Triarylmethanes and 9-arylxanthenes as prototypes amphihydric compounds for relating the stabilities of cations, anions and radicals by C-H bond cleavage and electron transfer

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPES

AMPHIHYDRIC COMPOUNDS FOR RELATING THE STABILITIES OF

CATIONS, ANIONS AND RADICALS BY C–H BOND CLEAVAGE AND

ELECTRON TRANSFER

EDWARD M. ARNETT,* ROBERT A. FLOWERS, II, RICHARD T. LUDWIG, ALISON E. MEEKHOF AND

STUART A. WALEK

Department of Chemistry, Duke University, Box 90346, Durham, North Carolina 27708-0346, USA

Thermodynamic stability properties of 11 p-substituted trityl and seven 9-phenylxanthyl carbocations are reported in

sulfolane and of their conjugate carbanions in DMSO. The cations are compared by calorimetric heats of hydride

transfer from cyanoborohydride ion, their ﬁrst and second reduction potentials, their pK_R⁺s in aqueous sulfuric acid, ¹³

C

chemical shifts and free energies of methoxy exchange. Carbanions are compared by their heats and free energies (pK_HA

)

of deprotonation and their ﬁrst and second oxidation potentials. Radicals are compared by their oxidation and reduction

potentials. Their bond dissociation energies are derived by alternative routes: from the carbocation and its reduction

potential and from the carbanion and its oxidation potential. The various properties are correlated against each other

and against appropriate Hammett-type substituent parameters. Correlations between the different measured properties

reported here range from fair to excellent. Despite their importance as historic prototypes for the three trivalent

oxidation states of carbon, trityl and xanthyl systems are atypical models for comparing transmission of electron

demand in other series of carbocations, radicals or carbanions with signiﬁcantly different structures. The 9-arylxanthyl

series is especially poor because of its insensitivity to substituent effects. The effects of substituents on various properties

which represent the stabilities of R⁺s correlate surprisingly well against those for corresponding R^Ϫs. Accordingly,

compensating effects on the oxidation and reduction of a series of related R^•s may lead to a nearly constant electron

transfer energy and absolute hardness for the series. In contrast, the free energies for interconversion of the

carbocations and carbanions which determine the gap between pK_R₊and pK_HAare very sensitive to structural change.

J. Phys. Org. Chem. 10, 499–513 (1997) No. of Figures: 2 No. of Tables: 8 No. of References: 55

Keywords: trianylmethanes; 9-arylxanthenes; ion and radical staleclitres; amphihydric compounds

Received 21 August 1996; accepted 21 August 1996

INTRODUCTION

may form stable, isolable R^•, R⁺or R^Ϫbut the tri-

arylmethyl systems, of which the 9-arylxanthanes are a

subset, are nearly alone³in allowing the study of all three

stable conjugates in the same types of solvent under

conditions which differ only by the presence of acids or

bases. Electrochemical studies of R^+/•/Ϫshow that these

three species are closely interrelated,⁴although they are

usually treated as belonging to three widely separated and

unrelated ﬁelds.

Triarylmethyl halides were the ﬁrst source of stable

carbocations whose trivalent ionic structure was inferred

correctly by Gomberg^5–7only 2 years after his momentous

discovery of triarylmethyl free radicals.⁸Later, the tria-

rylcarbinols were used to establish the most commonly used

free energy scale by which the stabilities of R⁺s are

expressed as the pK_R₊s for their equilibria with ROHs in

aqueous acid.^{9, 10}Triarylmethanes were identiﬁed as weak

Since the ﬁrst year of the twentieth century the three

conjugate trivalent oxidation states of carbon (henceforth

R^+/•/Ϫ) derived from triarylmethanes have played a leading

role in the development of the theory of organic chemistry.

No other class of compounds has provided such well

recognized stable prototypes of radicals, carbenium ions

and carbanions in solution where their structures can be

demonstrated unequivocally and thus serve as bona ﬁde

models for the more reactive intermediates and transition

structures which are now at the heart of modern mechanistic

thinking and which have often required extreme conditions

such as the gas phase¹or superacid at Ϫ100 °C for their

preparation and authentication.²Other types of compounds

* Correspondence to: E. M. Arnett.

CCC 0894–3230/97/070499–15 $17.50

500

E. M. ARNETT ET AL.

organic acids in 1925¹¹and their conjugate bases were ﬁrst

named ‘carbanions’ in 1933.¹²However, the development of

a reliable thermodynamic pK scale for ranking strong to

weak organic Brønsted acids was delayed through the lack

of appropriate media (analogous to aqueous sulfuric acid for

studying carbocations) which were both highly basic, and

also able to dissociate organoalkalis into various types of

ion pairs^13–15and ‘free ions’.¹⁶

Following Streitweiser et al.’s adoption of cyclohexy-

lamine and its cesium salt as a strongly basic medium,¹⁷

Bordwell’s group solved the problem by use of K⁺

DMSYL^Ϫ, the potassium salt of DMSO, in this solvent.

They provided the pK_HAs for over 2000 compounds in this

medium, probably the largest database in organic chemistry

from a single laboratory under identical conditions.¹⁸

Fortunately, DMSO is a good solvent for most organic

compounds, and also their potassium salts, and is well

suited for UV–VIS spectrophotometry, electrochemistry

and reaction calorimetry as has been used in the present

study. The application of electrochemistry for interrelating

triaryl R^+(•/Ϫalso has a long history.¹⁹

oxidation potential of the anion in solution²⁴or the

equivalent redox potentials of the radical in solution.²⁵In

the gas phase the corresponding properties are I, the

ionization energy, and A, the electron afﬁnity, of the

radical.^{26, 27}In turn, these properties are related directly to

the HOMO–LUMO gap, the electronegativity (␹), and the

absolute hardness (␩) of the R^•,²⁸i.e. its resistance to giving

or gaining electrons. Since the stabilities of R^+/•/Ϫare the

primary determining factor for the energies required for the

three modes of bond-breaking of RH, ⌬G_ET, ␹ and ␩ are the

fundamental properties which drive the majority of

observed structure-dependent phenomena of organic chem-

istry.²⁹

An important feature of amphihydric compounds is that

both the one- and two-electron redox potentials of both R⁺

and R^Ϫcan be obtained directly under similar conditions so

that electrode potentials for oxidizing and reducing R^•to its

conjugate ions can be obtained by two independent routes

and errors from irreversibility of electrode processes be

assessed.³⁰Correspondingly, the (homolytic) bond dissocia-

tion energy, BDE, of an R–H bond can be calculated²⁶both

from the pK_HAof R^{Ϫ 31, 32}or from the pK_R₊of R^{+ 30}using the

appropriate redox potentials and equations thus providing

another check for self-consistency.

Amphihydric compounds provide the best opportunity to

establish all of these properties under similar conditions and

to compare them to other familiar criteria for bond-making

and -breaking energies. In principle, they also offer the

unique opportunity of measuring directly the heat of

heterolysis, ⌬H_het, of the R–R bond from the calorimetric

heat of reaction of R⁺with R^Ϫin solution²⁰(however, see

Ref. 33 for problems in deﬁning structures of symmetrical

coupling products). This may be converted readily into the

more familiar heat of homolysis, ⌬H_homo, for such radical-

coupled dimers (R–R) by combining the ⌬H_hetwith the

redox potentials of the conjugate ions.²⁴These relationships

are shown in Scheme 1 and were communicated earlier²as

a complete analysis for a series of p-substituted-9-phenyl-

The triarylmethanes are an important class of amphi-

hydric compounds,²⁰for which the bond-making and

-breaking energies of tetracovalent RH, ROH or RX to form

the three related trivalent states can be examined under

conditions which also permit their interconversion by

electrochemistry.

The fact that the trivalent oxidation states of the

triaylmethanes, or any other class of amphihydric com-

pounds, can be studied in a single solvent is a fortuity. It is

usually true that compounds whose R⁺s are stable enough to

afford pK_R₊s in aqueous acid will have R^Ϫs that are too

basic to allow direct determination of their pK_HAs with

K⁺DMSYL^Ϫ/DMSO. Correspondingly, RH which yield

R^Ϫthat are stable in that solvent usually have R⁺that are

such strong Lewis acids that they react with it. Triaryl-

methanes just happen to fall in the particular

solvent-determined acid–base window that permits their

study by both of the unrelated processes used to establish

these two pK scales. Although R^+/•/Ϫmay all be stabilized

by resonance-delocalization,^{21, 22}it is generally true that the

factors which help to stabilize R⁺s (e.g. electron-releasing

groups) tend to destabilize their conjugate R^Ϫs and to have

comparatively little inﬂuence on the stabilities of their

conjugate R^•s. Breslow and co-workers developed a power-

ful method for calculating the pK_R₊s of very unstable R⁺s

from the pK_HAs and the two-electron oxidation potentials of

their very stable conjugate R^Ϫs and vice versa.^{4, 23}Amphi-

hydric trityl compounds played an important role in

authenticating this method. The free energy for two-electron

reduction of the cation or oxidation of the carbanion is an

important fundamental property since it determines the

energy gap between the two conjugate ions and thus of the

pK_HAof the carbanion and the pK_R₊of the carbocation.

The important differences in stability between the

conjugate trivalent states R^+/•/Ϫare determined by the

(reversible) reduction potential of the cation and the

Scheme 1

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPE AMPHIHYDRIC COMPOUNDS

501

xanthenes (Figure 1).

The appropriate equations and symbols for the measured

properties depicted in Scheme 1 are as follows:

on relative stabilities of ground states of triphenyl methanes

should be virtually identical with those on their carbinols⁴)

through the energies for the three different types of C–H

bond-breaking. Indeed, because of the important relation-

ships between the redox energies and the HOMO–LUMO

gap, ⌬G_ET, ␩ and ␹, the electrochemical data may be of

more fundamental signiﬁcance than the BDE, pK_HAor pK_R₊

which are based on different processes with different

standard states and can only be related to each other through

the redox potentials of their trivalent states.

Recently, we communicated a complete thermodynamic

analysis, such as that in Scheme 1, for a series of p-

substituted-9-phenylxanthenes and their cations, radicals

and carbanions.²⁰The present paper includes a full develop-

ment of the procedures and results presented there.

R⁺+BH₃CN^Ϫ→RH+BH₂CN ⌬H_H_Ϫ(R⁺)

(1)

(2)

(3)

(4)

R⁺H₂O

R⁺+e^Ϫ

R^•+e^Ϫ

→ROH+H⁺pK_R₊

→R^•

⌬G_red(R⁺),E_red(R⁺)

→R^Ϫ

⌬G_red(R^•), E_red(R^•)

R⁺+2e^Ϫ

→R^Ϫ

⌬G_red2(R⁺), E_red2(R+) (5)

RH+DMSYL^Ϫ→R^Ϫ+DMSO pK_HAand ⌬H_dep

(6)

(7)

(8)

R^ϪϪe^Ϫ

R^•Ϫe^Ϫ

R^ϪϪ2e^Ϫ

RH

→R^•

⌬G_ox(R^Ϫ), E_ox(R^Ϫ)

⌬G_ox(R^•), E_ox(R^•)

→R⁺

⌬G_ox2(R^Ϫ), E_ox2(R^Ϫ) (9)

→R^•+H^•

BDE

(10)

Thus, if redox processes are reversible:

Although many other types of data are available for

substituted xanthylium cations, we have been unable to ﬁnd

a collection of pK_R₊values measured by the classical

titration technique.^{7, 9, 10}Accordingly, pK_R₊values in aque-

ous sulfuric acid are reported here for the present series of

xanthylium ions. Heats of deprotonation (⌬H_dep) have been

repeated and signiﬁcantly different values found from those

reported earlier.²⁰All redox potentials have been remea-

sured and derived properties have been recalculated.

⌬G_red(R⁺)=Ϫ⌬G_ox(R^•)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

⌬G_red(R^•)=⌬G_ox(R^Ϫ)

⌬G_red2(R⁺)=Ϫ⌬G_ox2(R^Ϫ

)

BDE=1·37 pK_HA+⌬G_ox(R^Ϫ)+C (Ref. 31)

BDE=1·37pK_R+⌬G_red(R⁺)+CЈ (Ref. 30)

⌬G_et=⌬G_red(R⁺)Ϫ⌬G_red(R^•)

␩=(IϪA)/2 Ϫ⌬G_ET/2

where I=ionization energy and A=electron afﬁnity;

Correlations are given below for the properties reported

here both against each other and also against published data

from various sources. Comparisons are made of the second

reduction potentials of the cations with the ﬁrst oxidation

potentials of their conjugate anions as tests for reversibility.

BDEs have been estimated through the ⌬H_deps and ⌬G_ox

␹=(i+A)/2

(18)

(19)

(20)

⌬G_red2(R⁺)=⌬G_red(R⁺)+⌬G_red(R^•)

⌬G_ox2(R^Ϫ)=⌬G_ox(R^Ϫ)+⌬G_ox(R^•)

If all processes are reversible:

⌬G_red2(R⁺)=Ϫ⌬G_ox2(R^Ϫ

(R^Ϫ)s and through the ⌬H_H_Ϫand ⌬G_red(R⁺)s of the

)

(21)

s

Scheme 1 summarizes the three processes by which R^+/•/Ϫ

are formed by loss of H^Ϫ, H^•, H⁺from RH and their

electrochemical interconversion. Scheme 1 indicates that if

data are available for any of the three modes of RH bond

breaking, it is possible to calculate the others by thermody-

namic cycles if the appropriate reversible electron transfer

properties may be obtained by electrochemistry.

Clearly, it is just as reasonable to use redox energies to

relate the stabilities of the three trivalent states of carbon to

each other as it is to relate them to their common

tetracovalent precursor RH (effects of remote substituents

conjugate cations.

EXPERIMENTAL

In order to obtain a complete analysis of the type

represented by Scheme 1, it is necessary to (a) have a

solvent which is resistant to attack by a wide range of

carbanions and carbocations, i.e. is both a relatively weak

Brønsted acid and also a relatively weak Lewis base, and (b)

have

a small enough two-electron redox difference,

⌬G_red2(R⁺) or ⌬G_ox2(R^Ϫ) between R⁺and R^Ϫso that the

hydride afﬁnity, ⌬H_H_Ϫ, of R⁺and the heat of deprotonation,

⌬H_dep, of RH to form R^Ϫmay be measured by appropriate

techniques.

Unfortunately, we have not yet found a single solvent in

which we can do all of the types of experiments shown in

Scheme 1. Sulfolane comes close to meeting the require-

ments. It is extremely resistant to attack by protonic and

Lewis acids and is only slightly more acidic than dimethyl

sulfoxide^18bwith which some of its important properties are

compared in Table 1. However, the potassium salt of

sulfolane is too insoluble to allow the preparation of a

concentrated basic medium equivalent to K⁺DMSYL^Ϫ.

But for this shortcoming, sulfolane could be used to

determine pK_HAs, ⌬H_deps and all of the other properties in

Scheme 1.

Figure 1. Structures of 4-␹-triphenylmethane and 4-␹-9-phenyl-

xanthene

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

502

E. M. ARNETT ET AL.

Table 1. Solvent properties of sulfolane and DMSO (all ⌬H values

CaH₂overnight and was ﬁltered before use. Potassium

hydride was puriﬁed by washing an oil dispersion of

potassium hydride (Aldrich) with dry hexane under an

argon atmosphere to remove the mineral oil. The resulting

hydride powder was dried under vacuum and sorted in an

argon-ﬁlled dry-box. All solvents were checked by Karl

Fisher titration for water content, which was only accept-

able if <25 ppm.

in kcal mol^Ϫ1at 25 °C)

Property

DMSO

Sulfolane

a

pK_HA

35

0

31·1 (Me₂SO₂)

Ϫ2·64

Ϫ9·6

Ϫ4·25

Ϫ14·3

b

⌬H_dep

c

⌬H_HSO3F

Ϫ26·5

Ϫ6·6

Ϫ29·8

Ϫ105·3

46·7

c

⌬H_p-F-_phenol

d

⌬H_SbCl

5

General procedure for the synthesis of 9-aryl-9-xan-

thanols. To a solution of 2·0 g (10·2 mmol) of xanthone in

⌬H_BF3^e(CH₂Cl₂=Ϫ10·0)

Dielectric constant^f

Dipole moment^f

Ϫ51·3

43·32

THF was added dropwise 11 ml (11 mmol) of a 1·0

M

4·81 D

3·96 D

solution of phenylmagnesium bromide. The solution was

stirred an additional 45 min after the addition was complete

and was then quenched with saturated NH₄Cl solution. The

reaction mixture was diluted with diethyl ether (100 ml) and

washed twice with water. The organic layer was dried over

anhydrous Na₂SO₄. Evaporation of the solvent under

reduced pressure left a crude residue, which was ﬁltered

through neutral alumina to give 9-phenyl-9-xanthanol as a

white solid (2·5 g; 82%). The melting points of known

9-substituted xanthanols were compared with the published

values of McClelland et al.³⁸This material was used without

further puriﬁcation for the preparation of 9-arylxanthyl

cations and 9-arylxanthenes after examination for impurities

by ¹H and ¹³C NMR spectroscopy.

^aRef. 18b.

^bRef. 20.

^cRef. 34.

^dRef. 35.

^eRef. 36.

^fRef. 37.

Accordingly, we have had to compromise and use

sulfolane to study carbocation properties and DMSO for

carbanions. The conjugate radicals can be produced in

either solvent by redox of the two types of conjugate ions.

Sulfolane and DMSO are closely similar in most of their

important solvent properties, the principal difference being

the relatively high basicity of DMSO towards both protonic

and Lewis acids (Table 1).

General procedure for the synthesis of triarylcarbi-

nols. These were prepared by the reaction of aryl acid

chlorides and aryl Grignard reagents.³⁹For example, to a

solution of 1·5 g (10·7 mmol) of benzoyl chloride in THF

As before,²⁴we used reaction calorimetry and electro-

chemistry

of

carbocations

in

sulfolane

(tetramethylenesulfone) containing 5% 3-methylsulfolane

to lower the freezing point for operation at 25 °C [equations

(1), (3) and (4)]. By analogy with Bordwell and co-

workers,³¹we combined ⌬H_depwith the ﬁrst oxidation

potential of R^Ϫin DMSO [equations (6) and (15)] to give

the BDE for C–H cleavage to give R^•and also derived BDEs

from the corresponding ⌬H_H_Ϫ(R⁺)s in sulfolane and the ﬁrst

reduction potential of R⁺for comparison.

was added dropwise 30·0 ml (30 mmol) of a 1·0

M

solution

of phenylmagnesium bromide. The solution was stirred for

an additional 50 min at reﬂux after the addition was

complete and was then quenched with saturated NH₄Cl

solution. The reaction mixture was diluted with diethyl

ether (100 ml) and washed twice with water. The organic

layer was dried over anhydrous Na₂SO₄. Evaporation of the

solvent under reduced pressure left a crude residue which

was subjected to ﬂash silica gel chromatography to give

1·1 g (40%) of triphenylmethanol. These samples were

determined to be pure by ¹H and ¹³C NMR spectroscopy.

Since calorimetric measurements are meaningless unless

the presumed reactions are clean and complete, all of the

reactants, products and processes discussed here were

checked by H and ¹³C NMR spectroscopy. We recognize

1

that these tools may not be sensitive enough to detect small

quantities of reactive impurities in reactants or of side-

products that would contribute to signiﬁcant calorimetric

errors.

General reduction procedure for carbinols. Both the

9-arylxanthan-9-ols and the triarylmethanols were con-

verted to their carbon acids by heating

a formic

acid–Na₂CO₃solution of the alcohol at reﬂux for 2–3 h³⁹as

in the following example. To a solution of 5·0 g (18 mmol)

of 9-phenyl-9-xanthanol in 60 ml of 90% formic acid was

added 10 g of sodium carbonate. The solution was reﬂuxed

for 1–1·5 h and then cooled at 5–10 °C for 4 h. The product

was isolated by ﬁltration and washed with water. The

product was recrystallized from hexane to give a 95% yield

Puriﬁcation of solvents. All solvent batches were

checked by Karl Fisher titration to ensure a water content

below 50 ppm. Sulfolane (Phillips Petroleum) was distilled

from butyllithium at 1 mm Hg pressure (b.p. 110–111 °C)

and stored under argon. The melting point of sulfolane was

lowered for work at 25 °C by addition of 3-methylsulfolane

(Aldrich), which was dried by stirring overnight at room

temperature with CaH₂under an argon atmosphere followed

by vacuum distillation (b.p. 101–102 °C at 1 Torr). The

puriﬁed solvent was transferred to an argon-ﬁlled dry-box

and a 5% (v/v) solution of 3-methylsulfolane–sulfolane was

prepared. This mixture was dried further by stirring over

1

of desired product, shown to be pure by H and ¹³C NMR

spectroscopy.

Triarylmethyl cations. The tetraﬂuoroborate salts were

prepared by treating the corresponding carbinol with

ﬂuoroboreic acid (48% in diethyl ether, Aldrich) in

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPE AMPHIHYDRIC COMPOUNDS

503

propionic anhydride as described by Dauben et al.⁴⁰For

example, triphenylmethanol (13·2 g, 50 mmol) was dis-

solved in propionic anhydride (150 ml) and cooled to 10 °C.

After adding 48% ﬂuoroboric acid (4·4 g, 50 mmol) the

temperature was kept between 10 and 12 °C for 10 min. The

solution was diluted with dry diethyl ether (300 ml) and the

resulting precipitate was collected by ﬁltration under argon

and washed thoroughly with dry diethyl ether. The resulting

solid was dried under vacuum (0·5 mmHg) overnight. The

anions and their conjugate acids was determined by direct

integration of signals representative of the four species in

equilibrium.

pK_R₊determination of xanthylium ions. Techniques for

determining carbocation–carbinol equilibrium constants in

aqueous acid by UV–VIS spectrophotometry are widely

documented^{7, 9, 10}and were followed carefully. Spectra of the

cations agreed well with published spectra and gave good

Beer’s law plots. However, the spectra of the carbinols

overlapped with the solvent in some regions, in which case

the ROH concentration at equilibrium was determined from

the difference between the stoichiometric concentration and

that of the cation.

trityl tetraﬂuoroborate was stored in

4,4Ј,4Љ-Tris(p-chlorophenyl)methylium tetraﬂuoroborate

salt could not be prepared using this method. The penta-

chloroantimonate salt was made by addition of

a

dry-box.

a

stoichiometric amount of antimony pentachloride to a

solution of the 4,4Ј,4Љ-tris(p-chlorophenyl)methyl chloride

in hexane as described by Freedman and co-workers.⁴¹

Relatively dilute aqueous sulfuric acid solutions covering

an H_R(10) range from Ϫ1·65 to 1·16 were adequate for all

pK_R₊measurements, which were based on linear plots of

2–6 independent measurements of log[R⁺]/[ROH] vs H_R

for ﬁve different acid solutions. As before,¹⁰the concentra-

1

These samples were determined to be pure by H and ¹³C

NMR spectroscopy.

Equilibration measurements. pK_HAs of 9-arylxanthenes

tions of R⁺and ROH varied in the range 10^Ϫ6–10^Ϫ4

.

M

1

(HA) were determined by means of H NMR spectroscopy

instead of UV–VIS spectrophotometry as employed by

Bordwell and co-workers.¹⁸Since the pK_HAof 9-phenyl-

xanthene, HA₀, was available already⁴²and each of the four

species (HA, HA₀, A^Ϫand A₀^Ϫ) had distinct and separable

signals in the ¹H NMR spectra for the acids under study, the

relative pK_HAs could be determined by equilibration. The

difference, ⌬pK_HA, could be derived from the response of

the integrated peak areas of the four species as their

stoichiometric concentrations were varied and pK_HAdeter-

mined using a method equivalent to the methoxy exchange

reaction applied by Freedman and co-workers⁴¹to the

relative stabilities of the same series of 9-arylxanthyl

carbocations in acidic solution. ⌬pK_HAwas then added or

Calorimetric measurements. All ⌬H_H_Ϫmeasurements

reported here were made using a Tronac 450 titration

calorimeter at 25 °C. In contrast to the previous²⁰heats of

hydride transfer, obtained with cycloheptatriene as hydride

donor, the results given here used a recently published⁴³

procedure employing calorimetric titration of sulfolane

solutions of the cation tetraﬂuoroboride into a 50 ml Dewar

calorimeter vessel containing a large excess of 1

M

sodium

cyanoborohydride containing an equivalent amount of

18-crown-6 polyether to complex the sodium ion and

eliminate possible small ion-pairing contributions to the

measured heat of hydride transfer. A variety of reasonable

hydride donors were tried before settling on this system.

In addition to titrations using a standard Tronac Model

450 calorimeter, many measurements were made with a

more automated system that employed a Tronac Model 900

computer interface for collecting and analyzing the data.

Each ⌬H_H_Ϫpresented here represents an average value from

at least seven calorimetric measurements on two independ-

ently prepared solutions. The relative error for reaction

calorimetry with sensitive, unstable solutions such as some

of these is 2–4% rather than the usual 0·5–1%. The absolute

error (accuracy) is unknown and is best tested for by

correlating data with a variety of other properties as

reported below. This, however, provides no protection

against systematic errors that persist throughout a series of

related compounds.

subtracted from pK_HA. The equilibrium between a substi-

0

tuted 9-arylxanthene (HA) and 9-phenylxanthene (HA₀) is

described by the equations

Ϫ

*

HA+A₀)A +HA₀

(19)

(20)

(21)

⌬pK_HA=Ϫlog[AϪ][HA₀]/[HA][A₀^Ϫ]

pK_HAЈ=pK_HA+⌬pK_HA

0

where pK_HAЈrefers to the compound of unknown acidity and

pK_HAis that for 9-phenylxanthene. Concentrations close to

0

1ϫ10^Ϫ2

M

were employed. Since threefold dilution did not

affect the results, aggregation was deemed to be insigni-

ﬁcant. The method was checked using 9-phenylxanthene

and 4-methoxy-9-phenylxanthene, both of whose pK_HAs are

known.^18c

The determination of the pK_HAfor 9-(p-methylphe-

nyl)xanthene is given as an example of the general

procedure. Equimolar quantities of 9-(p-methylphenyl)xan-

thene and 9-phenylxanthene were dissolved in DMSO-d⁶

(Aldrich, distilled from KH), and the relative molar ratio of

the two compounds was determined through direct integra-

tion of the ¹H NMR spectrum of the mixture. To this

mixture was added 0·8 equiv. of KH. The ratio of the two

Determination of ⌬H_depof triarylmethanes and 9-aryl-

xanthenes. Heats of deprotonation are well precedented in

this laboratory⁴⁴using titration calorimetry of a molar

solution of the carbon acid in DMSO into a large excess of

K⁺DMSYL^Ϫ/DMSO. All solutions were prepared and

handled in a dry-box using carefully dried DMSO whose

water content was <25 ppm by Karl Fischer titration. Each

⌬H_deprepresents an average value from at least seven

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

504

E. M. ARNETT ET AL.

calorimetric measurements on two independently prepared

solutions.

effects of substitution in the phenyl rings on the stabilities of

the conjugate R⁺and R^Ϫrelative to their R^•were compared

with each other and with published data by correlation with

Electrochemistry. All electrochemical measurements

employed a BAS 100-A electrochemical analyzer using

cyclic voltammetry (CV) and Osteryoung square-wave

voltammetry (OSWV).⁴⁵The speed and simplicity of CV

made it the ﬁrst choice for characterizing redox systems.

OSWV was also employed because of its ability to

discriminate against charging currents and to produce a

peaked curve for Faradaic processes. Computer display of

the forward and reverse responses for an OSWV experiment

provided a test for the reversibility of the redox processes.

All measurements were made using a three-electrode

arrangement consisting of a platinum disk working elec-

trode, a platinum wire auxiliary electrode and an Ag/AgNO₃

reference electrode with the ferrocenium/ferrocene redox

couple as the external standard. The measured values were

referred to the normal hydrogen electrode by adding 0·75 V

and converting to the corresponding standard free energy

terms by multiplying by Ϫ23·06 kcal V^Ϫ1. Solutions of

+

Ϫ

␴

and ␴ Hammett parameters.⁴⁸

However, none of the above solution studies, except

those from this laboratory,^{20, 43}have combined direct

measurement under comparable conditions of the stabilities

of R⁺[by ⌬H_H_Ϫ(R⁺)] and of R^Ϫ[by (H_dep)] with redox

potentials to provide both a complete analysis of the

energetics for bond cleavage to form the R^+/•/Ϫand also of

their stabilities relative to each other by electron transfer, as

shown in Scheme 1.

Table 2 presents a variety of experimental properties

reﬂecting the stabilities of the trityl and xanthyl cations in

sulfolane (⌬H_H_Ϫand ⌬G_red) and aqueous sulfuric acid

(pK_R₊). In addition, hydride afﬁnities calculated by Parker

and co-worker’s method⁴⁹from the pK_HAs (method 1) or

⌬H_dep(method 2) and the oxidation potential of the radical

are listed. All of the data shown in Table 2 go from the least

stable R⁺at the top to the most stable ones at the bottom,

although the sign of the property depends on conventions

for its deﬁnition.

cations and anions in 0·1

M

tetrabutylammonium

tetraﬂuoroborate supporting electrolyte were prepared under

an argon atmosphere just prior to electrochemical analysis.

Reduction potentials for cations were determined in 95%

sulfolane–5% 3-methylsulfolane. Oxidation potentials for

the anions were determined from DMSO solutions of the

Correlations of ⌬H_H_Ϫwith other carbocation properties

Table 3 provides a means for comparing these properties

with each other in terms of the statistics of their linear

correlations rather than attempting to present the correla-

tions graphically. In general, none of the correlations is

excellent (r>0·99); xanthyl and trityl are mostly differ-

entiated with respect both to slope and r with correlations of

the xanthyl data usually being poorer. Correlations of ⌬H_H_Ϫ

with other properties in Table 2 are mostly poorer than those

for ⌬H_deps (see below) because of the relatively high

potassium salts of the anion using 0·1

supporting electrolyte.

M

TBABF₄as the

RESULTS AND DISCUSSION

Cations

Carbocations have played a key role in the development of

mechanistic organic chemistry. Their proposed involvement

in solvolysis and electrophilic aromatic substitution demon-

strated the need for reliable thermodynamic scales as a basis

for structure–reactivity correlations. Especially relevant to

the present discussion are Deno et al.’s pioneering studies of

the acidity function, H_R, for determining K_R, the carbinol–

cation equilibrium constants in aqueous acid,⁹Arnett and

Bushick’s examination¹⁰of the temperature coefﬁcient of

H_Rand pK_Rs for a number of di- and triarylmethyl cations in

experimental

4·184 kJ)]

error

this

[±0·4–0·6 kcal mol^Ϫ1

cation property

(1 kcal=

compared

of

to the usual ±0·2–0·4 kcal mol^Ϫ1for reaction calorimetry.

Not surprisingly, the best correlation (3-1) for ⌬H_H_Ϫof

the trityl cations is against their ¹³C NMR shifts, which

provide a sensitive indicator of transfer of electron demand

at the cationic center to the para substituents.^46cThe next

best correlation (3-2) is with ⌬G_red(R⁺), their free energies

of reduction in sulfolane. Of particular importance is the

nearly unit slope of (3-2) for all of the 40 highly diverse

types of compounds for which we have measured both

properties but the very low slope (3-8) and lower correlation

coefﬁcient for eight xanthylium ions compared to 11 trityl

cations is noteworthy (see below).

+

aqueous H₂SO₄, Brown and co-workers’ development of ␴

substituent constants⁴⁶and Arnett and Hofelich’s measure-

ments of the heats of formation of a corresponding series of

cumyl cations from their alcohols in superacid at low

temperature.⁴⁷

Close correlations between C–H bond-making and

-breaking processes and the redox potentials of the resulting

ionic trivalent species formed are well precedented-

More recently, Wayner and co-workers²⁵have exploited a

photomodulation technique to generate unstable R^•s in

solution and derive important thermodynamic properties for

their conjugate R⁺s and R^-s which are too unstable to allow

direct determination of pK_R+or pK_HAs. The redox potentials

of a number of highly unstable substituted benzyl, cumyl

and benzhydryl radicals that were generated by photolysis

of the RH precursor in acetonitrile were determined and the

24d,e,h,i, 31c, 50

and might reasonably be expected since both

types of process involve the difference in energy between a

neutral species and an ion. However, bond-breaking pro-

cesses perforce involve conversion of a tetravalent species

into a trivalent one and so may be sensitive to steric effects.

In contrast, redox processes involve interconversion of

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

Table 2. Stability data for p-substituted triphenylmethyl and 9-phenylxanthene cations (all ⌬G and ⌬H in kcal mol^Ϫ1

)

f

Entry

No.

(2-)

Ϫ⌬G_h_Ϫ

Ϫ⌬H_h_Ϫ

Cation

(R⁺)

h

Series

Trityl

Substituent

4,4Ј,4Љ-Cl

None

4-F

Ϫ⌬H_H_Ϫ(R⁺) ⌬G_red(R⁺)[E_red1(R⁺)]^{a, b}⌬G_red(R^•)[E_red(R^•)]^{a, b}

pK_R₊¹³C shift (Parker 1) (Parker 2) ⌬G_ET

⌬G_red2(R⁺)

1

2

3

4

5

6

7

8

9

54·4±0·58 Ϫ17·34 (0·752±·009)

52·7±0·23 Ϫ12·49 (0·542±·011)

52·1±0·36 Ϫ15·26 (0·662±·007)

51·5±0·32 Ϫ12·86 (0·558±·013)

51·3±0·52 Ϫ12·33 (0·535±·011)

49·4±0·41 Ϫ11·6 (0·503±·017)

48·2±0·37 Ϫ11·2 (0·486±·007)

48·2±0·40 Ϫ10·4 (0·451±·

12·68 (Ϫ0·550±·013)

16·64 (Ϫ0·722±·015)

16·3 (Ϫ0·707)

Ϫ7·74^c212·7

Ϫ6·63^c212·0

Ϫ6·42^e210·0

Ϫ5·41^c209·0

Ϫ4·39^e208·2

Ϫ3·56^c206·7

Ϫ3·40^c201·0

99·94

96·04

86·74

101·18

95·87

72·81

97·1

94·89

93·14

94·71

88·27

Ϫ30·02

Ϫ29·1

Ϫ31·6

Ϫ30·2

Ϫ31·6

Ϫ32·9

Ϫ29·8

—32·1

Ϫ24·6

Ϫ26·7

Ϫ26·0

Ϫ25·8

Ϫ25·2

Ϫ26·4

Ϫ30·6

Ϫ27·0

Ϫ33·8

Ϫ26·4

Ϫ29·5

Ϫ33·8

Ϫ36·9

Ϫ4·65

4·15

1·04

4·43

6·99

9·75

7·38

11·25

5·01

7·70

7·75

4·85

7·86

9·25

11·37

10·38

31·13

11·03

19·3

4-Me

17·29 (Ϫ0·750±·011)

19·32 (Ϫ0.838±.010)

21·35 (Ϫ0·926±·009)

18·58 (Ϫ0·806±·007)

21·65 (Ϫ0·939±·005)

14·78 (Ϫ0·641±·017)

17·22 (Ϫ0·747±·008)

16·83 (Ϫ0·730±·013)

4,4Ј-Me

4,4Ј,4Љ-Me

4-OMe

4,4Ј4,Љ-tbu

9-(p-CF₃)Ph

9-(p-Cl)Ph

9-(p-F)Ph

None

95·81

Trityl

Ϫ6·5^c

Ϫ0·11^d174·2

0·13 174·7

Ϫ0·64^d175·1

206·0

Xanthyl

Thioxanthene 9-Ph

Xanthyl

Trityl

Xanthyl

Tropylium

Xanthyl

Trityl

Cyclopropene 1,2,3-Triphenyl

Trityl 4,4Ј,4Љ-NMe₂

48·1±0·33

46·7±0·32

46·3±0·45

Ϫ9·77 (0·424±·021)

Ϫ9·52 (0·413±·010)

Ϫ9·17 (0·398±·007)

10

87·21

88·8

94·53

87·91

87·26

85·91

88·56

93·88

87·26

86·49

92·64

87·02

11

12

13

14

15

16

17

18

19

20

21

46·1±0·75 Ϫ10·46 (0·454±·017)

15·31 (Ϫ0·664±0·026) Ϫ0·84^c

16·55 (Ϫ0·718±·010)

45·3±0·21

45·1±0·19

44·6±0·38

43·9±0·31

43·4±0·43

43·3±0·42

41·4±0·18

38·8±0·15

28·9±0·34

Ϫ8·69 (0·377±·004)

Ϫ8·57 (0·372±·007)

Ϫ9·63 (0·418±·008)

Ϫ8·32 (0·361±·011)

172·4

9-Ph

17·82 (Ϫ0·773±·007)

21·00 (Ϫ0·911±·009)

18·70 (Ϫ0·811±·008)

0·81^d174·6

Ϫ1·24^e201·0

0·81^d172·3

4·7^c

4,4Ј-OMe

9-(p-Me)Ph

None

9-(p-OMe)Ph

4,4Ј,4Љ-OMe

86·95

86·17

1·38 (Ϫ0·060±·024) 32·51 (Ϫ1·410±·024)

Ϫ7·67 (0·333±·008)

Ϫ5·05 (0·219±·004)

18·7 (Ϫ0·811±·007)

24·4 (Ϫ1·058)

1·00^d164·8

0·82^e194·0

3·1^c

85·55

86·73

4·05 (Ϫ0·176±·008) 29·74 (Ϫ1·290±·012)

8·44 (Ϫ0·366±·007) 28·54 (Ϫ1·238±·025)

33·79

33·98

9·4^e

179·0

76·03

^aIn sulfolane at 25 °C.

^b0·1

M TBATFB, values reported are relative to the NHE.

^cRef. 7.

^dThis report from thesis of A.E.M. (1994).

^eRef. 7b.

^fCalculated hydride afﬁnities using Parker’s method,⁴⁹Ϫ⌬G_H_Ϫ=1·37 (pK_HA)+⌬G_red(R⁺)+⌬G_red(R^•)+constant.

^gCalculated hydride afﬁnities using a variation on Parker’s method, Ϫ⌬H_H_Ϫ=Ϫ⌬H_dep+⌬G_red(R⁺)+⌬G_red(R^•)+constant.

^h⌬G_ET=⌬G_red(R⁺)Ϫ⌬G_red(R^•).

506

E. M. ARNETT ET AL.

Table 3. Correlations between ⌬H_H_Ϫand other measures of carbocation stability

Entry

No.

(3-)

Correlation

Slope

r

n

Notes

1

2

3

4

5

6

7

8

9

Cation ¹³C shift

⌬G_red(R⁺)

1·34

Ϫ1·04

0·989

0·923

11

40

11

Trityl

a

⌬G_red(R⁺)

Ϫ0·932 0·982

Ϫ0·905 0·975

Ϫ0·905

Ϫ0·888

Ϫ0·878

Trityl

⌬G_red(R⁺)

⌬G (methoxy exchange)

1·37 pK_R₊

19 Xanthyl and trityl

11

10 Trityl

20 Trityl and xanthyl

8

7

5

12 Trityl and xanthyl

10 Trityl

6

17 Xanthyl and trityl

7

b

0·982

0·989

0·957

1·37 pK_R₊

⌬G_red(R⁺)

Ϫ0·495 0·891

Ϫ0·336 0·920

Xanthyl

Trityl compounds

1·37 pK_R₊

10

⌬G_H_Ϫ(R⁺) (Parker 1)

⌬H_H_Ϫ(R⁺) (Parker 2)

⌬H_tropylium

0·793 0·865

0·817 0·841

0·889 0·947

0·910 0·972

0·931 0·898

11

12

13

14

15

16

Xanthyl

⌬H_H_Ϫ(R⁺) (Parker 2)

⌬G_H_Ϫ(R⁺) (Parker 1)

1·14

1·27

0·550

0·620

Xanthyl

^aAll data for cations from Table 2 plus a variety of 1,3-dioxo- and 1,3-dithiocations in

preparation.

^bRef. 41.

trivalent radicals and ions. These species may have different

conformational energies, but these should be smaller than

the compression changes which accompany formation or

destruction of a tetracovalent species. Other signiﬁcant

differences between the two types of experiments are: the

presence of the electrode surface and of supporting

electrolyte for determination of the redox potentials.

However, several ⌬H_H_Ϫmeasurements were shown to be

unaffected, within experimental error, by the presence of

added supporting electrolyte at its working concentrations,

so this factor is not signiﬁcant.

of the conjugate R^Ϫand R^•to their conjugate R⁺(i.e. the

two-electron process for oxidizing R^Ϫto R⁺) in DMSO:

Ϫ⌬G_hydride(R⁺)=2·303RTpK_HA(RH)

+FE°_NHE[(R^•/R^Ϫ)+(R⁺/R)]

ϪFE°_NHE[(H^•/H^Ϫ)+(H⁺/H^•)]

(22)

Because of the availability of the enormous range of

pK_HAs, BDEs and (E_oxR^Ϫ) data established by Bordwell’s

group, Parker’s approach can provide estimates of ⌬G_hydride

(R⁺) for a much wider range of R⁺than the stable ones

whose reactions with borohydride are presented here. The

necessary data are available (Table 2 and Ref. 4a) for

calculating ⌬G_hydride(R⁺) in DMSO for 12 of the stable R⁺

for which we have made direct measurements of ⌬H_H_Ϫin

sulfolane, and these can at least provide a test of his

estimated relative energies. The results are correlated by

(3-10, 11, 16).

Another appropriate criterion for comparing the stabili-

ties of the trityl and xanthyl cations is the free energies for

their exchange of methoxy groups at equilibrium in D₂CCl₂

and D₃CN as reported by Freedman and co-workers⁴¹using

NMR spectroscopy. Although his conditions were different

from ours, the correlation (3-5), between the heats of

hydride reduction and the free energies of methoxy

exchange is one of the best in Table 3. Correlation of heats

of hydride transfer from cyanoborohydride ion with those

from cycloheptatriene to six xanthyl cations (3-13)²⁰is

good, but not excellent. In view of the generally poor

correlations that are found for the xanthyl cation properties

(see below), it would be hard to tell whether this case

reﬂects a difference in steric factors between the reaction of

a cation with an anion vs reaction with a neutal molecule or

just experimental errors due to slower reactions with this

hydride source.

The estimated numerical value for ⌬G_hydride(R⁺) and its

measured ⌬H_H_Ϫ(R⁺) should differ because the standard

state for making the C–H bond by reaction with sodium

cyanoborohydride in sulfolane is entirely different from that

for C–H bond-breaking with K⁺DMSYL^Ϫin DMSO. Also,

the assumption, which we have frequently made and

justiﬁed,^{24i, 51}that redox entropies of R^•are insigniﬁcant so

that ⌬Hs can be interchanged with ⌬Gs may not hold

exactly here. However, it would be reasonable to expect a

rather good correlation between Parker’s values and ours

but only fair to poor ones are found (2-10, 11, 16) the worst,

as usual, involving 9-phenylxanthylium ions.

Parker and co-workers⁴⁹have proposed a method for

estimating the free energy hydride afﬁnity of a carbocation,

⌬G_hydride(R⁺), from the measured pK_HAs of the conjugate

RH/R^Ϫin DMSO and the electrode potentials for oxidation

Improvement results from using the heat of deprotonation

(⌬H_dep) in DMSO instead of the corresponding free energy

term in that solvent (2·303RTpK_HA). If this substitution is

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPE AMPHIHYDRIC COMPOUNDS

507

made in Parker’s equation, a better correlation (2-12, 14,

15) is seen. Again, the 9-phenylxanthyl cations give some of

the poorest correlations of ⌬H_H_Ϫ(R⁺) with the other

properties in Table 3 (see below).

trivalent R to be Lewis acid R⁺in contrast to the strong base

R^Ϫ. Table 2 shows an enormous structure-dependent range

for this property, which explains in part why so few classes

of compounds can be converted readily into both their

conjugate R⁺and R^Ϫin solution, i.e. are amphihydric.

⌬G_ETand ⌬G_red2(R⁺)

Hammett-type correlations of R⁺stability data

These two properties, derived by adding or subtracting the

free energies of reduction for R⁺and R^•or of oxidation for

R^Ϫand R^•[see equations (16), (19), (20) and (21)] are of

fundamental importance to all of organic chemistry.

⌬G_ETis the energy required to transfer one electron from

the carbanion to the carbocation to produce a pair of

radicals. It is a direct expression of the HOMO–LUMO gap

and absolute hardness of R^{•26, 29}and is the necessary

property from relating free energies or heats of heterolysis

to those of homolysis.^{24, 26, 31}In the absence of ground-state

effects, stabilities of R^+/•/Ϫdetermine relative bond-break-

ing and -making energies and thus play a major role in

observed rates and equilibria.

It has long been recognized^{9, 46}that substituent effects on

many processes that form/destroy carbenium ion centers are

modeled better by the log k_sfor solvolysis of substituted

cumenes (␴⁺) in non-aqueous solvents than they are by the

ionization of substituted benzoic acids in water (␴). The

+

slopes (␳⁺) for correlations of R⁺properties vs ␴ can

cover a wide range, which is reasonably interpretable in

terms of electron demand, through one mechanism or

another, from the cationic center, particular importance

being given to resonance interaction with alkoxy or amino

groups.

Table 4 includes correlations of results from Table 2 and

other data collected by Wayner and co-workers⁴⁸(note also

24 ␳⁺s for reactions in Ref. 46a) to test the proposed

As seen in Table 2, the ⌬G_ETs for the trityl radicals lie in

a small range around Ϫ30 kcal mol^Ϫ1as a result of strong

afﬁnities of the cations for electrons and repulsion of

electrons by their conjugate radicals leading to conﬂicting

trends in the stabilities of R⁺and R^Ϫin response to p-

substitution. ⌬G_ETs of xanthyl radicals lie in a slightly lower

range than trityls.

+

relationship between ␳ and charge development. If, in fact,

there is relatively little electron demand on ring substituents

from the carbenium ion center, one might expect a better

correlation with ␴. Accordingly, Table 4 compares slopes

and rs of correlations against both types of substituent

constants as a measure of interaction between the carbenium

center and the p-substituent. The electron demand at the

para position should be reﬂected both by the slopes of

⌬G_red2(R⁺) or ⌬G_ox(R^Ϫ) are the free energies for

transferring two electrons to the carbocation, that is, for

interconverting the carbocation and the carbanion or

removing them from the carbanion. These properties

determine the gap between the pK_R₊of ROH and pK_HAof

RH⁴and express the difference between the proclivity for

+

correlations with ␴ and ␴ and their correlation coefﬁcients.

Because substituent constants are unitless values derived

from log Ks or log ks, it is necessary to divide ⌬Hs or ⌬Gs

Table 4. Hammett plots for carbocation stability data^a

Entry

No.

(4-)

Correlation

Slope

r

n

Notes

Xanthyl

Xanthyl (see text)

Xanthyl

Trityl

Xanthyl

Xanthyl (see text)

Xanthyl

Ref. 46

Ref. 48

Data from Ref. 47

Ref. 48

+

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

⌬G_red(R⁺)/1·37 vs ␴

Ϫ1·16

Ϫ1·25

Ϫ1·57

Ϫ1·61

Ϫ1·86

Ϫ2·70

Ϫ3·15

Ϫ3·25

Ϫ3·45

Ϫ3·50

Ϫ3·61

Ϫ4·43

Ϫ4·62

Ϫ6·45

Ϫ6·56

Ϫ7·19

Ϫ9·3

0·926

0·940

0·951

0·909

0·938

0·94

0·999

0·986

0·995

0·954

0·972

0·980

0·960

0·973

0·944

0·99

6

5

6

5

6

pK_R₊vs ␴

pK_R₊vs ␴⑂

+

⌬G_red(R⁺)/1·37 vs ␴

+

⌬H_H_Ϫ(R⁺)/1·37 vs ␴

5

+

pK_R₊vs ␴

11

6

5

8

6

5

8

+

⌬G_red(R⁺)/1·37 vs ␴

⌬H_H_Ϫ(R⁺)/1·364 vs ␴

⌬G_red(R⁺)/1·37 vs ␴

⌬G_red(R⁺) vs ␴°

+

⌬H_H_Ϫ(R⁺)/1·364 vs ␴

+

Log k_solvof cumyl chlorides vs ␴

⌬G_red(R⁺) for DMP cations (kcal/1·37) vs ␴

+

⌬G_red(R⁺) for cumyl cations (kcal/1·37) vs ␴

+

⌬H_f/1·37 of cumyl alcohols vs ␴

⌬G_red(R⁺) for benzyl cations (kcal/1·37) vs ␴

9

+

^a⌬Hs and ⌬Gs are divided by 1·37 to make slopes unitless when plotted against substituent constants.

508

E. M. ARNETT ET AL.

by 2·303RT to normalize them to the same scale.

position. McClelland and co-workers^{38, 53b}arrived at the

same conclusion independently and found that the log ks for

reaction of a series of xanthylium ions with various

nucleophiles correlated well with ␴.

Table 4 is arranged in terms of increasingly negative

slope, i.e. increasing transfer of electron demand to the para

position of the various phenyl groups. The majority of the

correlations are good, with r_s>0·95 for 22 out of 36

correlations. In several cases involving xanthylium ions,

An imporant factor behind the relatively poor correlations

of the various data for xanthyium ions is a consequence of

the feeble transmission of electron demand to their para

positions. As a result, the range of values for all of their

measured properties, as shown in Table 2, is considerably

smaller relative to their experimental errors than for the

trityl cations. Thus, delocalization of electron demand to the

oxygen at the 4-position reduces demand at the 9-position,

which in turn is relayed only weakly to the para position of

the attached benzene ring and results in a compressed scale

of observed values.

+

correlation with ␴ is actually better than with ␴ but the

overall trends in Table 4 are in the direction to support

strongly the notion that the magnitude of the reaction

parameter is a sensitive reﬂection of electron demand to the

para position and that the fewer rings share the demand, the

steeper is the slope.

Use of substituent constants for trityl and xanthyl

cations

+

The limitations of the ␴ parameter were already well

Carbanions and radicals

documented⁵²15 years after its successful introduction and

its poor performance with reactions of 9-arylxanthylium

ions has been noted more recently by several authors.^{38, 53}

Table 5 presents several sets of properties for well

authenticated p-substituted trityl and 9-phenylxanthyl car-

banions in K⁺DMSYL^Ϫ–DMSO at 25 °C and Table 6

compares correlations of these data with each other and with

other properties in this paper. It should also be noted that the

oxidation potentials listed here for the carbanions and in

Table 2 for the cations generally agree well with those of

other workers^{4, 19, 23, 25, 31, 42, 49, 50}for the same ions when

allowance is made for differences in reference electrodes or

solvent. However, we draw attention to the suspiciously

discrepent value for ⌬G_ox(R^•) of the 4-F-trityl anion/radical

in Table 5 which is echoed to some extent by the anomolous

+

Freedman and co-workers⁴¹attempted to use ␴ to

correlate their free energies for methoxy exchange between

xanthyl cations and found, as we do (Table 4, entry 11)

versus (4-1 or 5), that the Taft ␴° parameter⁵⁴for systems

with unconjugated charge relay gave an excellent correla-

tion (r²=0·995). From this they inferred that overlap of the

9-aryl system with the adjacent xanthylium cationic center

was precluded by steric hindrance between the 1 and 8

hydrogens and the ortho-hydrogens of the 9-phenyl ring,

which restrained the two rings to a nearly perpendicular

Table 5. Stability dta for p-substituted triphenylmethyl and 9-phenylxanthanyl anions (⌬H, ⌬G, BDE in kcal mol^Ϫ1

)

a

b

Series

Substituent

⌬H_dep

pK_HA

⌬G_ox(R^Ϫ) [E_ox(R^Ϫ)]^c

BDE^e

BDE^Ϫf

BDE^+g

Trityl

Xanthene

Trityl

4,4Ј,4Љ-OMe

Ϫ3·38±0·45

Ϫ5·01±0·34

Ϫ6·20±0·37

Ϫ6·93±0·35

Ϫ7·15±0·42

Ϫ7·50±0·46

Ϫ9·00±0·25

Ϫ9·20±0·19

Ϫ9·50±0·21

Ϫ10·30±0·27

Ϫ11·63±0·35

Ϫ12·45±0·38

Ϫ12·50±0·20

Ϫ12·81±0·36

Ϫ13·29±0·39

Ϫ13·90±0·41

Ϫ15·41±0·38

—

32·80

—

Ϫ25·55 (1108±10)

Ϫ22·81 (989±12)

Ϫ21·91 (950±14)

Ϫ20·11 (872±10)

Ϫ19·65 (852±20)

Ϫ18·91 (820±11)

Ϫ19·14 (830±14)

Ϫ18·17 (814±17)

Ϫ23·22 (1007±9)

Ϫ21·63 (938±19)

Ϫ19·81 (859±15)

Ϫ20·39 (884±8)

—

80·63

—

78·60

79·24

—

75·29

75·07

74·35

—

76·50

74·80

76·65

74·70

—

78·14

79·25

78·96

80·70

80·27

80·66

78·96

79·13

74·35

75·17

75·66

74·26

—

76·35

74·97

77·8

77·0

79·0

78·64

80·2

77·2

77·8

77·0

75·6

77·1

77·0

73·5

76·6

77·2

78·33

4,4Љ-OMe

4,4Ј,4Љ-Me

4-OMe

4,4ЈMe

4-Me

None

4-F

4,4Ј,4Љ-tBu

None

9-(p-Me)Ph

9-(p-OMe)Ph

4,4Ј,4Љ-Cl

9-(p-F)Ph

9-Ph

9-(p-Cl)Ph

9-(p-CF3)Ph

—

30·60

30·80

—

30·00

28·50

28·40

27·00

27·86

27·90

27·36

26·73

—

Xanthene

Thioxanthene

Xanthene

Ϫ17·50 (759±9)

Ϫ19·26 (835±14)

Ϫ16·67 (723±11)

Ϫ17·76 (770±12)

Ϫ15·31 (664±15)

76·79

74·55

76·53

73·93

—

^aMeasured in DMSO; values are relative to 0·1

^bRef. 18.

M

K⁺DMSYLϪ at 25 °C.

^cRelative to NHE measured in 0·1

M DMSO in TBATFB at 25 °C.

^dRelative to NHE, measured in 0·1

˚

M sulfolane in TBATFB at 25 C.

^eBDE=1·364 pK_HA+⌬G_ox(R^Ϫ)+56.^{31c, df}BDE^Ϫ=⌬H_dep+⌬G_ox(R^Ϫ)+107·1.

^gBDE⁺=⌬H_H_Ϫ+⌬G_red(R⁺)+40·0.

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPE AMPHIHYDRIC COMPOUNDS

509

Table 6. Correlations of pK_HAvs other carbanion properties of p-substituted triphenylmethyl

and 9-phenylxanthyl carbanions^a

Entry

No.

(6-)

Correlation

Slope

r

n

System

1

2

3

4

5

6

7

8

9

pK_HAvs ⌬H_dep

0·93

0·70

1·12

0·956

1·08

1·02

0·67

0·75

0·91

1·42

1·03

0·963

0·997

0·970

0·965

0·866

0·983

0·981

0·857

0·996

0·966

12 Triphenylmethyl and 9-Ph-xanthyl

5

7

4

7

Triphenylmethyl

Xanthyl

pK_HAvs ⌬H_dep

pK_HAvs ⌬G_ox(R^Ϫ

pK_HAvs ⌬G_red(R^•)

)

Triphenylmethyl

Xanthyl

18 Triphenylmethyl⁴²

8

6

5

Xanthyl⁴²

Triphenylmethyl and 9-Ph-xanthyl

Triphenylmethyl

pK_HAvs ⌬G_red(R^•)

pK_HAvs pK_R₊

10

11

Triphenylmethyl

⌬G_red(R^•vs ⌬G_ox(R^Ϫ

)

0·960 16 Triphenylmethyl and 9-Ph-xanthyl

^aAll ⌬H and ⌬G values divided by 1·364 to make slopes unitless.

⌬G_red(R⁺), which should correspond to the same redox

process except for a slight difference in solvent and, of

course, the different original source of the radical. Other

measured and derived properties for the 4-F-trityl cation and

anion seem to ﬁt reasonably and solutions of the cation were

well behaved. However, the solutions of this carbanion were

less stable than the others and the PMR spectrum was less

clean despite extra care to repeat measurements for this

system.

The most reliable property for comparing the data in

Tables 5 and 6 is pK_HA, since it is based on a very large and

well tested database.¹⁸Also, it is derived from equilibrium

constants which are inherently easier to measure precisely

than are heats of reaction. Excellent correlations of ⌬H_depvs

pK_HAhave been reported before from this laboratory⁴⁴and

are found again in Table 6 (6-1, 2, 3) for the data in Table

5. As before, the slopes are close to unity for correlating free

energies of deprotonation vs the corresponding enthalpy

terms (⌬G_depvs ⌬H_depis equivalent to pK_HAvs ⌬H_dep/1·37)

and the correlation coefﬁcients are very good to excellent.

Good linear relationships between pK_HAs and oxidation

potentials of the resulting anions have been noted fre-

quently.⁵⁰Correlations (6-4, 5) for the data in Tables 5 and

6 contrast sharply the trityl vs xanthyl series. Bordwell and

co-workers reported a slope of 0·974 and r=0·994 for 11 3-

and 4-substituted trityl anions under conditions similar to

those used here.

derived from their R⁺s to their conjugate R^Ϫs in sulfolane

are really equivalent to the corresponding ⌬G_ox(R^Ϫ)s for

oxidation of the same R^Ϫs to their conjugate R^•s in DMSO,

then there should be good correlations between these redox

free energies in the different solvents. Correlation (6-11) for

16 trityl and 9-arylxanthyl provides some strong support for

this important assumption of the present study.

An alternative, but related, test is to calculate homolytic

BDEs from the pK_HAs and ⌬G_ox(R^Ϫ) values for carbanions

in Table 4 using Bordwell’s well established equation in

DMSO (note, however, Ref. 32 concerning limitations due

to solvent effects):

BDE=1·364pK_HA+⌬G_ox(R^Ϫ)+56

(23)

and from the heat of hydride transfer from cyanoborohy-

dride ion to the conjugate R⁺of R^Ϫand the ﬁrst reduction

potential of the cation in sulfolane by an analogous

equation:

BDE⁺=⌬H_H_Ϫ+⌬G_red(R⁺)+40

(24)

Our value for the BDE of triphenylmethane agrees well

with that determined by Bordwell’s group,⁴²who also have

demonstrated^31athat their method of combining pK_HAs with

oxidation potentials gives reliable BDEs directly relatable to

the gas-phase values.

The constant in equation (24) is the average of the

differences between each of the 10 values in Table 5 listed

under BDE, which were calculated by equation (23), and

⌬H_H_Ϫ+⌬G_red(R⁺) from Table 2. Equation (24) was then

applied to the other seven systems in Table 5 for which there

were insufﬁcient data to calculate a BDE using equation

(23).

A similar approach was used to calculate the column in

Table 5 labeled BDE^Ϫfrom data in DMSO according to

equation (25) from the average difference between

⌬H_deps+⌬G₁(R^Ϫ) and the corresponding value for BDE for

the same 10 compounds used to calculate the constant in

equation (24) for the ‘carbocation route’ in sulfolane.

Wayner and co-workers⁴⁸demonstrated the reversibility

of electrode potentials in highly unstable benzyl and

benzhydryl systems (see Table 4), by showing the equality

of the reduction potential of the cation to the oxidation

potential of the radical. An equivalent strategy provides a

test of our assumed interconvertability of sulfolane as a

medium for R⁺measurements and DMSO as a medium for

study of their conjugate R^Ϫ. A small difference in solvation

energies³for transferring the ions from DMSO to sulfolane

prevents equality of electrode potentials in the two media.

However, if the ⌬G_red(R^•)s for reduction of the series of R^•s

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

510

E. M. ARNETT ET AL.

Table 7. Correlation of carbanion stability properties with various

substituent constants^a

Substituent

Property

pK_HA

constant

Slope

r

n

System

Ϫ

␴

Ϫ6·78

Ϫ6·17

Ϫ6·16

Ϫ4·86

Ϫ5·37

Ϫ3·91

Ϫ3·86

Ϫ3·11

0·996

0·942

0·986

0·983

0·890

0·931

0·932

0·943

5

8

7

9

5

Trityl

Xanthyl

Ϫ

⌬G_ox(R^Ϫ)/1·37

⌬G_ox(R^•)/1·37

⌬H_dep/1·37

⌬G_ox(R^Ϫ)/1·37

pK_HA

Ϫ

⌬G_red(R^•)

␴

Ϫ

^aRef. 42.

BDE^Ϫ=H_dep+⌬G_ox(R^Ϫ)+107.1

(25)

shows a good correlation of these apparently unrelated

properties. Correlation (8-2) tests the same point.

Although this is not a severe test of the data because of

the method of deriving both BDE⁺and BDE^Ϫfrom BDE,

the fact that the average difference between the 16 values of

BDE⁺and BDE^Ϫis only 0·22 suggests again that there are

no large systematic differences between the use of enthalpy

data in the two solvents and the better established free

energy data in DMSO. Also, the agreement between the

BDE values derived by the ‘cation route’ and the ‘carbanion

route’ is further support for consistency and the near

reversibility of the redox potentials.

Table 7 compares the power of the three types of the ␴-

parameter to correlate the acidic properties, pK_HAand

⌬H_dep, and the closely related redox properties for a number

of substituted triphenylmethyl and 9-phenylxanthyl sys-

tems. Except for the ⌬H_depof the xanthyl systems,

correlations of all properties can be rated from ‘excellent’ to

‘fair’ with all three types of substituent parameters with no

general pattern of superiority for any one.

Correlation (8-3) of bond-breaking properties pK_HAvs

pK_R₊for ﬁve trityl cations in aqueous acid vs their conjugate

anions in DMSO shows a surprisingly good correlation and

correspondingly ⌬H_depvs ⌬H_H_Ϫmakes the point but less

forcefully. Comparison of ␴ with ␴ (8-5) gives almost

unit slope and a fair correlation coefﬁcient.

Finally, we draw attention again to Table 2 listing ⌬G_ET,

which is the difference [⌬G_red(R⁺)Ϫ⌬G_red(R^•)]. If the

effects of substituents on stabilizing the R^Ϫand R⁺from a

given R^•are nearly equal and opposite, one might expect

⌬G_ETfor a given series of related radicals to remain nearly

constant. In other words, the substituents should have little

effect on the absolute hardness, ␩,^27–29of the radicals. This

is well supported by the ⌬G_ETvalues in Table 2 for xanthyl

and trityl radicals, all of which except 4,4Ј,4Љ-NMe₂are

distributed around an average of 30·7 kcal mol^Ϫ1. Since the

dimethylamino group is especially effective in stabilizing

the cation through resonance, it is not surprising to ﬁnd it as

a strong exception to the rest of the series.

+

Ϫ

Correlations of energies for forming carbocations vs

forming carbanions

CONCLUSIONS

The data available in Tables 2 and 5 provide a means for

testing the reasonable notion that the effects of substitution

on carbocation stability are generally in the opposite

direction from those on comparable carbanions. Correlation

(8-1), made by plotting ⌬G_red(R⁺) vs ⌬G_red(R^•) in the same

solution under identical conditions for 16 carbocations,

Substituted trityl and xanthyl cations, radicals and carban-

ions are historically important because of their ready

accessibility and stability. They provide an unusual opportu-

nity to study a wide variety of properties for all three of

these trivalent states of carbon both by bond-making and

-breaking reactions and by electrochemistry and relate them

Table 8. Correlation of selected cation stability data vs anion stability data

Entry

No.

(8-)

Correlation

Slope

r

n

System

1

2

3

4

5

⌬G_red(R⁺) vs ⌬G_red(R^•)

⌬G_ox(R^Ϫ) v s ⌬G_red(R⁺)

pK_HAvs pK_R₊

0·97

1·083

1·42

Ϫ1·30

0·99

0·960

0·911

0·966

0·842

0·930

16 Trityl and xanthyl

9

5

Trityl

⌬H_de₊_pvs ⌬H_H_Ϫ

10 Trityl

8

Ϫ

␴

vs ␴

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPE AMPHIHYDRIC COMPOUNDS

511

Ions, edited by G. A. Olah and P. v. R. Schleyer, Vol. 4, Chap.

28. Wiley, New York (1975).

to each other quantitatively under similar conditions.

Although their unusual stabilities set them apart structurally

from other less stable and mechanistically more important,

R^+/•/Ϫthe thermodynamic stability properties for these trityl

species generally correlate well with each other and also

with Hammett-type substituent parameters based on models

that are structurally far removed from them.

Several tests for consistency which combine energies for

forming R⁺by bond-breaking reactions with reduction

potentials give good agreement with values obtained by the

alternative routes from combination of energies for forming

R^Ϫwith their oxidation potentials.

Studies of this type with amphihydric compounds provide

a clear demonstration of the overall interrelation of

carbocation, radical and carbanion chemistry, in contrast to

their customary treatment as fundamentally separate ﬁelds.

(Particular care is required in reaction calorimetry to

conﬁrm by suitable analysis that the course of the presumed

reaction is clean and complete, otherwise the highly

exothermic production of a by-product may dominate the

reported value. Equilibrium free energies obtained by

spectroscopy are less vulnerable to such error.)

8. M. Gomberg, J. Am. Chem. Soc. 22, 757 (1975).

9. (a) N. C. Deno, J. J. Jaruzelski and A. Schriesheim, J. Am.

Chem. Soc. 77, 3044 (1955); (b) N. C. Deno and A.

Schriesheim, J. Am. Chem. Soc. 77, 3051 (1955); (c) N. C.

Deno and W. L. Evans, J. Am. Chem. Soc. 79, 5804 (1957).

10. E. M. Arnett and R. D. Bushick, J. Am. Chem. Soc. 86, 1564

(1964).

11. C. A. Kraus and R. Rosen, J. Am. Chem. Soc. 47, 2739–2748

(1925).

12. E. S. Wallis and F. H. Adams, J. Am. Chem. Soc. 55,

3838–3851 (1933).

13. C. B. Wooster and N. W. Mitchell, J. Am. Chem. Soc. 52,

688–694 (1930).

14. (a) J. B. Conant and G. W. Wheland, J. Am. Chem. Soc. 54,

1212–1221 (1932); (b) W. K. McEwen, J. Am. Chem. Soc. 58,

1124–1129 (1936).

15. C. R. Noller, J. Chem. Educ. 26, 429–430 (1949).

16. A. A. Morton, Chem. Rev. 35, 1–49 (1944).

17. (a) A. Streitwieser, Jr, E. Juaristi and L. L. Nebenzahl, in

Comprehensive Carbanion Chemistry, edited by E. Buncel and

T. Durst. Elsevier, New York (1980); for other important

reviews of carbanion chemistry, see (b) D. J. Cram, in Organic

Chemistry, edited by A. T. Blomquist, Vol. 4. Academic Press,

New York (1965); (c) O. A. Reutov, I. P. Beletskaya and K. P.

Butin, in CH-Acids, translated by T. R. Crompton. Pergamon

Press, Oxford (1978).

18. (a) W. S. Mathews, J. E. Bares, J. E. Bartmess, F. G. Bordwell,

F. J. Cornforth, G. E. Druken, Z. Margolin, R. J. McCallum and

N. R. Vanier, J. Am. Chem. Soc. 97, 7006 (1975); (b) F. G.

Bordwell, Acc. Chem. Res. 21, 456–463 (1988); (c) F. G.

Bordwell, printout of pK_HAdata, personal communication

(1992).

ACKNOWLEDGMENTS

The senior author (E.M.A.) is grateful to the Editor for

providing this special opportunity to express his apprecia-

tion to Professor Bordwell for his long-standing friendship

and frequent intellectual guidance in dealing with our

common interests acid–base chemistry and electron trans-

fer.

We acknowledge with much appreciation support of this

work by National Science Foundation Grant CHE-9218375

and the assistance of Drs Daniel Stoelting and Prakash

Reddy and Mrs Marjorie Richter.

19. (a) J. B. Conant and B. F. Chow, J. Am. Chem. Soc. 55, 3752

(1933); (b) E. D. Jenson and R. W. Taft, J. Am. Chem. Soc. 86,

116 (1964).

20. E. M. Arnett, R. A. Flowers, II, A. Meekhof and L. Miller, J.

Am. Chem. Soc. 115, 12603 (1993).

21. L. Pauling, The Nature of the Chemical Bond, 3rd ed. Cornell

University Press, Ithaca, NY (1960).

REFERENCES

22. G. W. Wheland, Resonance in Organic Chemistry. Wiley, New

York (1955).

23. R. Breslow and K. Balasubramanian, J. Am. Chem. Soc. 91,

5182 (1969); (b) R. Breslow and W. Chu, J. Am. Chem. Soc.

95, 411 (1973).

1. (a) A. G. Harrison, P. Kebarle and F. P. Lossing, J. Am. Chem.

Soc. 83, 777–780 (1961); (b) M. T. Bowers, Gas Phase Ion

Chemistry, Vol. 2. Academic Press, New York (1979); (c) S. G.

Lias, J. F. Liebman and R. D. Levin, J. Phys. Chem. Ref. Data

13, 695 (1984); (d) F. P. Lossing and J. L. Holmes, J. Am.

Chem. Soc. 106, 6917–6920 (1984); (e) J. C. Schultz, F. A.

Houle and J. L. Beauchamp, J. Am. Chem. Soc. 106,

3917–3927 (1984).

2. (a) G. A. Olah and P. von R. Schleyer, Carbonium Ions (4 vols).

Wiley–Interscience, New York (1970); (b) G. A. Olah, G. K. S.

Prakash and J. Sommer, Superacids. Wiley, New York (1985);

(c) G. A. Olah, Angew. Chem. 34, 1393–1405 (1995).

3. D. T. Stoelting, R. T. Ludwig and E. M. Arnett, Heteroatom

Chem. 7, 481 (1996).

4. R. Breslow, Pure Appl. Chem. 40, 493–509 (1974).

5. (a) M. Gomberg, Ber. Dtsch. Chem. Ges. 35, 2397 (1902); (b)

P. Walden, Ber. Dtsch. Chem. Ges. 35, 2018–2031 (1902).

6. C. D. Nenitzescu, in Carbonium Ions, edited by G. A. Olah and

P. v. R. Schleyer, Vol. 1. Interscience, New York (1968).

7. (a) D. Bethell and V. Gold, Carbonium Ions, an Introduction.

Academic Press, London (1967); (b) for an excellent review of

carbocation stabilities, see also H. H. Freedman, in Carbonium

24. (a)E. M. Arnett, K. E. Molter, E. C. Marchot, W. H. Donovan

and P. Smith, J. Am. Chem. Soc. 109, 3788 (1987); (b) E. M.

Arnett, L. G. Whitesell, Jr, J.-P. Cheng and E. Marchot,

Tetrahedron Lett. 29, 1507 (1988); (c) E. M. Arnett, N. G.

Harvey, K. Amarnath and J.-P. Cheng, J. Am. Chem. Soc. 111,

4143 (1989); (d) E. M. Arnett, K. Amarnath, N. G. Harvey and

J.-P. Cheng, J. Am. Chem. Soc. 112, 344 (1990); (e) E. M.

Arnett, K. Amarnath, N. G. Harvey and S. Venimadhavan, J.

Am. Chem. Soc. 112, 7346 (1990); (f) E. M. Arnett, K.

Amarnath, N. G. Harvey and J.-P. Cheng, Science 247, 423

(1990); (g) N. J. Pienta, R. J. Kessler, K. S. Peters and E. D.

O’Driscoll, J. Am. Chem. Soc. 113, 3773 (1991); (h) E. M.

Arnett and S. Venimadhavan, J. Am. Chem. Soc. 113, 6967

(1991); (i) S. Venimadhavan, K. Amarnath, N. G. Harvey and

E. M. Arnett, J. Am. Chem. Soc. 114, 221 (1992); (j) E. M.

Arnett, S. Venimadhavan and K. Amarnath, J. Am. Chem. Soc.

114, 5598 (1992).

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

512

E. M. ARNETT ET AL.

25. (a) D. D. M. Wayner and D. J. Griller, J. Am. Chem. Soc. 107,

7764–7765 (1985); (b) D. J. Griller and D. D. M. Wayner, Rev.

Chem. Intermed. 7, 31–44 (1986); (c) J. A. Hawari, J. M.

Kanabus-Kaminska, D. D. M. Wayner and D. J. Griller, NATO

ASI Series C, Vol. 189, p. 91. Reidel, Dordrecht (1986); (d) T.

J. Burkey, M. Majewski and D. J. Griller, J. Am. Chem. Soc.

108, 2218–2221 (1986); (e) D. D. M. Wayner, D. J. McPhee

and D. J. Griller, J. Am. Chem. Soc. 110, 132–137 (1988); (f)

D. J. Griller and J. M. Kanabus-Kaminska, J. Mol. Struct.

(Theochem 40) 163, 125–131 (1988); (g) D. J. Griller and D.

D. M. Wayner, Pure Appl. Chem. 61, 717–724 (1989); (h) D.

Griller, J. A. Martinho Simoes, P. Mulder, B. A. Sim and D. D.

M. Wayner, J. Am. Chem. Soc. 111, 7872–7876 (1989); (i) D.

D. M. Wayner and D. J. Griller, in From Atoms to Polymers,

Molecular Structure and Energetics Series, Vol. 11, edited by J.

F. Liebman and A. Greenberg, pp. 109–128. VCH, New York

(1989); (j) D. D. M. Wayner and D. J. Griller, in Advances in

Free Radical Chemistry, edited by D. D. Tanner, pp. 159–192.

JAI Press, Greenwich, CT (1990); (k) P. H. Milne, D. D. M.

Wayner, D. P. DeCosta and J. A. Pinock, Can J. Chem. 70,

121–127 (1992).

26. J. I. Brauman, in Frontiers of Free Radical Chemistry, edited

by W. A. Pryor, pp. 23–30. Academic Press, New York

(1980).

27. (a) R. G. Pearson, J. Org. Chem. 52, 2131–2136 (1987); (b) R.

G. Pearson, J. Org. Chem. 54, 1423 (1989); (c) R. G. Pearson,

J. Am. Chem. Soc. 110, 7684 (1988); (d) R. G. Pearson, J. Mol.

Struct. (Theochem) 255, 261–270 (1992); (e) R. G. Pearson,

Acc. Chem. Res. 26, 250–255 (1993).

28. (a) R. G. Parr and R. G. Pearson, J. Am. Chem. Soc. 105, 7512

(1983); (b) R. G. Parr and W. Yang, Density Functional Theory

of Atoms and Molecules. Oxford University Press, New York

(1989); (c) P. K. Chattaraj, H. Lee and R. G. Parr, J. Am. Chem.

Soc. 113, 1855 (1991); (d) P. K. Chattaraj and R. G. Parr,

Struct. Bonding (Berlin) 80, 11–26 (1993); (e) R. G. Parr and

Z. Zhou, Acc. Chem. Res. 26, 256 (1993).

35. V. Gutmann, Coordination Chemistry in Non-Aqueous Solu-

tions, p. 15. Springer, New York (1968).

36. P.-C. Maria and J.-F. Gal, J. Phys. Chem. 89, 1296–1304

(1985).

37. A. K. Covington and T. Dickinson (Eds), Physical Chemistry of

Organic Solvent Systems, p. 5. Plenum Press, London (1973).

38. R. A. McClelland, N. Banait and S. Steenken, J. Am. Chem.

Soc. 111, 2929–2935 (1989).

39. R. L. Shriner and C. N. Wolf, J. Am. Chem. Soc. 73, 891

(1953).

40. H. J. Dauben Jr, L. R. Honnen and K. N. Harmon, J. Org.

Chem. 25, 1442 (1960).

41. S. V. McKinley, J. W. Rakshys Jr, A. E. Young and H. H.

Freedman, J. Am. Chem. Soc. 93, 4715–4724 (1971).

42. F. G. Bordwell, J.-P. Cheng, A. V. Satish and C. L. Twyman, J.

Org. Chem. 57, 6542–6546 (1992).

43. (a) E. M. Arnett, R. A. Flowers II, A. E. Meekof, A. Pourjavadi

and S. A. Walek, J. Phys. Org. Chem. 7, 663–671 (1994); (b)

E. M. Arnett, R. A. Flowers II, R. T. Ludwig, A. E. Meekoff

and S. A. Walek, Pure Appl. Chem. 67, 729–734 (1995).

44. (a) E. M. Arnett, T. C. Moriarity, L. E. Small, J. P. Rudolph and

R. P. Quirk, J. Am. Chem. Soc. 95, 1492 (1973); (b) E. M.

Arnett, D. E. Johnston and L. E. Small, J. Am. Chem. Soc. 97,

5598 (1975); (c) E. M. Arnett and K. G. Venkatasubramaniam,

J. Org. Chem. 48, 1569 (1983); (d) E. M. Arnett and K. D.

Moe, J. Am. Chem. Soc. 113, 7068 (1991).

45. J. Osteryoung, Acc. Chem. Res. 26, 77 (1993).

46. (a) Y. Okamoto and H. C. Brown, J. Org. Chem. 22, 485–494

(1957); (b) H. C. Brown and Y. Okamoto, J. Am. Chem. Soc.

79, 1913–1917 (1957); (c) H. C. Brown, D. P. Kelly and M.

Periasamy, Proc. Natl. Acad. Sci. USA 77, 6956–6960 (1980).

47. E. M. Arnett and T. C. Hofelich, J. Am. Chem. Soc. 105, 2889

(1983).

48. B. A. Sim, P. H. Milne, D. J. Griller and D. D. M. Wayner, J.

Am. Chem. Soc. 112, 6635–6638 (1990).

49. (a) V. D. Parker, Acta Chem. Scand. 46, 1133–1134 (1992); (b)

J.-P. Cheng, K. L. Handoo and V. D. Parker, J. Am. Chem. Soc.

115, 2655–2660 (1993); (c) J.-P. Cheng, K. L. Handoo, J. Xue

and V. D. Parker, J. Org. Chem. 58, 5050–5054 (1993); (d) K.

L. Handoo, J.-P. Cheng and V. D. Parker, J. Am. Chem. Soc.

115, 5067–5072 (1993).

50. (a) See, for example, R. E. Dessy, W. Kitching, T. Psarras, R.

Salinger, A. Chen and T. Chivers, J. Am. Chem. Soc. 88, 460

(1996); (b) M. R. Feldman and W. C. Flythe, J. Org. Chem. 43,

596 (1978); (c) Y. Okamoto, J. Org. Chem. 39, 4011 (1983); (d)

D. D. M. Wayner, J. Am. Chem. Soc. 112, 6635 (1990); (e) E.

M. Arnett and J. A. Harrelson, Gazz. Chim. Ital. 117, 237

(1987); (f) V. D. Parker, J. Org. Chem. 58, 5811–5815

(1993).

29. E. M. Arnett and R. T. Ludwig, J. Am. Chem. Soc. 117, 6627

(1995).

30. D. D. M. Wayner and V. D. Parker, Acc. Chem. Res. 26,

287–294 (1993).

31. (a) F. G. Bordwell, J.-P. Cheng, G.-Z. Ji, A. V. Satish, X.

Zhang, J. Am. Chem. Soc. 113, 9790–9795 (1991); (b) F. G.

Bordwell, X.-M. Zhang and R. Filler, J. Org. Chem. 58,

6067–6071 (1993); (c) F. G. Bordwell and X.-M. Zhang, Acc.

Chem. Res. 26, 510–517 (1993); (d) F. G. Bordwell, A. V.

Satish, S. Zhang and X.-M. Zhang, Pure Appl. Chem. 67,

735–740 (1995).

32. D. D. M. Wayner, E. Lusztyk, D. Page, K. U. Ingold, P. Mulder,

L. J. J. Laarhoven and H. S. Aldrich, J. Am. Chem. Soc. 117,

8737–8744 (1995).

33. (a) H. Lankamp, W. T. Nauta and C. MacLean, Tetrahedron

Lett. 2, 249–254 (1968); (b) H. A. Staab, K. S. Rao and H.

Brunner, Chem. Ber. 104, 2634–2636 (1971); (c) W. D.

Hounshell, D. A. Dougherty, J. P. Hummel and K. Mislow, J.

Am. Chem. Soc. 99, 1916 (1977); (d) J. M. McBride,

Tetrahedron 30, 2009–2022 (1974) (this paper reviews the

erroneous use of the symmetrical hexaphenylethane structure

for trityl dimer until 1968,^33aeven though the correct quinoid

structure had been demonstrated in 1904); (e) B. Kahr, D. Van

Engen and K. Mislow, J. Am. Chem. Soc. 108, 8305–8307

(1986).

51. E. B. Troughton, K. E. Molter and E. M. Arnett, J. Am. Chem.

Soc. 106, 6726 (1984).

52. N. B. Chapman and J. Shorter (Eds), Advances in Linear Free

Energy Relationships, p. 33. Plenum Press, London (1973).

53. (a) P. Wan, K. Yates and M. K. Boyd, J. Org. Chem. 50,

2881–2886 (1985); (b) R. A. McClelland, V. M. Kanagasabap-

athy, N. S. Banait and S. Steenken, J. Am. Chem. Soc. 111,

3966 (1989); (c) M. R. Valentino and M. K. Boyd, J. Org.

Chem. 58, 5826–5831 (1993); (d) R. Karaman, J.-T. L. Huang

and J. L. Fry, J. Org. Chem. 56, 188–195 (1991).

54. R. W. Taft Jr, J. Phys. Chem. 64, 1805–1815 (1960).

55. X.-M. Zhang and F. G. Bordwell, J. Am. Chem. Soc. 114,

9787–9792 (1992).

34. E. M. Arnett, E. J. Mitchell and T. S. S. R. Murty, J. Am. Chem.

Soc. 96, 3875 (1974).

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

TRIARYLMETHANES AND 9-ARYLXANTHENES AS PROTOTYPE AMPHIHYDRIC COMPOUNDS

APPENDIX

513

NMR 7·30 (2H, d, J=7·1 Hz), 7·22–7·15 (4H, m),

7·11–6·99 (6H, m), 5·39 (1H, s); ¹³C NMR 151·7, 146·8,

132·8, 130·5, 130·4, 129·9, 129·2, 129·0, 125·1, 125·6,

117·4, 43·9; 9-(p-ﬂuorophenyl), ¹H NMR 7·22 (2H, d,

J=7·1 Hz), 7·10–7·05 (4H, m), 6·99–6·889 (6H, m), 5·42

(1H, s); ¹³C NMR 151·7, 146·8, 132·8, 130·5, 130·4, 129·9,

129·2, 129·0, 125·1, 124·6, 117·4, 43·9; 9-phenylthiox-

Characterization data

Triphenylmethanes

1

Triphenylmethane, H NMR 8·60 (2H, t, J=5·7 Hz), 8·44

(2H, d, J=7·2 Hz), 8·17 (2H, d, J=7·2 Hz), 7·99 (2H, t,

J=5·7 Hz), 7·84 (3H, m), 7·72 (2H, d, J=7·2 Hz); ¹³C NMR

174·6, 157·9, 143·9, 131·5, 130·7, 129·7, 129·0, 128·6,

1

anthene, H NMR 7·55 (2H, d, J=7·8 Hz), 7·35 (2H, d,

J=7·8 Hz), 7·25 (2H, d, J=7·8 Hz), 7·18 (3H, m), 7·09 (2H,

d, J=8·4 Hz), 7·01 (2H, d, J=8·4 Hz); ¹³C NMR 142·7,

138·1, 133·4, 130·8, 130·6, 129·1, 128·9, 128·5, 128·3,

128·1, 128·0, 53·1.

1

123·3, 119·3, 118·7; 9-(p-methoxyphenyl), H NMR 8·54

(2H, t, J=6·6 Hz), 8·36 (2H, d, J=6·9 Hz), 8·26 (2H, d,

J=6·9 Hz), 7·96 (2H, t, J=6·6 Hz), 7·72 (2H, d, J=7·2 Hz),

7·39 (2H, d, J=7·2 Hz) 4·56 (3H, s); ¹³C NMR 175·6, 164·8,

159·1, 144·8, 134·9, 132·8, 129·7, 124·2, 120·2, 118·6,

115·9, 56·7; 9-(p-methylphenyl), ¹H NMR 8·44 (2H, t,

J=6·9 Hz), 8·22 (2H, d, J=7·2 hz), 8·15 (2H, d, J=7·2 Hz),

7·81 (2H, t, J=6·9 Hz), 7·52 (2H, d, J=7·2 Hz), 7·41 (2H, d,

J=7·2 Hz), 2·34 (3H, s); ¹³C NMR 172·3, 165·1, 158·1,

145·2, 133·9, 131·7, 129·2, 126·8, 121·3, 117·9, 115·8, 22·4;

9-(p-triﬂuoromethylphenyl), ¹H NMR 8·67 (2H, t,

J=8·4 Hz), 8·50 (2H, d, J=8·7 Hz), 8·14 (4H, m), 8·03 (2H,

d, J=8·4 Hz), 7·90 (2H, d, J=8·7 Hz); ¹³C NMR 174·2,

159·7, 145·9, 132·7, 132·1, 130·3, 127·3, 126·7, 125·0,

120·5; 9-(p-chlorophenyl), ¹H NMR 8·62 (2H, t, J=5·7 Hz),

8·45 (2H, d, J=8·7 Hz), 8·17 (2H, t, J=6·0 Hz), 8·00 (2H, t,

J=6·8 Hz), 7·87 (2H, d, J=8·7 Hz), 7·72 (2H, d, J=8·7 Hz);

¹³C NMR 174·7, 159·6, 145·5, 139·1, 132·8, 132·6, 131·3,

Triaryl cations

4,4Ј,4Љ-(Methoxy)triphenylmethylium tetraﬂuoborate, ¹H

NMR 7·55 (6H, d, J=6·9 Hz), 7·32 (6H, d, J=6·9 Hz), 4·07

(9H, s); ¹³C NMR 194·11. 171·23, 144·14, 133·14, 118·40,

57·74; 4,4Ј-(dimethoxy)triphenylmethylium tetraﬂuoborate,

¹H NMR 8·08 (1H, t, J=6·9 Hz), 7·52 (4H, d, J=6·6 Hz),

7·78 (2H, t, J=6·3 Hz), 7·67 (2H, d, j=6·3 Hz), 7·44 (4H, d,

J=6·9 Hz), 4·00 (6H, s); ¹³C NMR 201·0, 177·19, 149·20,

148·77, 140·39, 134·61, 130·53, 130·18, 119·31, 59·29;

4-(methoxy)triphenylmethylium tetraﬂuoborate, ¹H NMR

8·07 (2H, t, J=6·0 Hz), 7·87 (2H, d, J=6·0 Hz), 7·78 (4H, t,

J=5·7 Hz), 7·51 (4H, d, J=6·0 Hz), 7·47 (2H, d, J=6·0 Hz),

4·11 (3H, s); ¹³C NMR 202·02, 177·64, 149·22, 148·72,

140·37, 134·59, 130·58, 130·15, 119·14, 59·24;

4,4Ј,4Љ-(methyl)triphenylmethylium tetraﬂuoborate, ¹H

NMR 7·10 (6H, d, J=7·5 Hz), 7·01 (6H, d, J=7·5 Hz), 2·25

(9H, s); ¹³C NMR 196·37, 173·54, 147·38, 133·83, 118·12,

1

130·5, 124·9, 120·9, 120·4; 9-(p-ﬂuorophenyl), H NMR

8·61 (2H, t, J=7·2 Hz), 8·44 (2H, d, J=8·7 Hz), 8·19 (2H, t,

J=8·7 Hz), 8·00 (2H, t, J=7·2 Hz), 7·77 (2H, d, J=8·1 Hz),

7·63 (2H, d, J=8·1 Hz); ¹³C NMR 175·1, 159·5, 146·5,

144·3, 135·4, 133·2, 131·4, 129·2, 121·8, 119·4, 116·2;

1

23·44; 4,4Ј-(methyl)triphenylmethylium tetraﬂuoborate, H

1

9-phenylthioixanthene, H NMR 8·81 (2H, d, J=6·3 Hz),

NMR 7·93 (1H, t, J=6·9 Hz), 7·64 (2H, t, J=7·5 hz), 7·45

(2H, d, J=6·0 Hz), 7·41 (4H, d, j=6·3 Hz), 7·24 (4H, d,

J=6·3 Hz), 2·72 (3H, s); ¹³C NMR 203·1, 179·91, 151·45,

149·99, 143·01, 135·22, 131·51, 130·83, 118·95, 23·53;

4-(methyl)triphenylmethylium tetraﬂuoborate, ¹H NMR

8·23 (2H, t, J=6·6 Hz), 7·86 (4H, t, J=7·2 Hz), 7·74 (2H, d,

J=6·0 Hz), 7·68 (4H, d, J=6·6 Hz), 7·24 (2H, d,

J=6·60 Hz), 2·72 (3H, s); ¹³C NMR 203·3, 180·01, 151·57,

150·21, 143·40, 135·39, 131·54, 130·89, 119·03, 23·54;

4-(ﬂuoro)triphenylmethylium tetraﬂuoborate, ¹H NMR 8·29

(2H, t, J=7·5 Hz), 7·92–7·79 (6H, m); ¹³C NMR 210·31,

175·38, 172·11, 144·21, 141·03, 132·79, 131·33, 130·92,

8·37 (2H, t, J=6·6 Hz), 8·23 (2H, d, J=6·3 Hz), 8·02 (2H, t,

J=6·9 Hz), 7·81 (2H, t, J=6·6 Hz), 7·55 (2H, d, J=6·9 Hz);

¹³C NMR 172·4, 149·3, 138·3, 136·8, 136·1, 132·4, 131·8,

131·4, 129·53, 128·7, 128·6.

9-Arylxanthenes

9-Phenyl, ¹H NMR 7·23 (2H, d, J=7·2 Hz), 7·16–7·06

(11H, m), 5·40 (1H, s); ¹³C NMR 152·7, 145·5, 137·2, 132·9,

131·9, 128·9, 128·5, 127·3, 124·9, 119·0, 117·2, 43·9;

9-(methoxyphenyl), ¹H NMR 7·16 (2H, d, j=7·2 Hz),

7·08–6·98 (8H, m), 6·72 (2H, d, J=7·2 Hz), 5·19 (1H, s),

3·65 (2H, s); ¹³C NMR 151·79, 139·9, 130·6, 129·8, 129·7,

128·8, 125·9, 124·4, 124·3, 117·2, 115 ·0, 55·6, 43·9; 9-(p-

methylphenyl), ¹H NMR 7·72 (2H, d, J=6·9 Hz), 7·12

(4H, m), 7·04 (6H, m), 5·32 (1H, s), 2·24 (2H, s); ¹³C NMR

152·3, 145·1, 137·4, 132·7, 132·1, 129·1, 128·9, 127·5,

124·3, 118·5, 116·9, 43·7, 22·5; 9-(p-chlorophenyl), ¹H

´

119·85; 4,4,4Љ-(tert-butyl)triphenylmethylium tetraﬂuoror-

ate, ¹H NMR 7·94 (6H, d, J=8·7 Hz), 7·60 (6H, d,

j=8·7 Hz), 1·47 (27H, s); ¹³C NMR 206·03, 169·07, 142·68,

138·55, 129·56, 37·32, 30·76; 4,4Ј,4Љ-(dimethyl-

1

amino)triphenylmethylium tetraﬂuoroborate H NMR 7·32

(6H, d, J=9·0 Hz), 6·96 (6H, d, J=9·0 Hz), 3·29 (18H, s);

¹³C NMR 179·57, 156·49, 140·32, 127·07, 112·96, 40·69.

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 499–513 (1997)

Article Doi

Triarylmethanes and 9-arylxanthenes as prototypes amphihydric compounds for relating the stabilities of cations, anions and radicals by C-H bond cleavage and electron transfer

DOI: 10.1002/(SICI)1099-1395(199707)10:7<499::AID-POC896>3.0.CO;2-2

Source and publish data:

Authors:

Article abstract of DOI:10.1002/(SICI)1099-1395(199707)10:7<499::AID-POC896>3.0.CO;2-2

Full text of DOI:10.1002/(SICI)1099-1395(199707)10:7<499::AID-POC896>3.0.CO;2-2

Products guided by the article

R&D Labs maybe for 147246-89-7

Relevant to this article

Hot Product