Reactivity of δ‐Functionalized Para‐Quinone Methides in Nucleophilic Addition Reactions

Abstract

The electrophilic reactivities of para‐quinone methides (pQMs) with functional groups (FG) at the exocyclic polarized carbon–carbon double bond were determined by photometrically monitoring the kinetics of their reactions with carbanions in dimethyl sulfoxide (DMSO) at 20 °C. The experimental second‐order rate constants k ₂ were evaluated by the Mayr‐Patz equation, that is, the linear free energy relationship lg k ₂ = s _N(N + E), which was leveraged to determine the electrophilicity descriptors E of the pQMs. These electrophilicity parameters E were subsequently used to successfully predict the scope of the pQM reactions with C‐, H‐, N‐, O‐, and S‐centered nucleophiles. Moreover, the electrophilicity parameters E correlate linearly with a linear combination of quantum‐chemically calculated methyl anion affinities (MAAs) and buried volumes (%V _bur). While MAA values mainly reflect the thermodynamic driving force of the carbon–carbon bond formation, %V _bur values take account of the variable steric effects of substituents at the electrophilic δ‐position of the pQMs. Knowledge of MAA and %V _bur thus enables chemists to tailor novel pQMs with predictable reactivity properties.

The reactivities of δ‐functionalized para‐quinone methides (pQMs) were characterized by kinetic studies of their reactions with carbanions in DMSO. Evaluating the second‐order rate constants (k ₂) by the Mayr‐Patz equation gave the electrophilicities E of pQMs. Reactivities of unseen δ‐functionalized pQMs can be predicted by using the linear relationship of E with a combination of quantum chemically calculated methyl anion affinities (MAA) and buried volumes (%V _bur).

1. Introduction

Para‐quinone methides (pQMs) are a subclass of cyclic Michael acceptors, in which an exocyclic methylene group is in conjugation with an α,β‐unsaturated ketone.^{[ 1 ]} This combination of functionalities results in a strong polarization of the exocyclic π‐bond and a high reactivity toward nucleophiles at the δ‐position (Scheme 1a).^{[ 2 ]} Besides the significant role of pQMs in biological processes^{[ 3} , ⁴ , ^{5 ]} they are frequently utilized in synthetic transformations where pQMs undergo 1,6‐conjugate nucleophilic additions^{[ 6} , ⁷ , ⁸ , ^{9 ]} or a multitude of ring‐forming reactions.^{[ 10} , ¹¹ , ¹² , ^{13 ]}

Scheme 1.

a) Nucleophiles attack δ‐functionalized para‐quinone methides (δ‐FG‐pQMs) at the polarized exocyclic π‐bond. b) Structures of δ‐FG‐pQM electrophiles and reference nucleophiles used in this work (counterion: K⁺ or K⁺/18‐crown‐6). Nucleophilicity parameters N (and s _N) refer to reactivities in DMSO; refs. [20, 26].

Owing to their relevance as vinylogous Michael acceptors, several attempts were made to characterize their electrophilicity, in particular by pH‐dependent kinetic measurements in aqueous solution.^{[ 14} , ¹⁵ , ¹⁶ , ¹⁷ , ^{18 ]} Furthermore, δ‐aryl‐substituted pQMs (1 with FG = aryl, such as 1 g in Scheme 1b) resemble structural analogues of benzhydrylium ions and were, therefore, used by H. Mayr and coworkers as reference electrophiles to extend their comprehensive reactivity scales toward highly reactive nucleophiles.^{[ 19} , ²⁰ , ^{21 ]}

The Mayr reactivity scales are based on the linear free energy relationship in Equation (1), that uses three parameters to calculate the second‐order rate constant k ₂ of a given electrophile‐nucleophile reaction at 20 °C.^{[ 22} , ²³ , ²⁴ , ^{25 ]}

$$\lg k_2=s_{\mathrm{N}}\left(N+E\right)$$ (1)

Nucleophiles are characterized by two solvent‐dependent parameters, s _N and N. The reactivity of electrophiles is described by a single electrophilicity parameter E. So far, Equation (1) has been utilized to characterize the reactivity of > 1300 nucleophiles and 350 electrophiles, including aryl‐substituted ortho‐ and para‐quinone methides.^{[ 26 ]}

Though syntheses and spectroscopic characterization of simple δ‐functional group‐substituted para‐quinone methides (δ‐FG‐pQMs) 1 (Scheme 1) have been reported since the 1960s,^{[ 27} , ²⁸ , ²⁹ , ^{30 ]} systematic reactivity studies of these versatile electrophiles with synthetically relevant nucleophiles in organic solvents are still scarce.^{[ 18} , ²⁷ , ³¹ , ^{32 ]}

This lack of knowledge is currently retarding a more comprehensive exploitation of the synthetic potential of δ‐FG‐pQMs and motivated us to quantify their electrophilicity by studying the kinetics of their reactions with carbanions 2 as reference nucleophiles. The experimentally determined rate constants k ₂ of these model reactions provide the fundament for embedding δ‐FG‐pQMs 1 in Mayr's reactivity scales. The location of δ‐FG‐pQMs 1 in the electrophilicity scale will then provide a powerful tool to enhance the synthetic space of these electrophiles by the straightforward identification of novel nucleophilic reaction partners. Additionally, we will show that DFT calculations can be efficiently used to reliably predict the electrophilicities of further δ‐FG‐pQM derivatives.

2. Results and Discussion

2.1. Product Studies

First, we studied the products of the reactions between pQMs 1 and the potassium salts of the carbanions 2 that we selected as potential reference nucleophiles for the kinetic measurements. To do so, we generated the parent δ‐FG‐pQM 1a (FG = H) by oxidation of 3,5‐di‐tert‐butyl‐4‐hydroxytoluene (BHT) with silver(I) oxide in tetrachloromethane as originally reported by Winstein.^{[ 27} , ^{30 ]} The NMR spectroscopic analysis showed that BHT was quantitatively converted into the pQM 1a within 20 minutes under Winstein's conditions.^{[ 33 ]} After filtration from solids, the thus obtained CCl₄ solution of 1a was mixed with a DMSO solution of potassium diethyl malonate (2j). These reaction conditions did not furnish simple Michael adducts, however, but a mixture of the spirocyclopropane 3 and the bis‐spiro cyclopentane 4. After aqueous workup and separation by chromatography, 3 and 4 were isolated in yields of 31% and 51%, respectively (Scheme 2a). Analysis of 3 and 4 by single crystal X‐ray diffraction (scXRD) confirmed the structural assignments (Figure 1a,b),^{[ 34 ]} which were derived from the NMR spectra for both reaction products. We rationalized the formation of both 3 and 4 by initial Michael additions followed by ring‐forming reactions involving radical intermediates.^{[ 35} , ^{36 ]} The ring systems in the solid‐state structures of both 3 and 4 are characterized by one C─C bond which is significantly longer than the average of the other C─C bond lengths in the same ring. In the three‐membered ring of 3, d(C1‐C3) = 1.5711(13) Å is longer than d(C1‐C2) = 1.4913(13) Å or d(C2‐C3) = 1.5217(13) Å.^{[ 34 ]} Even more obvious, d(C1‐C2) = 1.6044(17) Å in the cyclopentane ring of 4 deviates from the lengths of the other C─C bonds in the same ring, which are in a narrow range from 1.5479(19) to 1.5552(19) Å.^{[ 34 ]}

Scheme 2.

Reactions of 1a with the nucleophile 2j: depending on the solvent used to preform pQM 1a either a) cyclic products 3 and 4 from Michael addition/oxidative radical cyclization sequences or b) Michael adducts 5 and 6 were isolated (conditions: 1 hour, r.t.).

Figure 1.

Crystalline products from the reactions of the pQM 1a with potassium diethyl malonate (2j): scXRD structures of a) the cyclopropane 3, b) the cyclopentane 4, and c) the double Michael adduct 6. Thermal ellipsoids are shown on the 25% probability level (at 173 K).^{[ 34 ].}

A screening of conditions for the generation of pQM 1a by oxidation of BHT with silver(I) oxide identified pentane as a solvent, which made it possible to carry out the Michael reaction of 1a and 2j without competing intramolecular radical cyclizations. As shown in Scheme 2b, the 1:1 Michael adduct 5 as well as the 2:1 Michael adduct 6 were isolated (Figure 1c).^{[ 36 ]} Thus, polar reactions of the parent 1a with nucleophiles could be carried out without being disturbed by electron transfer reactions, presumably induced by traces of oxidizing metal ions.

A rapid fading of the colored pQMs 1 ^{[ 33} , ^{37 ]} was observed when they were mixed with the carbanions 2 (counterion: K⁺) in DMSO or n‐pentane/DMSO solvent mixtures. The reaction mixtures were worked up and purified by chromatography to isolate the Michael adducts in unoptimized yields of 60–98% (Scheme 3). For example, pQM 1a (FG = H) and 1e (FG = Me) reacted with carbanions 2 to furnish after aqueous workup in good yields the simple 1,6‐addition products 7–11, which were spectroscopically characterized. In CDCl₃ solution, 9 was characterized as a 1:1 keto‐enol mixture. In crystalline state, however, the scXRD analysis of 9 (Scheme 3d) showed exclusively the enol form of the 1,3‐dicarbonyl groups of the dimedone moiety.

Scheme 3.

Michael additions of carbanionic nucleophiles 2 (counterion: K⁺) to δ‐substituted pQMs 1a and 1e (a), 1c (b), and 1d (c) in DMSO or pentane/DMSO mixtures (1 hour at r.t.). The quinone methides 1 were generated from the corresponding phenols by oxidation with silver(I) oxide in n‐pentane.^{[ 33 ]} Yields refer to isolated products after aqueous workup. Single crystal structures with thermal ellipsoids on the 50% probability level (at 173 K) are shown for 9 (d), 12 (e), and 13 (f).^{[ 34 ]} [a] Owing to a high degree of disorder in crystalline 14 only low‐quality scXRD data were obtained.^{[ 33} , ^{34 ].}

Also, the reactions of pQM 1c (FG = CN) with deprotonated Meldrum's acid 2a as well as with the malononitrile‐derived carbanion 2h generated the Michael adducts 12 and 13, respectively, in decent yields of 79% and 86%. In agreement with the structural data for 7, NMR spectroscopic data and scXRD analysis of 12 (Scheme 3e) showed the bis‐lactone structure of the Meldrum's acid moiety. Indications for the alternative enol forms were not found. Phenol 13 carries three adjacent nitrile groups in the side chain at the 4‐position (Scheme 3b,f). The analogous reaction of pQM 1d (FG = CO₂Me) with the diethyl malonate‐derived nucleophile 2j yielded a 1,6‐addition product with three ester groups at the side chain of the phenol 14 (Scheme 3c).

Scheme 4 illustrates that reactions of the δ‐methoxy‐pQM 1f with carbanions took another course. Additions of the C‐nucleophiles 2e and 2j to 1f were accompanied by subsequent methanol eliminations to give the benzylidene pentan‐2,4‐dione 15 and the benzylidene malonate 16, respectively. Nevertheless, formation of both 15 and 16 is rationalized by an initial σ‐bond formation through an electrophile‐nucleophile combination. The similar Brønsted basicities of the phenolate oxygen (pK _a = 16.8 for 2,6‐di‐tert‐butylphenol in DMSO) and the acceptor‐stabilized carbanions which emerge from the acetylacetone or malonate part of the adducts (pK _a = 15.1 for 2‐methylacetylacetone; pK _a = 18.0 for dimethyl 2‐methylmalonate in DMSO)^{[ 38 ]} facilitate a subsequent proton shift from the CH acid to the phenolate oxygen to give carbanions, which then eliminate methoxide ions to furnish the isolated benzylidenes 15 and 16, respectively. An analogous course was described by Tsuri and colleagues for the reaction of 1f with α‐lithiated N‐ethyl‐γ‐sultam, which gave the corresponding benzylidene compound considered as a drug candidate for the treatment of arthritis.^{[ 39 ]}

Scheme 4.

The δ‐methoxy‐substituted pQM 1f reacted with the carbanions 2e and 2j (counterion: K⁺) in an addition‐elimination sequence to form 15 and 16, respectively. Yields refer to isolated products after aqueous workup.

2.2. Kinetics

The product studies showed that formation of all isolated products can be rationalized by electrophile‐nucleophile combinations of 1 and 2, in which one new σ‐bond is formed in the first step of the reaction. Thus, a crucial prerequisite for the application of the Mayr‐Patz equation (Equation (1)) was fulfilled and we, therefore, set out to characterize the electrophilicity of the δ‐FG‐pQMs 1 by kinetic methods.

Figure 2 exemplifies the general procedure for determining the kinetics of the 1 + 2 reactions. The kinetics of the reactions of pQMs 1 with carbanions 2 in DMSO at 20 °C were monitored with stopped‐flow and conventional UV‐Vis spectroscopy by following the decay of the absorbance of the colored pQMs 1. Initial concentrations of the colorless carbanions 2 were at least 10‐fold higher than the initial concentrations of the pQMs 1 to fulfill pseudo first‐order reaction conditions (Figure 2a). Thus, first‐order rate constants k _obs (s⁻¹) were determined by least squares fitting of the mono‐exponential decay function A_t = A ₀ exp(−k _obs t) + C to the time‐dependent absorbances A_t during the reaction of 1 with 2 (Figure 2b). The second‐order rate constants k ₂ ^exptl (M⁻¹ s⁻¹) were then calculated as the slopes of the linear correlations of k _obs with the carbanion concentrations, as exemplified in Figure 2c. Table 1 lists the second‐order rate constants k ₂ of all investigated electrophile‐nucleophile combinations (see Supporting Information and ref. [40] for details of the kinetic measurements).

Figure 2.

a) Carbon–carbon bond‐forming reaction of an electrophilic pQM 1 with a C‐centered nucleophile 2 in DMSO at 20 °C. b) The time‐dependent decay of the absorbance of 1a ([1a]₀ = 2.50 × 10⁻⁵ M) at 310 nm in the course of the reaction with 2a ([2a]₀ = 2.40 × 10⁻³ M) was used to determine the first‐order rate constant k _obs (s⁻¹). c) The slope of the linear correlation of k _obs with [2a] corresponds to the second‐order rate constant k ₂ ^exp (M⁻¹ s⁻¹) for the 1a + 2a addition reaction.

Table 1.

Experimental and calculated second‐order rate constants k ₂ of the reactions of pQMs 1 with the reference nucleophiles 2 (DMSO, 20 °C).

δ‐FG‐pQM	2	k 2 exptl [M−1 s−1]	k 2 Eq.(1) [a] [M−1 s−1]	k 2 exptl/k 2 Eq.(1)
1a (FG = H)	2a	1.17 × 102	2.27 × 102	0.51
E = −11.17	2b	1.73 × 103	7.16 × 102	2.4
	2c	6.78 × 103	8.45 × 103	0.80
1b (FG = CF3)	2f	7.92 × 104 [b]	1.11 × 105	0.71
E = −11.68 [b]	2h	1.03 × 105 [b]	1.40 × 105	0.74
	2i	2.73 × 105 [b]	2.09 × 105	1.3
	2j	5.38 × 105 [b]	3.56 × 105	1.5
1c (FG = CN)	2a	2.29 × 101	5.57 × 101	0.41
E = −11.88	2c	4.98 × 103	2.40 × 103	2.1
	2d	2.68 × 103	4.24 × 103	0.63
	2h	8.90 × 104	1.03 × 105	0.87
	2i	3.71 × 105	1.53 × 105	2.4
1d (FG = CO2Me)	2c	2.22 × 103	1.91 × 103	1.2
E = −12.01	2e	9.55 × 103	1.29 × 104	0.74
	2g	7.58 × 104	5.00 × 104	1.5
	2h	2.53 × 104	8.40 × 104	0.30
	2i	1.04 × 105	1.26 × 105	0.83
	2j	6.68 × 105	2.17 × 105	3.1
1e (FG = Me)	2c	2.73 × 102	9.87 × 101	2.8
E = −13.68	2g	3.52 × 103	3.52 × 103	1.0
	2h	4.55 × 103	6.39 × 103	0.71
	2i	5.05 × 103	9.55 × 103	0.53
	2j	1.48 × 104	1.78 × 104	0.83
1f (FG = OMe)	2e	4.85 × 101	7.15 × 101	0.68
E = −15.10	2g	4.29 × 102	3.69 × 102	1.2
	2h	1.04 × 103	7.15 × 102	1.5
	2j	1.89 × 103	2.13 × 103	0.89
1g (FG = Ph)	2c	2.63	3.40	0.77
E = −15.58	2e	3.26 × 101	3.19 × 101	1.0
	2g	1.93 × 102	1.72 × 102	1.1
	2h	2.70 × 102	3.41 × 102	0.79
	2i	5.56 × 102	5.09 × 102	1.1
	2j	1.37 × 103	1.04 × 103	1.3

^[a] The second‐order rate constants k ₂ ^Eq.(1) were calculated by using Equation (1), reported nucleophile‐specific parameters N and s _N (from ref. [26]) as well as the electrophilicity parameter E of the pQMs 1 (this work).

^[b] From ref. [32].

2.3. Correlation Analysis

To determine the Mayr electrophilicity descriptors E of the δ‐functionalized pQMs 1 we used the second‐order rate constants k ₂ ^exptl (Table 1) and the previously reported nucleophile‐specific reactivity parameters (N and s _N) of the reference nucleophiles 2 (see Scheme 1). A least‐squares analysis to minimize Δ² as defined in Equation (2)^{[ 21 ]} by adjusting E as the only variable gave the electrophilicity E for each pQM 1.

$${\mathrm{\Delta }}^2=\mathrm{\Sigma }{\left(\lg {k_2}^{\mathrm{exptl}}-s_{\mathrm{N}}\left(N+E\right)\right)}^2$$ (2)

The ratios k ₂ ^exptl/k ₂ ^Eq.(1), which are listed in the 5^th column of Table 1, show that using Equation (1) and the reported N and s _N parameters along with the electrophilicities E determined in this work, results in a deviation of k ₂ ^Eq.(1) from k ₂ ^exptl lower than a factor of 3.3. For practical applications in organic synthesis (see below), this error margin is of sufficient precision, in particular if one considers that both electrophilicity (E) and nucleophilicity (N) scales currently cover 40 logarithmic orders of magnitude.^{[ 26 ]} Usually, an accuracy of 1 kcal mol⁻¹ (= 4.18 kJ mol⁻¹) is targeted in high‐level quantum‐chemical calculations for polar reactions in solution.^{[ 41 ]} It is noteworthy, therefore, that the Gibbs energies of activation (ΔG ^‡) calculated by using the three‐parameter Equation (1) agree within ± 2.9 kJ mol⁻¹ with the experimentally determined energetic barriers for the electrophile‐nucleophile additions in Table 1.

Figure 3 visualizes the results of the kinetic data evaluation by Equation (1) and illustrates that the experimentally determined second‐order rate constants (lg k ₂ ^exptl)/s _N correlate linearly with the nucleophilicity parameters N of the C‐centered nucleophiles 2. The slopes of these lines are enforced to unity, as required by Equation (1), and their intercepts with the abscissa [that is, (lg k ₂ ^exptl)/s _N = 0] correspond to E = −N.

Figure 3.

Plot of (lg k ₂)/s _N for the reactions of the pQMs 1a, 1c, and 1e−1g with the reference nucleophiles 2 against the nucleophilicity parameters N of 2 (DMSO, 20 °C). The slopes of the correlation lines were enforced to unity as required by Equation (1). The correlation for 1d is shown in Figure S4 (Supporting Information).

On the basis of the Mayr E parameters, we compared the reactivity of the δ‐FG‐pQMs 1 with those of analogously substituted phenylogous pQMs. Figure 4 highlights that the electrophilicity of the simpler δ‐FG‐pQMs 1 is generally 1 to 4 orders of magnitude higher than that of the δ‐aryl‐substituted pQMs. The electronic substituent effects are obviously much stronger if the functional group is directly bound to the electrophilic center than being attached at a more remote position of the phenyl ring. For methyl‐ and methoxy‐substituted pQMs 1e and 1f, moderate increases in electrophilic reactivity by two and one orders of magnitude, respectively, are observed. The pQMs 1c and 1d carrying electron‐withdrawing groups, as well as the previously characterized δ‐trifluoromethylated 1b ^{[ 32 ]} are by roughly three units on the E scale more reactive than analogously aryl‐substituted pQMs.

Figure 4.

Comparison of the electrophilicities E of δ‐FG‐pQMs 1 with those of phenylogous pQMs. [a] The E values for X = CN, CF₃, and CO₂Me were extrapolated by a Hammett correlation with data from ref. [21].

Additionally, the position of pQM 1a in Figure 4 is interesting, as it outperformed the δ‐methoxy‐pQM 1f by four orders of magnitude on the E scale. Moreover, 1a is an even stronger electrophile than the δ‐acceptor‐substituted pQMs 1b (FG = CF₃), 1c (FG = CN), and 1d (FG = CO₂Me). The sequence of δ‐FG‐pQMs 1 in Figure 4 clearly demonstrates, therefore, that electronic effects cannot be the only decisive factors that determine the reactivity of these electrophiles. In order to develop a model that reliably predicts the reactivity of δ‐FG‐pQMs 1 further dimensions have to be considered, as will be discussed below in the section on quantum‐chemical calculations.

2.4. Scope of pQM Reactions with Further Types of Nucleophiles

After quantifying the reactivity of δ‐FG‐pQMs 1 by their Mayr electrophilicities E, these reactivity descriptors can now be used to rationalize reported reactions and to predict new reactions. Figure 5 shows a combination of electrophilicity and nucleophilicity scales, in which nucleophiles and electrophiles located on the same horizontal level (E + N = 0) combine with rate constants of k ≈ 1 M⁻¹ s⁻¹ at 20 °C, corresponding to half‐reaction times of 10 seconds for 0.1 M solutions. The δ‐FG‐pQMs 1 are located in a reactivity range of −11.1 < E < −15.6. It can, therefore, be predicted that the δ‐FG‐pQMs 1 form products easily with nucleophiles that exceed a nucleophilicity of N > 7 to 10.

Figure 5.

The δ‐FG‐pQMs 1 studied in this work and further Michael acceptors are ranked according to their electrophilicity parameters E in the right‐hand part of the combined reactivity scales for nucleophiles (left‐hand side) and electrophiles.^{[ 26 ].}

It is in accord with these predictions that reactions of pQMs 1 with diazomethane (N/s _N = 10.48/0.78),^{[ 42 ]} sulfonium ylides^{[ 43 ]} (for example, SY with N/s _N = 13.95/0.69), the anion of α‐bromo‐malonate^{[ 44 ]} (N/s _N = 18.19/0.74 for α‐chloro malonate), deprotonated α‐bromo Meldrum's acid^{[ 45 ]} (N/s _N = 13.91/0.86 for the parent 2a), phosphorus nucleophiles,^{[ 46 ]} such as triphenylphosphine (N/s _N = 14.33/0.65 in CH₂Cl₂) and trimethyl phosphite (N/s _N = 9.04/0.70 in MeOH/MeCN), have been reported in the literature.^{[ 26 ]}

On a more quantitative basis, the unexpectedly high electrophilicity E of 1a is corroborated by reported first‐order rate constants for the hydrolysis^{[ 31 ]} and methanolysis^{[ 27 ]} of 1a. Table 2 shows that using the solvent nucleophilicity parameters N ₁ (and s _N) for 50/50 (v/v) water/acetonitrile and methanol^{[ 47 ]} along with E(1a) = −11.17 from this work in Equation (1) gives calculated rate constants, which agree within one order of magnitude with the experimentally determined first‐order rate constants (k ^exptl/k ^Eq.(1) = 4.2 for hydrolysis; k ^exptl/k ^Eq.(1) = 0.20 for methanolysis). In addition, also the reactivity of 1a toward hydroxide ions,^{[ 31} , ^{48 ]} that is, a negatively charged O‐nucleophile, is excellently reflected by the Mayr reactivity parameters E, N, and s _N in Equation (1) (entry 3 in Table 2).

Table 2.

Comparison of reported and predicted rate constants for reactions of O‐nucleophiles with the pQM 1a (E = −11.17).

Entry	Nucleophile	k exptl [a]	N (s N)	k Eq.(1) [b]	k exptl/k Eq.(1)
1	50W50AN [c]	1.5 × 10−5 s−1 [d]	5.05 (0.89) [e]	3.6 × 10−6 s−1	4.2 [f]
2	MeOH	9.4 × 10−5 s−1 [g]	7.54 (0.92) [e]	4.6 × 10−4 s−1	0.20 [f]
3	HO‐ (in 50W50AN [c] )	0.18 M−1 s−1 [d]	10.19 (0.62) [h]	0.25 M−1 s−1	0.72 [f]

^[a] Experimentally determined first‐order (s⁻¹) or second‐order rate constants (M⁻¹ s⁻¹) at 25 °C.

^[b] First‐order (s⁻¹) or second‐order rate constants (M⁻¹ s⁻¹) at 20 °C as predicted by Equation (1) from E, N, and s _N.

^[c] 50W50AN = 50/50 (v/v) water/acetonitrile.

^[d] At 25 °C, from ref. [31].

^[e] Reactivity parameters N ₁ and s _N as reported in ref. [47].

^[f] We considered the temperature difference, that is, 25 °C in the experiments and 20 °C as standard in Equation (1), to be negligible for this comparison.

^[g] At 25 °C, from ref. [27].

^[h] Reactivity parameters N and s _N as reported in ref. [48].

To enhance the scope of nucleophilic reaction partners for the δ‐FG‐pQMs 1, we explored further carbon‐ or heteroatom‐centered nucleophiles of different reactivity in the range from N = 11 to N = 22 (Scheme 5).

Scheme 5.

Extending the scope of reactions between δ‐FG‐pQMs 1 and carbon‐ or heteroatom‐centered nucleophiles.

The reactions of pQMs 1d and 1f with strong nucleophiles such as the N‐heterocyclic carbene IMes (17, N/s _N = 21.72/0.45 in THF),^{[ 49 ]} the anion of benzimidazole 19 (N/s _N = 19.13/0.55 in DMSO),^{[ 50 ]} or the thiophenolate 21 (N/s _N = 22.50/0.78 in DMSO)^{[ 51 ]} gave rapidly the Michael adducts 18, 20, and 22, respectively. Also, the enamine 25 (N/s _N = 14.91/0.86)^{[ 26 ]} and the primary amine 27 (N/s _N = 15.28/0.65)^{[ 26 ]} reacted with pQMs 1 within short times to give after aqueous workup the ketone 26 and the secondary amine 28, respectively. Due to the relatively high reactivity of δ‐FG‐pQMs 1, which are significantly more reactive than typical Michael acceptors such as methyl vinyl ketone, methyl acrylate, or exo‐methylene δ‐valerolactone (Figure 5), it was also possible to observe product formations from rather weak nucleophiles, such as the mild hydride donor sodium cyanoborohydride 23 (N/s _N = 11.52/0.67).^{[ 52 ]} It can therefore be concluded that the characterized electrophilicities E of δ‐FG‐pQMs 1 have proven useful to select in an informed way novel nucleophilic reaction partners for the pQMs 1.

2.5. Quantum‐Chemical Calculations

We used quantum‐chemical calculations to unravel the seemingly unsystematic δ‐FG effects on the electrophilicity of pQMs 1. Previously, the linear correlation of Mayr electrophilicities E with methyl anion affinities (MAAs)^{[ 53} , ⁵⁴ , ^{55 ]} was shown to enable the prediction of electrophilic reactivities of typical Michael acceptors, including a series of δ‐aryl‐pQMs and ortho‐quinone methides when involving the continuum solvation model (SMD) in the MAA calculations.^{[ 56} , ^{57 ]} Analogously, we calculated^{[ 58} , ^{59 ]} the MAAs as the Gibbs reaction energies ΔG _R for the 1,6‐Michael addition of a methyl anion to the δ‐FG‐pQMs 1a–1g at the SMD(DMSO)^{[ 60 ]}/B3LYP/6–311++G(3df,2pd)//B3LYP/6–31G(d,p) level of theory^{[ 61} , ⁶² , ^{63 ]} (Figure 6a).

Figure 6.

a) Definition reaction for the quantum‐chemical calculation of MAAs of pQMs 1 at the SMD(DMSO)/B3LYP/6–311++G(3df,2pd)//B3LYP/6–31G(d,p) level of theory. b) Definition of the viewing direction for the determination of buried volumes (%V _bur). c) Topographic steric maps for 1a and 1b indicate the variable steric demand at the electrophilic position in δ‐FG‐pQMs 1. d) The electrophilicity E of 1 correlates with a linear combination of MAA (in kJ mol⁻¹) and %V _bur.

The calculated MAAs (in kJ mol⁻¹) are tabulated in Figure 6d along with the electrophilicity parameters E of the δ‐FG‐pQMs 1. Attempts to establish a linear correlation of Mayr E with the MAAs of the pQMs 1 led, however, only to a relationship of inferior quality (R ² = 0.7725).^{[ 33 ]} This weak correlation deviates from previous results for a total of 52 Michael acceptors whose electrophilicities E were strongly correlated with MAA (R ² = 0.8890).^{[ 57 ]} Already the entries for the δ‐FG‐pQMs 1a–1c indicate that the experimentally determined reactivities (Mayr E), which differ by less than one order of magnitude, do not follow the trend of the thermodynamic driving force for the carbon–carbon bond formation that is reflected by the quantum‐chemically calculated MAAs, which differ by more than 20 kJ mol⁻¹ for 1a–1c.

Obviously, the steric demand of the δ‐FG at the pQMs 1 significantly affects their reactivity. We, therefore, used the optimized geometries from the quantum‐chemical MAA calculations as input data to assess buried volumes (%V _bur, Figure 6b,c), from which we expected that they are a useful measure of steric effects at the reaction centers of the δ‐FG‐pQMs 1.^{[ 64} , ⁶⁵ , ^{66 ]} Comparison of the %V _bur of 1a with that of 1b shows that the change from a δ‐H‐ to a δ‐(trifluoromethyl)‐substituted pQM enhances the buried volume from 38% to 58%, which counteracts the Lewis acidity (MAA) that is significantly higher for 1b than for 1a (Figure 6d).

Thus, electronic and steric effects as expressed by MAA and %V _bur allow the qualitative interpretation of the electrophilic reactivity ordering of δ‐FG‐pQMs 1. Also, quantitative predictions become possible though we have to admit that the use of only two parameters might be an oversimplification. Yet, experimentally determined electrophilicity parameters E for the 2,6‐di‐tert.‐butyl‐substituted pQMs showed an excellent linear relationship (R ² = 0.9271) with a linear combination of MAAs and %V _bur. The graph in Figure 6d will, therefore, also be significantly helpful for predicting the reactivity of further δ‐FG‐pQMs, for which kinetic data are currently not available.

3. Conclusion

In summary, we quantified the Mayr electrophilicity parameters E of δ‐functional group substituted para‐quinone methides through following the kinetics of their reactions with carbanions as reference nucleophiles in DMSO. Product studies support that simple electrophile‐nucleophile reactions gave rise to the photometrically observed decay of the absorbance of the quinone methides. We demonstrate that embedding and locating δ‐FG‐pQMs in Mayr's reactivity scales facilitates the informed selection of novel nucleophilic reaction partners. Thus, the data from this work will make it possible to systematically enhance the synthetic scope of δ‐FG‐pQMs toward currently unseen electrophile‐nucleophile combinations.

Variation of FG within the series of δ‐FG‐pQMs gave rise, however, to a relative reactivity ordering that could not be predicted straightforwardly. Counterintuitively, the simple 2,6‐di‐tert.‐butyl quinone methide 1a (FG = H) was found to be a stronger electrophile than structurally analogous quinone methides with electron‐withdrawing groups at the electrophilic center (δ‐position). Quantum‐chemically calculated methyl anion affinities and buried volumes (%V _bur) were, therefore, used to rationalize the relative reactivities of the studied δ‐FG‐pQMs. We show that the electrophilicities E of δ‐FG‐pQMs correlate linearly with a linear combination of methyl anion affinities and buried volumes (%V _bur). Thus, we established a simple relationship that will support the tailored development of additional novel pQMs with predictable properties.

4. Experimental Section

Chemicals

Supporting Information contains procedures for the preparation of δ‐FG‐pQMs 1 and the details for the reactions of 1 with anionic and neutral nucleophiles, which led to the isolated products 3–16, 18, 20, 22, 24, 26, and 28.^{[ 34 ]}

Kinetics

The kinetics of the reactions of the δ‐FG‐pQMs 1 with the (reference) nucleophiles 2 in DMSO at 20 °C were followed by using conventional photometric or UV‐Vis stopped‐flow techniques. Details of the kinetic experiments are given in the Supporting Information and in ref. [40].

Quantum‐Chemical Calculations

Details are reported in the Supporting Information.

Supporting Information

The authors have cited additional references within the Supporting Information.^{[ 67} , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ^{80 ]}

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Data Availability Statement

The raw data of kinetic measurements that support the findings of this study are openly available in Open Data LMU at DOI: 10.5282/ubm/data.582, ref. [40]. Further data available in article supplementary information.