Donor Strengths Determination of Pnictogen and Chalcogen Ligands by the Huynh Electronic Parameter and Its Correlation to Sigma Hammett Constants.

The suitability and accuracy of the Huynh electronic parameter (HEP) was further tested to reveal remote substituent effects in pyridines, which are located five or six bonds away from the reporter probe. These values show an excellent correlation to Hammett σ-constants of the respective substituents with coefficients of R2 =0.9856 (σm ) and R2 =0.9857 (σp ). Based on this observation, a methodology for the re-evaluation of certain Hammett constants with larger uncertainties has been proposed and demonstrated. Moreover, the scope of HEP was extended to various neutral pnictogen and chalcogen donors during which "transphobia effects" were revealed for mixed NHC complexes containing phosphites, arsine and stibine for the first time.

Introduction

The properties of n class="Chemical">metal complexes are determiclass="Chemical">ned by the class="Chemical">nature of the class="Chemical">n class="Chemical">metal center and the stereoelectronic signatures imposed by the ligands. Often the steric bulk of a particular ligand can be easily estimated by its Lewis structure drawing, while its electronic properties are more difficult to judge. This is especially so, when the ligand has multiple substituents of different inductive and mesomeric effects, which could enhance or oppose each other. In order to compare the donating abilities of ligands experimentally, a few electronic parameters have been developed (Figure 1).

Figure 1

Selected experimental electronic parameters.

The ligand electrochemical parameter E L (n class="Gene">LEP) has beeclass="Chemical">n iclass="Chemical">ntroduced by Lever based oclass="Chemical">n redox poteclass="Chemical">ntials of for example, RuII/III class="Chemical">n class="Chemical">metal complexes.1 The E L values of a large number of Werner‐type ligands have been tabulated, which reflect their relative capacity to stabilize a metal in a certain oxidation state. Ligands with smaller E L values can therefore stabilize the RuIII state in the RuII/III couple better than those with larger E L values. Although E L values are not a direct measure for the donating ability of a ligand, they are nevertheless related to its donating power.2 The drawback is that only complexes with reversible redox chemistry can be considered for the determination of E L values. The requirement for less common electrochemical setups and the exclusion of non‐innocent ligands is a further limitation.

In organon class="Chemical">metallic chemistry, the most commoclass="Chemical">nly used parameter is the so‐called Tolmaclass="Chemical">n electroclass="Chemical">nic parameter (class="Chemical">n class="Chemical">TEP) developed in 1970, which compares the A1 carbonyl IR stretching frequencies of [Ni(CO)3L] complexes, in which L denotes the ligand of interest.3 This methodology evaluates the amount of π‐backdonation from the nickel(0) center to the carbonyl ligands, which is in turn influenced by the donor/acceptor power of the ligand L. For phosphines, for which σ‐donor and π‐acceptor strengths are approximately inversely proportional, TEP has become a valuable tool of assessment. For ligands that do not show such behavior, interpretation of TEPs is less straightforward, since deconvolution into donor and acceptor contributions is not possible. Another disadvantage of TEP and its Rh/Ir variants4 is the requirement for highly toxic materials, that is, [Ni(CO)4] or carbon monoxide. The largest drawback, however, is that the majority of Werner‐type ligands cannot be systematically probed using carbonyl‐based methods. For example, amines and pyridines do not react with [Ni(CO)4].

We have introduced a new electronic parameter, which addresses these shortcomings and allows direct comparison of Werner‐type and organon class="Chemical">metallic ligaclass="Chemical">nds oclass="Chemical">n a uclass="Chemical">nified scale without the use of highly toxic materials.5 Our parameter, that is, Huyclass="Chemical">nh electroclass="Chemical">nic parameter (class="Chemical">n class="Chemical">HEP), evaluates the influence of a particular trans‐ligand L on the 13Ccarbene NMR shift of the Pr2‐bimy reporter ligand (i.e. HEP signal) in square‐planar complexes of the type trans‐[PdBr2(Pr2‐bimy)L]. It was found that a weaker donor would lead to a relative upfield shift, whereas a stronger donor would lead to a downfield shift of the HEP signal. Different types of NHCs,6 phosphines, isocyanides, and N‐donors have been evaluated.7, 8, 9, 10 Furthermore, much smaller electronic differences among similar ligands can be resolved, providing useful information for the fine‐tuning of complexes. To test its suitability further, we herein report a systematic evaluation of remote substituent effects by using meta‐ and para‐substituted pyridines and correlated the HEP results to the respective σ p and σ m Hammett constants.11 Subsequently, the HEP scale was extended to various neutral pnictogen and chalcogen ligands such as phosphites, arsines, stilbenes, etc. to provide insights into their electronic differences.

Results and Discussion

HEPs of substituted pyridines and correlation to σ‐Hammett constants

One advantage of the n class="Chemical">HEP is the ease of complex probe preparatioclass="Chemical">n. Thus, simple mixiclass="Chemical">ng of the dimeric complex [class="Chemical">n class="Chemical">PdBr2(Pr2‐bimy)]2 (I)12 with two equivalents of the substituted pyridines in dichloromethane at ambient temperature already results in rapid reactions. In all cases, the initial orange suspensions clear into bright‐yellow solutions (Table 1). To ensure completion, the mixtures were allowed to stir for 2 h. Overall, 26 new trans‐[PdBr2(Pr2‐bimy)(Py‐R)] complexes were prepared with 13 meta‐ (2–14) and 13 para‐substituted pyridines (15–27) and compared to the previously reported parent complex trans‐[PdBr2(Pr2‐bimy)(Py)] (1).13, 14

Table 1

Synthesis of trans‐[PdBr2(Pr2‐bimy)(Py‐R)] complexes 2–27, in which Py‐R=meta or para‐substituted pyridines, and their characteristic 1H NMR signals [ppm], HEP values [ppm],[a] and Hammett σ‐constants in CDCl3.

R‐group	δNCH	δPyH2,6	HEP	σ m/p
3‐CN (2)	6.25	9.53, 9.43	157.33	0.56
3‐NO2 (3)	6.28	10.10, 9.55	157.38	0.71
3‐CHO (4)	6.31	9.67, 9.43	158.54	0.35
3‐Br (5)	6.29	9.26, 9.11	158.63	0.39
3‐I (6)	6.29	9.36, 9.12	158.68	0.35
3‐Cl (7)	6.30	9.18, 9.07	158.72	0.37
3‐F (8)	6.30	9.12, 9.02	158.73	0.34
3‐CO2H (9)	6.33	9.81, 9.37	159.04	0.37
3‐OH (10)	6.32	8.63, 8.54	159.60	0.12
3‐Ph (11)	6.37	9.39, 9.12	159.80	0.06
3‐Et (12)	6.35	8.94, 8.92	160.33	−0.07
3‐Me (13)	6.34	8.91, 8.89	160.41	−0.07
3‐NH2 (14)	6.33	8.51, 8.44	160.47	−0.16
4‐CN (15)	6.25	9.42	157.79	0.66
4‐CF3 (16)	6.29	9.40	158.38	0.54
4‐CHO (17)	6.30	9.43	158.76	0.42
4‐Cl (18)	6.29	9.08	158.98	0.23
4‐CO2H (19)	6.31	9.36	159.00	0.45
4‐Br (20)	6.29	8.94	159.03	0.23
4‐I (21)	6.28	8.79	159.21	0.18
4‐Ph (22)	6.38	9.16	160.16	−0.01
4‐OMe (23)	6.33	8.92	160.39	−0.27
4‐Me (24)	6.33	8.93	160.43	−0.17
4‐Et (25)	6.34	8.96	160.58	−0.15
4‐NH2 (26)	6.34	8.61	161.35	−0.66
4‐NMe2 (27)	6.36	8.61	161.97	−0.83

[a] HEP values are given with the second decimal in subscript for comparison. Detailed discussion on standard deviations of HEP can be found in References [5], [7] and [8].

Synthesis of trans‐[n class="Chemical">PdBr2(Pr2‐bimy)(Py‐R)] complexes 2–27, iclass="Chemical">n which Py‐R=class="Chemical">n class="Chemical">meta or para‐substituted pyridines, and their characteristic 1H NMR signals [ppm], HEP values [ppm],[a] and Hammett σ‐constants in CDCl3.

n class="Chemical">HEP

3‐n class="CellLine">CHO (4)

3‐n class="Chemical">CO2H (9)

4‐n class="CellLine">CHO (17)

4‐n class="Chemical">CO2H (19)

4‐n class="Gene">NMe2 (27)

[a] n class="Chemical">HEP values are giveclass="Chemical">n with the secoclass="Chemical">nd decimal iclass="Chemical">n subscript for comparisoclass="Chemical">n. Detailed disclass="Chemical">n class="Chemical">cussion on standard deviations of HEP can be found in References [5], [7] and [8].

All complexes are air‐ and moisture‐stable and were characterized by n class="Chemical">1H aclass="Chemical">nd class="Chemical">n class="Chemical">13C NMR spectroscopies, ESI mass spectrometry and elemental analysis. Single crystals for six complexes were also obtained as representatives by slow evaporation of their solutions, and X‐ray diffraction analyses confirm their desired compositions (vide infra). The 1H NMR spectra of all compounds show the presence of the Pr2‐bimy and the pyridine derivative in an expected 1:1 ratio. Upon coordination, the ortho‐proton signals of the pyridine derivatives shift downfield by 0.4–0.6 ppm (i.e. δPyH2,6 in Table 1). Another striking feature is the downfield shift of 1.04–1.17 ppm observed for the isopropyl C−H protons (i.e. δNCH in Table 1) of the complexes in comparison to those in the 1,3‐diisopropyl‐benzimidazolium bromide salt at 5.21 ppm. These suggest the presence of anagostic C−H⋅⋅⋅Pd interactions, which are quite common for d8 metal complexes with Pr2‐bimy ligands.15 Similar spectral observations were also made for the related compounds with the parent pyridine and bridging bipyridine ligands.13

More important for our purpose are the n class="Chemical">13C class="Chemical">n class="Chemical">NMR chemical shifts of the Pr2‐bimy reporter ligand, that is, HEP values. The good solubility of the complexes allows an easy detection of these signals, and they are summarized in Table 1. In general, electron releasing substituents lead to small downfield shifts, while electron‐withdrawing ones induced an upfield trend with respect to the parent unsubstituted pyridine complex (R=H), which has an HEP value of 160.01 ppm.8 Notably, remote changes five (meta) or six (para) bonds away can be differentiated by the Pr2‐bimy reporter signal in line with chemical intuition. In this respect, it is worth mentioning that a standard deviation of SD=0.01 ppm can be estimated by the full‐width‐at‐half‐maximum (fwhm≈0.02 ppm) for typical 13C NMR signals.5, 7, 8

In organic chemistry, electronic substituent efn class="Chemical">fects have beeclass="Chemical">n quaclass="Chemical">ntified by the Hammett σ‐coclass="Chemical">nstaclass="Chemical">nts, which have beeclass="Chemical">n deficlass="Chemical">ned usiclass="Chemical">ng class="Chemical">n class="Chemical">meta and para substituted benzoic acids, while no constants were considered for ortho derivatives due to steric interferences.16 They are typically based on conductance measurements of ionization constants of the appropriately substituted benzoic acids [Eq. (1)], in which K H is the ionization constant for benzoic acid in water at 25 °C, and K X is the ionization constant for a meta‐ or para‐substituted benzoic acid.11, 16

Electron‐withdrawing substituents would increase the acidity giving rise to positive values of Hammett σ‐constants and vice versa. In addition, these constants have also been shown to reflect the electron densities of substituted n class="Chemical">pyridines.17 Heclass="Chemical">nce, the class="Chemical">n class="Chemical">HEP values obtained for complexes 2–27 were correlated to the well‐accepted and widely used Hammett constants in order to further evaluate our methodology in terms of accuracy (Table 2).

Table 2

Selected NMR spectroscopic data of complexes 32–34 and trans‐36–39 measured in CDCl3.

Complex	Ligand	1HMe [ppm]	HEP [ppm]	2 J(C,P) [Hz]
32	TBD	1.74	165.74	–
33	DBU	1.77	166.32	–
34	NEt3	1.74	157.96	–
trans‐36	P(OiPr)3	1.77	175.25	287.1
trans‐37	P(OPh)3	1.51	171.65	289.8
trans‐38	P(O‐2,4‐tBu‐Ph)3	1.59	171.48	289.6
trans‐39	AsPh3	1.79	169.17	–

n class="Chemical">Notably, aclass="Chemical">n excelleclass="Chemical">nt correlatioclass="Chemical">n was already observed usiclass="Chemical">ng all data with liclass="Chemical">near regressioclass="Chemical">n coefficieclass="Chemical">nts of R 2=0.9773 aclass="Chemical">nd R 2=0.9816 for the class="Chemical">n class="Chemical">meta and para series, respectively. In an early paper, McDaniel and Brown highlighted that there were considerable variations in the σ‐values for the 3‐ and 4‐CO2H substituents (0.37 and 0.45) with large estimated limits of uncertainty of around 0.1.18 Based on this statement, the correlation study was repeated with the exclusion of the carboxylic acid group, which led to further improvements of the coefficients to R 2=0.9856 (meta) and R 2=0.9857 (para). The linear regression graphs and equations are depicted in Figure 2.

Figure 2

Graphs and equations for the regressions of (i) HEP vs. σ m Hammett constants (upper) and (ii) HEP vs. σ p Hammett constants (lower).

Graphs and equations for the regressions of (i) n class="Chemical">HEP vs. σ m Hammett coclass="Chemical">nstaclass="Chemical">nts (upper) aclass="Chemical">nd (ii) class="Chemical">n class="Chemical">HEP vs. σ p Hammett constants (lower).

The resulting equations allow interconversion of n class="Chemical">HEP iclass="Chemical">nto Hammett σ coclass="Chemical">nstaclass="Chemical">nts or vice versa [e.g., Eqs. (2), (3)]. For example, usiclass="Chemical">ng the reported class="Chemical">n class="Chemical">HEP value for 3‐ethynylpyridine (3‐Py‐C2H) of 159.2 ppm8 and Equation 2, we can predict a σ m value of 0.20 for the ethynyl substituent, which is very close to the reported value of 0.21.11

Since Hammett σ‐values are determined from acidity constants of n class="Chemical">benzoic acids, it is actually uclass="Chemical">nderstaclass="Chemical">ndable why the uclass="Chemical">ncertaiclass="Chemical">nties for the class="Chemical">n class="Chemical">carboxylic acid substituent are rather large. The increased acidities of isophthalic (pK a1=3.70) and terephthalic acid (pK a1=3.54) compared to that of the parent benzoic acid (pK a=4.19) cannot be solely attributed to the electron‐withdrawing nature of the CO2H group. Due to the presence of two identical acid functions in these molecules, the proton dissociation is statistically enhanced. Thus, the electronic influence of the CO2H group is likely to be overestimated resulting in larger σ‐values. On the other hand, the HEP values of the pyridine carboxylic acid complexes 9 and 19 reflect the intrinsic electronic property of the molecules without such interferences (Figure 3). Application of Equations 2 and 3 with these HEP values gives lower calculated constants of σ m*=0.25 and σ p*=0.29.

Figure 3

Determination of the Hammett constant and HEP for the carboxylic acid substituent.

Determination of the Hammett constant and n class="Chemical">HEP for the class="Chemical">n class="Chemical">carboxylic acid substituent.

McDaniel and Brown also noted that ionic substituents have Hammett constants with larger uncertainties.18 An example is the n class="Chemical">carboxylate group (class="Chemical">n class="Chemical">CO2 −), which has been assigned a zero value for σ p without any consideration of cation effects.18 However, we suppose that the electronic influence of this group cannot be cation‐independent. To test this notion, we subjected the complex trans‐[PdBr2(Pr2‐bimy)(Py‐4‐CO2H)] (19) to common alkali metal hydroxides to obtain HEP probes of pyridyl‐4‐carboxylates with different alkali metal ions (Scheme 1).

Scheme 1

Synthesis of HEP probes with different alkali pyridinecarboxylate ligands.

Synthesis of n class="Chemical">HEP probes with difclass="Chemical">n class="Chemical">ferent alkali pyridinecarboxylate ligands.

Upon deprotonation, an increased donating ability is expected for the resulting n class="Chemical">pyridinecarboxylate ligaclass="Chemical">nd, which is iclass="Chemical">ndeed coclass="Chemical">nfirmed by the larger class="Chemical">n class="Chemical">HEP values. As anticipated, different cations induce a different response, and the downfield shifts are smallest for Li+ and largest for Na+ and K+. This observation implies that the negative charge is best stabilized by the lithium cation, which is in line with its largest charge/size ratio leading to the best hard/hard interactions with the carboxylate group. The HEP value peaks with sodium and gradually decrease with increasing size of the cation. The difference between potassium and rubidium is larger as the latter is known to form compounds with higher coordination numbers. Scheme 1 also depicts the calculated σ p* constants for the different alkali metal carboxylates using Equations 2 and 3.

With the availability of n class="Chemical">HEP data summarized iclass="Chemical">n Table 1, it is lastly also of iclass="Chemical">nterest to study if the class="Chemical">n class="Chemical">HEP value of a meta‐substituted pyridine could be predicted by that of its para‐analogue and vice versa. In order to test such a possibility, HEPs of meta‐ and para‐substituted pyridines bearing identical substituents were plotted against each other where available. With inclusion of all ten substituents, a very good linear regression coefficient of R 2=0.9701 was obtained. Exclusion of the NH2 group as an outlier led to a further improvement to R 2=0.9846 (Figure 4).

Figure 4

Regression graphs and equation of HEP (meta) vs. HEP (para).

Regression graphs and equation of n class="Chemical">HEP (class="Chemical">n class="Chemical">meta) vs. HEP (para).

Based on this encouraging result, it appears that substituents in n class="Chemical">meta‐ aclass="Chemical">nd para‐positioclass="Chemical">n exhibit a similar (almost liclass="Chemical">near) efclass="Chemical">n class="Chemical">fect on the HEP, potentially allowing for predictions from each other. This could be particularly powerful for groups with primarily inductive effects, while those with dominant mesomeric contributions have to be treated with more caution. The latter is understandable as mesomeric effects have a drastically different impact in meta‐ compared to para‐position.

Determining the HEPs of selected pnictogen donors

Having established an excellent correlation of n class="Chemical">HEP with Hammett coclass="Chemical">nstaclass="Chemical">nts, we foclass="Chemical">n class="Chemical">cused on the extension of our method to previously unaccounted ligands. In particular, triethylamine (NEt3), 1,5,7‐triazabicyclo[4.4.0]dec‐5‐ene (TBD) and 1,8‐diazabicyclo[5.4.0]undec‐7‐ene (DBU) are widely used organic bases, but they can also stabilize active metal species by coordination. It is therefore surprising that their donating abilities have not been explicitly compared yet. Hence, the respective HEP probes were routinely prepared by exposure of [PdBr2(Pr2‐bimy)]2 (I)12 to the free amine bases. The yellow products trans‐[PdBr2(Pr2‐bimy)(TBD)] (32) and trans‐[PdBr2(Pr2‐bimy)(DBU)] (33) formed rapidly, while prolonged stirring with an excess of triethylamine was required to fully convert complex I into trans‐[PdBr2(Pr2‐bimy)(NEt3)] (34). TBD is a bicyclic guanidine containing an imine and a secondary amine function that are available for coordination. However, the ligand shows selective coordination of the imine nitrogen to metal centers such as Pd, Cu, Ag, Fe and Li,19, 20 which is also confirmed by single crystal X‐ray diffraction of TBD complex 32 (vide infra). Coordination via the imine function is also expected for the DBU ligand. The three base adducts 32–34 have good solubilities in most organic solvents, and no cis‐isomers were detected. In the 13C NMR spectra, the three carbene carbon atoms resonate in the order of 34 (NEt3, 157.96 ppm)<32 (TBD, 165.74 ppm)<33 (DBU, 166.32 ppm, Table 2).21 This trend indicates that the aliphatic amine NEt3 is a significantly weaker donor compared to the unsaturated imines, which is in agreement with a greater availability of the lone pair of sp 2‐hybridized imine donors. The weaker donating effect of NEt3 comparable to the weakest pyridines (Table 1) also explains its sluggish reaction with dimer I. In addition, DBU is a stronger ligand than TBD, which is reasonable by considering the additional −I effect from the third nitrogen atom in the TBD ligand.

Selected n class="Chemical">NMR spectroscopic data of complexes 32–34 aclass="Chemical">nd traclass="Chemical">ns‐36–39 measured iclass="Chemical">n class="Chemical">n class="Chemical">CDCl3.

n class="Chemical">1HMe [ppm]

n class="Chemical">HEP [ppm]

n class="Chemical">TBD

–

n class="Chemical">DBU

–

n class="Gene">NEt3

–

P(OPr)3

n class="Chemical">P(OPh)3

P(O‐2,4‐Bu‐Ph)3

n class="Chemical">AsPh3

–

In addition to n class="Chemical">nitrogen class="Chemical">n class="Species">donors, HEP has also been used to evaluate different types of phosphines.8, 22, 23 It is thus of interest to extend the scope to other pnictogen donors such as phosphites, arsines and stibines. This study was initiated with P(OMe)3, P(OPr)3, P(OPh)3 and P(O‐2,4‐Bu‐Ph)3, which are diverse in terms of stereoelectronic properties. Instant color changes from orange to pale yellow were observed for most reaction mixtures indicative of a generally strong donating ability of these ligands and the fast formation of mixed NHC/phosphite complexes, which are surprisingly rare.24, 25 Similar to their phosphine counterparts, electronically driven trans–cis isomerizations were observed. The thermodynamically preferred cis‐arrangement in these complexes is due to the so‐called “transphobia effect” of phosphorus donors.12, 26, 27 This term was introduced to describe the general difficulty of placing a phosphorus trans to a carbon donor.28 The isomerization is fastest for the P(OMe)3 and P(OPr)3 complexes containing the smallest phosphites. Hence, only the cis‐[PdBr2(Pr2‐bimy){P(OMe)3}] (cis‐35) and cis‐[PdBr2(Pr2‐bimy){P(OPr)3}] (cis‐36) complexes were isolated in the standard procedures. The process is slower for P(OPh)3 complex, and signals for both trans/cis‐[PdBr2(Pr2‐bimy){P(OPh)3}] (trans‐/cis‐37) complexes were captured in the NMR spectra. The complex trans‐[PdBr2(Pr2‐bimy){P(O‐2,4‐Bu‐Ph)3}] (trans‐38) resists isomerization due to the enhanced steric bulk, and a HEP value of 171.48 ppm was detected for the P(O‐2,4‐Bu‐Ph)3 ligand without problems (Table 2).

Since trans‐configured complex probes are required for the n class="Chemical">HEP method, direct class="Chemical">n class="Chemical">NMR scale reactions were carried out to capture the signals of the initially formed trans‐isomers for the less bulky phosphites before complete isomerization. Upon mixing the reactants in NMR tubes, the samples were immediately measured after addition of CDCl3. Indeed, only one set of signals due to the trans‐[PdBr2(Pr2‐bimy){P(OPr)3}] (trans‐36) or trans‐[PdBr2(Pr2‐bimy){P(OPh)3}] (trans‐37) complexes was observed in the respective 1H NMR spectra. 31P and 13C NMR data were measured immediately thereafter. This approach was successful for complex trans‐37, and a HEP of 171.65 ppm was obtained for the P(OPh)3 ligand. However, signals due to both trans‐ and cis‐isomers were detected in the 13C NMR spectrum of 36. To resolve the carbene signal of trans‐36 in a shorter time, the NMR reaction was repeated with 13C2‐labeled complex I,5 which allowed detection of the HEP signal for P(OPr)3 at 175.25 ppm with a single scan. Unfortunately, the signals of trans‐35 complex could not be captured even within such a short time due to its very rapid trans–cis isomerization. Overall, the donating ability of the phosphites decreases in the order P(OPr)3≫P(OPh)3>P(O‐2,4‐Bu‐Ph)3 (Table 2). Generally, they are also weaker donors compared to phosphines with the same substituents. No HEP value was obtained for P(OMe)3 due to its strong transphobia. However, it is intuitive to place it in between P(OPr)3 and P(OPh)3, which is reflected in the trends of the available cis complexes (Table 3).

Table 3

Selected NMR spectroscopic data of complexes cis‐35–37, cis‐39 and cis‐40 measured in CDCl3.

Complex	Ligand	1HMe [ppm]	HEP [ppm]	2 J(C,P) [Hz]
cis‐35	P(OMe)3	1.74, 1.66	170.79	21.4
cis‐36	P(OiPr)3	1.71, 1.69	171.98	22.9
cis‐37	P(OPh)3	1.57, 1.14	169.27	22.9
cis‐39	AsPh3	1.66, 0.89	169.34	–
cis‐40	SbPh3	1.66, 1.04	167.46	–

Selected n class="Chemical">NMR spectroscopic data of complexes cis‐35–37, cis‐39 aclass="Chemical">nd cis‐40 measured iclass="Chemical">n class="Chemical">n class="Chemical">CDCl3.

n class="Chemical">1HMe [ppm]

n class="Chemical">HEP [ppm]

n class="Chemical">P(OMe)3

P(OPr)3

n class="Chemical">P(OPh)3

n class="Chemical">AsPh3

–

n class="Chemical">SbPh3

–

Similar trans–cis isomerizations were also observed for mixed n class="Chemical">NHC/class="Chemical">n class="Chemical">AsPh3 and NHC/SbPh3 complexes. The process is particularly rapid for the antimony compound, and only cis‐[PdBr2(Pr2‐bimy)(SbPh3)] (cis‐40) could be detected by all means. Complex trans‐[PdBr2(Pr2‐bimy)(AsPh3)] (39), on the other hand, isomerizes slower and could be fully characterized by direct NMR reaction, giving a HEP value of 169.17 ppm. This arsine is therefore a weaker donor compared to phosphines and phosphites. Comparison of the cis‐39 and cis‐40 could indicate that the stibine is a weaker donor than the arsine (Table 3), which is in agreement with TEP studies.29

After having obtained the n class="Chemical">carbene sigclass="Chemical">nals of the traclass="Chemical">ns‐coclass="Chemical">nfigured pclass="Chemical">nictogeclass="Chemical">n complexes 36–39, the isomerizatioclass="Chemical">n processes were purposely moclass="Chemical">nitored by class="Chemical">n class="Chemical">1H NMR spectroscopy in CDCl3. Differentiation of trans‐ versus cis‐configured products can be best achieved by analyzing the 1H NMR signals of methyl groups in the Pr2‐bimy ligand. For the symmetrical trans‐isomers, only one doublet is observed, while non‐equivalence in the cis‐isomers leads to two doublets of equal integration. In addition, the phosphite complexes can be distinguished by the 31P coupled carbene signals. The trans‐isomers exhibit 2 J(C,P) coupling constants of >280 Hz (Table 2), while <23 Hz are observed for cis‐isomers (Table 3).

The P(OPr)3 complex trans‐36 fully converted to cis‐36 within two hours. For the n class="Chemical">P(OPh)3 complex 37, oclass="Chemical">nly 17 % was coclass="Chemical">nverted to the cis‐isomer after eight days. However, the class="Chemical">n class="Chemical">1H NMR spectrum of the product mixture obtained from the initial lab‐scale reaction (30 min stirring in CH2Cl2) shows a trans–cis conversion of 50 %. Besides better mixing, the different solvent would be the key factor in influencing the isomerization rate. The increased polarity of CH2Cl2 compared to CDCl3 should favor the formation of the cis‐isomer with a larger dipole moment. Hence, by repeating the reaction in the even more polar organic solvent MeOH, cis‐[PdBr2(Pr2‐bimy){P(OPh)3}] (cis‐37) was exclusively obtained after 10 hours of stirring at ambient temperature. The P(O‐2,4‐Bu‐Ph)3 complex trans‐38 did not change at all after ten days. Heating the NMR sample in an oil bath (65 °C) for several days led to decomposition to palladium black, but no cis‐isomer was detected. The isomerization of the AsPh3 complex 39 seems to reach equilibrium after six days giving rise to a trans/cis mixture with a ratio of 1:3.5. Due to transphobia, only single crystals of the cis‐adducts could be obtained (vide infra).

Determining the HEPs of selected neutral chalcogen donors

n class="Chemical">Neutral chalcogeclass="Chemical">n class="Chemical">n class="Species">donors form weaker bonds with metal centers compared to their pnictogen counterparts. Their weak donor capability makes them suitable candidates for the design of hemilabile hybrid ligands for catalysis. For example, complexes of thioether‐functionalized NHC ligands have shown promising catalytic activities and an interesting structural dynamics.30, 31, 32, 33, 34 However, less attention has been paid to the measurement of their electron donating abilities compared to those of the other ligands. Herein, the electron donating ability of two common thioethers, that is, dimethylsulfide (DMS) and tetrahydrothiophene (THT) are assessed. We were also interested in studying the coordination chemistry and electron donating properties of analogous neutral oxygen donors. Unfortunately, dialkyl ethers are even weaker donors and were not able to cleave the dimeric complex I. Pyridine N‐oxide (PNO) as a formally neutral oxygen donor was included instead. The increased dipole moment due to charge separation leads to a stronger nucleophilic property of the oxygen atom, which may allow the cleavage reaction to proceed. The two thioethers smoothly cleaved complex I, affording complexes trans‐[PdBr2(Pr2‐bimy)(DMS)] (41) and trans‐[PdBr2(Pr2‐bimy)(THT)] (42). However, an excess of PNO was required to afford the desired complex trans‐[PdBr2(Pr2‐bimy)(PNO)] (43). The two thioethers are expected to have very similar electron donating properties. However, 13C NMR spectroscopy could still resolve the small difference as indicated by the HEP values of complexes 41 and 42 at 163.46 and 163.63 ppm, respectively. The cyclic THT ligand was found to be slightly stronger donating than the DMS ligand. The PNO complex 43 shows the most upfield HEP of 155.74 ppm indicating a weak donor property of the PNO ligand, which explains the difficulty met in the preparation of this complex.

Overall, the ligands investigated in this work cover a n class="Chemical">HEP raclass="Chemical">nge of 20 ppm (Figure 5). The class="Chemical">n class="Chemical">phosphites are among the strongest donors, but they are still weaker than their direct phosphine counterparts,8 which in turn are inferior to common NHCs in terms of electron donation.7 The HEP value of triphenylarsine is smaller than all P‐donors, but significantly larger than all N‐donors. The latter can be differentiated into sp 2‐ and sp 3‐hybridized N‐donor atoms and aromaticity of the heterocycle. Generally, sp 2‐hybridized N‐donors are stronger due to the less hindered lone pair, and aliphatic imines such as DBU and the guanidine TBD are stronger than aromatic pyridines. We also note that the donating ability of pyridines can be fine‐tuned over a range of 5 ppm. Thioethers are found in between aliphatic imines and pyridines, and pyridine‐N‐oxide is the weakest donor on the HEP scale thus far.

Figure 5

The donor strengths of selected organometallic and Werner‐type ligands on the HEP scale.

The n class="Species">donor streclass="Chemical">ngths of selected orgaclass="Chemical">noclass="Chemical">n class="Chemical">metallic and Werner‐type ligands on the HEP scale.

Solid‐state molecular structures

Single crystals of complexes 4, 9, 15⋅n class="Chemical">CH2Cl2, 19⋅2 class="Chemical">n class="Chemical">CH2Cl2, 26⋅0.5 CHCl3, 27⋅0.5 CHCl3⋅0.5 CH2Cl2, 32, cis‐35⋅CH2Cl2, cis‐36⋅CHCl3, cis‐37⋅2 CHCl3, cis‐39, cis‐40, 42 and 43⋅CHCl3 were obtained by slow evaporation of their solutions (see Supporting Information), and X‐ray diffraction reveals square‐planar geometries expected for PdII complexes. In all cases, the Pr2‐bimy ligand is perpendicular to the [PdCBr2L] coordination plane.

The structures of the n class="Chemical">pyridine adducts 4, 9, 15, 19, 26 aclass="Chemical">nd 27 closely resemble those of previously published aclass="Chemical">nalogues13 aclass="Chemical">nd do class="Chemical">not require additioclass="Chemical">nal commeclass="Chemical">nts (Figure S1). The remaiclass="Chemical">niclass="Chemical">ng complexes, oclass="Chemical">n the other haclass="Chemical">nd, are rare examples aclass="Chemical">nd their moleclass="Chemical">n class="Chemical">cular structures are depicted in Figure 6.

Figure 6

Molecular structures of 32, cis‐35⋅CH2Cl2, cis‐36⋅CHCl3, cis‐37⋅2CHCl3, cis‐39, cis‐40, 42 and 43⋅CHCl3 showing 50 % probability ellipsoids; hydrogen atoms and solvent molecules are omitted for clarity. Selected bond length [Å] and bond angles [°]: 32, Pd1−C1 1.967(4), Pd1−Br1 2.4380(7), Pd1−Br2 2.4341(7), Pd1−N3 2.075(4); C1‐Pd1‐Br1 86.3(1), C1‐Pd1‐Br2 89.2(1), Br1‐Pd1‐N3 93.4(1), Br2‐Pd1‐N3 91.2(1); PdCBr2N/NHC dihedral angle θ 88.1°. cis‐35, Pd1−C1 1.980(7), Pd1−Br1 2.4914(9), Pd1−Br2 2.472(1), Pd1−P1 2.204(2); C1‐Pd1‐Br1 84.8(2), C1‐Pd1‐P1 89.7(2), Br1‐Pd1‐Br2 94.21(3), Br2‐Pd1‐P1 91.30(6); θ 89.6°. cis‐36, Pd1−C1 1.988(3), Pd1−Br1 2.4853(4), Pd1−Br2 2.4779(4), Pd1−P1 2.2133(8); C1‐Pd1‐Br2 85.90(8), C1‐Pd1‐P1 90.41(8), Br1‐Pd1‐Br2 92.22(2), Br1‐Pd1‐P1 91.33(2); θ 83.8°. cis‐37, Pd1−C1 1.992(4), Pd1−Br1 2.4751(5), Pd1−Br2 2.4767(5), Pd1−P1 2.204(1); C1‐Pd1‐Br1 88.1(1), C1‐Pd1‐P1 90.4(1), Br1‐Pd1‐Br2 94.64(2), Br2‐Pd1‐P1 86.88(3); θ 82.8°. cis‐39, Pd1−C1 1.970(2), Pd1−Br1 2.4818(3), Pd1−Br2 2.4624(3), Pd1−As1 2.3568(3); C1‐Pd1‐Br2 85.44(6), C1‐Pd1‐As1 90.40(6), Br1‐Pd1‐Br2 93.92(1), Br1‐Pd1‐As1 90.09(1); θ 88.4°. cis‐40, Pd1−C1 1.971(3), Pd1−Br1 2.4699(4), Pd1−Br2 2.4771(4), Pd1−Sb1 2.4967(3); C1‐Pd1‐Br1 86.44(9), C1‐Pd1‐Sb1 95.66(9), Br1‐Pd1‐Br2 95.55(1), Br2‐Pd1‐Sb1 82.49(1); θ 84.2°. 42, Pd1−C1 1.969(3), Pd1−Br1 2.4286(4), Pd1−Br2 2.4342(4), Pd1−S1 2.3784(7); C1‐Pd1‐Br1 86.85(8), C1‐Pd1‐Br2 87.63(8), Br1‐Pd1‐S1 87.13(2), Br2‐Pd1‐S1 98.76(2); θ 87.1°. 43, Pd1−C1 1.935(2), Pd1−Br1 2.4213(3), Pd1−Br2 2.4201(3), Pd1−O1 2.113(2); C1‐Pd1‐Br1 88.42(6), C1‐Pd1‐Br2 86.46(6), Br1‐Pd1‐O1 88.86(4), Br2‐Pd1‐O1 96.39(4); θ 88.2°.

Molen class="Chemical">cular structures of 32, cis‐35⋅class="Chemical">n class="Chemical">CH2Cl2, cis‐36⋅CHCl3, cis‐37⋅2CHCl3, cis‐39, cis‐40, 42 and 43⋅CHCl3 showing 50 % probability ellipsoids; hydrogen atoms and solvent molecules are omitted for clarity. Selected bond length [Å] and bond angles [°]: 32, Pd1−C1 1.967(4), Pd1−Br1 2.4380(7), Pd1−Br2 2.4341(7), Pd1−N3 2.075(4); C1‐Pd1‐Br1 86.3(1), C1‐Pd1‐Br2 89.2(1), Br1‐Pd1‐N3 93.4(1), Br2‐Pd1‐N3 91.2(1); PdCBr2N/NHC dihedral angle θ 88.1°. cis‐35, Pd1−C1 1.980(7), Pd1−Br1 2.4914(9), Pd1−Br2 2.472(1), Pd1−P1 2.204(2); C1‐Pd1‐Br1 84.8(2), C1‐Pd1‐P1 89.7(2), Br1‐Pd1‐Br2 94.21(3), Br2‐Pd1‐P1 91.30(6); θ 89.6°. cis‐36, Pd1−C1 1.988(3), Pd1−Br1 2.4853(4), Pd1−Br2 2.4779(4), Pd1−P1 2.2133(8); C1‐Pd1‐Br2 85.90(8), C1‐Pd1‐P1 90.41(8), Br1‐Pd1‐Br2 92.22(2), Br1‐Pd1‐P1 91.33(2); θ 83.8°. cis‐37, Pd1−C1 1.992(4), Pd1−Br1 2.4751(5), Pd1−Br2 2.4767(5), Pd1−P1 2.204(1); C1‐Pd1‐Br1 88.1(1), C1‐Pd1‐P1 90.4(1), Br1‐Pd1‐Br2 94.64(2), Br2‐Pd1‐P1 86.88(3); θ 82.8°. cis‐39, Pd1−C1 1.970(2), Pd1−Br1 2.4818(3), Pd1−Br2 2.4624(3), Pd1−As1 2.3568(3); C1‐Pd1‐Br2 85.44(6), C1‐Pd1‐As1 90.40(6), Br1‐Pd1‐Br2 93.92(1), Br1‐Pd1‐As1 90.09(1); θ 88.4°. cis‐40, Pd1−C1 1.971(3), Pd1−Br1 2.4699(4), Pd1−Br2 2.4771(4), Pd1−Sb1 2.4967(3); C1‐Pd1‐Br1 86.44(9), C1‐Pd1‐Sb1 95.66(9), Br1‐Pd1‐Br2 95.55(1), Br2‐Pd1‐Sb1 82.49(1); θ 84.2°. 42, Pd1−C1 1.969(3), Pd1−Br1 2.4286(4), Pd1−Br2 2.4342(4), Pd1−S1 2.3784(7); C1‐Pd1‐Br1 86.85(8), C1‐Pd1‐Br2 87.63(8), Br1‐Pd1‐S1 87.13(2), Br2‐Pd1‐S1 98.76(2); θ 87.1°. 43, Pd1−C1 1.935(2), Pd1−Br1 2.4213(3), Pd1−Br2 2.4201(3), Pd1−O1 2.113(2); C1‐Pd1‐Br1 88.42(6), C1‐Pd1‐Br2 86.46(6), Br1‐Pd1‐O1 88.86(4), Br2‐Pd1‐O1 96.39(4); θ 88.2°.

Compound 32 is the first mixed n class="Chemical">NHC/class="Chemical">n class="Chemical">guanidine complex. The N‐heterocycle is also oriented perpendicular to the coordination plane and coordinates via the imine instead of the two amine moieties, confirming once again the superior donating ability of sp 2‐N‐donors over sp 3‐hybridized counterparts. As common for N‐ligands, a trans arrangement is observed making them ideal candidates for HEP analyses.

In contrast, only cis‐isomers of the n class="Chemical">phosphite adducts 35–37 crystallized as a coclass="Chemical">nsequeclass="Chemical">nce of rapid traclass="Chemical">ns–cis isomerizatioclass="Chemical">n due to traclass="Chemical">nsphobia. Previously reported mixed class="Chemical">n class="Chemical">NHC/phosphite complexes all have a trans arrangement, since they contain the very bulky NHCs IPr and SIPr.25 Compared to these, the averaged Pd–C separation of 1.987 Å in our cis complexes is significantly shorter indicating stronger bonds. They are also identical within error margin. The cis‐configuration also leads to shorter and stronger Pd−P bonds with an average length of 2.207 Å.

Similar observations were made for the n class="Chemical">arsine aclass="Chemical">nd class="Chemical">n class="Chemical">stibine complexes cis‐39 and cis‐40, which are the first structurally characterized cis‐isomers. The Pd−C bonds of 1.970(2) and 1.971(3) are identical and equal to that observed for the cis‐[PdBr2(Pr2‐bimy)(PPh3)]12 complex within 3σ. On the other hand, bonds to palladium steadily elongate going from phosphorus {2.2624(8) Å}12 to arsenic {2.3568(3) Å} to antimony {2.4967(3) Å}, which is due to the increasing atom size going down the group. All these bonds are significantly shorter and stronger than those observed in related trans complexes.35

Complexes 42 and 43 are also the first of their kind. The n class="Chemical">thioether aclass="Chemical">nd class="Chemical">n class="Chemical">N‐oxide donors adopt preferably a trans orientation with respect to the NHC. The Pd–C distance for the thioether complex of 1.969(3) Å is substantially longer than that in the pyridine‐N‐oxide complex Pd1‐C1 1.935(2) Å, which is reflective of their different donor strengths. The latter is equal to that observed for complex formed with acetonitrile,12 which is also a very weak donor.

Conclusions

The study of 26 complex probes of the type trans‐[n class="Chemical">PdBr2(Pr2‐bimy)(Py‐R)] usiclass="Chemical">ng the Huyclass="Chemical">nh electroclass="Chemical">nic parameter (class="Chemical">n class="Chemical">HEP) have revealed remote substituent effects in meta‐ and para‐substituted pyridines five and six bonds away from the NMR probe. Moreover, the influence of these substituents on the 13Ccarbene NMR signal of the Pr2‐bimy reporter ligand shows an excellent correlation to their respective sigma Hammett constants (σ m and σ p), providing further support for the suitability of HEP as an efficient and accurate electronic parameter. The resulting regression equations could also be used for the re‐evaluation of certain Hammett constants that carry larger uncertainties. This has been demonstrated for the CO2H and CO2M (M=alkali metal) substituents. In addition HEPs of meta‐ and para‐substituted pyridines bearing the same substituent also showed an excellent linear correlation. In principle, this allows for HEP predictions of para‐substituted pyridines from that of the respective meta‐isomer and vice versa. Finally, attempts were made to extend the library of HEP values to various ligands that hitherto have not been covered yet. In particular, these include phosphites, arsines, stibines, thioethers and N‐oxides. The mixed NHC/ligand complexes of the pnictogens show an interesting and electronically‐driven trans–cis isomerization. The so‐called “transphobia” effect previously described for phosphorous ligands have thus been extended to include arsenic and antimony compounds. In addition to reporting new types of complexes, this work also expands the toolbox for chemists and provides valuable electronic data for the proper comparison of different ligand classes on a unified scale.

Supporting Information

Supporting Information for this article is given via a link at the end of the don class="Chemical">cumeclass="Chemical">nt. The Supporticlass="Chemical">ng Iclass="Chemical">nformatioclass="Chemical">n coclass="Chemical">ntaiclass="Chemical">ns a detailed descriptioclass="Chemical">n of the syclass="Chemical">nthetic work, the characterizatioclass="Chemical">n of all aclass="Chemical">nalytes as well as selected crystallographic data.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.