2-halo-1.4-dimethoxybenzene.pdf

FULL PAPER

Multistep Reversible Redox Systems, LXIII [e]

2,5-Disubstituted N,N 9 -Dicyanoquinone Diimines (DCNQIs)

Syntheses, and Redox Properties

Siegfried Hünig* a , Robert Bau b , Martina Kemmer a , Hubert Meixner a[2] , Tobias Metzenthin b , Karl Peters c ,

Klaus Sinzger a[3] , and Juris Gulbis a[4]

Institut für Organische Chemie der Universität Würzburg a ,

Am Hubland, D-97074 Würzburg, Germany

University of Southern California, Chemistry Department b ,

Los Angeles, CA 90089, USA

MPI für Festkörperforschung c ,

Heisenbergstr. 1, D-70506 Stuttgart, Germany

Received August 4, 1997

Keywords : DCNQIs / Quinones / Substituent effects / Voltammetry / Crystal structures

Quinones 1a

q and DCNQIs 2a

g have been synthesized

voltammetry). Linear correlations have been found between

E 2 (OX/SEM) data of 1 and 2 with (ó m + ó p )/2 and between

in order to investigate substituent effects. It was necessary to

employ novel synthetic routes for the introduction of iodine

E 2 of 1 and 2 . Correlations have also been found between E 2

and the LUMO energies of 1 and 2 . The crystal structure of

into 1f ( 7 ), the trifluoromethyl group into 1g

i , deuterium

into 1m

p , and especially for the chloride/fluoride exchange

quinone 1i shows some special interactions due to the two

CF 3 groups, whereas the structures of DCNQIs 2d and 2g

of 1j to 1k , and 1l . With few exceptions both 1 and 2 undergo

reversible electron transfer in two steps including thermody-

link up with those evaluated earlier.

namically very stable radical cations (lg K SEM > 10, cyclic

Within the series of new quinoid derivatives [5]

2,5-disub-

in reasonable yields. Quantitative demethylation of 4a 2 c

with subsequent oxidation then afforded the expected qui-

stituted

- dicyano-1,4-benzoquinone diimines 2

(DCNQIs) have gained special importance due to the ex-

N, N

nones 1a

c in high purity and yield (Scheme 2). The use

of a superior iodination reagent, benzyltriethylammonium

ceptional conducting properties of their radical anion

salts [6] . In addition to the derivatives already reported [5] we

dichloroiodate (BTMA-ICl 2 ), has recently been reported [8] .

Although a yield of 90 % of 1c from 1,4-dimethoxybenzene

now describe syntheses and properties of DCNQIs 2 , to-

gether with, where necessary, their precursors, quinones 1

in acetic acid was originally reported, we were unable to

reproduce these results (ca. 59 % yield) , however, isolated

(Scheme 1). The charge-transfer (CT) complexes and rad-

ical anion salts derived from these DCNQIs will also be

79 % of 1c using dichloromethane as the solvent.

presented.

Scheme 2

Scheme 1

Syntheses of Haloquinones 1a

The

halogen/methoxy-substituted

quinones

1d [9]

and

1e [10]

have already been described. However, we replaced

Iodonation

the

substituted

1,4-dimethoxybenzenes

the cumbersome route to 1d [9]

by a simple one-step ap-

3a 2 c , using the well-established reagent iodine/mercury ox-

proach from the inexpensive amine 5 . The yield was rather

ide [7]

yielded the expected dihalogenated derivatives 4a

low, but this was compensated by the high purity of 1d (no

[

]

Part LXII: Ref. [1] .

isomers, Scheme 3).

Eur. J. Org. Chem. 1998 ,335

348

335

WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998

1434

193X/98/0202

0335 $ 17.50

.50/0

FULL PAPER

S. Hünig et al.

Scheme 3

Scheme 4

Direct iodination of 1,2,4-trimethoxybenzene by both of

the methods discussed above was not successful. It is al-

ready known [11] that consecutive formation of the corre-

sponding biphenyl derivative prevails. A convenient alterna-

tive is based on halogen exchange 6

7 and direct oxida-

tive demethylation [12]

of 7 to quinone 1f (Scheme 3).

Syntheses of Trifluoromethyl-Substituted Quinones 1g

The synthesis of quinones 1g 2 i requires appropriate aro-

matic precursors carrying one or two trifluoromethyl

groups. For these precursors two general routes [13]

have to

be considered: (a) Transformation of a CX 3 (X

Cl, SMe)

Synthesis of Halomethyl Quinones 1j and 1l

group into CF 3 by e.g. SF 4 or ET 2 NSF 3 (DAST). Since this

The methoxy group in 1 and 2 is the only substituent

approach requires rather unusual starting materials this

without rotational symmetry that fits into the general crys-

route was excluded. (b) Introduction of the CF 3 group into

tal lattice I 4 a/a of the radical anion salts of 2 [6] . However, a

the aromatic ring by ipso substitution of haloaromatics

CH 2 F group, which may be introduced via

CH 2 Cl, may

using (trifluoromethyl)copper. This short-lived intermediate

also yield radical anion salts of 2 with the same crystal

may be produced either from gaseous CF 3 I (using an auto-

structure.

clave)

and

copper,

conveniently [14]

from

The easily accessible bis(chloromethylated) 1,4-dimeth-

CF 3 CO 2 Na/CuI by decarboxylation. We found the latter

oxybenzene 17 [16]

is smoothly oxidized to quinone 1j [17] .

procedure to be suited for our purpose (Scheme 4).

For the preparation of 1l we therefore chose to synthesise

Starting from 8 , substitution to 9 occurs rather smoothly,

18 by exchange of the chlorine in 17 by fluorine. Of the

as does oxidation of 9 to 1g . The complications associated

several reported methods for this halogen exchange treat-

with 7 (vide supra) mean that use of the bromoderivative

ment of 17 potassium fluoride in polyethylene glycol in the

10 is advisable for the introduction of the trifluoromethyl

presence of potassium iodide [18]

worked best (Scheme 5).

group. The expected product 11 could indeed be isolated

Oxidation of 18 affords the expected quinone 1l in an equiv-

and easily oxidized to 1h . However, in the first step ( 10

alent process to 17 .

11 ) a complex reaction mixture (57 %) was formed which

contained besides 11 (65 %), the starting material (16 %), the

halogen exchange product 8 (10 %) and the corresponding

Scheme 5

diphenyl derivative (5 %). From these results it is not sur-

prising that trifluoromethylation of 2,5-diodo-1,4-dimeth-

oxybenzene ( 4c ) affords a mixture of 12 , 13 , and 14 .

Unfortunately, demethylation of 12 either using boron

tribromide or by oxidative cleavage with Ce IV is not pos-

sible. Substituting 4c by the corresponding dibenzyl ether

15 paves the way to hydroquinone 16 and quinone 1i .How-

ever, debenzylation does not proceed by catalytic hydrogen-

ation although this procedure does remove impurities (de-

halogenation) of the crude product. Instead, the benzylic

protecting groups 15 are smoothly cleaved to hydroquinone

16 by the established reagent EtSH/Et 2 OBF 3 [15] .

336

Eur. J. Org. Chem. 1998 , 335

348

FULL PAPER

Multistep Reversible Redox Systems, LXIII

Syntheses of Quinones 1m

q Labelled with Isotopes

Syntheses of DCNQIs 2a

According to our general procedure [5] outlined in Scheme

Deuterated quinones 1m 2 p are important precursors for

the production of DCNQI copper salts which have unique

7 quinones 1a

h , 1j and 1l

q are smoothly transformed

properties [6] [19] . Inspired by these results, all possible iso-

mers of deuterated 2,5-dimethylbenzoquinone-1,4 have re-

into the corresponding DCNQIs 2 .

Scheme 7

cently been prepared, albeit by a different approach [20] .

According to Scheme 6 the most reliable method for in-

troducing either deuterium (> 95 %) or CD 3 groups (via

ICD 3 ) into aromatic rings is halogen/lithium exchange

which works smoothly not only for 21 and 23 but also for

the persubstituted derivatives 19 and 27 [21] . Dilithiation of

e.g. 25 [22] is less rapid and resulted only in 85 % deuteration

of 26 . Hydrolysis of e.g. 26 is also rather difficult to achieve

(cf. ref. [23] ).

Scheme 6

Difficulties are only observed with the rather sensitive 2i

[í(C

1555 cm 21 ] which is

contaminated with either the reduction product N, N

;

2180 cm 21 , í(C

-dicy-

ano-1,4-diaminobenzene 29 or the 1,4-addition product 30

[í(C

1525 cm 21 , cf. ref. [25] ].

Even after careful recrystallization, 2i (16 %) still contains

some of these impurities. The very low energy LUMO of

2i (vide infra) points to both the strong reducibility and

nucleophilicity of 2i .

DCNQI 2q was prepared from 2,5-dimethyl-1,4-benzo-

quinone by the same route (70 %) except that 13 C-enriched

bis(trimethylsilyl)carbodiimide was employed. The latter is

easily accessible by the established route [26] from 13 C-cyana-

mide.

;

2255 cm 21 , í(C

Although

the

published

preparation

for

this

re-

agent [27]

reproduced [28] ,

13 C-

could

not

reaction

bromocyan with ammonia in ethanol starting from

afforded a nearly quantitative yield of

13 C-cyanamide.

Redox Properties of Quinones 1 and DCNQI 2

In an extension of earlier work [6] both quinones 1 and

DCNQIs 2 were studied using cyclic voltammetry under the

same conditions. According to Scheme 8, two reversible

one-electron transfers at E 2 and E 1 are expected. This ideal

behavior is followed by nearly all quinones 1 and DCNQIs

2 , including the highly sensitive 2i (R 1 ,R 3

CF 3 )asexem-

plified by Figure 1.

However, in accordance with earlier results [5] , the second

reduction step of quinones 1 is not fully reversible, probably

due to aggregational effects. Therefore, in Table 1, where all

Oxidative cleavage of the hydroquinone dimethyl ethers

20 , 22 , 24 , and 28 afforded the corresponding quinones

relevant data are collected, only approximated semiquinone

formation constants K SEM are given.

1j 2 m in high yield. The amount of deuterium in the reagent

CD 3 I (> 99.5 %) is fully transferred to the products (MS).

CH 2 Cl and CH 2 F cause decompo-

sition on introduction of the second electron into 2j SEM .

Obviously, R 1 /R 3

The dideuterated quinone 1m has already been prepared

from perdeuterated p -xylene using deuterated reagents in

This is especially true for semiquinone radical anions of 1j

and 1l where the working electrode (Pt) is already blocked

all steps [24] .

Eur. J. Org. Chem. 1998 , 335

348

337

FULL PAPER

S. Hünig et al.

Scheme 8

Table 1. Potentials E 1 and E 2 from cyclovoltammograms of quino-

nes

and

DCNQIs

CH 2 Cl 2

versus

Ag/AgCl/MeCN,

n Bu 4 N 1 BF 4 2 ; n

number of formally transferred electrons, log

K SEM

semiquinone formation constant (ferrocene

539 mV)

R 1 /R 2

E 2 [V]

E 1 [V] [a]

lg K SEM

Cl/I

0.01

0.56

0.74

0.45

Br/I

0.03

0.37

0.77

0.30

I/I

0.02

0.43

0.77

0.25

Cl/MeO

0.25

0.9

1.01

0.21

Br/MeO

0.22

0.33

1.06

0.16

Figure 1. Cyclovoltammogram of DCNQI 2i (R 1 ,R 3

I/OCH 3

0.25

0.59

1.01

0.27

CF 3 )in

CH 2 Cl 2

versus

Ag/AgCl/CH 3 CN;

n Bu 4 N 1 BF 4 2 ,

scan

rate

CF 3 /CH 3

0.13

0.47

0.97

0.19

CF 3 /OCH 3

0.13

0.66

1.00

0.30

100 mV/s (ferrocene

539 mV)

CF 3 /CF 3

0.23

0.35

0.68

0.18

j [b]

CH 2 Cl/CH 2 Cl

0.36 [c]

0.42

irr.

l [b]

CH 2 F/CH 2 F

20.31

0.50

irr.

R 1 /R 3

E 2 [V]

E 1 [V]

lg K SEM

Cl/I

10.64

0.59

10.01

0.59

10.7

Br/I

0.65

0.54

0.02

0.56

10.7

I/I

0.63

0.74

±0.00

0.84

10.7

Cl/MeO

0.43

0.74

0.19

0.84

10.5

Br/MeO

0.44

0.33

0.20

0.54

10.9

I/OCH 3

0.40

0.74

0.20

0.74

10.2

CF 3 /CH 3

0.55

0.74

0.14

0.74

11.6

CF 3 /OCH 3

0.53

0.74

0.15

0.74

11.4

CF 3 /CF 3

0.83

0.79

0.06

0.79

13.1

j [b]

CH 2 Cl/Cl 2 Cl

0.35

0.91

irr.

l [b]

CH 2 F/CH 2 F

0.33

0.98

0.35

0.42

11.5

[a]

Quasi reversible.

[b]

In MeCN.

[c]

Not fully reversible.

10 3 compared to the K SEM values of the corresponding qui-

nones 1 . These differences are probably mainly due to

smaller Coulomb repulsions in the more extended DCNQI

after only one scan. It seems probable from Scheme 9 that

ð-systems. It has already beeen demonstrated [30]

that the

a halide ion is expelled from 1j RED , 1l RED and 2j RED .

Coulomb integral J mm of the two electrons in the HOMO

Scheme 9

of the reduced form is basically responsible for Ä H R of the

equilibrium OX

RED v 2 SEM. A quantitative corre-

lation between J mm from SCF calculations and log K SEM

from CV data has already been derived for other two-step

redox systems [31] .

Figure 2 demonstrates an excellent linear correlation ( r

0.993) between E 2 (quinone) and E 2 (DCNQI) for 1a

and 2a

o together with all corresponding derivatives pub-

lished so far.

The gradient a

1,2-Quinonemethides 31j and 31l or the 1,2-quinonei-

mines 32l are formed in this way. Both classes of com-

0.82 indicates that the differences for

the first reduction potentials ( E 2 ) become smaller with in-

creasing acceptor strength of the substrates. This substitu-

pounds are highly reactive (dimerization, polymerization

etc. [29] ). In 1 RED with X

O 2 the negative charge is bet-

ent effect may either originate from differing solvation ener-

gies of OX and SEM (ÄÄ G solv ) or from the differing elec-

ter available for the extraction of

Y 2 than with X

CN 2 . Therefore quinones 1j and 1l are expected to

tron affinities. Since experimental data are not available for

the latter, Koopman

decompose more readily than DCNQIs 2j (or 2l )onre-

duction. Furthermore, fluoride is a much worse leaving

s electron affinities (

å LUMO [eV])

were calculated. In Figure 3 these data are correlated with

the corresponding E 2 potentials.

group than chloride, especially in aprotic solvents. This is

consistent with the relative decomposition rates 1 RED >

The two linear correlations shown in Figure 3 are both

very good. The gradients differ by 0.08 mV/eV and demon-

1l RED and 2j RED >> 2l , although even the latter shows

strongly diminished concentration in the second reduction

strate again the weaker substituent effect with increasing

electron affinities. In the region where the two curves nearly

wave.

In full accord with earlier observations [5] [6] , K SEM values

meet (at 2.2

2.3 eV) their difference amounts to only 0.05

V, well within the limits of error for these semiempirical

of the DCNQIs 2 are smaller by a factor of approximately

338

Eur. J. Org. Chem. 1998 , 335

348

FULL PAPER

Multistep Reversible Redox Systems, LXIII

An excellent correlation is obtained again even for such

Figure

Correlation

the

potentials

E 2

(quinone)

and

E 2

(DCNQI) for 1a 2 o and 2a 2 o (h) together with those of all other

extreme electron attracting substituents as CF 3 /CF 3 .

2,5-disubstituted derivatives (O) published

Figure 4. Plot of the first reduction potentials E 2 of 2,5-disubsti-

tuted quinones versus the Hammett parameter (ó m

ó p )/2; (

)

former data [6] ,(

) data from this paper, (

) excluded from the

calculated correlation

Figure

Correlation

the

potentials

E 2

(23

quinones

and

DCNQIs) with Koopman

s electron affinities (negative LUMO

energies) calculated with the AM1 program

Figure 5. Plot of the first reduction potentials E 2 of the correspond-

ing DCNQIs versus the Hammett parameter (ó m

ó p )/2; (

) for-

mer data [6] ,(

) data from this paper, (

) excluded from the calcu-

lated correlation

correlations. From these results different solvation energies

for the systems under discussion become very improbable,

leaving intrinsic (electronic) factors as the most plausible

On the other hand stronger deviations are observed for

the combinations CH 3 /F, allyl/allyl, MeS/MeS, MeO/MeO

reason for the observed effects. As already demonstrated

there exists a linear correlation between the first reduction

which had to be excluded from the correlation E 2

a (ó m

potentials E 2 of disubstituted quinones together with their

DCNQI derivatives and (ó m

ó p )/2

b to obtain r

0.993. With the exception of

ó p )/2 [32] . The results pres-

CH 3 /F, the deviating substituents are bent and may prefer

to adopt different conformations in OX and SEM as sug-

ented in this paper are collected in Figure 4 and 5 and

complement this previously published data.

gested by AM1 calculations [3] . The dependence upon the

Eur. J. Org. Chem. 1998 , 335

348

339

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: