Chemistry and catalysis advances in organometallic chemistry and catalysis
REDOX POTENTIAL–STRUCTURE RELATIONSHIPS AND
Download 11.05 Mb. Pdf ko'rish
|
- Bu sahifa navigatsiya:
- TABLE 50.1 Values of the P L Ligand Parameter for Selected Ligands
|
50 REDOX POTENTIAL–STRUCTURE RELATIONSHIPS AND PARAMETERIZATION IN CHARACTERIZATION AND IDENTIFICATION OF ORGANOMETALLIC COMPOUNDS M. F ´atima C. Guedes da Silva *
Armando J. L. Pombeiro *
The properties of coordination compounds in general, and organometallic ones in particular, depend on the metal, ligands, and coordination geometry, features that influence, for example, their electrochemical behavior, catalytic performances, and biological activity. The redox potential of metal complexes has been correlated [1–17] with a number of their properties, for instance, IR stretching frequencies, charge-transfer bands energy, ligand field stabilization energy, and Hammett’s σ and
related constants, nuclear magnetic resonance (NMR) spectroscopy parameters, macrocyclic ligand structural parameters, gas-phase ionization potential, or the highest energy occupied molecular orbital (HOMO) energy. Since these properties, which correlate with the redox potential, are dependent on the electron-donor/acceptor abilities of the ligands and metal center, the redox potential would be expected to provide a tool to quantify the latter features and to be used as an important characterization parameter. Following preliminary recognition of ligand additivity effects on the redox potential [1, 2] in a series of closely related octahedral-type 18-electron first-row transition metal complexes of the general type [M(CO) 6 −x L x ]
+ undergoing a one-electron reversible metal-centered oxidation, Pickett [18–20], Lever [21–24], and Bursten [25–27] have proposed systematic approaches for redox potential–structure relationships (or redox potential parameterization). The Bursten proposal is not as easy to apply as the others, has been much less explored, and is not discussed here. Further strategies have been oriented, namely, by the present authors, toward the extension of the proposed models to other geometries and electron counts, and to a wider variety of ligands and metal centers, as illustrated below. However, parameterization of the redox potentials of organometallic complexes other than those of carbonyl or isocyanide ones has seldom been attempted, regardless of the great development of their chemistry. Gathering of electrochemical parameters for these and other metal-ligating species relevant in organometallic chemistry has been presented [21–32], but the topic still requires further expansion. Advances in Organometallic Chemistry and Catalysis: The Silver/Gold Jubilee International Conference on Organometallic Chemistry Celebratory Book, First Edition. Edited by Armando J. L. Pombeiro. © 2014 John Wiley & Sons, Inc. Published 2014 by John Wiley & Sons, Inc.
678 REDOX POTENTIAL – STRUCTURE RELATIONSHIPS AND PARAMETERIZATION 50.2 PARAMETERIZATION OF LIGANDS AND METAL CENTERS 50.2.1 Octahedral-Type Complexes According to Pickett’s model [18], an electrochemical ligand parameter, P L , is defined as the difference between the oxidation potentials ( E ox 1 /2 ) of the complexes [Cr(CO) 5 L] and [Cr(CO) 6 ] (Eq. 50.1). Consequently, P L is a measure of the effect of the change of a ligand CO (for which P L = 0) by a ligand L on the redox potential, and the more negative value it adopts, the stronger is the electron-donor ability of L. It is noteworthy to mention that the P L ligand parameter (similar to the parameter E S for the metal centers—see below) reflects the overall (net) π-electron-acceptor minus σ -donor character of the ligand and not the differentiated electron acceptance and electron-donor abilities. The stronger the net electron-acceptor character, the higher the P L . For simplicity, the “net” or “overall” feature is commonly omitted in the discussion. The P L ligand parameter is linearly dependent on the redox potential, within series of closely related 18-electron [M S L] metal complexes—therefore possessing a common 16-electron square pyramidal metal center, {M S }—as expressed by the simple Eq. 50.2 involving two other parameters, E S and β. E S represents the value of E ox 1 /2 of the carbonyl complex [M S (CO)]
and is termed the electron richness of the metal center {M S }, adopting the lowest (more negative) values for the richest {M S } sites. The parameter β (slope of the linear relationship of Eq. 50.2) concerns the sensitivity of the redox orbital energy (commonly based on the metal center) to a changing of the ligand L. It is called the polarizability of the metal site. P L
ox 1 /2 [Cr (CO)
5 L] − E ox 1 /2 [Cr (CO
6 )] (50.1) E ox 1 /2 [M S L] = E
S + βP
L (50.2)
The parameter P L can be derived either directly by definition from Eq. 50.1 when the [Cr(CO) 5 L] complex and its oxidation potential are known, or from Eq. 50.2 if the oxidation potential of the complex [M S L] and the parameters E S and
β for the {M S } metal site are known. Examples of ligands with proposed P L value [9, 18, 33–66] are given in Table 50.1. E S and β values have also been reported [18, 33, 36–39, 41, 53, 66–70] for a number of metal centers {M S }. Pickett’s parameterization method was quite useful is several cases, such as (just to mention a few) (i) in the identification of unstable complexes in solution, for example, [Mo(N 2 )(NH
3 )(dppe)
2 ] [18], formed in situ upon reaction of the dinitrogen TABLE 50.1 Values of the P L Ligand Parameter for Selected Ligands a L P L , V References NO + 1.40 18 Carbynes 0.24–0.21 40 η 2 -Vinyl (
η 2 –C(CH 2 )CH
2 Ph)
0.22 44 Aminocarbyne (CNH 2 + ) 0.09 44 CO 0 21 Isocyanides (bent) b −0.07 to −0.18 33, 34, 36, 47 Vinylidines 0 to −0.6
28, 40 η 2 -Allene −0.21
44 Ferricinium isocyanides −0.22 to −0.28 49, 50
Allenylidenes ( =C=C=CR
2 ) 0 to −0.8 28, 52
Isocyanides (linear) c −0.33 to −0.44 33, 34, 36, 53 CN–BPh
3 − −0.51 56 Oxocarbenes [ =C(OR)Y] −0.51 to −0.64 28 Thiocarbenes [ =C(SR)Y] −0.66 to −0.68 28 Aminocarbenes [ =C(NRR )Y] −0.69 to −0.80 28, 85
−0.70 to −0.9 28 C
−c −1.00
18 Alkynyls( −C≡CR −
−0.9 to −1.7 28, 48, 62, 63, 85 Aryl, alkyl, NO − −1.7 to −1.9 28 a Ordered generally from higher to lower P L values; for multidentate ligands the values refer to each ligating arm. b At an electron-rich metal center such as trans- {MX(dppe) 2 } (M = Re, Tc; X = Cl, H). c At an electron-poor metal center such as {Cr(CO) 5
PARAMETERIZATION OF LIGANDS AND METAL CENTERS 679 analog [Mo(N 2 )
(dppe) 2 ] with ammonia; (ii) for discriminating the geometry (linear or bent) of an isocyanide ligand, as in [ReCl(CNR)(dppe) 2 ] and [Re(CNR) 2 (dppe)
2 ] [33–36] with P L values dependent on the geometry of CNR; (iii) in the mathematical expressions for the prediction of E ox 1 /2 values for 18-electron octahedral compounds of the types [M S L 2 ] or [M S LL ] with a square-planar 14-electron {M S } metal site or for compounds [M S L 3 ] with a 12-electron {M S
[35–37]. We have also extended the model to open-shell 17- and 16-electron octahedral-type complexes [38, 39]. The Pickett and other derived models are only partially additive, as the effects on the redox potential of the complex [M S
n ] are assumed to concern only the {M S
Eq. 50.2) corresponds to the minimum additivity requirement. Since additivity of the ligand effects can fail in some cases, for example, owing to ligand synergisms, the model can be more appropriate than any other, with more extensive additivity requirements, but it presents a relatively limited scope, being valid only for rather closely related complexes of a series with a common {M S
accord with the limited number of {M S } centers with available values of E S and β , Pickett’s model has not found a wide application. In contrast, a full additivity model has been proposed by Lever [21–24] where additivity has been extended to all the ligands. For such a purpose, he selected the Ru III /Ru
II redox pair, for which a wide number of complexes are known, as the standard one, and defined a novel, fully additive ligand parameter, E L , normally obtained through a statistical analysis of the known redox potentials of the complexes with such redox couple (E L values can also be estimated from Hammett’s relationships [24, 34–37, 71, 72]). From the thus determined E L values, he proposes that for each particular M n /M
−1 redox
pair, its complexes with a certain stereochemistry and spin state satisfy the linear relationship given by Eq. 50.3. E = S
M (SE
L ) + I
M (50.3)
in which the potentials (E) are in V versus the normal hydrogen electrode (NHE), and S M (slope) and I M (intercept) are parameters characteristic of the M
/M
−1 redox couple, assuming values of 1 and 0, respectively, for the ruthenium standard. A slope (S M ) greater than unity concerns a greater sensitivity, relative to the Ru III /Ru
II standard, of the redox potential of the complexes bearing the M
/M
−1 redox pair on the ligands contributions. The intercept (I M , also in V vs NHE) depends on the gas-phase ionization energy, the electronic repulsion energy both in the ligand and in the metal, and also on the difference in solvation energies for M n and M
n −1 . Lever’s model has been applied to numerous ligands with wide ranges of electron-donor and π-electron-acceptor characters and binding various types of metal centers; extensions thereof have also been achieved. Table 50.2 lists the values of S M and I M for a variety of metal redox couples, whereas E L values for a diversity of ligands with relevance in organometallic chemistry are presented in the subsequent tables, some of them were obtained from the literature but others were estimated through application of Eq. 50.4 (see below) from the already known parameter P L .
data. Nevertheless, its full additivity character confers a higher risk of failure than a more limited model (such as Pickett’s) based on a lower additivity. The P L and E L parameters are a measure of the net (overall) electron-donor ability of the ligands and are expected to follow parallel trends in accord with the empirical Eq. 50.4 [21] but exceptions were found for strong π-acceptor ligands, namely, CO [21, 69], carbynes [40], and isocyanides [21, 41], which present relevant π-stabilization effects on the HOMO. For these cases, apart from corrections to the E L parameters, improvements to the general expression 50.3 have been required; for a complex with two very strong π-acceptor ligands (e.g., CO and CNR), the novel expression 50.5 was proposed [21, 24] in which c and c are empirical corrections and x and x are the number of π* orbitals from such ligands that interact with the HOMO. Depending on the presence of other different π-acceptors, additional positive corrections could be added to Eq. 50.5. Such corrections to the potential values can reach values as high as 0.3 V per ligand [21, 24, 41]. P L = 1.17 E L − 0.86 (50.4) E = S
M (SE
L ) + I
M + cx + c x (50.5) Other limitations of the model have been recognized, in particular due to failure of ligand additivity [42], to isomeric effects [21, 22] and to deviations from the linearity expressed by Eq. 50.3 for a wide range of E L [22, 38, 40], eventually suggesting the need for specific S M and I M values for a particular metal center or even the consideration of a curved relationship between the redox potential and
L or of different expressions along the E L scale. The possible dependence of the E L ligand parameter on the metal center is also a difficulty which was recognized in the cases of isocyanides [33, 34, 36, 37, 47], cyanide [43], nitriles [38], cyanamides [37, 57, 59], vinylidines [28], allenylidenes [28], and alkynyls [28]. 680 REDOX POTENTIAL – STRUCTURE RELATIONSHIPS AND PARAMETERIZATION TABLE 50.2 Values of the S M and I M Parameters for Selected Metal Redox Pairs, in Organic Medium a Redox Pair S M
M (V) versus NHE References Cr III /Cr II (LS) 1.18 −1.72
21 Cr III /Cr II (HS) 0.84 −1.18
21 Cr I /Cr 0 0.5 −1.75 21 Fe IV /Fe
III (specific) b 1.49
−0.18 47 Fe III /Fe
II (LS)
1.10 −0.43
21 Fe III /Fe II (specific) c 1.07
−0.3 37 Fe III /Fe
II (specific) d 1.32
−0.57 47 Fe III /Fe
II (HS)
0.89 −0.25
21 Mn II /Mn I 0.81 −1.76 21 Mo I /Mo
0 0.74
−2.25 21 Mo II /Mo
I 0.81
−1.76 21 Nb V /Nb
IV 0.76
−1.24 21 Nb IV /Nb
III 0.75
−0.12 21 Os III /Os
II 1.01
−0.40 21 Re IV /Re
III 0.86
0.51 22 Re III /Re
II 1.17
−0.88 22 Re II /Re
I (upper)
0.76 −0.95
22 Re II /Re I (lower) 0.27 −1.43
22 Ru IV /Ru III
1.03 1.68
73 Ru III /Ru II 0.97 0.04 21 Ta V /Ta
IV 0.79
0.66 21 Tc IV /Tc
III 1.00
0.65 22 Tc III /Tc
II 1.28
−0.89 22 Tc II /Tc
I 1.42
−2.09 22 Rh II /Rh
Ie 1.68
−0.87 39 Rh II /Rh
If 0.80
−0.95 76
a Six-coordinate metal center, except when stated otherwise. b Established for the series trans-[FeBrL(depe) 2 ]
+/3+ (L = CO, aryl and alkyl N≡CR). c Established for the series trans-[FeL 2 (depe)
2 ] 2 +/3+ (L = CO, aryl and alkyl N≡CR, N ≡C–NR
2 ). d Established for the series trans-[FeBrL(depe) 2 ] +/2+ (L = CO, aryl and alkyl N≡CR, Br − ). e Four-coordinate metal center. f Five-coordinate metal center. The estimate of the E L value of a particular ligand (L) can be achieved provided one knows (i) the redox potential of a complex with that ligand L bound to a M n +1/n metal redox couple with known I M and S M parameters and (ii) the E L values
of the other ligands. Following this procedure, the E L data set has been greatly enlarged to a wide variety of ligands and further extensions are expected as more metal complexes are being synthesized and their redox potentials measured. The availability of various complexes with the same ligand should also allow the refinement of its E L value and checking of its nondependence on the metal center. E L can also be obtained from Hammett’s relationships [24, 34–37, 71, 72]. Once the E L , I M , and S M electrochemical parameters are established, they can be used to predict the redox potential of a species and enable the in situ identification of an (un)known complex, product, or intermediate, thus avoiding its isolation and eventual decomposition. Examples include some chloro/azole-ruthenium [73, 74] or chloro-osmium [75] complexes derived upon ligand replacement induced by electron transfer. The identification of such species, generated in situ, has allowed the establishment of the detailed mechanisms of the reactions by digital simulation of cyclic voltammetry. The Lever model has typically been applied to octahedral-type six-coordinate complexes, with metal-centered redox processes, but extensions to other types of complexes have also been proposed, namely, to square-planar four-coordinate and five-coordinate Rh II /Rh I complexes [39, 76, 77], to sandwich complexes [23, 24, 31, 32], to Ru clusters [78–80], and also to complexes with ligand-centered reduction processes [24]. Relevant or representative cases are discussed in the following sections. PARAMETERIZATION OF LIGANDS AND METAL CENTERS 681 50.2.2 Tetra- and Pentacoordinate Rh II/I Complexes The S M and I M values (1.83 and −1.22 V vs NHE, respectively) for the square-planar Rh II /Rh I couple were firstly estimated [77] through the knowledge of the E ox 1 /2 of the complexes [RhCl(L)(P i Pr 3 ) 2 ] [L = ethylene (0.94 V), CO (1.36 V)] and the known [21] E L values for Cl − , ethylene, CO, and trimethylphosphine instead of the unknown triisopropylphosphine (E L =
through the application of the same method to a broader series of rhodium(I) complexes and the values of S M = 1.68 and I M = −0.87 V were those lastly achieved (Table 50.2). The value of 1.68 for the S M (slope) parameter is higher than those reported for the hexacoordinate redox pairs (Table 50.2) and suggests a higher susceptibility of the square-planar Rh II /Rh I metal center to a change of the ligands. However, the generalization of this observation would need an even broader number of Rh I
M and I M parameters enabled the estimate of the E L values for a set of unsaturated organic ligands of the type =C(=C) n Ph 2 (n = 0, 1, 2; Tables 50.3–50.5). Expectedly, distinct S M and I M parameters were found for the pentacoordinated Rh I compounds [RhH(CO)L 3 ] and
[RhHL 4 ] [L = P(NC 4 H 4 ) 3 , PPh 3 , P(OPh) 3 or P(OC
6 H 4 Me-4) 3 ; L = PPh 3 or P(OC 6 H 4 Me-3) 3 ]: 0.80 and −0.95 V, respectively [76] (Table 50.2), but these values should be considered with great care in view of the limited number of compounds used in their estimate.
Download 11.05 Mb. Do'stlaringiz bilan baham: 
|