the experimentally collected results.
Nevertheless the overall trend and selectivity is successfully explained. In order to validate the employed methodology, the mechanism indicated in Scheme 16 has been also computed for the reaction of the fluorinated compounds on entries 9 and 10 in Table 3. As stated before, for these systems the formation of the five-membered platinacycles of type 5C is not possible, thus only pathway II has been computed ( Table   9). Table 9. Computed relative free energies (in kJ mol -1 and in toluene solution at 110 °C) for the reaction leading to Pt II compounds of type 7C from complexes of type A with chelate N (amino) -N (imine) ligands (entries 9 and 10 of Table 3).

TABLE 9 HERE
The calculated energy requirements for the process producing the compounds of type 7C, although very similar for both complexes, are slightly lower (ca. 1 kJ mol -1 ) for the system containing the chlorido ligand (entry10 on Table 3). The experimental data collected, as indicated in the previous section indicate that the rate of formation of the chlorido compound of type 7C is ca. four times faster than that of the analogous bromido complex (entry 9 on Table 3). 11 The qualitative kinetic model built for these processes totally agrees with this observation. Figure 11 shows the concentration-time profile for the formation of seven-membered platinacycles of type 7C from complexes of type A in entries 9 and 10 of Table 3; as predicted, the product ratio obtained is 1:4 in favor for the chlorido species. Figure 11. Qualitative product concentration evolution over time for Pt II complexes with chelate N (amino) -N (imine) ligands (Entries 9 and 10 in Table 3) in arbitrary time scale.

Concluding remarks
In this contribution, we have illustrated how a series of oxidative addition and reductive elimination reactions, occurring on platinum(II) organometallic complexes, can be resolved experimentally by the use of an appropriate interlock between preparative and kinetic methodologies. Namely, the reaction of bis(aryl)platinum(II) precursors with imine or amino-imine ligands leads to the formation of platinacyclic compounds with five-or seven-membered metallacycles via a sequence that involves C-X bond oxidative addition plus reductive C-C coupling followed by C-H oxidative addition with a final C-H reductive elimination. Given the fact that the process includes a series of non-simple and fairly slow processes, several intermediates have been proposed and fully characterized in the reaction media. Nevertheless, with the precision of the approach some of these characterized species are found to be completely irrelevant in the sequential reaction pathway; that is they are de facto false intermediates or dead-end compounds.
The kinetico-mechanistic studies carried out on this time-resolved reactivity has provided series of thermal and pressure derived activation parameters that has allowed a fairly comprehensive description of the reaction sequence involved in the full process.
Nevertheless, a careful and comprehensive analysis has had to be applied, since the rate sequence is not uniformly applicable for the full series of complexes. That is, the combination of two reactions having only slightly different activation energies produces dramatic changes when taken jointly, and the rate-determining step of the full process could differ within a series of very similar reactions.
In this respect, the synergy between experiments and DFT calculations has been found to be very important in order to improve the understanding of both the mechanisms involved and the observed reactivity. DFT calculations have not only ascertained the elementary reaction steps of the process, they have also explained the kinetic preference for the less thermodynamically stable products, found in the majority of cases.
Furthermore, the presence and incidence in the overall process of dead-end false intermediate compounds can be also assessed through calculations.
It is thus clear that computational studies, when fully supported by the empirical results, are able to fill in the areas of the puzzle that cannot be experimentally explored and resolved. In this respect, the set of studies presented in this contribution constitute a perfect example of how a close collaboration between experimentalists and theoreticians, with their specific perspectives when observing and interpreting data, provides remarkable value-added results to the work carried out.