Thermodynamic stabilities of complexes of synthetic macrocyclic valinomycin mimics with alkali, barium, and guanidinium cations

Trrruhrdron Vol 43. No Pnnted m Great Bntam

2. ,p

0(MO4020:87

397 to 404. I987

Pergamon

T¶iElW3DYNAUICSTAF3ILITIBSOP VALINCUYCIN

UIMICS

WITH -1,

CQlPWXBS BARIW,

op smEr1c

55

WI+00

Journals

Ltd

f4immCYCLIc

AND GUANIDINIIJR CATIONS

PETER D. J. GRODTENHUIS, PETER D. VAN DER ktAL, and DAVID N. REINHOUUT*

Laboratory of Organic Chemistry, Twente University of Technology, 7500 AE Enschede, The Netherlands.

(Rrcelcrdin UK 30Scprember

1986)

Association constants of the 1:l complexes of a series of pyrido crown ethers 1 (n=O-6) and a number of large ring (27-33 ring atoms) benzo and dibenzo crown ethers 2,3 with alkali, barium and guanidinium chlorides have been determined potenticmetrically in MeOH. For most of the ions besides a global maximum for la-membered rings, a local maximum for the complexation with 30membered rings is observed. The appearance of local maxima is attributed to the tennisball-seam like conformations that large macrqyclfs can adopt. The larger crown ethers show a remarkbly high K /Na selectivity similar to the antibiotic valinanycin, a structurally related 36membered macrocycle. Therefore, these large ring crown ethers can be regarded as synthetically easy to modify valinomycin mimics. The effects of the solvation of the hosts, guests and complexes on the overall (co-)ccmplexation is discussed. Comparison of data from extraction experiments with association constants clearly shows that differences in the distribution of the complexes between the aqueous and organic phases may be a dominating factor for the selective extraction of cations. Abstract-

The ability of biochemical systems to recognize and selectively bind guest species, represents one of their most

important properties. In order to mimic the selective

binding properties of the natural ionophores, several classes of macrocyclic ligands have been designed and synthesized. Cram and co-workers have recently compared these classes and they concluded that there seems to be a relation between the thermodynamic stability of host-guest complexes and the binding selectivity of the hosts towards the guests.' The following order was observed: spherands > ct-yptahemispherands> cryptands > hemispherands >

chorands

(crown ethers) >

podands, for both

the

stability of

the

complex and

selectivity of the host. Most of the work in the host-guest chemistry has been focussed on the relatively small crown ethers with 12-21 ring atoms and an overwhelming number of thermodynamic data on the canplexation of spherical cations by these ligands

has

been

collected.2 For the

optimal ccmplexation of polyfunctional cations such as (thio)uronium,3'4 and guanidinium salts,' we have shown that larger ligands with 27-33 ring atoms are needed, because otherwise relatively weak perching complexes6 are formed. We are currently interested in the selective transport of such polyfunctional cations through membranes in competition with other, spherical cations.7 Therefore, we decided to study the thermodynamics of complexation of macrccycles with 27 to 33 ring atoms towards several cations because hardly any data are available in the literature.8*g Other reasons to perform a systematic study originate from our work on the design of ion-selective field-effect transistors (ISFETS)." The aim of that work is the selective conversion of chemical information into electrical signals. Until now natural ionophores exhibiting a high binding selectivity for one ion amongst other ions (e.g. valinanycin: KK/KNa = 6700 ,MeOH)2a were used in the sensing components of ISFSTS.~~ Since valinomycin (36 ring atoms) is structurally related to large ring crown ethers12 we wanted to explore the possible selective ccmplexation behaviour of the large ring crown ethers, hoping to 397

P. D. J. GROOTENHUIS et al.

398 barium Table

cat ions

6

d 0

generally

show the

following

KLi < KNa < KK,

trend:

KRb, Kc’

< KBa

( see

I).

b

9 \ 0

N

\ 1 n=O

: Bl8C6 : B27C9

P21C7

5

: B3OClO

P24C8

6

: B33Cll

5

2

5

: [2,71 cB27C9 : [3,71 DB3OClO : [4,71 DB33Cll

: P27c9

0

6

:

: P3OClO : P33Cll

1

1

: I3,31 B18C6

3

3

:

ligands

33membered

2 and 3 the association form relatively

macrocycles.

and crown ethers [4,71CB33Cll

Association are

rather

Table

Surprisingly

2 and 3 are

exhibits

crown ethers is

1

5

30-membered ligands

cation

n=O ,m=5

4 6 For

4

n=l

3

2

the

P15C5 P18C6

: : : :

1

50-800

a relatively

constants also

given

of

the fold

strong the

data

smaller

K valuesa

crown ethers

and

than with

of

for

Rb+

the

(Gu+) of

cations

for

1:l

cations.

complexes

B18C6b

I

4.2

5.28

4.86

3.88

5.35

I I

1.88C

4 .oo

3.88C

4.57

3.93

1.93

4.28

4.49

4.35

5 .Ol

1.83

224

I I I I I

1.91

3.57

4.27

3.88

4.30

1.37

46

1.90

4.36

4.31

4.35

2.83

(1

288

1 .98e

4.47

4.6ge

4.38

3.94

1.50f

309

1.82

3.79

4.55

4.72

>5.5

1.37

93

1.62

437

B3OClO B33Cll

[2,71

DB27c9

[3,71 DB3OClO c4.71 DB33Cll [2,81 DB3OClO I3,31 DB18C6b [5,51 DB30ClOb

I I I I

with

of

M Et4NClJ at 25.0 OC cs+

27-

and

of

Na+

Ba2+

1.98

4.62

4.89

4.79

2.99

4.36

4.96

4.23

3.23

4.28

2.11

4.57

4.64

4.2

several

the polyfunctional

I

B27C9

the

and Cs+ cations.

1 igand

Na+

For most ions

complexes

the K+, Rb+,

The binding

K+/Na+ selectivity

K’

of

the centrosymmetrical

2 and 3 in MeOH (0.1

II.

Ba2+ and Cs+.

guanidinium

I and II. of

in Table

[5,51 DB3OClO

when compared with

constants

preference

complexes

in both Table

collected

complexes

association

weak compared to the binding

II. kg

are

stable

l2,81 DB3OClO

Gu+

KK/KNa 10

l.77d

132

4 288

(8.d. aThe log K values for Na+, K+, and Rb+ were determined directly and the others in competition experiments (s.d. 0.2). The 0.041, taken as the average of the association values for these ligands, in the literature, inclqjed dfo; constants rqported Ref +2a, were Ref 19, by polarcgrapp: comparison. 2.01 (Na 1, and 3.&8 (Rb 1. at 20.0 C: +1.7# (GuSCN). Ref 19, by conductometriy 19, by Ref 19, by conductonmtry at pola&ography: 2.18 (Na 1, and 5.09 (Rb ). 20.0 C: 2.00 (GuCl).

Gu+

399

Synthetic macrocyclic valinomycinmimics

find valincmycin mimicry and with that a much higher cation selectivity than was found for small ring crown ethers. Finally it is noted that in order to fully canprehend the ion canplexating properties of

synthetic

ionophores in

correlate the relative themdynamic

transport or

extraction systems one

has

to

stabilities of complexes and the partition of the

complexes and its components over aqueous and organic phases. It is known from X-ray data that [5,51DB3OClO 3 (n=m=3) wraps around K+ and Bb+ ions.13 The resulting conformation of the complex resembles to some extent a tennisball-seam. The complexes formed in this way show a more pronounced lipophilicity than the corresponding complexes with 18-crown-6, in which the guest is still partly solvated as theoretical studies indicate.l4 Therefore, the degree of desolvation of the guest cation , which is an important contribution to the overall free energy of canplexation,2b is expected to be very large in the case of the 21-33 membered macrocyles. This paper describes the thermodynamics of canplexation of 15 to 33membered crown ethers both with centrosynnnetricaland with polyfunctional cations. The association constants were determined potentiometrically in methanol (0.10 M Et4NCl) at 25.0 OC, using direct15 and competition16 methods.

In order to study systematically the effect of an increasing ring size of the crown ethers on the stabilities of the complexes, we have measured the association constants (K) of alkali cations with a series of pyridc crown ethers 1 (n=0-6)5b (see Table I). Table I. Log K valuesa and K+/Na+ selectivities for 1:l ccanplexes of 2,6-pyrido crown ethers 1 in MeOH (0.1 M Et4NC1) at 25.0 OC ligand

)

Li+

PlSC5

2.19

3.34

3.01

2.63

2.44

3.01

4.25'

5.30'

4.66'

4.08


1.91

3.79

4.36

2.81

5.37

P24C8

I I I I


2.27

2.89

3.25

3.28

2.91

P27C9

I

2.02

1.60

3.23

3.08

3.20

5.45

1.18

42

( 1

2.12

1.71

3.15

3.44

3.51

>5.5

1.32

27

2.55

1.44

3.00

3.28

3.24

2.97

1.44

36

P18C6 P21C7

P3OClO P33Cll

Na+

K+

Rb+

cs+

Ba*+

Gu+

KK/KNa

5.05

b

0.5

>5.5

b

11

b

76

b

4

aThe leg K values for Na+, K+, and Rb+ were determined directly (8.d. 0.04), an$ the others in competition experiments 4s.d. 0.2)+ Not measured* Ref 17, by calorimetry at I=O.OOS: 4.09 (Na ), 5.35 (K ), and 4.56 (Rb ). From the data in Table I it is concluded that for all the cations a global maximum in the stability constants of

the complexes is observed for the

18membered

ligand. This

observation seems to be rather surprising since the ionic radii of the guests vary from 0.78 R (Li+) to 1.65 ft (Cs+).18 According to the so-called 'hole-size cation-diameter' relationship2p16f one would expect that the larger ligands would preferentially bind the larger cations. However, it is known that besides the ring size more factors contibute to the binding of cations, e.g. (de)solvation, conformational flexibility, the number, spatial arrangement, and electron density of the binding sites, and the type of the anion. When the ring size is gradually increased to 24 ring atans the association constants of the complexes decrease. Further increasing the ring size leads to an enlargement of the association constants for most of the canplexes. Noteworthy a sort of local maximclm in the stability constants of the complexes with ligands possessing 27 to 33 ring atans is observed. The association constants of the pyrido crown ethers 1 with the alkali and

P. D. J. GROOTENHUIS ef al.

400

RiclgSize. The stability ring

sizes

data

set

of

15-33

to verify

relatively

ring

maximum in binding the the symmetric explained

by

30membered

the

ion

in

solution

are

able

way that

interactions

crown ethers

in which the size

of

in Table

atoms possess could and

one

the

the

ligand

reported

complexes

27 to

conformations. a

atom -

salts

cation It

l3

rings

This

is

due to

formed.

conformational

Previously

we

a

proton

ring

size

can

Possibly

able

have

to

a

and

that

the

also

with

Recently

with

a cation

ethers

and

Stoddart

crown ethers

explanation

crown

factor

to

30 ring

complex

pyrido

seem

respect

An alternative

for

in

conformations

crown ethers

stabilizing

dependent.2l

and 36-membered dibenzo

the

with

simultaneously

shown

be

of

wrap around

the 30-membered cycles

properties.

are

studies

minimalized

ligands

be

complexes

twisted the cation. 20

of

can

X-ray

the

can adopt

to the size

of the canplexes

to

In

are

the 27- and 33membered

30-

ligand.

known from NMR studies

may be deduced that

is

with

The global

atoms,

the macrocycles

distances

is

(heterolaromatic

of

studies

that

33 ring

According

conformation.

to

that

of

enthalpy

it

solvation

of

donor

favourable

only

the la-membered

binding

1 having

an excellent

performed.16f

I and II

molecule.

macroring-assisted

for

possessing

such

30-membered crown ethers

solvent

basicity

hosts

is adjusted

stabilities

provide

Hitherto been

in comparison

relatively

be that

found

and rubidium13b

the

have

crown ethers

the ligand.2a’c

for adopt

possessing

to be somewhat favoured the themdynamic

is

are most favourable.

the cavity

From the data

atoms)

a favourable

like

to

potassium13atcrd

such

electrostatic

of

we observe

‘tennisball-seam’

ligands

[5,51 LB3OClO with guest

geometry

of

pyrido cations

relationship.

ring

terms

of

and barium

the cations

in

and planar

the complexes alkali

(12-24

towards

explained

maxima that

of

the

cation-diameter

macrocycles

ability

been

Local

constants

atoms , with

the hole-size

small

has previously

ring

that

leads

to

a

and coworkers22

lithium

picrate

and

cations

and ligands

water.

Cation Selectivity. If would

be

the

known,

complexes

and the

the

strength

of

the

solvation

energies

for

are

available

for

ligands.

MeGH increase for

Na+

more

than

energies

the

between

coordinating associated

to

preference for

the

K+,

the

performed

by

Oven

reveals

that

authors

suggest

-one sodium

of

in

and

the that

the

ion would

by

of

the

host and guest. Although data known, 23 no quantitative

solvation

be the

(-AGO)

reason

to

Kollman (1:l)

with

replacement

the larger

the

and

of

the

cations

The very

that

large

in complexes 22 The

cation.

coworkers.

complex Kf,

due

Na+.

of

It

actually

to

the

the solvation

macrocycles

of a number of

solvent

might

of

appeared

is

in

value Li+

differences

in

that

solution in

differences

a in

was and the less

solvation

of

Li+,

the

free

Na+,

and

energies

of

Truter,24 the

ligand

unfavourable

be placed

in

the

a

Gu+. is

the

for

KK/KNa = 437 for

central

a

somewhat of

is

ligands are

the

interactions

cavity

cation

by the entropy

rather 2

and

3-4 kcal/mol

the

strong 3

this

smaller

30-membered macrocycles

[2,81 Dl33OClO. X-ray

[5,51 LB3OClO .sodium possesses

the

crown ethers

For

of ccmplexation

non-bonding

of

upon complexation.

pyrido

The K+/Na+ selectivity of

sphere

be due to the favourable

molecules

and Sa2+ by the

Rb+, Cs+,

canplex very

solvation

may dominate the overall complexation process as 14a and distance gecmetry14b studies of 18-crown-6

a maximum value

with

the

and the complexes.

ccmplexation

binding

could

coordinated

ccmplex

is even more clear:

the most clear,

of

on

between

in MeOH are

energies

18-crown-6 .Na+

atoms of of

the various

information

Cs+ < Rb+ < K+ << Na+ << Li+.

performed

for

to the release

compared

cations

stable

force

the cations

still

mechanics

K+, Na+,

&nor

The binding

than

the

for

yield

interactions

kcal/mol)

corresponding

The driving

mutual

order: are

K+ complexes,

intrinsically stable

of

energies

could

The free

following

by molecular and

the

most of

molecules

energies

illustrated

data

(AGo = -115.0

ions solvent

solvation its

the

in the

lithium

number of

solvation

association

thiocyanate twisted would ligand.

(1:2)

analysis, complex

conformation. be

introduced

Since

The when

we observed

401

Synthetic macrocychc valinomycm mimics only of

1:l

the

this

stoechianetries,

larger

ligands

Although

towards

might

the K+/N~+ selectivities

than the corresponding

selectivity

might

simple

expect

for

such

mDre structured

cryptands.2

as

mimics

valinanycin

linked

to solid

of

of

21) and

in

MACH is as

cation

is

Pyrido

the

the

units

decrease

pyridine

than

association

is

the

than

15-65

abilities

ring

for

smaller than one

measured

macrocycles and

times

higher

for

the

can be regarded

instance

All

(1.43

~a~+

energy

the

cation

is

of

A)

covalently

of

Ba2+ show large

is

in between

this

tightly

the

doubly the

that

charged

monovalent

of

the

barium

coordinated

The

rather

by

the

small

However,

better

by the

isarters

with

the

and this

TI-systems

of

than with

seems hard to rationalize in

oxygen

with

addition

atoms. in

the

benzo

However,

MeOH of

these

the

two benzo moieties,

The differences that

they

in

the

following

Na+

the

[2,81

two benzo

the

binding

1 (n=4-6)

basic

of H-bonds.

ligands

might

show smaller

and

solvation

indicating

with

this

oxygen

available.

decreases

canplexes the

sight

due to the formation

binding

study ligands

catecholic

4, and 5 are

are

this

atom than

of

for

ligands

in

At first

donor

densities

used

membered pyrido

2 (n=7-9).

[5,5lDB3OClO.

between

ccmplexation

of

solvation

guests 27-33

a better

3 (n=l),

the

constant.

the

the solvation energies for 23b Therefore, desolvation

For the 30membered

isomers

Cs+

for

complex

the

compensated

form more stable

interactions

3 are

substantially to the values

large

radius

solutions.

in the

with

the free

the

and

K values

However

if

I-rim.

[3,71LX33OClO > other

binding

the ligand.

electron

isomers

for

Fb+

functionalized

aqueous

ligands

atom

more

different

With

close

these

The ionic

Benz0 Moiety.

benzo

moiety of

constants

are

2 -

are

reason

much larger

in

canplexes

the

structural three

case

nitrogen

are

they

easily

The kg

be achieved

the corresponding

effects

probably

versus

for

since

and they

be

(l.49).18

donor atoms of

constants

can

up to >5.5.

Rb+

the

can only

macro ring

the reduced

the macrocycles

valincmycin,

For this

Barim.

from 2.83

cations,

for

supports.25

mlexation K+ (1.33

of

of

chorands

that

fluctuations cation

be an explanation

Na+.

order:

moieties

arcnnatic

form similar

isomer

be due to rings

canplexes.

exhibits

the

the

to

of

free

>

lowest

close

elimination

in

for

I2,81~~30ClO

spatially

the

data

in canplexation

each

repulsive

ligands

upon

complexation. Comparison competitive appeared

that

Zl-membered organic Li+

Li+

but

phase

was

Gu+.

The

The larger These

observations

than

with

should

(the

be

found

yet).

in

the

one or

to

be more lipophilic

two solvent

complexes polarity

increase of

the

molecules

solvent

to CDc13 phases

caution

is

required effect

are

considerably

aqueous

which this

than

is

the

in

not evaluated.

ethers

the

distribution the

larger

are

be

Gu+ This

is

the

ethers

determined

the

not extract which

Li+

are

case

since

of with

solubilities Li’

of

complexes

no

the

two

Gu+ are probably the

Gu+

when the

extracted clearly

by extraction

show

larger

contradiction

in which

preferentially

specific

to

phases

complexes, the

water

determined

Li+

of

and

apparent the

It

15-

did

with

crown

As a result

Therefore,

this

between

solubilities

crown ethers. constants

not of

(n=O-6)

the

data,

1 (n=4-6)

pyridc

corresponding

by

transfered

potentianetric of

1

previously.5b

atcms (n=3-6)

CDC13 could

co-ccnnplexed.

decreased.

when association is

of

crown

a CDc13 phase

24 ring

an interpretation

more than the

by the pyrido

to

the canplexes

and

canplexes

pyrido

preferably

contradict

constants

solubilities

an aqueous at least

of

that

the

Li+ and Gu+ has been studied

macrocycle with

seem to

We feel

canplexes. The encapsulated expected

fran

constants

association

available

of

18membered

in MeOH the association Gu+

best

macrocycles

that

methods are

transfer

extracted

ligands.

phase.

For

with !zxtraction &xperiments.

liquid-liquid

from

shows that

methods,26

in

P. D. J.GIWTENHUIS efal.

402

Conclusions The results described in this paper reveal that crown ethers with 27-33 ring atoms show unexpected high selectivities for K+, Rbf, and Cs+ amongst other ions like Na+ and Li+. They may therefore be seen as valinanycin mimics , which can be chemically modified rather simple. A careful reexamination of the hole-size cation-diameter relation shows that two maxima in binding abilities towards a number of cations are encountered: a global one, for la-membered macrocycles, and a local one for 30-membered macrocycles. Experimental Section Materials. The pyrido,5b benxo,7 and dibenso7 crown ethers were prepared according to literature procedures. The alkali, barium, and guanidinfum chlorides were of Suprapur quality (Merck). Tetraethylarmnoniumchloride (Merck) was recrystallized three times from chloroform/ether and dried for 24 hours at 120°C under low pressure (lo-12 mm Hg). Triethylamine (Merck) and Methanol (Merck, 99.8%) were p.a. grade. mtentimtric

Measu-nts.

The stability constants were determined by potentio-

metric measurements using a Radiometer Copenhagen FHM 26. The K values of the Na+,

K+,

and Rb+ canplexes could be determined directly by the Frensdorff method,16 using a Na+ (Metrotxa,6.0501.100) or a K+(Rb+) (Ingold, pK271) selective electrode, and a @/AgCl electrode as a reference electrode. The electrodes were calibrated daily in standard cation solutions. All solutions were prepared using a methanolic stock solution of constant ionic strength (0.10 M Et4NCl), to which 5.05 g/L triethylamine was added in order to prevent interference of H+ ions. The titrations were carried out in a thermostated cell at 25.020.1 OC under a nitrogen atmosphere. In a typical experiment 0.05 M crown ether dissolved in MeoH was added in lo-15 increments frcnna 250 DL Hamilton syringe to 20.00 mL of a 0.001 M salt solution. Between every addition a 2 minute waiting time was maintained. Each emf measurement (20.1 mV) was obtained at constant stirring speed. In this way at least lo-12 data points were collected which were used for the final iterative calculation of the association constants by the SUPEFQUAD program27 on a DEC 2060 ccmputer. 1:l canplexation was observed in all cases and indications for 2:l or other stoechicanetrieswere absent (SUPERQUAD offers facilities to test various models). For the Li+, Cs+, Ba2+, and Gu+ ions a competition method16 was used with K+ as indicator ion in solutions with approximately eguimolar amounts of cation and potassium chlorides. In the case of the guanidinium salts a lo-fold excess of this salt was used because of the relatively small association constants of the ccmplexes. In general log K values between 1 and 5.5 can be determined by the method described in this paper. Acknowledgement Ws warmly thank dr. E. J. R. Sudhalter and dr. ir. M. Bos for stimulating discussions and drs. C. J. van Staveren for providing the pyrido crown ethers. This investigation was supported by the Netherlands Foundation for Chemical Research (SON) with financial aid from the Netherlands Organization for the Advancement of Pure Research (ZWO). mferences and Notes (11

D. J. Cram and S. Peng Ho, J. Am. Chem. SCX. 108, 2998 (1986).

(2) For recent reviews see: (a) R. M. Izatt, J. S. Bradshaw, S. A. Nielsen, J. D. Lamb and J. J. Christensen, Chem. Rev. 85, 271 (1985). (b) Y. Inoue and T. Hakushi, J. Chem. Sot., Perkin Trans. 2 935 (1985). (cl F. de Jong and D. N. Reinhoudt, Adv. Phys. Org. Cbmn. 17, 279 (1980).

Synthetic macrocyclic vahomycin (3)

J. w. H. M. Uiterwijk,

S.

mimics

403

Harkema, D. N. Reinhoudt, K. Daasvatn, H. J. den Hertog

Jr. and 3. Geevers, Angew. Chem., Int. Ed. Engl. 21, 450 (1982). Anqew. Chem. supp1. 1100 (1982). (4)

D. N. Reinhoudt, J. A. A. de Boer, a

(5)

J.

W.

H. M. Uiterwijk and S. Harkema, J. Org.

50, 4809 (1985).

(a) J. A. A. de Boer, J.

W.

H. M. Uiterwijk, J. Geevers, S. Harkema and D. N.

Reinhoudt, J. Org. Chem. 48, 4821 (1983). (b) J. W. H. M. Uiterwijk, C. J. van Staveren, D. N. Reinboudt, H. 3. den Hertog Jr., L.Kruise and S. Harkema, 3. Org. Chem. 51, 1575 (1986). (6)

S. Harkema, G. J. van Hummel, K. Daasvatn and D. N. Reinhoudt. J. Chem. Sot., Chem. Cmun.

368 (1981).

(7) T. B. Stolwijk, P. D. J. Grcotenhuis, P. D. van der Wal, E. J. R. Sudholter, D. N. Reinhoudt, S. Harkema, J. W. H. M. Uiterwijk, and L. Kruise, J. Org. Chem., in press. (8)

See Ref. 2a for sane data for

f4.51UB27C9, B3OC10, [5.5]Df33OClO,and dimethyl

(5,51DB3OClO. Literature data for sane of the macrocycles that we have studied are included in Table II. (9) Whenever was

possible codes were used

in order

to name the macrocycles and

distinguish between the various structural isomers. Abbrevations: D = di, B = benzo, P = pyrido, C = crown. In the presence of several functionalities the numbers of hetero atoms between the functionalities (p,q) are indicated as follows: IP&ll.

(10) A. van den Berg, P. Bergveld, D. N. Reinboudt and E. J. R. Sudtilter, Sensors and Actuators 8, 129 (1985). (11) S. D. Moss, J. Janata and C. C, Johnson, Anal. Chem. 47, 2238 (1975). (121 R. A. Schultz, B. D. White, D. M. Disbong, K. A. Arnold, G. W. Gokel, J. Am. Chsm. Sot. 107, 6659 11985), and references cited herein. (13)

(a) M. A. Bush and M. R. Truter, J. Chem. Sot., Per-kinTrans. 2 345 (1972). (b) J. Hagek,

K.

Hum1 and D. Hlavata, Acta Crystallcgr., Sect. B 35, 330 (1979).

fc) J. Hagek, D. Hlavata and K. Huml, Acta Crystallogr., Sect. B 36, 1782 (1980). Id) J. D. Owen, M. R. Truter, and J. N. Wingfield, Acta Crystallcgr., Sect. C 40, 1515 (1984). (14)

(a) G. Wipff, P. Weiner and P. Kollman, J. Am. Cbem. Sot. 104, 3249 (1982). (b) P. K. Weiner, S. Profeta Jr., G. Wipff, T. Havel, I. D. Kuntz, R. Langridge and P. A. Kollman, Tetrahedron 39, 1113 (1983).

(15) H. (16)

K.

(a) J.

Frensdorff J. Am. Chem. Sot. 93, 600 (1970). M.

Lshn

and J. P. Sauvage, J. Am. Chem. Sot. 97, 6700 (1975).

fb) J. Gutknecht, H. Schneider and J. Stroka, Inorg. Chem. 17, 3326 (1978). (c) B. G. Cox, N. van Truong and H. Schneider, J. Fun.Chem. Sot. 106, 1273 (1984). (d) P. Fux, J. Lagrange and P. Lagrange, Anal. Chem. 56, 160 (1984). (e) M. Brighli, P. Fux, J.

Lagrange and P. Lagrange, Inorq. Chem. 24, 80 (1985).

ff) G. W. Gokel, D. M. Goli, C. Minganti and L. Echegoyen, J. Am. Chem. Sot. 105, 6786 (1983). (17)

J. D. Lamb, R. M. Izatt, S. W. Swain and J. J. Christensen, J. run.Cbem. Sot. 102, 475 (1980).

(18)

8. Dietrich, J. Chem. Ed. 62, 954 (1985).

(19)

D. Ph. Zollinger, Thesis, Twente University of Technology, Enschede (1984).

(20)

(a) D. N. Reinhoudt, R. T. Gray, F. de Jong and C. J. Smit, Tetrahedron 33, 563 11977).

(21)

(b) M. Shamsipur and A. I. popov, J. Fun.Cbem. Sot. 101, 4051 (1979) P. D. J. Grootenhuis, J. W. H. 13. Uiterwijk, D. N. Reinhoudt, C. J. van Staveren, E. J. R. Sudhtilter, M. Bos, J. van Eerden, W. T. Klooster, L. Kruise and S. Harkema, J. Am. Chem. Sot. 108, 780 (1986).

P.

404

D. J.GROOTENHUISet

al.

(22) S. M. Doughty, 3. F. Stoddart, H. M. Colquhoun, A. M. 2. Slawin and D. J. Williams, Polyhedron 4, 567 (1985). (23)

(a) H. Strehlow, In "The Chemistry of Non-Aqueous Solvents", Part I; Lagowski, J. J., Ed.; Academic Press: New York, 1966; pp 129-171. (b) Y. Marcus, Ion Solvation, John Wiley and Sons, Chichester (1985).

(24)

J. D. Owen and M. R. Truter, J. Chem. Sot., Dalton Trans. 1831 (1979).

(25) A. G. Talma, H. van Vossen, E. J. R. Sudtilter, J. van Eerden and D. N. Reinhoudt, Synthesis, 680 (1986). (26)

Y. Takeda, Top. Curr. Ghan. 121, 1 (1984).

(27)

P. Gans, A. Sabatini and A. Vacca, Inorg. Chim. Acta. 79, 219 (1985).