Nucleophilicity parameters for strong nucleophiles in dimethyl sulfoxide. A direct alternative to the s(E + N) equation T. William Bentley

To cite this version:

T. William Bentley. Nucleophilicity parameters for strong nucleophiles in dimethyl sulfoxide. A direct alternative to the s(E + N) equation. Journal of Physical Organic Chemistry, Wiley, 2011, 24 (4), pp.282. ￿10.1002/poc.1747￿. ￿hal-00625939￿

HAL Id: hal-00625939 https://hal.archives-ouvertes.fr/hal-00625939 Submitted on 23 Sep 2011

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Journal of Physical Organic Chemistry

Nucleophilicity parameters for strong nucleophiles in dimethyl sulfoxide. A direct alternative to the s(E + N) equation For Peer Review

Journal: Journal of Physical Organic Chemistry

Manuscript ID: POC-10-0063.R2

Wiley - Manuscript type: Research Article

Date Submitted by the 30-Apr-2010 Author:

Complete List of Authors: Bentley, T.; UW Swansea, Chemistry

nucleophilicity scales, Swain-Scott equation, nucleophiles, Keywords: benzhydrylium cations, quinone methides

http://mc.manuscriptcentral.com/poc Page 1 of 35 Journal of Physical Organic Chemistry

1 2 3 4 Nucleophilicity parameters for strong nucleophiles in dimethyl 5 6 sulfoxide. A direct alternative to the s(E + N) equation 7 8 9 10 11 12 T. William Bentley* 13 14 15 16 17 18 A scale of nucleophilicityFor Peer(N′′′) for relatively Review strong nucleophiles (e.g. carbanions 19 20 21 and amines), spanning over 5 orders of magnitude was constructed directly from 22 23 experimental rate constants for reactions of 34 nucleophiles with a benzhydrylium 24 25 + cation, (lil)2CH (log k = N′′′ in dimethyl sulfoxide at 20 ˚C). The equation log k = 26 27 28 (E′′′ + sEN′′′ +c), where E′′′ is an electrophilicity parameter, sE is a Swain-Scott type 29 30 of response parameter (to variation in nucleophilicity) for the electrophile, and c is a 31 32 residuals term, is used to correlate second order rate constants for reactions of 33 34 35 nucleophiles with other benzhydrylium cations, quinone methides (QM) and 36 37 Michael-acceptor electrophiles. Contrary to published claims (Mayr et al. Angew 38 39 40 Chem. Int. Edn. 2002, 41, 92, and later work), sE increases as the reactivity of the 41 42 QM decreases. The N′′′ scale was extended a further 3 orders of magnitude by an 43 44 extrapolation involving a QM electrophile. In contrast, published procedures 45 46 47 involve over 40 reference electrophiles and over 100 adjustable parameters obtained 48 49 from the equation log k = sN(E + N), where k is the rate constant, and sN is a 50 51 nucleophile parameter. Values of N – N′′′ in DMSO increase by 8 log units, as the 52 53 54 reactivity of the nucleophile increases, because N is a floating scale whereas N′′′ is a 55 56 fixed scale. 57 58 59 60 1

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 2 of 35

1 2 3 * Chemistry Unit, Grove Building, Swansea University, Singleton Park, Swansea SA2 4 5 6 8PP, Wales, UK 7 8 Email: [email protected] 9 10 11 12 13 KEYWORDS: nucleophilicity scales, Swain-Scott equation, nucleophiles, 14 15 benzhydrylium cations, quinone methides 16 17 18 For Peer Review 19 20 INTRODUCTION 21 22 23 24 25 26 The general equation (1) has recently been proposed to correlate logarithms of rate 27 28 constants (log k) at 20 ˚C for a huge range of reactions of electrophiles (electrophilicity E) 29 30 and nucleophiles (nucleophilicity N); in Eqn (1), s is referred to as a ‘nucleophile- 31 N 32 [1,2] 33 specific’ parameter and sE as an ‘electrophile-specific’ parameter. It is then assumed 34 35 that sE = 1 for many reactions, so ‘deriving’ Eqn (2) which had already been used in plots 36 37 38 of log k vs. E to evaluate values of sN (from slopes) and N (from the intercept on the 39 [3,4] 40 abscissa) for over 100 nucleophiles. 41 42 43 Originally[1-4] the symbol s was chosen for Eqn (2), but it is not the same as the s (or 44 45 [5] 46 sE) parameter in the well established Swain-Scott equation (3). Consequently sN is 47 48 more appropriate than s for Eqn (2). Equation (4) fits many cation – anion 49 50 recombinations,[6] and the absence of an response parameter is consistent with the 51 52 [1,2] 53 assumption that sE = 1. In Eqns (3) and (4), n or N+ quantify nucleophilicity, and k0 54 55 refers to a reference reaction, so relative rates are correlated. 56 57 58 59 60 2

http://mc.manuscriptcentral.com/poc Page 3 of 35 Journal of Physical Organic Chemistry

1 2 3 log k = s s (E + N) (1) 4 E N 5 6 log k = sN (E + N) (2) 7 8 log k/k0 = sn (3) 9 10 11 log k/k0 = N+ (4) 12 13 14 15 The question ‘How constant are Ritchie’s constant selectivity relationships’ was the 16 17 [7] 18 title of a recent Forpaper, concluding Peer that sEReview is constant if anomalous data for water are 19 20 excluded. Recognising this problem earlier, Ritchie proposed[6] that the reference 21 22 nucleophile (rate constant k0) be hydroxide, instead of water which was chosen in earlier 23 24 [8] [6] 25 work in line with Eqn (3). There are various N+ scales, and ideally the reference 26 27 electrophile should be specified when N+ correlations are reported. 28 29 Published work[3,4] on N scales uses values of E for a ‘basis set’ of 23 reference 30 31 32 electrophiles; the electrophilicities of seven benzhydrylium cations, ranging in values of E 33 34 from 0 to 6 were defined by rate constants for reactions with 2-methylpent-1-ene,[3] and 35 36 37 values of E down to -10 (along with N values for 38 nucleophiles) were then obtained by 38 [3] 39 a multi-parameter extrapolation procedure (MPC1, Scheme 1). Plots of log k vs. E for 40 41 individual nucleophiles then gave new values of N. This complex procedure (upper half 42 43 44 of Scheme 1) leads to a floating scale when sN ≠ 1, referring to dichloromethane (DCM) 45 46 as solvent at 20 ˚C: (i) which deviates from a fixed scale by 5 orders of magnitude; (ii) in 47 48 which none of the >200 parameters are defined by directly-determined experimental 49 50 [9] 51 data. 52 53 54 55 56 57 58 59 60 3

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 4 of 35

1 2 3 IN OUT 4 Experimental Calculated 5 Computation 6 kinetic data data 7

8 extrapolate k 9 define s = 1 E from 0 to 6 + Ar CH+ 10 2 define E = 0 11 at low temp 12 13 200 data points MPC1 38 N values 14 + 15 16 E values data for log k vs. E 16 individual plots 17 nucleophiles additional 18 For PeerN values Review 19 solvent changes from dichloromethane to DMSO

20 E values data for MPC2 of 1a - 1f 21 reaction of 22 2a - 2g (key 23 carbanions) E values of key with 1a - 1m quinone methides 24 1g - 1m 25 N values of 26 MPC3a key carbanions 27 data for various general types MPC3b 28 of electrophiles 29 MPC3c additional 30 E values 31 data for log k vs. E 32 individual 33 nucleophiles plots 34 additional 35 N values 36 37 38 39 40 Scheme 1. An outline of the data processing by multi-parameter correlations (MPC) and 41 42 log k vs. E plots, previously employed to obtain E and N values from Eqn (2); MPC1 is in 43 44 Reference [3] and MPC2 is based on References [11, 12]; there are 6 MPC3 correlations 45 46 47 (3a, 3b etc) in References [13 – 18] 48 49 50 51 An alternative design was reported recently;[9] two benzhydrylium cations were 52 53 54 chosen to define N′ and N′′ directly from experimental data (Eqns (5) and (6)), and the 55 56 57 58 59 60 4

http://mc.manuscriptcentral.com/poc Page 5 of 35 Journal of Physical Organic Chemistry

1 2 3 two fixed scales were linked by the formula N′ – N′′ = 6.6. To avoid complications from 4 5 [10] [9] 6 solvent effects, only kinetic data in DCM were evaluated. 7 8 9 10 11 log k = N′ for decay of 1n, at 20 ˚C (5) 12 13 log k = N′′ for decay of 1a, at 20 ˚C (6) 14 15 16 17 18 To furtherFor extend the Peer N scale to reactions Review of carbanions and other nucleophiles in 19 20 dimethyl sulfoxide (DMSO), six cations (1a – 1f) and seven quinone methides (QM, 1g – 21 22 1m) were chosen[11,12] as reference electrophiles (the codes 1a – 1m in Scheme 2 are the 23 24 [11] 25 same as in the key paper ). Kinetic data for reactions of these electrophiles with key 26 27 nucleophiles (Scheme 3) underwent another correlation (MPC2, Scheme 1), further 28 29 extrapolating the E scale from -10 to -18.[11] The derived N values were then the input for 30 31 [13- 32 MPC3a, MPC3b etc, characterising E values for other electrophiles from E = -8 to -23. 33 34 18] Finally, log k vs. E plots gave additional values of N for stronger nucleophiles.[19-25] 35 36 37 Using Eqn (2) the slope defines sN and intercept on the abscissa defines N, so 38 39 residual errors are incorporated into the parameters, and when sN ≠ 1 the N scale ‘floats’ 40 41 (the reference point is variable and hypothetical). Consequently, it is virtually impossible 42 43 44 to define the parameters, and revision of one input rate constant in MPC1 would affect 45 46 hundreds of other parameters. This procedure is indirect (N values are obtained from the 47 48 E scale) and requires long extrapolations, exacerbating errors inherent in a simple model. 49 50 51 A much simpler alternative design, avoiding a floating scale and requiring much 52 53 shorter extrapolations, will now be presented using the same published kinetic data.[7,11-25] 54 55 The results will complete the alternative design,[9] will lead to a N′′′ scale in DMSO (fixed 56 57 58 59 60 5

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 6 of 35

1 2 3 not floating), and will show (contrary to recent claims[1,2,7]) that the electrophile parameter 4 5 6 (sE) varies significantly. Selection of the mid-range electrophile (1a) as reference for a 7 8 unified N′′ scale (linked to N′′′ and N′) will also be discussed. 9 10 11 12 13 RESULTS 14 15 16 17 18 N′ (Eqn (5)) refersFor to the reactionsPeer of the relativelyReview reactive dianisyl cation (1n, Scheme 19 20 2), for which E = 0, giving N′ values for very weak nucleophiles (e.g. arenes).[9] The less 21 22 reactive dimethylaminobenzhydrylium cation (1a) gives N′′ (Eqn (6)) for more reactive 23 24 [9] 25 nucleophiles (e.g. enamines). N′′′ for even more reactive nucleophiles including 26 27 carbanions will now be evaluated from cation (1f) using Eqn (7) and from additional 28 29 Swain-Scott type correlations (Eqn (8)) for the auxiliary reference QM (1m). 30 31 32 33 34 log k = N′′′ for decay of 1f at 20 ˚C (7) 35 36 37 log k = E′′′ + sEN′′′ + c (8) 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 6

http://mc.manuscriptcentral.com/poc Page 7 of 35 Journal of Physical Organic Chemistry

1 2 3 X 4 + X = Y = NMe2 (dma)2CH , E = -7.02

5 1a = 6 HC + reference substrate for N 7 + 8 X = Y = N(CH2)4 (pyr)2CH , E = -7.69 9 Y 1b 10 11 NMe NMe N N 12

13

+ + + 14 HC HC HC + HC 15 16 17 NMe NMe N N

18 (thq) CH+ + + + 2 (ind)For2CH (jul) 2CPeerH (lil)2CH Review 19 E = -8.22 E = -8.76 E = -9.45 E = -10.04 20 1c 1d 1e 1f 21 _ = 22 reference substrate for N

23 H X C+ X 24 _ 25 Y O Y O 26 X X

27 1g, Y = OMe, X = Br, E = -8.63 1j, Y = Me, X = But, E = -15.83 28 1h, Y = OMe, X = Ph, E = -12.18 1k, Y = OMe, X = But, E = -16.11 29 t 1i, Y = NMe2, X = Ph, E = -13.39 1l, Y = NMe2, X = Bu , E = -17.29 30 31 But 1m, E = -17.90 32 33 N O auxiliary reference substrate But 34

35 H C+ 1n, E = 0 36 primary reference substrate 37 MeO OMe _ 38 for N and N 39 40 41 Scheme 2. Codes and electrophilicities for the 6 benzhydryl cations (1a – 1f,) and 7 42 43 quinone methides (QM, 1g – 1m), required for the multi-parameter correlation (MPC2) fit 44 45 [11] 46 to Eqn (2); E = 0 for 1n 47 48 49 50 51 52 53 54 55 56 57 58 59 60 7

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 8 of 35

1 2 3 O O O O 4 O O 5 O O

6 2a 2b 2c 7 O O O O 8 CN CN 9 OEt EtO OEt CN CO2Et 10 2d 2e 2f 2g 11 12 13 14 15 Scheme 3. The 7 key carbanions (2a – 2g) required for the multi-parameter correlations 16 17 (MPC2, MPC3a, 3b etc, Scheme 1) for Eqn (2)[11,13-18] 18 For Peer Review 19 20 21 [11] 22 The data including QMs (1g – 1m) were processed by MPC2, assuming the same 23 24 E values of cations (1a – 1f) from previous work[11] in DCM; 21 adjustable parameters (E 25 26 values of 7 quinone methides (QM, 1g – 1m) and s and N values for 7 enolates, 2a -2g, 27 N 28 [11] 29 Scheme 3) were obtained from 70 data points, (additional rate constants for 30 31 - [11] MeCHNO2 reacting with 1k. 1l and 1m were included in MPC2 ). 32 33 34 In contrast, values of N′′′ (Eqn (7)) can be obtained directly from experimental 35 36 data;[11,12] considering Eqn (8) when N′′′ = 0, log k = E′′′ + c = 0, so no adjustable 37 38 parameters are required at this stage. Values of N′′′ and N (Eqn (2)) show that N - N′′′ 39 40 [9] 41 increases as N increases (Table 1), as expected for a floating N scale. Values of N′′′ for 42 43 27 additional nucleophiles (Scheme 4) were also calculated (Table 2) from data published 44 45 later,[7,19-23] giving a total of 34 values of N′′′ from Eqn (7) in Tables 1 and 2. Correlations 46 47 48 using Eqn (8, Tables 3 and 4) require 24 adjustable parameters (two per electrophile, 49 50 excluding 1f), whereas Eqn (2) requires many more (13 + (2 x 34) = 81). 51 52 53 54 55 56 57 58 59 60 8

http://mc.manuscriptcentral.com/poc Page 9 of 35 Journal of Physical Organic Chemistry

1 2 3 4 SO2CF3 NO2 SO2Ph 5 Z Z Z 6 3 4 5 7 NO NO 8 NO2 2 2

9 H H H CH3 H3C CH3 10 6a 6b 6c 11 12 CN CN 13 Z Z O N NO 14 2 2 7 8 9 15 16 17 18 For Peer Review 19 Scheme 4. Additional carbanions (3 - 9) for which values of N′′′ are known (Tables 2 and 20 21 5) 22 23 24 25 26 The data in Table 3 were obtained using Microsoft Excel, and show slopes, 27 28 intercept, correlation coefficients (r) and number of data points (n). The slopes show that 29 30 31 sE is very close to unity, but the data vary from extensive (n = 19) for 1e, for which the 32 33 correlation is very good, to less convincing for 1a and 1b because n ≤ 5 (also, despite 34 35 their small number the range of nucleophiles is diverse (Table 3 footnotes d and f)). 36 37 38 Correlations using Eqn (8) for QMs (1g – 1m) reacting with carbanions (Table 4) 39 40 include rate constants for 1l and 1m covering less than 2.5 log units; the data show 41 42 ‘scatter’ when the additional nucleophiles are included; the distribution of data points also 43 44 45 increases uncertainties in the values of intercepts (E′′′ + c), because extrapolations are 46 47 required from typical values of N′′′ (> 3.0) to N′′′ = 0. Another indication of the reliability 48 49 50 of the correlations is the standard errors – the worst is small (0.27 for 1l). In contrast to 51 52 the data in Table 3, the results show a substantial increase in sE (selectivity) as E′′′ 53 54 (reactivity) decreases. 55 56 57 58 59 60 9

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 10 of 35

1 2 3 From the equation for 1d (Table 3), it is predicted that log k = 2.22 (log k = 4 calc obs 5 [21] 6 2.02 ) for reaction of 1d with the ylid Ph3PCHCO2Et; however, reaction of the ylid with 7 [21] 8 1h is predicted to be slower than observed (log kcalc = -1.04, log kobs = 0.03 ). 9 10 11 Eqn (9), based on the first entry for 1m in Table 4, was then applied to estimate N′′′ 12 13 values for 10 more reactive nucleophiles by extrapolation (Table 5). Provisional data for 14 15 3 other nucleophiles are estimated from correlations for other QMs (1h. 1j and 1k). 16 17 18 For Peer Review 19 20 N′′′ = (log k (for 1m) – 8.0)/1.43 (9) 21 22 23 24 25 Having obtained values of N′′′ for 34 nucleophiles from the same reference 26 27 nucleophiles and electrophiles as LLM (Schemes 2 and 3),[11] correlations using Eqn (8) 28 29 were carried out for 19 other electrophiles (Table 6); these all involve Michael additions 30 31 32 to structurally-similar substrates (10 – 18), Scheme 5. Correlations for 5 additional QMs 33 34 (Scheme 6) reacting with key carbanions (Scheme 3) are shown in Table 7. 35 36 37 Intercepts refer to N′′′ = 0, and errors for neutral substrates (Tables 6 and 7) are 38 39 relatively large because long extrapolations from N′′′ ~ 5 are often required. Calculated 40 41 values agree better than expected from the errors in individual parameters: e.g. for 42 43 44 Michael addition of 9 to 10, Z = OMe, the two entries (Table 6) for Eqn (8) predict log k 45 46 = 6.03 and 6.09, in agreement with the Eqn (2) prediction of 6.14; the experimental value 47 48 is only 5.26,[24] and as noted earlier [22,24] Michael acceptors react more slowly than 49 50 51 expected from parameters based on cations and QMs (e.g. see Table 6, footnotes h and j). 52 53 Nucleophilicity parameters for amines (Table 2) can be used to predict rate 54 55 constants for Michael additions (kinetic data for amines were not included in the 56 57 58 59 60 10

http://mc.manuscriptcentral.com/poc Page 11 of 35 Journal of Physical Organic Chemistry

1 2 3 correlations shown in Table 6). Predicted rate constants by either Eqn (2) or Eqn (8) for 4 5 [14] 6 reactions of amines with 12, Z = H agree satisfactorily with experimental values. 7 8 However, primary and secondary amines react with 16 about 100-fold faster than 9 10 [16] 11 predicted, perhaps because of transition state stabilisation by H-bridges. 12 13 14 15 CN O 16 CN 17 Z 10 N O 18 O For Peer11 Review 19 20 O Z O 21 NMe 22 12 N O N X 23 O Me 13 24 NMe O 25 Z O N O Me NMe 26 14 Me N O N S 27 O 2 Me 28 O 15 29 O Z O O 30 16 O CO Et 31 2 N O O 32 CO2Et 17 33 Z 18 34 35 36 Scheme 5. Additional electrophiles for correlations shown in Table 6 37 38 39

40 H X C+ X 41 _ O 42 Y Y O 43 X X

44 19a, Y = 3-F, X = But, E = -15.03 19d, Y = H, X = Ph, E = -11.87 45 19b, Y = 3,5-F, X = But, E = -14.50 19e, Y = 4-OMe, X = OMe, E = -16.38 46 t 19c, Y = 4-NO2, X = Bu , E = -14.36 47 48 49 Scheme 6. Additional quinone methide (QM) electrophiles for correlations shown in 50 51 Table 7; values of E from Reference [18] 52 53 54 55 56 57 58 59 60 11

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 12 of 35

1 2 3 DISCUSSION 4 5 6 Comparisons of sE with  and of N′′′ with N 7 8 9 10 11 N′′′ correlations (Eqn (8)) for over 300 rate constants in DMSO at 20 ˚C show good 12 13 correlations for benzhydrylium ions (Table 3) and QMs (Tables 4 and 7), but are less 14 15 satisfactory for Michael additions (Table 6). As for other cations,[6,8] the parameter s is 16 E 17 18 approximately constantFor (s EPeer =1.00, Table 3),Review but sE varies between 1.0 and 1.5 for QMs 19 20 (Tables 4 and 7). Nucleophilic attack on activated 9-methylenefluorenes is known to give 21 22 [26] 23 slopes > 1 in plots vs. N+ , and the tri-p-anisylmethyl cation has a lower response to 24 [27] 25 changes in N+ than less reactive cations (crystal violet or malachite green). 26 27 [28] In general sE ≠ 1.00, so Eqn (4) is a special case, and a selectivity parameter 28 29 [26,27,29,30] 30 should be incorporated into the Ritchie Eqn (4). Consequently, it is not 31 [1,2] 32 appropriate to ‘derive’ Eqn (2) by setting sE =1.00 in Eqn (1), and to then apply Eqn 33 34 (1) to other electrophiles.[11,12] 35 36 37 Despite the questionable derivation, Eqn (2) is a useful way to predict structural 38 39 effects on reactivity using the E scale of electrophilicity and the substituent effect 40 41 parameter (s ), assuming that the ‘electrophile specific’ parameter s = 1.00.[1-4,7,11-25] 42 N E 43 44 The alternative assumption that sN = 1.00 leads to a variation on the Swain-Scott equation 45 46 (Eqn (8)) and gives more reliable values for nucleophilicity parameters.[9,10] The two 47 48 49 approaches are different ways to locate a point in 3D space, and sE and sN are 50 [9,10] 51 interdependent parameters. 52 53 N′′′ correlates satisfactorily with pKa for 27 nucleophiles in DMSO (Figure 1). Two 54 55 56 correlation lines are shown: six cyano- stabilised carbanions, including arylacetonitriles 57 58 59 60 12

http://mc.manuscriptcentral.com/poc Page 13 of 35 Journal of Physical Organic Chemistry

1 2 3 (7, 8) but also 2e and 2f, fit one line; nine sulfonyl-stabilised carbanions (3, 5), and four 4 5 [35] 6 enolates (2a, 2b, 2c and 2g) fit a parallel line (such family behaviour is typical ). The 7 8 remaining eight nucleophiles (all nitro-stabilised) are close to the two correlation lines, 9 10 11 but show distinct curvature possibly due to the failure of pKa to model the kinetic 12 [35] 13 phenomena of desolvation. 14 15 16 17 18 11 For Peer Review 19 10 20 cyano family 21 9 22 8 23 N''' 7 sulfonyl family 24 nitro 25 6 26 nitro 5 27 28 4 carbonyl 29 3 30 5 10 15 pK 20 25 31 a 32 33 34 35 Figure 1. Brønsted plot (N′′′ at 20 ˚C vs. pK at 25 ˚C ) for 27 carbanions in DMSO. N′′′ 36 a 37 38 from Tables 1, 2 and 5, and pKa data from References [31-34]; separate correlation lines 39 40 with  = 0.35 ± 0.04 (r = 0.978) are shown for six cyano-stabilised carbanions (2e, 2f, 7 41 42 43 and 8), and with  = 0.35 ± 0.02 (r = 0.991) for nine sulfonyl-stabilised carbanions (3, 5) 44 45 46 47 48 The value of  = 0.35 ± 0.04 (Figure 1) for reactions of 1f with nucleophiles, is the 49 50 same within errors or slightly lower than the value of = 0.40[33] for reactions of n-butyl 51 52 53 chloride reacting with benzyl phenylsulfonyl carbanions. Both  and sE show the effect of 54 55 varying the nucleophile for a fixed electrophile, so an SN2 reaction is responding more 56 57 58 than a cation-anion recombination. A correlation of N+ (Eqn (4)) with n (Eqn (3)) in 59 60 13

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 14 of 35

1 2 3 water has a slope of 2.0[36] or 2.1.[37] So, in contrast to the results in DMSO, cation-anion 4 5 6 recombinations (N+) in water are twice as sensitive as SN2 reactions (n) to variations in 7 8 nucleophilicity. 9 10 [22] 11 For nucleophilic additions to the QM (1l) in DMSO,  = 0.58 (or ~ sE x 0.35 = 12

13 0.50), again illustrating that sE > 1 for QMs (1, 19). QM (20) may be treated as a strongly 14 15 resonance-stabilised carbocation;[36] for nucleophilic additions to 20 in water, an N 16 + 17 [36] 18 correlation has Forslope (sE) ofPeer only 0.92, fittingReview the trend that more reactive QMs have 19 20 lower values of sE (Tables 3 and7). 21 22 23 24 F C + CF 25 F3C CF3 3 3 26 27 _ 28 O O 29 20 30 31 32 33 34 When N′′′ is defined directly by Eqn (7), values of N = N -N′′′ for the key 35 36 37 carbanions (Table 1) vary by 3.1 log units, whereas the total range of values of N′′′ is only 38 39 3.2; the additional nucleophiles in Table 2 also show substantial variations in N (Table 40 41 42 2). When the extrapolated data for more reactive carbanions is included, values of N 43 44 range from 10.5 (Table 1) or 10.8 (Table 2) to 18.9 (Table 5). The floating N scale varies 45 46 from 12.15 (Table 2) to 28.95 (Table 5), and all of the correlations (MPC2, 2, 3a 3b, 3c 47 48 49 etc, Scheme 1) are needed. For the same range of nucleophiles, the fixed N′′′ scale varies 50 51 only from 1.35 (Table 2) to 10.1 (Table 5). The above calculations indicate that during 52 53 the complex sequences of processing of the kinetic data (Scheme 1), the N scale floats by 54 55 56 8 orders of magnitude away from the initial reference point (1n). 57 58 59 60 14

http://mc.manuscriptcentral.com/poc Page 15 of 35 Journal of Physical Organic Chemistry

1 2 3 It was mentioned over a decade ago that N is a floating scale,[38] but the large extent 4 5 [9] 6 of floating has only recently been highlighted. Theoretically, it should be possible to 7

8 correct from the floating N scale to a fixed scale sNN. However, this requires a high level 9 10 [39,40] 11 of confidence in a long extrapolation of complex substituent effects, and neglect of 12 13 solvent effects. In practice, sNN does not give a satisfactory trend with pKa for 14 15 [23] arylacetonitriles (Table 8), and, these carbanions also deviate from a plot of N vs. pKa. 16 17 18 in contrast to NFor′′′ (Figure 1).Peer Review 19 20 21 22 Constant selectivity plots and electrophilicity parameters 23 24 [9,10] 25 The chosen methodology (Case 1: fix E, vary N, extrapolate to estimate E) is 26 27 designed to lead to a reliable scale of nucleophilicity, with most of the errors incorporated 28 29 into the extrapolated values of E′′′ + c. In contrast, published procedures (Case 2: fix N, 30 31 [3,7,19-25] 32 vary E, extrapolate to obtain N ) lead to less reliable values of N, which include 33 34 most of the errors. Both cases are based on the assumptions that log k depends on 35 36 37 parameters for electrophilicity and nucleophilicity. Consequently, published values of E 38 39 should be closely related to the extrapolated values of E′′′ + c, as observed (Figure 2). 40 41 The slopes of the correlation lines for reactions of the key carbanions with 42 43 44 benzhydrylium ions (Table 3) and QMs (Table 4) are both 1.00 within calculated errors 45 46 (Figure 2), but there is a separation between the two lines corresponding to 0.5 units on 47 48 the x-axis. These results, coupled with the variations in s for QMs (Table 4), are 49 E 50 51 unexpected. Therefore the constant selectivity plots, which are the foundations of Eqn 52 53 (2),[41] were reinvestigated. The constant selectivity plots for QMs[12] are based on data 54 55 for only 4 QMs (1j – 1m), and slopes (based on limited data) vary from 0.81 for 2a to 56 57 58 59 60 15

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 16 of 35

1 2 3 1.12 for 2c and 2d, relative to log k for 2f as the model nucleophile for an E scale. These 4 5 [12] 6 results were shown in a graph omitting numerical values of the slopes, and the 7 8 interpretation was made that the correlation lines are ‘parallel’.[12] 9 10 11 12 13 4 14 cations 15 2 16 0 17 18 -2 For Peer Review 19 -4 indan-1,3-diones 20 + c E''' 21 -6 quinone methides excluded from MPC2 22 -8 23 quinone methides 24 -10 25 -20 -18 -16 -14 -12 -10 -8 -6 E (Mayr et al) 26 27 28 29 30 Figure 2. Plots of E (Scheme 2) vs. E′′′ + c for benzhydrylium cations (1a-1f, Table 3) 31 32 and quinone methides (1g -1m, Table 4); the slopes of the correlation lines are 0.996 ± 33 34 35 0.050 (n = 6,excluding 1a, r = 0.996) and 0.984 ± 0.034 (n = 6, r = 0.997) respectively; 36 37 data from Tables 6 and 7 are plotted for comparison (not fitted to correlation lines) 38 39 40 41 [11,12] 42 Combining two sources of experimental data leads directly (without 43 44 ‘optimisation’) to an E scale spanning 7 orders of magnitude in which QMs are for the 45 46 47 first time included on the same selectivity plots as benzhydrylium ions. The ‘E scale’ 48 49 (Figure 3) is defined by log k for 2b, the only nucleophile for which kinetic data for all 50 51 cations and QMs (1a – 1m) are available. 52 53 54 The slopes of correlation lines vary from 0.96 to 1.13 (Figure 3), are dominated by 55 56 QMs, and cations show deviations (e.g. for 2g). Despite the idea that QMs are activated 57 58 59 60 16

http://mc.manuscriptcentral.com/poc Page 17 of 35 Journal of Physical Organic Chemistry

1 2 3 benzhydrylium ions,[11,12,36] there is not a seamless connection between these reference 4 5 [11] 6 electrophiles (see also Figure 2). MPC2 is able to adjust E to enforce a fit to Eqn (2), 7 8 but this procedure does not give confidence in extrapolations needed to obtain N. 9 10 11 12 13 8 14 15 16 6 17 18 4 slopeFor 0.96 Peer Review 19 2g cations 20 k log 2 21 quinone methides 22 0 23 2c 2a slope 0.99 24 slope 1.13 25 -2 26 -2 0 2 4 6 8 log k for 2b 27 28 29 30 31 Figure 3. Log k for reactions of 2a, 2c and 2g vs. corresponding data for 2b, reacting 32 33 with cations (open symbols) and quinone methides (QMs, filled symbols) in DMSO at 20 34 35 36 ˚C; slopes (errors: ± 0.02, r = 0.999) refer only to QMs; data from References [11,12] 37 38 39 40 Sufficient data are available to permit the calculation of s and N for 4 key 41 N 42 [3,7,19-25] 43 nucleophiles (2a -2d) from standard plots of log k vs. E for benzhydrylium ions; 44 45 values of N require a long extrapolation, and are significantly different (Table 9) from 46 47 those obtained by MPC2 which includes data for QMs. Because of a leverage effect, s is 48 N 49 50 determined mainly by QMs, not benzhydrylium cations, and the differences in slopes 51 52 (Figure 3 and sN, Table 9) show that sN is not exactly ‘nucleophile specific’. Also, unlike 53 54 55 Eqn (8) which utilises a different equation to calculate parameters, the differences in 56 57 Table 9 are due solely to the exclusion of QMs from the data processing to fit Eqn (2). 58 59 60 17

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 18 of 35

1 2 3 Towards a unified N′′ scale of nucleophilicity 4 5 [6,8] 6 As expected from N+ correlations, slopes (Table 3) are close to unity for reactions of 7 8 cations (1a – 1f) with nucleophiles, and this allows connections to be made between the 9 10 11 N′, N′′ and N′′′ scales (Eqns (5 - 7)). Data for five nucleophiles (Table 3, footnote f) are 12 13 available in DMSO for both 1a and 1f, giving N′′ - N′′′ = 2.2 ± 0.2 in DMSO. However, 14 15 the scope of a unified scale for addition of nucleophiles to cations in DMSO is limited by 16 17 [25] 18 the relatively highFor nucleophilicity Peer of DMSO Review (N = 9.75 in acetonitrile ), so by this 19 20 measure DMSO is more nucleophilic than water or typical alcohols, but less nucleophilic 21 22 than amines (Table 2). Therefore, changes of solvent and consideration of solvent effects 23 24 25 on rates is also required. 26 27 Cation-anion recombinations are the reverse of SN1 reactions, and solvent effects 28 29 are likely to be significant, particularly for protic solvents.[9,42] Considering only data for 30 31 32 aprotic solvents, 22 values of N′′ - N′′′ are available in DCM, giving N′′ - N′′′ = 2.4 ± 0.4, 33 [43] 34 a slightly greater ratio than in DMSO. Reactions of the ylid Ph3PCHCO2Et, are faster 35 36 37 in DCM than in DMSO, and the kinetic effect is larger for 1a, N′′(14-fold) than 1f, N′′′(5- 38 [44] 39 fold); reaction of PPh3 with 1a is also faster (4-fold) in DCM than in DMSO. 40 41 Carbanion 3, Z = NO2 reacts about 10-fold faster in acetonitrile than DMSO, so 42 43 [20] 44 presumably reactions in DCM will also faster. 45 46 In contrast, other reactions are favoured by solvents with a higher donor number 47 48 (e.g. ethyl diazoacetate reacting with 1a[45]); also reactions of amines are faster in DMSO 49 50 51 than in acetonitrile (e.g. only 1.4 fold for piperidine and 1f, but 22-fold for CF3CH2NH2 52 53 and 1a),[7,18] so presumably reactions in DCM will be slower than in DMSO. 54 55 56 57 58 59 60 18

http://mc.manuscriptcentral.com/poc Page 19 of 35 Journal of Physical Organic Chemistry

1 2 3 It appears that corrections for solvent effects in aprotic media will be relatively 4 5 6 small, but not insignificant. No other generalisations will be attempted at this stage, 7 8 because it is not easy to predict even whether rates will increase or decrease. If as a first 9 10 [3] 11 approximation solvent effects are ignored (as proposed earlier ), connections (Scheme 7) 12 [9] 13 can be made from N′′ to N′′′ and to N′ (= N, when the N scale is correctly attached to the 14 15 fixed reference point 1n in DCM). 16 17 18 A unified,For scale (N′′) Peeris suggested, basedReview on the less reactive electrophile (1a) in 19 20 DCM at 20 ˚C as the fixed reference.[9] The N′′ scale spans over 22 orders of magnitude 21 22 from 12.3 for 8, Z = H (N′′′ = 10.1, Table 5) to -10.1 for m-xylene (N′ = -3.5[9]), so 1a is 23 24 [46] [23] [3] 25 located centrally; the corresponding N values are 28.9 and -3.54 respectively, 26 27 ranging over 32 orders of magnitude. Consequently, the N scale floats by over 10 orders 28 29 of magnitude relative to the fixed N′′ scale!! 30 31 32 33 + / 34 (ani)2CH N 35 1n 6.6 36 37 + // 38 (dma)2CH N 39 1a 40 2.2 2.4 41 /// (lil) CH+ N 42 2 1f 43 DMSO dichloromethane 44 45 46 47 48 Scheme 7. Interconversion factors (increments of log k) for N′′′ in DMSO (this work) 49 [9] 50 with N′ or N′′ in dichloromethane 51 52 53 54 55 In general, it is not necessary to use a floating scale such as Eqn (2) to design a 56 57 reactivity scale spanning 20 orders of magnitude. Three reference substrates (e.g. 1a, 1f, 58 59 60 19

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 20 of 35

1 2 3 1n) are sufficient; e.g. a unified data set for bridgehead reactivity spanning 1023 in first 4 5 6 order rate constants was established for tosylates, linked to triflates for less reactive 7 8 substrates and to halides for more reactive substrates.[47] Second order rate constants have 9 10 11 the advantage that they can be obtained from pseudo-first order rate constants, and 12

13 reactivities can be controlled by varying the concentrations of nucleophiles. 14 15 Simple additivity rules[48] for estimating orders of magnitude of rates of reactions of 16 17 18 electrophiles withFor nucleophil Peeres can be devised Review using Scheme 7: e.g. Eqn (10), based on 19 20 the N′′′ scale and the established E scale.[3,11] Values of (N′′′ + 9) for 7 and 8 are 2 - 4 21 22 orders of magnitude larger than sNN (Table 8), the hypothetical values of log k when E = 23 24 25 0. 26 27 28 29 log k ~ E + N′′′ + 9 (10) 30 31 32 33 34 CONCLUSIONS 35 36 [3,11,13-18] 37 Up to 8 multi-parameter correlations, 3 of which are consecutive (Scheme 38 39 1), are required to extend the E scale of electrophilicity from a well defined region 40 41 between E = 6 and E = 0, all the way back to E = -23.8.[17] Subsequent calculations of N 42 43 44 for strong nucleophiles give values which deviate from a fixed scale (N′′′, Eqn (8)) by up 45 46 to 8 log units (see values of N - N′′′ in Tables 1. 2 and 5), an extent of floating 47 48 49 amounting to half of the whole range of N values in DMSO. In contrast, the N′′′ scale 50 51 based on directly-defined values (Eqn (8)), can be extended by a much shorter 52 53 extrapolation (Eqn (9)). Good N′′′ correlations using Eqn (8) are observed for cations 54 55 56 57 58 59 60 20

http://mc.manuscriptcentral.com/poc Page 21 of 35 Journal of Physical Organic Chemistry

1 2 3 (Table 3) and quinone methides (Tables 4 and 7), but precision is lower for Michael 4 5 6 additions (Table 6). N′′′ also gives a satisfactory Brønsted plot (Figure 1). 7 [27-29] 8 Future work based on Eqns (6 - 8), with greater emphasis on sE, should reveal 9 10 11 solvent effects on electrophilicity for strong electron pair donor solvents such as water 12 [3,11] 13 and alcohols. Currently, sE = 1.00 is assumed incorrectly (Table 4), and solvent 14 15 effects on electrophilicity are ignored. 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 22 of 35

1 2 3 4 5 6 REFERENCES 7 8 [1] T. B. Phan, M. Breugst, H. Mayr, Angew. Chem. Int. Ed. 2006, 45, 3869-3874. 9 10 [2] H. Mayr, A. R. Ofial, J. Phys. Org. Chem. 2008, 21, 584-595. 11 12 [3] H. Mayr, T. Bug, M. F. Gotta, N. Hering, B. Irrgang, B. Janker, B. Kempf, R. Loos, 13 A. R. Ofial, G. Remennikov, H. Schimmel, J. Am. Chem. Soc. 2001, 123, 9500-9512. 14 15 [4] H. Mayr, B. Kempf, A. R. Ofial, Acc. Chem. Res. 2003, 36, 66-77. 16 17 [5] C. G. Swain, C. B. Scott, J. Am. Chem. Soc. 1953, 75, 141-147. 18 For Peer Review 19 [6] C. D. Ritchie, Can. J. Chem. 1986, 64, 2239-2250. 20 21 [7] S. Minegishi, H. Mayr, J. Am. Chem. Soc. 2003, 125, 286-295. 22 [8] C. D. Ritchie, Acc. Chem. Res. 1972, 5, 348-354. 23 24 [9] T. W. Bentley. J. Phys. Org. Chem. 2010, DOI 10.1002/poc.1670. 25 26 [10] T. W. Bentley, J. Phys. Org. Chem. 2010, 23, 30-36. 27 28 [11] R. Lucius, R. Loos, H. Mayr, Angew. Chem. Int. Ed. 2002, 41, 92-95. 29 [12] R. Lucius, H. Mayr, Angew. Chem. Int. Ed. 2000, 39, 1995-1997. 30 31 [13] T. Lemek, H. Mayr, J. Org. Chem. 2003, 68, 6880-6886. 32 33 [14] S. T. A. Berger, F. H. Seeliger, F. Hofbauer, H. Mayr, Org. Biomol. Chem. 2007, 5, 34 35 3020-3026. 36 37 [15] F. Seeliger, S. T. A. Berger, G. Y. Remennikov, K. Polborn, H. Mayr, J. Org. Chem. 38 2007, 72, 9170-9180. 39 40 [16] O. Kaumanns, H. Mayr, J. Org. Chem. 2008, 73, 2737-2745. 41 42 [17] O. Kaumanns, R. Lucius, H. Mayr, Chem. Eur. J. 2008, 14, 9675-9682. 43 44 [18] D. Richter, N. Hampel, T. Singer, A. R. Ofial, H. Mayr, Eur. J. Org. Chem. 2009, 45 3203-3211. 46 47 [19] T. Bug, T. Lemek, H. Mayr, J. Org. Chem. 2004, 69, 7565-7676. 48 49 [20] S. T. A. Berger, A. R. Ofial, H. Mayr, J. Am. Chem. Soc. 2007, 129, 9753-9761. 50 51 [21] R. Appel, R. Loos, H. Mayr, J. Am. Chem. Soc. 2009, 131, 704-714. 52 53 [22] F. Seeliger, H. Mayr, Org. Biomol. Chem. 2008, 6, 3052-3058. 54 [23] O. Kaumanns, R. Appel, T. Lemek, F. Seeliger, H. Mayr, J. Org. Chem. 2009, 74, 55 56 75-81. 57 58 59 60 22

http://mc.manuscriptcentral.com/poc Page 23 of 35 Journal of Physical Organic Chemistry

1 2 3 [24] S. T. A. Berger, T. Lemek. H. Mayr, Arkivoc. 2008(x), 37-53. 4 5 [25] T. B. Phan, C. Nolte, S. Kobayashi, A. R. Ofial, H. Mayr, J. Am. Chem. Soc. 2009, 6 7 131, 11392-11401. 8 9 [26] S. Hoz, D. Speizman, J. Org. Chem. 1983, 48, 2904-2910. 10 11 [27] K. Hillier, J. M. W. Scott, D. J. Barnes, F. J. P. Steele, Can. J. Chem. 1976, 54, 3312- 12 3314. 13 14 [28] P. Denton, C. D. Johnson, J. Chem. Soc., Perkin Trans. 2, 1995, 477-481. 15 16 [29] S. Hoz, Acc. Chem. Res. 1993, 26, 69-74. 17 18 [30] J. P. Richard,For T. L. Amyes, Peer T. Vontor, Review J. Am. Chem. Soc. 1992, 114, 5626-5634. 19 [31] F. G. Bordwell, M. J. Bausch, J. Am. Chem. Soc. 1986, 108, 1979-1985. 20 21 [32] F. G. Bordwell, J.-P. Cheng, M. J. Bausch, J. E. Bares, J. Phys. Org. Chem. 1988, 1, 22 23 209-223. 24 25 [33] F. G. Bordwell, J. C. Branca, T. A. Cripe, Isr. J. Chem. 1985, 26, 357-366. 26 [34] Supplementary information, page S33 of Reference [20]. 27 28 [35] F. G. Bordwell, T. A. Cripe, D. L. Hughes, Adv. Chem, Series, 1987, 215, 137-153. 29 30 [36] J. P. Richard, M. M. Toteva, J. Crugeiras, J. Am. Chem. Soc. 2000, 122, 1664-1674. 31 32 [37] J. W. Bunting, J. M. Mason, C. K. M. Heo, J. Chem. Soc., Perkin Trans. 2, 1994, 33 34 2291-2300. 35 [38] H. Mayr, O. Kuhn, M. F. Gotta, M. Patz, J. Phys. Org. Chem. 1998, 11, 642-654. 36 37 [39] I. Lee, Adv. Phys. Org. Chem. 1992, 27, 57 -117. 38 39 [40] Y. Tsuno, M. Fujio, Adv. Phys. Org. Chem. 1999, 32, 267-385. 40 41 [41] H. Mayr, M. Patz, Angew. Chem. Int. Ed. 1994, 33, 938-957. 42 [42] D. N. Kevill, M. J. D’Souza, J. Chem. Res. 2008, 61-66. 43 44 [43] R. Appel, R. Loos, H. Mayr, J. Am. Chem. Soc. 2009, 131, 704-714. 45 46 [44] H. Mayr, B. Kempf, Chem. Eur. J. 2005, 11, 917-927. 47 48 [45] T. Bug, M. Hartnagel, C. Schlierf, H. Mayr, Chem. Eur. J. 2003, 9, 4068-4076. 49 50 [46] T. W. Bentley, M. S. Garley, J. Phys. Org. Chem. 2006, 19, 341-349. 51 [47] T. W. Bentley, K. Roberts, J. Org. Chem. 1985, 50, 5852-5856. 52 53 [48] T. W. Bentley, Chem. Eur. J. 2006, 12, 6514-6520. 54 55 56 57 58 59 60 23

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 24 of 35

1 2 3 Table 1. Comparisons of N and N′′′ for seven key carbanion nucleophiles in DMSO at 20 4 5 a 6 ˚C 7 8 9 10 b c d e 11 Nucleophile N N′′′ N 12 13 14 15 2a 13.91 3.41 10.50 16 17 18 2b For 16.27Peer 4.78 11.49Review 19 - 20 CH(COMe)2 (2c) 17.64 5.52 12.12 21 22 23 2d 18.82 6.08 12.74 24 - 25 CH(CN)2 (2e) 19.36 6.25 13.11 26 27 2f 19.62 6.38 13.24 28 29 - 30 CH(CO2Et)2 (2g) 20.22 6.61 13.61 31 32 33 34 a Kinetic data from References [11,12]. 35 36 b See Scheme 3. 37 c 38 From Eqn (2) and Reference [11]. 39 d 40 From Eqn (7), this work. 41 e N = N - N′′′. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 24

http://mc.manuscriptcentral.com/poc Page 25 of 35 Journal of Physical Organic Chemistry

1 2 3 Table 2. Comparisons of N and N′′′ (Eqn (7)) for additional nucleophiles in DMSO at 20 4 5 6 ˚C 7 8 9 10 a b c 11 Nucleophile N N′′′ N 12 13 14 15 CF CH NH d 12.15 1.35 10.80 16 3 2 2 17 e 18 ylid Ph3PCHCOFor2Et Peer12.21 1.40 Review10.81 19 d 20 H2NCH2CO2Et 14.30 2.89 11.41 21 22 f 23 EtCH(NH2)CH2OH 14.39 2.92 11.47 24 f 25 PhCH2NH2 15.28 3.40 11.88 26 27 f MeCH(OH)CH2NH2 15.47 3.52 11.95 28 29 n d 30 Pr NH2 15.70 3.59 12.11 31 f 32 NH2CH2CH2OH 16.07 3.67 12.40 33 34 f (HOCH2CH2)2NH 15.51 3.79 11.72 35 36 g 37 3, Z = NO2 14.49 3.83 10.66 38 39 morpholine d 16.96 4.66 12.30 40 41 4, Z = 4-NO h 16.29 4.71 11.58 42 2 43 g 44 3, Z = CN 16.28 4.83 11.45 45 46 piperidine d 17.19 5.05 12.14 47 48 h 49 4, Z = 4-CN 16.96 5.21 11.75 50 g 51 3, Z = CF3 17.33 5.37 11.96 52 53 - e (EtO)2P(O)CH CN 18.57 5.56 13.01 54 55 - e 56 (EtO)2P(O)CHC(O )OEt 19.23 5.76 13.47 57 58 59 60 25

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 26 of 35

1 2 3 3, Z = H g 18.67 5.81 12.86 4 5 h 6 4, Z = 3-NO2 18.06 5.88 12.18 7 8 4, Z = H h 18.29 5.93 12.36 9 10 d 11 azide 6.23 12 g 13 3, Z = Me 19.35 6.27 13.08 14 15 - e Ph2P(O)CHC(O )OEt 19.20 6.27 12.93 16 17 - e 18 Ph2P(O)CH CNFor Peer18.69 6.30 Review12.39 19 20 4, Z = 4-Me h 18.31 6.35 11.96 21 22 i 5, Z = 4-NO2 18.5 6.45 12.05 23 24 25 26 27 a From Eqn (2) and kinetic data referenced for each nucleophile. 28 29 b Calculated from Eqn (7) and kinetic data referenced for each nucleophile. 30 31 c 32 N = N - N′′′. 33 d 34 Reference [7]. 35 36 e Reference [21]. 37 38 f Reference [25]. 39 40 g 41 Reference [20]. 42 43 h Reference [19]. 44 45 i 46 Reference [22]. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 26

http://mc.manuscriptcentral.com/poc Page 27 of 35 Journal of Physical Organic Chemistry

1 2 3 TABLE 3. Correlations (Eqn (8)) of second order rate constants for reactions of 4 5 a 6 benzhydrylium ions (1a-1f) with carbanions and amines in DMSO at 20 ˚C with N′′′ 7 8 9 b 10 Cation sE E′′′ + c r n 11 12 13 (lil) CH+ (1f) 1.00c 0.0c 14 2 15 + (jul)2CH (1e) 0.96±0.02 0.62±0.08 0.999 7 16 17 0.98±0.01 0.51±0.03 0.999 19 18 + For Peer Review 19 (ind)2CH (1d) 0.95±0.03 1.13±0.13 0.999 4 20 21 1.00±0.02 0.84±0.10 0.997 15 22 + (thq)2CH (1c) 0.89±0.05 1.88±0.24 0.997 4 23 24 0.99±0.04 1.32±0.19 0.992 11 25 + 26 (pyr)2CH (1b) 0.91±0.06 2.33±0.29 0.998 3 27 d 28 1.10±0.07 1.46±0.28 0.994 5 29 (dma) CH+(1a) 0.97e 2.45e 2 30 2 31 1.10±0.05 1.87±0.16 0.996 5f 32 33 34 a 35 The first entry for each electrophile refers to kinetic data for reactions of key carbanions 36 37 in Scheme 3 from References [11,12]; the second entry includes mostly amines (from 38 References [7, 25]) and additional carbanion nucleophiles. 39 40 b Structures are given in Scheme 2. 41 c 42 By definition. 43 d 44 The nucleophiles are 2a, 2b, 2c, (3, Z = NO2), and ylid Ph3PCHCO2Et. 45 e Errors not known because n = 2. 46 f 47 The nucleophiles are 2a, 2b, CF3CH2NH2, H2NCH2CO2Et, and ylid Ph3PCHCO2Et. 48 49 50 51 52 53 54 55 56 57 58 59 60 27

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 28 of 35

1 2 3 TABLE 4. Correlations (Eqn (8)) of second order rate constants for reactions of quinone 4 5 methides (1g–1m) with carbanions in DMSO at 20 ˚C with N′′′ a 6 7 8 b 9 Quinone sE E′′′ + c r n 10 11 12 1g 1.00c 1.08±0.05 3 13 14 1h 1.25±0.02 -2.94±0.13 0.999 7 15 d 16 1.23±0.08 -2.76±0.46 0.982 11 17 18 1i 1.25±0.0For2 -3.82 Peer±0.11 0.999 Review 7 19 1.22±0.08 -3.74±0.42 0.978 14 20 21 1j 1.43±0.01 -6.56±0.07 0.999 7 22 23 1.34±0.10 -6.03±0.57 0.969 14 24 25 1k 1.41±0.01 -6.68±0.05 0.999 7 26 1.33±0.10 -6.17±0.61 0.965 14 27 28 1le 1.42±0.03 -7.52±0.20 0.999 6 29 30 1.20±0.15 -6.22±0.87 0.916 15 31 f 32 1.28±0.12 -6.76±0.72 0.955 13 33 34 1m 1.43±0.07 -8.00±0.44 0.995 6 35 1.34±0.14 -7.40±0.84 0.949 12 36 37 1.45±0.06 -8.13±0.36 0.993 10f 38 39 40 a 41 First entry for each electrophile refers to kinetic data for reactions of the key carbanions 42 in Scheme 3 from References [11,12]; a second entry includes additional carbanion 43 44 nucleophiles (Table 2), usually covering a relatively small range of values of N′′′. 45 46 b Structures are given in Scheme 2. 47 c 48 Assumed sE = 1.00. 49 d 50 Excluding the data point for ylid Ph3PCHCO2Et. 51 e The rate constant for reaction with 2g was assumed to be 69.9 M-1 s-1 (not 699, shown in 52 53 Table 2[12]), in accordance with the constant selectivity plot (Figure 2 of Reference [12]). 54 f - - 55 After deletion of the data points for (EtO)2P(O)CH CN and (EtO)2P(O)CHC(O )OEt. 56 57 58 59 60 28

http://mc.manuscriptcentral.com/poc Page 29 of 35 Journal of Physical Organic Chemistry

1 2 3 Table 5. Values of N and N′′′ (Eqn (9)) for additional nucleophiles in DMSO at 20 ˚C 4 5 6 7 Nucleophile N a N′′′ b N c 8 9 10 d 11 8, Z = NO2 19.61 6.2 13.4 12 7, Z = NO d 19.67 (6.5) 13.2 13 2 14 e 15 9 19.92 (6.5) 13.4 16 17 6a f 20.71 6.7 14.0 18 For Peer Review 19 f 20 6c 20.61 6.9 13.7 21 f 22 6b 21.54 7.2 14.3 23 24 5, Z = 4-CN g 22.6 7.5 15.1 25 26 g 27 5, Z = 4-CF3 24.3 7.9 16.4 28 29 7, Z = CN d 25.11 8.3 16.8 30 31 5, Z = 3-Cl g 8.4 32 33 d 34 8, Z = CN 25.35 8.7 16.7 35 d 36 7, Z = CF3 27.28 8.9 18.4 37 38 8, Z = H d 28.95 10.1 18.9 39 40

41 42 a 43 From Eqn (2) and kinetic data in references shown for each nucleophile. 44 b Based on Eqn (9) and kinetic data in references shown for each nucleophile, except for 45 46 the 3 values in parentheses which are estimated from correlations for quinone methides 47 48 (1h, 1j and 1k). 49 c 50 N = N - N′′′. 51 d 52 Reference [23]. 53 e Reference [24]. 54 55 f Reference [19]. 56 g 57 Reference [22]. 58 59 60 29

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 30 of 35

1 2 3 TABLE 6. Correlations (Eqn (8)) of second order rate constants for reactions of neutral 4 5 electrophiles (10-18) with carbanions in DMSO at 20 ˚C with N′′′ a 6 7 8 b b 9 Electrophile sE E′′′ + c r n 10 11 12 10, Z=Hc 1.00d 0.4 3 13 14 10, Z=OMec 1.38±0.05 -2.88±0.34 0.999 3 15 16 1.18±0.14 -1.64±0.88 0.987 4 17 c 18 10, Z=NMe2 1.32±0.02For -Peer4.23±0.14 0.999 Review 3 19 1.31±0.01 -4.17±0.08 0.999 4 20 21 11e 1.10±0.11 -3.76±0.62 0.986 5 22 23 1.16±0.10 -4.21±0.59 0.978 8 24 25 0.97±0.08 -3.19±0.51 0.962 15 26 12, Z=NMe e 1.03±0.10 -2.54±0.56 0.987 5 27 2 28 1.07±0.12 -2.89±0.73 0.962 8 29 30 0.92±0.08 -2.11±0.55 0.955 14 31 e d 32 12, Z=OMe 1.0 -0.8 3 33 34 (0.64±0.11 1.13±0.66 0.955 5) 35 1.0f,g -0.8 8 36 37 12, Z=He 1.0d,h -0.1 3 38 i 39 13, X = O 1.09±0.16 -3.05±0.96 0.959 6 40 41 1.22±0.09 -3.97±0.54 0.976 11 42 0.89±0.09 -2.06±0.57 0.944 14 43 44 i 14, Z=NMe2 0.98±0.15 -1.67±0.88 0.957 6 45 46 1.18±0.16 -2.95±0.94 0.948 8 47 48 0.82±0.11 -0.93±0.71 0.908 13 49 i 50 14, Z=OMe 1.00±0.07 -0.20±0.38 0.989 7 51 0.99±0.07 -0.20±0.42 0.983 8 52 53 0.84±0.18 0.47±1.04 0.857 10j 54 i 55 13, X = S 0.90±0.14 -0.70±0.82 0.957 6 56 i 57 15 0.99±0.09 -0.49±0.48 0.985 6 58 59 60 30

http://mc.manuscriptcentral.com/poc Page 31 of 35 Journal of Physical Organic Chemistry

1 2 3 16, Z=Hk 1.0d 0.7 2 4 5 16, Z=OMek 0.93±0.08 0.27±0.46 0.981 7 6 7 0.92±0.09 0.27±0.49 0.975 8 8 k 9 16, Z=NMe2 0.99±0.20 -1.77±1.18 0.928 6 10 11 1.01±0.21 -1.93±1.23 0.909 7 12 17k 1.16±0.21 -3.59±1.27 0.939 6 13 14 1.06±0.18 -3.08±1.09 0.925 8 15 l 16 0.72±0.12 -1.11±0.77 0.901 11 17 m 18 18,Z=4-NO2 1.54For±0.13 -Peer8.52±0.85 0.996 Review 3 19 1.40±0.26 -7.74±1.73 0.951 5 20 21 1.36±0.10 -7.47±0.72 0.983 8 22 m 23 18,Z=4-CN 1.57±0.42 -9.1±2.8 0.934 4 24 n 25 1.37±0.13 -7.81±0.93 0.979 7 26 18,Z=3-Clm 1.56±0.38 -9.6±2.5 0.946 4 27 28 1.43±0.13 -8.83±0.94 0.984 6n 29 30 31 a 32 Italic entries refer to the same data set of nucleophiles (Tables 2 and 5), as that used to 33 obtain values of E from the MPC3 optimisations fitting Eqn (2); kinetic data are from the 34 35 references shown for each electrophile; if an earlier entry is shown, it refers to kinetic 36 data for reactions of the key carbanions (Scheme 3); a later entry refers to the 37 incorporation into Eqn (8) of subsequently-published data from References [22-24]. 38 b 39 Structures are given in Scheme 5; values of E are in the References shown below. 40 c Reference [13]. 41 d Assumed. 42 e 43 Reference [14]. f 44 Assuming sE = 1.0 gives more plausible results than Eqn (8) for nucleophiles 7, 8, Z = 45 NO2 and 9. 46 g 47 5, Z = NO2 is predicted to react 20-fold faster than observed. h 48 This equation also fits data for 8, Z = NO2. 49 i 50 Reference [15]. j 51 5, Z = NO2 is calculated to react 13-fold faster than observed. 52 k Reference [16]. 53 l 54 The additional nucleophiles are 5, Z = 3-Cl, 4-CF3 and 4-CN. m 55 Reference [17]. 56 n Only two key nucleophiles (Scheme 3); additional kinetic data from Reference [23]. 57 58 59 60 31

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 32 of 35

1 2 3 TABLE 7. Correlations (Eqn (8)) of second order rate constants for reactions of 4 5 additional quinone methide (QM) electrophiles (19) with key carbanions (Scheme 3) in 6 7 DMSO at 20 ˚C with N′′′ 8 9 10 a 11 QM sE E′′′ + c r n 12 13 14 19a 1.50±0.07 -6.49±0.42 0.997 5 15 16 19b 1.51±0.09 -6.15±0.54 0.995 5 17 18 19c 1.45±0.08For -5.66±0.47 Peer 0.994 Review 6 19 19d 1.20±0.05 -2.42±0.24 0.998 5 20 21 19e 1.36±0.04 -6.55±0.23 0.999 4b 22 23 24 a 25 The quinone methide electrophiles (19, Scheme 6) are structural variations on those 26 given in Scheme 2, but the kinetic data were obtained[18] years after MPC1;[11] values of 27 28 E are in Scheme 6 29 30 b Data includes carbanion 6b. 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 32

http://mc.manuscriptcentral.com/poc Page 33 of 35 Journal of Physical Organic Chemistry

1 2 3 Table 8. Comparisons of s N and N′ with pK values for arylacetonitriles in DMSO 4 N a 5 6 a b c d 7 Carbanion pKa sNN N′′′ 8 9 10 11 7, Z = NO2 12.3 13.38 6.5 12 7, Z = CN 16.0 13.56 8.3 13 14 7, Z = CF3 18.1 13.64 8.9 15 16 8, Z = H 23.0 16.79 10.1 17 18 For Peer Review 19 a Structures are in Scheme 4. 20 21 b Refers to the conjugate acid; data at 25 ˚C from References [31,32]. 22 c 23 From Eqn (2), sNN = hypothetical log k estimated for electrophile 1n (E = 0.0) in DMSO 24 25 at 20 ˚C; data from Reference [23]. 26 d From Table 5. 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 33

http://mc.manuscriptcentral.com/poc Journal of Physical Organic Chemistry Page 34 of 35

1 2 3 Table 9. Comparisons of s and N (Eqn 2) calculated from log k vs. E plots (first entry) 4 N 5 with values from the multi-parameter correlation (second entry) a 6 7 8 b 9 Nucleophile sN N 10 11 12 2a (6) 0.806±0.033 0.856 14.24 13.91 13 14 2b (6) 0.787±0.036 0.767 16.12 16.27 15 16 2c (5) 0.755±0.026 0.729 17.30 17.64 17 18 2d (4) 0.623±0.027For 0.688Peer 19.81 Review18.82 19 20 21 a Kinetic data and parameters from the multi-parameter correlation (MPC2) are from 22 23 Reference [11]. 24 b 25 The nucleophiles are key carbanions (Scheme 3) and the number of data points for the 26 log k vs. E plots is also shown. 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 34

http://mc.manuscriptcentral.com/poc Page 35 of 35 Journal of Physical Organic Chemistry

1 2 3 4 5 Graphical Table of Contents 6 7 8 9 10 Nucleophilicity parameters for strong nucleophiles in dimethyl sulfoxide. A direct 11 alternative to the s(E + N) equation 12 13 14 T. William Bentley 15 16 17 N N 18 For Peer Review 19 + 20 H A nucleophilicity scale (N′′′) is defined for 34 strong nucleophiles 21 amines (e.g. carbanions and amines) and is linked to two previously-defined log k = N''' 22 carbanions nucleophilicity scales (N′ and N′′) including 96 nucleophiles. The 23 22 24 N N unified N′′ scale spans 10 in rate constant, whereas the floating N H 25 scale spans 32 orders of magnitude. 26 Nuc 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 35

http://mc.manuscriptcentral.com/poc