~ubstituent Eject on the

Jomation oJ 28@ complexes

CHAPTER - I

CHARGE - TRANSFER PROCESS AND EDA COMPLEX FORMATION

f . f . INTRODUCTION

Eversince the formulation of a theory for the intermolecular

charge-transfer (CT) transitions in the electron donor-acceptor

complexes by Mulliken [I], the topic received a great attention from

the chemists and a lot of work in this field has since been done. The

theory, as such, now stands well tested. Most of the Mulliken's

predictions have been found to be generally true. Beginning with the

first investigation of benzene - iodine system by Benesi-Hildebrand

[ 2 ] , there has been considerable interest in the spectral and

thermodynamical properties of CT complexes [3-61.

Electron donor-acceptor (EDA) complexes, also known as

charge-transfer (CT) complexes, are organic molecular complexes

formed between two molecules, which possess regions differing

considerably in electron density. One molecule acts as an electron

donor and the other as an electron acceptor. Their formation is usually

accompanied by the appearance of a new absorption spectrum that is

not simply an addition of the spectra due to the donor and acceptor

alone. The EDA complexes are formed almost instantaneously and

exist in a reversible equilibrium with the donor and the acceptor.

Donor Acceptor Complex

The equilibrium or formation constant K, is also known as the

association constant of the complex. Usually, a new spectral

absorption band appears in the near UV-visible region of the

spectrum indicating the presence of the new species, the EDA

complex. To the naked eye, this often manifests itself as a distinct

colour change, which takes place on mixing the donor and the

acceptor solutions. They are generally characterized by a 1:l

stoichiometry.

EDA complexes play a role in all realms of science. They are

gaining importance as potential high efficiency, second-order non-

linear optical materials [7]. The importance of EDA complexes in

biological systems have been suggested by Pullman [8], Szjent-

Gyorgyi [9] and others. It has been established that many antibiotics

and other drugs exert their influence by initial formation of EDA

complexes with the amino acids of body proteins [lo]. Charge-

transfer interactions of proteins, amino acids and amines in polar and

non-polar solvents were studied by Slifkin [ll-131. It has also been

suggested that CT interactions play a role in respiratory processes

[14], hormone activity [15-171 and the carcinogenicity of some

compounds [I 8,191. A study of EDA complexes has permitted the

indirect determination of ionisation potentials of less volatile aromatic

compounds, which would not have been possible with other available

techniques [20]. EDA complexes play important role in the fields of

analytical chemistry [21] and organic semi-conductors [22]. It was

found that the electrical conductivity of the complex formed between

tetracyanoquinomethane and tetrathiofulvalene is comparable to that

of copper [23]. EDA complexes also appear as intermediates in

various reactions.

1.2. THEORY OF EDA COMPLEXES

The nature of the bonding between the donor and the acceptor

moieties in these complexes was quite puzzling to start with. A

number of reviews and monographs have appeared on the subject of

electron donor-acceptor complexes [6,25,26]. A number of theories

were proposed, including those involving a normal covalent bond, an

ionic interaction and various types of dipole-dipole and dipole-

induced dipole interactions. When x-ray diffraction results on crystals

of the complexes which can be prepared in the pure solid state,

showed that the donor and the acceptor are usually separated by

distances of over 30 pm, which are much larger than the usual

covalent bond lengths of about 15 pm, and are only slightly less than

van der Waals distances, it became apparent that the force of

interaction between the donor and the acceptor moieties in the

complex is of longer range than the usual binding forces.

With the development of the valence bond theory, attempts

were made by Brackman [27] and others to interpret the charge-

transfer complex as a resonance hybrid of "no bonded" and bonded

canonical structures. The theory of bonding in these complexes were

finally put on a firm footing by Mulliken using the molecular orbital

treatment in a classical series of papers, now available as a lecture and

reprint volume [ 6 ] , where in he proposed that while electrostatic

interactions are mainly responsible for linking the molecules in the

ground state, charge-transfer bonding is the dominant mechanism in

the excited state. He attributed the new spectral bands as arising

essentially due to charge-transfer forces. He later agreed with the

view of Briegleb [4] that the stability of certain n-n complexes may

be due to predominant electrostatic forces and hence preferred the

name electron donor-acceptor complexes over the designation charge-

transfer complexes which he had initially employed.

The wave-function of the complex consists of contributions

from the wave-function of this charge-transfer dative state as well as

the state in which there is no bond between the two. For weak

complexes (i.e., for which K < 1) , the contribution of the dative state

to the ground state wave-hnction of the complex is small. There is an

excited state, the charge-transfer state. For the loose complex, the

ground state is mostly no-bond, while the excited state is mainly

dative. Excitation therefore essentially amounts to the transfer of an

electron from donor D to the acceptor C. Theory shows that the

spectroscopic absorption will be of high intensity. The energy level

diagram is in fact much more complicated, involving a number of

locally excited states of D and of C, and a number of charge-transfer

states. Mulliken classified donors such as the arenes which donate

their n- electrons, as x-donors, and molecules lacking n-electrons but

possessing a lone pair of electrons, such as the aliphatic amines, as

n-donors. Aniline can act both as a x-donor because of the n-electrons

in the phenyl ring, and as an n-donor through the lone pair of

electrons on the nitrogen atom.

1.3. RECENT TRENDS IN THE THEORY

Mulliken's theory considers the complex as a hybrid resonating

between a non-polar structure and a polar one, resulting from the

transfer of one electron. In Dewar's theory [28] the transfer is

supposed to take place between the highest occupied orbital of the

donor (HOMO) and the lowest unoccupied one of the acceptor

(LUMO). More recently Flurry [29] studied the EDA complexes as

new molecules which are a linear combinations of the HOMO and

LUMO from the donor to the acceptor respectively. Using

perturbation theory Murrel et al. [30] calculated the energy as the

total of the contributions from different types of interactions like Van

der Waals, electrostatic etc. Chesnet and Moseley [3 11 consider these

complexes as a 'super molecule' made up of donor and acceptor

molecules.

According to Nagy [32], Mulliken's theory exaggerates the

importance of electronic transfer neglecting other types of

interactions. On the other hand Murrel's theory gives good results for

stabilization energies and equilibrium distances but does not account

for the new band that appears as a characteristic of these complexes.

The CND0/2 calculations on the geometry and stabilization energies

of certain selected EDA complexes by Yanez et al. [33] showed that

the 'super molecules' approach gives fairly consistent values for the

stabilization energies. The variation of the dipole moment of the

complex indicated that the electron transfer is important in stabilizing

these complexes.

Chiu [34] developed an unified molecular charge transfer

theory which embraced all ranges of molecular interactions. In the

limit of strong electronic effect (uv and visible region) it reduces to

Mulliken's theory of CT molecular complexes and in the limit of

dominating vibrational effect (IR and F-IR region), it reduces to

intervalence CT theory.

f .4. EXPERIMENTAL METHODS TO STUDY EDA COMPLEXES

Mulliken's charge transfer theory predicts the existence of an

intense absorption spectrum corresponding to the transition YN+YE

and its total absolute intensity can be approximately computed. If YN

is nearly pure no-bond character and YE nearly pure ionic character,

the spectra associated with the transition is called the intermolecular

charge transfer spectrum. Light absorption causes an electron jump

from the ground state to the excited state. Large amount of data is

available in literature regarding the spectral characteristics of CT

complexes. In a majority of systems the characteristic CT band

appears in the uv-visible region and can be easily characterised.

However, frequently intramolecular and charge transfer spectra may

overlap or sometimes also interfere quantum mechanically. In such

cases it may not be possible to identify the CT spectra uniquely.

One other important method is the study of the dipole moment

changes on formation of EDA complexes. Since an electron jumps to

excited state the dipolar character of the niolecule undergoes a

change, hence the measurement of dipole moment may prove to be a

diagnostic tool in the interpretation of the EDA complexes.

Several other physico-chemical methods are also available for

the study of EDA complexes [ 5 ] . Several authors have observed

small changes in the chemical shift in the NMR spectra of mixtures in

which the acceptor-donor type interactions occur. A simple

interpretation of the direction of the shift due to complexation was

given by Pople et al. [ 3 5 ] . It was shown by Hanna and Aashbogh

[36] that in a donor acceptor interaction to form 1 :1 complex the

chemical shift of the acceptor protons is related to the strenght of the

donor acceptor interaction. This technique was used very

successfully by Foster and Fyfe [37] in the stydy of nitrobenzene

complexes with several aromatic compounds. Wehrnan and Popov

[38] studied the donor ability of tetrazoles using this technique.

Similar studies were made by Muralikrishna and Krishnamurty [39]

to study the interaction between piperidine and TCNQ. Dielectric

polarization studies were successfully used to determine the

formation constants of several EDA complexes 140-421. X-ray

crystallography can be used for the study of EDA complexes. In

general the intermolecular distance between the partners of an EDA

complex where only non-normal inter-crystalline forces operate have

slightly larger intermolecular distances. Thus Wall work [43] has

suggested that inter atomic distances of less than 3.4 A are

characteristic of charge transfer bonded molecules.

The Infrared spectroscopic studies of EDA complexes are rare.

It was sho~vn that quinone complexes are characterised by the shift in

carbonyl group frequency [44]. In the extreme case where charge

transfer is more or less complete in the ground state, one can observe

the infrared spectrum of the ionised donor and the acceptor [45].

Mulliken's theory suggested that EDA complex should be

paramagnetic due to displacement of the charge from the donor to the

acceptor. This has been proved experimentally. But no extensive

work has been done.

Direct measurement of the ionization potential (ID) of aromatic

molecules have been made for relatively few substances. Price

et al.[46] have determined the ID of some donors by direct photo-

ionization experiments. Watanable [47] has measured ID for a range

of substituted benzenes and for some aromatic hydrocarbons by the

same method. The electron impact method has also been used to

obtain the values for a few aromatic hydrocarbons. However, values

obtained for ID by different workers from the electron impact method

are often not in agreement with one another. Baker, May & Turner

[48 ] have determined the ID of substituted benzenes by photoelectron

spectroscopy. An alternative method for obtaining ID is from

measurements on the CT absorption maxima.

The comparison of the position of absorption bands with the ID

of the donor was found by Mc.Connel1, Ham & Platt [49] to be

facilitated by a relationship (eqn.2) that is linear for I D values between

7 and 12 ev. n

where ET is the charge transition energy, C1 and C2 are perhaps

essentially constant for a series of complexes.

It is evident that there exists a quadratic dependence on the ID

and the actual theoretical curve should be a parabola and not a

straight line. Mulliken and Person [6] have shown that the

experimentally observed linear relationship are not so conflicting with

the above equation as long as the ID of the donor does not in general

fall much below 7 ev, with the consequence that the segment of the

parabola described by the eqn.(3) over which the results have been

applied approximates to a straight line. In general, the equation

hvCT = a ID + b . . . (3)

has been used, where 'a' and 'b' are constants for a given acceptor.

The value of 'a' in the above equation has no direct significance since

the line itself is only an approximation. Foster [ 5 ] has shown that the

constant 'b' contains terms such as othe electron affinity of the

acceptor and coulombic interaction terms between the nobond and

dative states. 'a' and 'b' are assumed to be constants for a series of

similar donors reacting with a given acceptor in the same solvent.

1.6. EQU1LIBNUh.I CONSTANT OF EDA COMPLEXES

Theoretically, K can range from zero (implying no complex

formation) to infinity (implying complete conversion of

stoichiometric quantities of the donor and the acceptor into the

complex). Generally, complexes with formation constants less than

unity are classified as weak, while for strong complexes D l . In

many situations, these complexes are non-isolable in the pure state,

thus ruling out the possibility of the direct measurement of their

molar extinction coefficients at the charge-transfer maximum and,

therefore, the direct evaluation of their formation constants. In 1949,

however, Benesi and Hildebrand [2] reported a graphical method for

the simultaneous evaluation of the formation constant, K, and the

molar extinction coefficient, E, of 1:l EDA complexes formed

between some aromatic hydrocarbons (acting as electron donors) and

iodine (an electron acceptor) in inert solvents like carbon tetrachloride

and heptane:

ArH -t Iz f ArH.1,

By measuring the absorbance, Z, at the charge-transfer

maximum, of a series of solutions in which the initial concentrations

of both the donor and the acceptor were varied (while holding the

concentration of the donor very much greater than that of the

acceptor, i.e., D>>C) and using the relationship

(where I is the path length or the cell size), they were able to estimate

both K and E for each donor-acceptor system fkom the linear plot of

C/Z versus 11D.

The following years saw a spate of similar publications by a

number of ~iorkers employing a variety of donors and acceptors.

Even as the limitations of the Benesi-Hildebrand approach became

apparent, a number of modifications and improvements, including

other methods of plotting the spectrophotometric data on the basis of

linear equations similar to Benesi-Hildebrand's, were suggested.

1.6.a. DETERMINATION OF EQUILIBRIUM CONSTANT BY

METHODS OTHER THAN UV-VIS. SPECTROMETRY

The reason for the preference of workers to the

spectrophotometric method of evaluation of formation constants of

charge-transfer complexes lies in the fact that spectrophotometers of

fairly high accuracy have been available for decades, as a standard

instrument in most laboratories. The spectrophotometer, especially

the computerized one, is easy to operate and from the absorbance

readings, K and E can be evaluated quickly, on an electronic

calculator in the linear regression (LR) mode using a Benesi-

Hildebrand type of equation.

Besides the uv-visible spectrophotometric technique, many

other methods of evaluating the formation constants of charge-

transfer complexes, have been developed over the years:

1. Infra-red spectroscopy, using the same principles as uv-visible

spectrophotometry;

2. Nuclear magnetic resonance spectroscopy; from the change in

the chemical shifts leading to a relationship similar to Benesi-

Hildebrand's;

3. Distribution methods, partitioning one of the components of

the complex between two immiscible liquids;

4. Polarography, by measuring the shift in the half-wave

potential as the donor is added to the solution of the acceptor;

5. Calorimetry, which allows simultaneous measurements of the

formation constant as well as the heat of formation of the complex;

6. Dielectric constant measurements for determining the

formation constants of hydrogen-bonded complexes;

7. In the case of reactive charge-transfer complexes, from the

rates of their reaction following their (nearly instantaneous) formation

in equilibrium; and,

8. Relaxation methods.

The principles of these methods have been reviewed by Foster

[24]. Many of them suffer from infirmities similar to those

encountered with the spectrophotometric method.

1.6.b. SPECTROPHOTOMETRIC DETERMINATION OF 'K' AND

'E' BY BENESI-HILDEBRAND METHOD (1949)

Iodine has long been known to dissolve in different solvents to

produce various colours, ranging from violet in carbon tetrachloride

to brown in alcohols, Early colorimetric investigations in 1909 by

Hildebrand and Glascock (cited in [Z]) indicated the formation of a

1 : 1 compound in equilibrium between iodine and ethanol dissolved

together in carbon tetrachloride. Attempts to calculate the formation

constant of the complex from the intensity of the colour were

hampered by the impossibility of the direct determination of its molar

extinction coefficient. A workable solution was first devised by

Benesi and Hildebrand [2] in 1949. They carried out a

spectrophotometric investigation of the interaction of iodine acting as

acceptor with the aromatic compounds, benzotrifluoride (C6H5CF3),

benzene, toluene, o- and p-xylenes and mesitylene (acting as donors)

in carbon tetrachloride and in heptane at room temperature (22'~). On

mixing the donor and the acceptor, dissolved separately in the inert

solvent, there was a colour change and the spectrum of the mixture

showed a new absorption band with maximum at a wavelength of

about 500 nm, a region in which neither the donor nor the acceptor

has any significant absorption. They rationalized this as being due to

the formation in equilibrium of a 1 : 1 complex between the donor and

the acceptor:

K

Arene -t I2 t + Arene.12

(Donor) (Acceptor) (Complex)

If x is the equilibrium concentration of the complex, D and C

are the initial concentrations of the donor and the acceptor, then

Since D>>C it follows that D>>x and the tern (D-x) in the

denominator of the above equation can be replaced by D to obtain

whence KDC-KDx = x

Dividing throughout by KD x

or, -=- ] + 1 x KD

Since the observed absorbance Z is due to the complex alone,

Z= ExE . . . (8)

where 1 is path length of the light in the cell, and E the molar

extinction coefficient of the complex; whence, from eqn. (4.)

CZ 1 1 1 -=- .- + - Z K E D E

which is the Benesi-Hildebrand equation. This is of the type

Y = B X + A . . . (9)

A plot of Y, i.e., C11Z versus X (i.e., I D ) must be linear with the

slope, B = 1/KE, and intercept, A = 1/E. Benesi and Hildebrand

obtained good linear plots for all the systems. From the intercept, they

obtained an estimate of E and substituting it in eqn. (4) they obtained

estimates of K.

In employing eqn. (5 ) , Benesi and Hildebrand, as well as

Keefer and Andrews, who published their findings on the interactions

of these donors with bromine [SO] and iodine monochloride [5 11 soon

after Benesi and Hildebrand [2], chose to represent x in the numerator

in moles per litre, (C -x) also in moles per litre, but in the term (D-x)

they represented both D and x in mole fractions. The final result is

that D in eqn. (4) remains to be represented in mole fractions of the

donor in the mixture, so that K is the formation constant of the

complex in per mole fraction units. In fact, the equation which, by

some authors, is referred to as the Keefer-Andrews equation is

identical with the Benesi-Hildebrand eqn. (4).

Most workers, however, prefer to represent the concentration of

all the three species, viz., the donor, the acceptor and complex in

molarities, that is mole per litre, so that the value of K obtained by

them is in the more conventional litre per mole units. These two differ

in magnitude, being related to each other in dilute solutions (in which

ideal behaviour can be assumed to hold good) as follows:

where d is the density of the solvent and M, its molecular weight.

The Benesi-Hildebrand eqn. (4) is valid for all scales of

concentration, but the magnitude of K evaluated will depend on the

scale employed. In this thesis, the formation constant is based on the

molarity scale, except where specifically stated otherwise.

The Benesi-Hildebrand equation reduces to the form

C 1 1 1 -=- .-f - Z K E D E

when all absorbance measurements are made in 1 cm. cells. It must be

noted that the function Y comprises of two variables, C and Z, the

concentration of the acceptor and the absorbance, respectively. C, as

well as, D, were varied in their work, which should be the preferred

design of the experiments. But many subsequent workers preferred to

keep C constant while changing the concentration of the donor, D, in

a series of mixtures. This leads to yet another form of the Benesi-

Hildebrand equation

1 1 1 1 - ---.-+- Z KEC D CE

A plot of 1/Z versus 1/D should be linear with the slope,

B =1/KEC and intercept, A = 1/EC, where C is the constant acceptor

concentration. This design has the advantage of making the

preparation of the series of mixtures a facile task, since two stock

solutions, one of the donors and another of the acceptor, alone need to

be prepared.

Some further remarks on the Benesi-Hildebrand equation

and its application are in order. Nearly all workers have employed

the condition D>>A in their studies because most of the electron

acceptors (though not iodine) have low solubilities in the solvents

employed. The first equation that is written for the formation constant

of the complex, viz.,

is symmetric with respect to D and C, i.e., the value of K is unaltered

if D and C are interchanged. The Benesi-Hildebrand eqn. (1) is valid

also under condition of C>> D, but in the form

t D 1 1 1 -=- .- + - Z K E C E

K and E have been determined in some cases from

spectrophotometric data obtained when the initial concentration of the

acceptor is much higher than that of donor (C>>D) and consistent

values have been obtained [52]. But Ross and Labes [53] found that

in the case of N, N-dimethylaniline interacting with 173,5-

trinitrobenzene in chloroform, such conditions lead to estimates of K

and E very different from those evaluated with D>>C. This they

attributed to the tendency for formation of 1 :2 and 2: 1 complexes in

this system.

The Benesi-Hildebrand procedure was graphical: they

plotted the chosen functions on a graph sheet, drew what was visually

the straight line of best fit and then proceeded to measure the

intercept to determine E. Similar procedures were followed by other

workers who employed similar linear equations to estimate K and E

(q. v.). However, after the advent of electronic calculators, linear

regression came to be increasingly employed to determine the slope

and the intercept of the line of best fit, whose "goodness fit" is

characterized by R, the correlation coefficient.

1.6.c. ASSUMPTIONS INHERENT IN THE

BENESI-HILDEBRAND MODEL

The Benesi-Hildebrand equation, and indeed the various linear

and non-linear equations to be reviewed below, and which are

employed for the purpose of evaluating the formation constant and

the molar extinction coefficient of EDA complexes rest on a number

of assumptions; any departure of the actual system to a significant

extent from any of these assumptions will lead to model failure with

the results that the estimates of K and E become unreliable.

1. The foremost assumption is that a 1 : 1 complex alone is formed,

and higher order complexes are absent, which needs to be confirmed.

For instance, Lofti and Roberts [54] confirmed this in the case of the

interaction between trimethylsilyltriptycene and tetracyanoethylene

by verifying that the Job plot had a maximum at 1:l stoichiometry

and showed no detectable asymmetry. Howevere, Foster [6] has

suggested that the Job's plot is not always capable of distinguishing

between instances of different stoichiometry. Dimicoli and HCl&ne

[55] have used it to show the presence of 2: 1 as well as 1 : 1 complexes

in a system studied by them. But not all workers have taken this

precaution of establishing the stoichiometry of the complex prior to

proceeding to evaluate K and E.

2. It is assumed that the solutions are ideal in behaviour, so

that activity coefficients of all the species can be assumed to be unity

and the concentration equilibrium constant evaluated by the method is

the same as the true or thermodynamic equilibrium constant (based on

activities of the species). This implies that there is no variation of K

with the concentration. But it is open to question whether, when

concentrations of reactants as high as 0.5 M or 1 M are used, some

departure fiom ideality would not occur.

3. It is assumed that the complex obeys the Beer-Lambert law.

This is a vital assumption, since any non-compliance with the law

would lead to a failure of the method. The molecules act as absorption

centres for the radiation and for the law to be valid, they should act

independently of each other. This restriction, that they should not

interfere with each other causes Beer's law to be a limiting law

applicable in dilute solution of concentrations less than 0.01M (of the

absorbing species). Failures of the law in ho~nogeneous systems, e.g.,

solutions, are unknown [56]. However, if the absorbing species tends

to associate or dissociate, deviations do occur. At concentrations

above 0.01M, deviations fiom Beer's law occur due to the change in

the refractive index of the solution as the concentration is changed.

4. Implicit in the concept of the EDA complex formation as

purely due to an interaction between a donor molecule and an

acceptor, is an assumption that the environments, including the

solvent molecules, have very little role to play. Hence the need to use

'inert' solvents. Carbon tetrachloride, along with heptane and hexane,

have long been solvents of choice. However, Anderson and Prausnitz

1571 in 1963 found by spectrophotometry that carbon tetrachloride

itself interacts with arenes forming weak complexes. For instance,

with benzene, the complex has a K value of about 0.009 L0.004 L 1

mole at 2 5 ' ~ . Evans [58] found that the interaction of iodine with

heptane in perfluoroheptane, which is perhaps one of the most 'inert'

of solvents, gives a good linear Benesi-Hildebrand plot, the straight

line passing almost exactly through the origin, i.e., with a near zero-

intercept. He interpreted it to indicate the formation of a very weak,

possibly of a collisional type, charge-transfer interaction between

iodine and heptane. The myth of 'inert' solvents has been exploded

long ago: no solvent can remain an innocent spectator in an

interacting system. How far the interaction between the solvent

(always present in a great excess) and the donor or the acceptor

affects the determination of the equilibrium constant of complex

formation between the donor and the acceptor is not easy to quantify.

But many workers tend to attribute anomalous values of K, at least in

part, to such solvent-reactant interactions.

5. One anomaly often encountered is that, for a series of related

donors interacting with a particular acceptor to form weak complexes,

one expects that, as the electron-donating capacity of the donor

increases and K increases, the transition moment should increase

resulting in an increase in E as well. Many cases of the opposite trend

have been found, including those for the interaction of iodine with

benzene and mesitylene in Benesi and Hildebrand's original work [ 2 ] .

Orgel and Mulliken [59], in order to account for these contradictory

trends, advanced the idea that, besides definite, discrete 1:l charge-

transfer complexes, absorption in the band may also arise whenever a

donor and acceptor are sufficiently close to one another during

collisions without forming a complex. These 'contact' charge-transfer

pairs, they estimated, may account for roughly three-fourths of the

charge-transfer absorption in the case of the benzene-iodine system,

with the actual complexes themselves contributing only one-fourth.

Carter et al. [60], however, differed and instead proposed a model in

which solvation by the solvent of the donor, the acceptor and the

complex all play a role and if, competition between complexing and

solvation is taken into account, there is no need to introduce two

kinds of charge-transfer absorption, one due to the complex and the

other due to 'contact pairs'; the behaviour of weak complexes can

then be fitted into the same theory as that of strong complexes. They

concluded that the Benesi-Hildebrand treatment, in the case of weak

complexes, leads to underestimates of K and overestimates of E.

Emslie et al. [61] soon countered the concept of solvation as a major

factor by pointing out that for the same, given complex, the values of

E evaluated by the Benesi-Hildebrand type of equations is fairly

constant in a series of solvents of widely varying solvating power

(while the formation constants vary). They instead suggested that the

observed anomalies may result from the complex not obeying Beer's

law as the concentration of the species present in excess, which is

usually the donor, is increased. They showed that even small

differences in E can lead to large differences in the calculated values

of K and E, but the product KE will remain unaltered.

Person [62] opines that the distinction between true complexes

and contact pairs appears to revolve around the charge-transfer

stabilization energy (which is greater than the ordinary van der Waals

attraction energy: if the stabilization energy is greater than the

translational kinetic energy of kT, the interaction must be considered

to be a complex; if the energy of formation is less than kT, then the

interaction is better described as a 'contact'. These questions, he

concludes, are difficult, if not impossible, to answer.

Orgel and Mulliken [59] have discussed the complications

arising from the formation of several geometrically and/or

electronically different 1: 1 complexes: they showed that if two or

more isomeric 1 : 1 complexes are formed, the spectrophotometric

method measures a total equilibrium constant, K (measured) = C Ki,

where Ki is the formation constant for the ith complex and a weighted

molar extinction coefficient, E (measured) = Ki CEi/K, where Ei is the

molar extinction coefficient of the i'" complex.

1.6.d. OTHER LINEAR EQUATIONS AVAILABLE FOR THE

EVALUATION OF 'K' AND 'E'

( 1 ) The Foster-Harnmick-Wardley Equation (1952):

Hammick was one of the first to foresee, as early as in 1936,

the possible use of absorbances to determine formation constants of

charge-transfer complexes [63], but it fell to Benesi and Hildebrand to

evolve a practical method. Foster, Hammick and Wardley [64] were

concerned with the general case of the formation of higher order

complexes, such as D,C and DC,, Considering the formation of D,C,

if D is the initial concentration of the donor. C that of the acceptor,

and x the concentration of the complex in equilibrium, then

If D>> C, since x < C, effectively, D - x = D; then

z x = - (for 1 cm path length) E

L or KCE - K Z =- D

... (17)

Putting Y = Z/D" and X = Z, this becomes a linear relationship

between Y and X, if C is kept constant in all the mixtures. A plot of

Z/D" versus Z should be linear with the slope, B = - K and intercept

A = KCE. The linear relationship obtains only when the proper choice

of n is made. The data are plotted for different values of n (=I, 2, 3,

etc ) and that value of n which gives a straight line is inferred to give

the stoichiometry of the complex. From the slope and intercept of this

line K and E are calculated. It may be noted that the slope of the

FHW plot is always negative. Using this method, Foster et al. found

that the acceptor 1,3,5- trinitrobenzene forms a 1: 1 complex with

diphenylamine but a 1:4 complex with dimethylamine. The

composition of the complexes were confinned by the method of

continuous variations due to Vosburgh and Cooper [65] .

If the complex is known to have 1:l stoichiometry then the

FHW equation (1 5 ) reduces to

z - = -KZ+KCE D

...( 18)

for constant acceptor concentration.

( 2 ) The Scott Equation (1956):

The Scott equation [66] ,

is readily obtained from the Benesi-Hildebrand equation (2). The

Scott plot of ZCDIZ versus D should be linear with the slope,

B = 1/E and intercept, A = 1/KE. E is given by the reciprocal of the

slope while it is obtained from the reciprocal of the intercept in the

Benesi-Hildebrand method. The roles of the slope and the Intercept

are thus interchanged in the two plots. In both treatments, E is

evaluated first and then K from the product KE. The Scott treatment

obviates the difficulties when the data are poor, as often happens in

the case of weak complexes: the intercept of the Benesi-Hildebrand

plot may be nearly zero (implying a nearly infinite molar extinction

coefficient for the complex), or, in worse cases, it may be negative (as

has been observed in some cases) leading to negative values of both

K and E, which would be absurd. In the Scott plot E is estimated from

the slope and turns out to be almost always positive (since the slope is

almost always positive, unless the data are atrociously bad) leading to

a positive estimate of K.

Scott pointed out that a great advantage of his method is that

one extrapolates through region of increasing dilution to the intercept,

and that, with precise experimental measurements, one can also

determine the initial slope at high dilution; moreover, if the points do

not define an exact straight line, this method gives a more reasonable

weighting to the different measurements. Apart from Benesi-

Hildebrand's, Scott's is the most widely employed procedure for the

treatment of spectrophotometric data of 1 : 1 charge-transfer

complexes. For a path length of 1 cm., the Scott equation becomes

CD 1 1 --- = -.D +- Z E KE

. . .(20)

and if C, the concentration of the acceptor, is kept constant in all the

mixtures of the series,

D 1 1 - =-.D+- Z EC KEC

A plot of D/Z versus D should be linear with the slope, B = 1/EC and

intercept, A =l/KEC .

( 3 ) The Seal-Sil-Mukherjee Equations (1982):

Seal, Sil and Mukherjee [67] rearranged the Benesi-Hildebrand

equation into the following two forms for absorption in a 1 cm cell.

Form I : 1 z

Z = --.-+EC . . .(22) K D

and Form I1 : D 1 D=EC-- - - Z K

Both the equations do not contain the product KE. A plot of Z

versus Z/D as well as a plot of D versus D/Z will be linear. For the

first plot, the slope, B =-lK and the intercept A = EC; for the second,

B = EC and A = -1lK. These two equations, the authors claim,

overcame the inherent defect present in the earlier linear equations, in

that a complete separation of K and E has been achieved: each can be

determined exclusively from the slope or the intercept. This, they

claim, should lead to a more accurate evaluation of K and E from a

given set of data than would be possible with the Benesi-Hildebrand,

Scott or Foster-Hamrnick-Wardley plots. Seal et al. employed

literature data on a number of systems to determine the values of K

and E as estimated on the basis of their two equations, but these do

not differ significantly from those reported by the original workers.

( 4 ) The El-Haty Equation (1991):

When the initial concentrations of the donor and the acceptor

are comparable, El-Haty [68] derived the following linear

relationship.

and used it to investigate solvent effects on charge-transfer complex

formation: a plot of (C+D) versus CDIZ should be linear with the

slope B = E and intercept, A = - 1K. This plot is, in fact, the

'inverse' of the Rose-Drago-Ayad linear plot: (C+D) is plotted along

the X-axis and CD/Z along the Y-axis in the Rose-Drago-Ayad plot,

while they are plotted along the Y -axis and the X-axis, respectively,

in the El-Haty plot. The El-Haty equation is not restricted by the

condition D >> C; also it does not contain the product KE and

achieves a separation of K and E.

( 5 ) The Rose-Drago-Ayad Equation (1994):

Rose and Drago [69] derived an absolute equation applicable

with no restriction on the relative concentrations of the interacting

species. In its most general from, it is also applicable to regions where

the charge-transfer band of the complex and the absorption band of

the acceptor overlap. It is generally applicable to 1: 1 complexes, but

similar equations can be derived for higher order complexes.

For the formation of a 1 : 1 complex,

D t C 3 DC

Ayad [70] modified the Rose-Drago equation and arrived at an

equation of the form

This leads to a plot of CD/Z versus (C+D) being linear with the slope,

B = 1/E and intercept A = IIKE.

( 6 ) The " Inverse" Benesi-Hilderband Method :

All the above workers have employed the concentration CD of

the donor as the x-variable and the observed absorbance Z due to the

complex as the y-variable in the regression. Ordinary linear

regression requires that the x-variable be free from error, while the

uncertainties lie in y-variable. The regression minimized the sum of

the squares of the deviations (of y) to obtain optimal estimates of the

slope and the intercept. It has been pointed out by Baskara raju [71]

that the present day double beam spectrophotometers incorporating

computerized devices, can measure absorbances accurately with

deviations of less than i 0.002, if the absorbance scale of the

instrument has been calibrated using accurately prepared standard

solutions of potassium chromate.This for a maximum absorbance

reading say, 1.000 for the solution with the highest initial donor

concentration, implies an error of less than 0.2%. It is shown that the

errors in the concentration of the donor and the acceptor in the

mixtures are usually much higher than this due in part to uncertainties

in their assay. So it is the concentration term which must be taken as

the y-variable, (though it is still the "independent variable") for

regression purposes. Therefore another linear equation has been

proposed [71].

This equation demands a plot of 1/D vs 112 to be linear from which

slope one can get K = - intercept and E = - C .intercept

This equation is the inverse form of the B-H equation, since x-and y-

co-ordinates have been interchanged.

Due to the presence of experimental errors these eight

equations invariably lead to different estimates of K (and E), the

difference getting larger as the levels of error gets higher. The

unreliability of the estimates of K, especially for week complexes

(defined as this for which K < I), as obtained fkom

Benesi-Hilderband type of plots was realized quite early;

Person [62] has shown that for reliable estimates of K the

concentration of donor in the most concentration solution of the series

must be greater than about O.l/K.

1.7. NON-LINEAR TECHNIQUES OF DETERMINATION OF 'K'

Non-linear regression methods broadly fall into two categories:

those which employ derivatives and those which do not require the

computation of derivatives. Both use a cycle of computations which

is repeated a number of times (iterations) until the estimates of the

parameters reach a pre-designed level of precision. Methods which

employ derivatives or gradients are generally quicker to converge to

these regression estimates than iterative methods which do not

employ gradients. All, however, require initial, rough inputs of the

parameters that are sought to be optimized. Usually these are

obtained from a linear plot, such as Benesi-Hildebrand's or Scott's.

The limitations of the Benesi-Hildebrand type of linear

equations to estimate the formation constant of complexes have long

been apparent and therefore non-linear regression methods have also

been attempted to arrive at more reliable estimates of K from

absorbance data. Since they employ iterative procedures, they are

well suited for programming on computers. Among the early non-

linear methods are those due to Wentworth et a1 (1967) [72], and

Rosseinsky and Kellawy (1969)[73], both of whom have used the

Deming routine [74], which requires the computation of partial

derivatives of a trial function. Subsequently, Farrell and Ng6 [75]

used the Rosenbrock [76] routine which does not use derivatives but

is slower to converge.

( 1 ) The Wentworth / Rosseinsky-Kellawy Method :

Without imposing restrictions on the relative initial concentrations

of the donor and the acceptor, the following equation, when solved

for x, yield the concentration of the 1:l complex present in the

equilibrium.

Then for absorbance in a 1-cm cell

Equation (27) is non-linear in C and D. This is an iterative

procedure which, with approximate initial input of K and E, rapidly

converges to the non-linear regression estimates. Rosseinsky and

Kellawy [73] using independently the same routine, showed that the

absorbance data of many complexes, reported by earlier workers,

gave non-linear regression estimates of K and E significantly different

from those calculated using linear plots. The fact remains that non

linear regression must be superior, since the dependence of Z and C

and D essentially non-linear. A few workers have since adopted

Wentworth's [72] techniques for evaluation of K and E.

( 2 ) Deming Method :

Much earlier, in 1938, Deming [74] had evolved an iterative

method for the minimization of sum of squares of deviations of a

variable which is non-linearly dependent on two or more variables.

This method is actually an adaptation of the well-known Newton-

Raphson iterative procedure for finding the minimum of a function in

numerical analysis. It employs a first-order Taylor expansion to

obtain a linear function of the parameters. It is an elegant routine and

generally converges smoothly to the optimized estimates of the

parameters in a few iterations if the data are not very erratic and the

initial inputs of the parameters are fairly close to their true values.

Otherwise, no convergence may result. The method also estimates

the standard errors in the computed parameters.

Baskara Raju [71] contended that the errors are more likely to

be present on the concentration terms rather than in the absorbance

readings, and he derived from first principles the expression

EZ+KEC z- KZ* D = ... (28)

K E ~ C - KEZ

in which the initial concentration of the donor, D is expressed as a

function of the other two variables, C and 2. If the experimental

conditions are such that the relative error is higher in the donor

concentration that in the acceptor concentrations or in the absorbance

units (as can arise in the case of highly pure acceptor like iodine,

employing good spectrophotometer), it is the sum of squares of the

derivations in CD which must be minimized rather than those in Z

Baskara raju has adopted the Deming [74] routine for evaluation of K

(and E) from spectrophotometric data using equation.(28). These

estimates of K and E will vary from those obtained by the

WentwortMRosseinsky Kellawy method since the objective function

being minimized are different.

( 3 ) Nelder - Mead Method :

The Nelder-Mead downhill simplex function minimization

routine was first formulated for the asymmetric simplex by Spendley

et a1 [77], but was improved for the more general simplex by Nelder

and Mead [78] in 1965. It is a direct method of function

minimization which does not require the computation of derivavatives

as in the Deming routine. It is slower to converge than the gradient

methods, but has a remarkable elegance and simplicity about it. In

the present case, it can be given in the form

C = EZ + KEDZ - KZ*

K E ~ D -KEZ

If CA is constant in the series of solutions, the Demings routine

is inappropriate since it is meant for the regression of one variable on

another. The appropriate procedure here is to employ a function

minimization technique such as Rosenbrock's. Baskara Raju et

a1. [79] have preferred to employ the Nelder-Mead downhill simplex

routine. These estimates will differ from those obtained in the

WenworthiRosseinsky Kellawy method using eqn (27) and also from

those obtained by the Deming method using eqn (28). Like

Rosenbrock's and unlike the WenworthRossinsky Kellawy and

Deming's method the Nelder Mead method routine does not give

estimates of the standard errors in K and E.

8 STRUCTURE - RELATED INTERACTIONS AND

CORRELATION ANALYSIS

Structural variation in the vicinity of the reaction centre of a

compound results in an almost continuous variation on its

electrophilic or nucleophilic character. This capacity may be used as

a delicate probe into the effects which electron perturbation produces

upon reaction affinity and from which the electronic demands of the

reaction may be inferred.

Correlation analysis involves relating empiricallly the

reactivities of a series of compounds in which the structure of the

substrate is varied by the introduction of substituents, while the

solvent and the reagent remain constant. A linear relationship,

involving the logarithm of rate coefficient (k) or equilibrium constant

(K), and referred to as a linear free-energy relationship (LFER), is

generally observed and is interpreted using the slope and intercept of

the simple linear regression. The polar effect of a substituent

comprises all the processes whereby the substituent may modify the

electrostatic forces operating at a reaction centre, relative to the

standard substituent, which is usually the hydrogen atom. The most

familiar of these is the Hammett relationship [SO] which takes the

form

~ O ~ K = I O ~ K O + o p . . . (30)

where KO is the equilibrium constant of unsubstituted or parent

compound. The substituent constant o, measures the polar effect

(relative to hydrogen) of the substituent in a given position, meta or

para, and is independent of the nature of the reaction. The reaction

constant p is a measure of the extent of susceptibility of the reaction

to polar effects.

1.9. SCOPE OF THE PRESENT WORK

Nogami et a1 [81] studied the interaction of aniline with

chloranil and the formation of the product was rationalized via EDA

and o complexes as intermediates. In the present investigations, we

have tried to understand the effect of substitution on the donor aniline

molecule through the formation of EDA complexes with chloranil as

the acceptor.

CI 0 0

Aniline Chloranil EDA Complex

(or substituted (or its bromo aniline ) analogue, bromanil)

To get an insight into the changes brought about on the electron

affinity by the presence of different groups on the quinoid skeleton,

the spectrophotometric study has been extended to another acceptor

bromanil. The shift in CT transition due to substitution, the changes

on the ionization energy (ID) of the differently substituted donor

molecules, the changes in the thermodynamic parameter K and the

Gibbs free energy of formation AGO for chloranil and bromanil

systems are reported. The different spectroscopic and thermodynamic

parameters are correlated. A comparison is made between systems

performed with the two acceptors.

CHAPTER - I1

MATERIALS AND METHODS

2.1. PURIFICATION OF MATERIALS

All the materials used for the study were purified by following the

standard procedures available in the literature [82]. The purified

materials were all used immediately after the final step of purification

without any storage except in the case of chloranil and bromanil, which

were sufficiently stable.

2.1.a. ACCEPTORS :

1 p-Chloranil : Commercially available synthesis grade sample

('Loba Chemie') had 97% assay. It was recrystallised from AR. glacial

acetic acid and dried in air. Its purity was confirmed by determining the

melting point (290' C). The bright yellow coloured pure crystalline

sample was stored in an amber coloured sample tube.

2. p-Bromanil : The compound was prepared by the method of

Torrey and Hunter [83].

Hydroquinone (10 g) was suspended in glacial acetic acid and

bromine (90 g) was added. The solution was allowed to stand overnight.

An equal volume of water and concentrated nitric acid were added and

heated in a water bath. The bromanil formed was recrystallized from

acetic acid until its m.p. agreed with the literature value (299 - 300'~).

2.l.b. DONORS:

1. Aniline : The analar grade aniline (Fischer sample) was dried by

keeping in contact with potassium hydroxide pellets for several hours. It

was then distilled at 184" and the middle portion of the fraction was

collected.

2. m-Chloro Aniline: The procured reagent grade CDH sample of

m-chloro aniline was dried over potassium hydroxide pellets and

distilled to get pure sample boiling at 230' C.

3. p-Chloro Aniline: The commercially available crystalline (m.pt.

72' C) 'Burgoyne', LR sample of p-chloro aniline was purified by

distillation at 232' C (its boiling point) with hot water circulation in the

condenser. Care was taken to avoid solidification of the material in the

condenser. The oily distillate on cooling gave colourless pure crystals.

4. p-Bromo Aniline : It was prepared by hydrolysis of p-bromo

acetanilide which had been obtained by the bromination of analar grade

acetanilide (BDH) by following the prescribed procedure [82]. The

compound was then recrystallised from ethanol and the purity was

checked by finding its melting point (66 C).

5. p-Iodo Aniline : Pure aniline was used for the synthesis of p-iodo

aniline by iodination following the procedure available in the literature

[82]. The crude product was dried in air and refluxed with light

petroleum (60 - 80 C boiling fraction) in a flat bottomed flask fitted

with a double surface condenser over water bath maintained at 75-80' C.

After about 15 minutes, the contents are emptied into a beaker set in a

freezing mixture of ice and salt and stirred constantly. The crystallized,

almost colourless needles of p-iodo aniline was filtered under suction,

dried in air. The purity was confirmed by finding its melting point

(63 C).

6. m-Toluidine : The Fluka, pract. sample was first dried by treating

with potassium hydroxide pellets for several hours and then distilled at

204 O C just before use.

7 . p-Toluidine : The synthesis grade (Merck, zur synthesi) crystalline

sample was purified by distillation at 200 O C with hot water in the

condenser, taking special care to avoid crystallization in the condenser

itself. The middle portion of the distillate on cooling yields colourless

pure crystals of p-toluidine (m.pt. 44 O C).

8. p-Anisidine: The synthesis grade sample (S.D's) with an assay of

98% was further purified by recrystallization from benzene, employing

activated carbon to remove traces of coloured impurities. The crystals

obtained were used after checking the purity by determing the melting

point (57.2 O C).

9. N-Methyl Aniline: The commercial (L.R) BDH sample was

purified following the standard procedure [82]. Requisite quantities of

pure commercial N-methyl aniline, concentrated hydrochloric acid and

crushed ice are taken in a 500 ml beaker and the sodium nitrite solution

was added little by little with stirring, taking care not to increase the

temp above 10 O C. Stirring was continued for another hour and the

separated oily layer was washed once with water and dried over

anhydrous magnesium sulphate. It is then distilled at 196 O C to collect

the pure sample of N-methyl aniline.

10. N,N-Dimethyl Aniline: Commercial BDH sample of N,N-

dimethyl aniline and acetic anhydride are heated under reflux for about

3 hours and then cooled. The solution was distilled and the pure

colourless liquid boiling at 193 - 194 O C was collected.

11. Diphenyl amine: The analar BDH sample was found to be very

pure and was used as such after checking its melting point (54 O C).

2 .1 .~ . SOLVENT:

Benzene : The analytical reagent grade, thiophen-fiee benzene

(Qualigen's sample) was procured. It was shaken well with water in

order to remove any traces of acid and the benzene layer was separated.

It was then dried by first treating with anhydrous calcium chloride,

filtered and then placed over sodium wire . The solvent was then

refluxed for about one hour, distilled and the middle portion of the

distillate at its boiling point (80 " C) was collected.

2.2. ANALYTICAL TECHNIQUE

Spectrophotometric studies are carried out on a Shimadzu UV-

Vis. Spectrophotometer (UV-160). It is a micro computer controlled

double beam recording instrument (Kyoto, Japan). The instrument

combines a monochromator, keyboard, CRT and Graphic Printer. It has

a scanning speed upto 2400 nm per minute. It scans from 200 - 1100

nm and permits various spectral processings such as

expansion~compression of spectra, peak-pick, derivatives, smoothing,

data storage and arithmetic calculations between spectra. It incorporates

many standard calculation programmes such as data determination at a

fixed wavelength, automatic quantitative analysis by 2/3 - wavelength

calculation or derivative values, kinetic measurements and

multicomponent analysis. The instrument was kept in an AJC room and

the measurements were done at 298 K.

The required quantities of the pure acceptor and donor substances

are weighed accurately in calibrated standard measuring flasks and made

up to the mark with benzene solvent to get 0.0015 M and 1.5 M stock

solutions of the acceptor and donor respectively. 2 ml of the acceptor

solution and 1 ml variation (with pure solvent) of donor solution are

mixed to get a total constant volume of 3 ml of the mixture. A uniform

concentration (0.001 M) of the acceptor and nine different

concentrations of the donor solutions varied uniformly in the range 0.1

to 0.5 M were used for all the systems except with p-anisidine for which

half of the stock and final concentrations were used.

Matched pair of quartz cells of 1 cm path length and 4 ml capacity

were used. The freshly prepared donor and acceptor solutions were

mixed in the cell itself very quickly with the help of a calibrated 2 ml

syringe in which acceptor solution was taken. A trial spectrum was

recorded in the 'spectrum mode' to fix the wavelength of maximum

absorption in the CT region i.e., the charge transfer maxima (ACT). The

absorbance values at the 3LCT were recorded in the 'photometric mode'

within 3 seconds of mixing in all the cases, taking the pure acceptor

solutions of final concentration as the reference. The absorbance values

of the pure donor solutions of varying concentrations used were recorded

against pure solvent as the reference. These values were subtracted from

the respective absorbance values for the mixture, in order to get the

absorbance due the complex alone. The increase in the absorbance with

increasing concentration of the donor was noted. The 3LCT values were

found to be sensitive to substitution.

To confirm the 1 : 1 stoichiometry of the complexes formed, Job's

continuous variation method [84] was carried out . Solutions of

equimolar concentration (0.045 M)of the donor and acceptor were

prepared and the concentration of the fractional molarities of the solution

were varied uniformly so that the overall molar concentration was kept

constant. The absorbance values at the C.T maxima were recorded and

the contribution due to the complex alone is found out by subtrac,ting the

absorbances of the pure donor and acceptor solutions of respective

concentration 1851. The complex absorbance values were plotted against

the mole fraction of the donor to arrive at the result.

The formation constant (K) and molar extinction coefficient (E)

values for different systems were computed by eight linear and three

non-linear methods as described in Ch.1, using the program in QBASIC

given in the Annexure I.

2.3, CONDITIONS ADOPTED TO ARRIVE AT THE RELIABLE

ESTIMATES OF 'K'

As the spectroscopic method of evaluation of K is liable for wrong

estimation by even very small errors present in the absorbance readings

or concentrations of donorlacceptor solutions, much care was taken to

get the most accurate values of absorbance for every concentration in all

the systems. This was achieved by strictly following the conditions

given below during our entire experimental work.

1. Best grades of chemicals and reagents available were used after

subjecting to rigorous purification, following the standard procedures.

2. Calibrated weights, microburettes and volumetric flasks were used.

3. The donor substances and the solvents were purified freshly before

starting every experiment. Storage for more than 24 hours was

strictly avoided.

4. Before using the purified solvent, it was ascertained that peroxides

were totally absent by testing with the Reisenfeld-Lebafsky reagent

(2% potassium iodide in water buffered to a pH of 7 with 5% sodium

hydrogen carbonate): the non-development of a yellow colour, (or a

blue colour on subsequent addition of starch solution) indicating that

iodine is not released, vouches for the absence of peroxides and

hydroperoxides.

5. Mixing of the donor and acceptor solutions in the cell itself, kept

inside the spectrophotometer, using a calibrated syringe so that the

readings were taken quickly after mixing (in just three seconds).

6. Absorbance readings were taken in the 'photometric' mode so that

quick scanning became possible.

7. The absorbance due to the complex alone is found out by subtracting

the absorbances of the pure donor and acceptor solutions of

respective concentration and the net absorbance value was used for

the calculation of K and E as proposed by Anunziata et al. [85].

8. A minimum of nine different donor concentrations were used for

collection of data.

9. Experiments were repeated several times using newly prepared

solutions until constant absorbance values were obtained.

10.The absorbances were measured at more than one wavelength near

the charge-transfer maxima, in order to confirm the reliability of the

K values obtained.

1 1 .The absorbance versus concentration data had been analysed by eight

linear and three non-linear methods of evaluation of K. The

correlation coefficient (R) in the case of linear regression methods

and the (minimized) sum of deviations squares (S) in the case of non-

linear methods were determined.

12.Temperature change during the experiments was kept minimal by

using an effective air-conditioner.

CHAPTER - 111

RESULTS AND DISCUSSION

3.1. RESULTS

3.1.1 Studies with ring substituted aryl amine complexes:

The molecular complexes formed between electron donors of low

ionisation potential and acceptors of high electron affinity have their CT

absorption at relatively longer wavelengths, often well separated &om

the absorption bands of the components themselves. It is well known

that p-chloranil is a strong n: acid and forms EDA complexes with a

variety of Lewis bases. The strong acidic properties of p-chloranil are

due to the higher electron affinity conferred by the electron withdrawing

effect of the extended 7c orbital of the quinone and the planarity of its

structure. The aliphatic amines and the substituted anilines are the ones

which are often used as electron donors.

Foster and coworkers [86,87] examined chloranil complexes

formed with various methylbenzenes in cyclohexane and

carbontetrachloride media. Plots of CT band wavelength versus free

energy of complex formation were found to be linear. Foster [88] has

also plotted ionisation potential Vs charge transfer band wavelength for

some chloranil complexes in carbon tetrachloride. The agreement was

fair. Slifkin [12] has investigated the CT band arising at 600 - 800 nm

region representing complexes betweeen chloranil and donors

triethylamine, diethylamine and ethylamine. He attributed a charge

transfer band to a n -+ .n* transition from the lone pair of nitrogen in the

amino group to an empty chloranil orbital. Datta et al. [89] and Sarkar

et al. [90] have investigated the interaction of p-chloranil with a series of

phenols. It was found that all these EDA complexes (except with

p-quinol) were stable. In contrast, o-chloranil forms unstable complexes

with all phenols [91].

Arylamines, by virtue of their aromatic .n-electron system and the

arnine group, offer the possibility of functioning either as n- and lor as

n-electron donors towards a given electron acceptor, giving 1:l

complexes in the former two cases or a 1:2 complexes if both sites

function at the same time. These possibilities do not appear to have been

thoroughly considered hitherto.

In the present study, the following EDA complex formation

systems were investigated with p-chloranil (CA) and p-bromanil (BA) as

the acceptors, different substituted anilines as donors and benzene as the

solvent. Generally, quinones are known as good electron acceptors.

Hammond [92] has shown that the electron afficinity of p-benzoquinone

changes with the nature of the substituent at the quininoid nucleus.

Since a study of the electronic spectra of such molecules and the spectra

of the EDA complexes formed by them will provide a wealth of

information regarding the nature of inter molecular interactions. We

have chosen chloranil (tetrachloro-p-benzoquinone) and bromanil

(tetrachloro-p-benzoquinone) as acceptors. We report our results and

conclusions on the nature of interaction of these EDA complexes by

performing several correlations between the spectral and thermodynamic

parameters. We tried for the establishment of a linear free energy

relationship between the formation constant of these complexes and the

substituent constant o of the donor amines.

/ Sl.No / Sys tern j ~ 1 . ~ 0 / System 1 1 1. CA + Aniline 1 9. 1 BA t Aniline I

Intense blue to violet colours developed immediately on mixing

solutions of chloranil (or bromanil) and the arylamines, ascribable to the

formation of molecular complexes of the EDA type. The absorption

spectra of the complexes were well separated from the absorption

spectra of either of the components. Figures 3.1 - 3.8 shows the overlay

spectra of three resolved bands in the solvent benzene, one arising from

the local excitation of donor (D), the other from the local excitation of

the acceptor (C) and the third one from the CT excitation of the EDA

complex. The stoichiometry of the complexes have been verified by

Job's [84] method and the results collected for two representative

systems are presented in Table 3.1.

The 1 : 1 stoichiometry of the complexes formed by chloranil and

bromanil with all the amines was hrther confirmed by the perfectly

linear B.H plots and non-linear FHW-2 plots (cf. eqn.(l7) with n =2)

obtained in all the cases.

The increase in CT absorbance values with increase in

concentration of the donor has been recorded and the results are

summarized in Table 3.2 to 3.17. The last column of the tabulated

results gives the correlation coefficient (R) in the case of linear

regression methods and the (minimized) sum of deviations squared (S)

in the case of non-linear methods. The wavelength versus absorbance

spectrum of the donor-acceptor complex for all the concentrations are

obtained as overlay picture directly from the spectrophotometer, of

which a few are illustrated in Figs.3.9 to 3.12.

The pK, values of various ring substituted aromatic amines under

study are collected from the literature [93] and are presented in Tables

3.18 and 3.19. The given Hammett [80] a constant values are those

considered by Mc Daniel and Brown [94] for all the substituted anilines

except p-iodoaniline, the value for which is taken from van Bekkum,

Verkade and Wepster [95] .

To determine the ionization potential of the donor as per the

equation (31, hvCT = a ID -t b, Fanell and Newton [96] have used the

values 0.80 and -4.2 for 'a' and 'b' respectively for chloranil system.

Srinivasan and Uma Maheshwari [97] have found out these values to be

0.8058 and -4.2094 in their studies with chloranil acceptor. We have

used these values of 'a' and 'b' and calculated the ionization potential

values of the donors (Tables 3.18 and 3.19).

The following procedure is adopted to arrive at the values of K by

different methods suggested in chapter-I through a computer program

written in QBASIC [71]. The program inputs include the number, N, of

datum points which may vary from experiment to experiment. The

donor concentrations in each mixture need not be evenly spaced but it is

a matter of experimental convenience to space them in that manner. The

donor concentrations D (I) and the corresponding absorbances Z (I) are

entered as data towards the end of the program. It calculates the eight

linear and the three non-linear estimates of K and E and displays the

summarized results on the screen with detailed intermediate results as

output in the file "PRO DAY.

The values of K obtained by eight linear and three non-linear

methods for all the systems performed are given in Table 3.2 to 3.17. As

per the findings of Bhaskara Raju, Srinivasan and Sivaramakrishnan

[79], the Nelder-Mead routine returns the best estimates of K and is

more reliable than those returned by the other linear and non-linear

methods, especially for systems in which acceptors like chloranil and

bromanil are used. Hence the values of K obtained by Nelder Mead

method alone are collected and presented in the summarized form in

Tables 3.18 and 3.19 and are used for correlation purposes. The

observed molar extinction coefficient (E) values and the calculated

Gibb's free energy of formation (AGO) obtained from the relationship

AGO = - RT In K ... (31)

(where R is the gas constant and T is the temperature in absolute units)

are all collected and presented in Tables 3.18 and 3.19.

3.1.2 Studies with N-substituted aryl amine complexes :

One of the important properties of N-substituted aryl amine bases

is their ability to form EDA complexes with acceptors like chloranil and

bromanil. To understand the effect of structure on the reactivity of the

charge-transfer reactions, one can think of changes in the reaction center

brought about by substitution in the ring. Similarly one can think of a

change in the reaction center by substitution in the N-atom of aniline

substrate molecule. Hence for such an investigation, we took chloranil I

bromanil acceptors in benzene solvent and varied the donors by N-

substitution to constitute primary, secondary and tertiary amines and the

following systems were investigated:

! 1 ATO. I SYS tern NO. 1 ~ y s tenz I

1 17. CA + N-Methyl aniline 1 20. / BA + N-Methyl aniline I I I

I i i 1 18. CA + N-Phenyl aniline , 21. a BA + N-Phenyl aniline I i I i

The absorbance versus concentration values are reported in Tables

3.20 to 3.25. The values for aniline in these tables are taken from Tables

3.2 and 3.1 0. The results of the observed spectral parameters LC-, ET

and ID, the thermodynamic parameters K (Nelder Mead) and AGO

values, along with the literature [93] values of pK, are collected and

presented in the summarized tables 3.26 and 3.27.

3.2, DISCUSSION

Spectral measurements have been carried out under identical

conditions at a single temperature for a closely related series of ring and

N-substituted aryl amine donors with chloranil and bromanil acceptor

EDA complexes. Hence it is reasonable to assume that the same

mechanism is operating throughout these reactions. With the results

obtained from our measurements, we have performed the following

analysis of the various parameters to arrive at the nature of

intermolecular interactions of the EDA complexes studied and to probe

into the effects of electronic perturbations produced by the substituents

on the reaction center.

The stoichiornetry of the EDA complexes have been confirmed by

the Job's method as illustrated for one system each for chloranil and

bromanil in Figures 3.13 and 3.14, in which the results presented in

Table 3.1 are plotted. It is found that the absorbance reaches a

maximum when the mole fraction is in the same ratio as that of the

stoichiometry, namely 1 : 1. The linearity of the B.H plots (as illustrated

for the two systems in Figures 3.15 & 3.16) and the non-linear F.H.W

plots obtained for higher order complexes @lot of Z I Dn Vs Z with

n > 1) further confirms the absence of higher order complexes. The

illustrative liner F.H.W plot (n = 1) for the 1:1 complex formed by

aniline with chloranil in benzene solvent is shown in Fig. 3.17 and that

with bromanil is shown in Fig. 3.18.

3.2.1. Influence of substituents on the spectral parameters:

The early studies of Foster and co-workers 186-881 on chloranil

complexes with methyl benzenes, nitro benzene, iodine and iodine

monochloride showed a linear relationship between CT bands and the

free energy of complex formation in non-linear solvents. Gore and

Wheals [98] and Mukherjee and Chandra [99] have studied the

interaction of chloranil with aniline and its methyl derivatives,

Foster and Hanson [I001 studied the interactions of chloranil with

indoles and reported a bathochromic shift with increase in methylation

of indoles. Generally in chloranil complexes, the absorption below 400

nm is ascribed to n+n* transition and absorptions above that

wavelength is attributed to n+n* transition [ I 0 1 - 1031. The influence of

the substituents in the present study has been revealed by the observed

spectral changes. Relative to the parent H in the series, the substituents

3-CI, 4-1, 4-Br and 4-C1 shift the wavelength of the charge transition to

the lower side while the substituents 3-Me, 4-Me and 4-Me0 shift the

charge transfer maxima to the higher wavelength side. Though the CT

maxima are lower in bromanil system, a perusal and Tables 3.18 and

3.19 shows the same trend as that of the chloranil system. For the arnine

X-C6&-NH2, the substituent group X, if electron withdrawing, produces

a blue shift while an electron donating X produces a red shift.. '

complex)

The measure of change in energy of the CT transition reaction

hvCT = ET are indicating the higher energy demand by electron

withdrawing groups and lower energy demand by electron donating

groups. The summarized results indicate the higher energy demand of

the bromanil complexes confirming the lesser electron affinity of the

acceptor compared to that of chloranil as observed by Hammond [92].

Chowdry and Basu [I041 reported that the CT interaction energies

and pKa values do not appear to correlate. Nag Chaudhury and Basu

[I051 considers the pKa value as a measure of the n-electron ionisation

energy. Figures 3.19 & 3.20 and the data in Table 3.18 and 3.19 show

the experimental values of CT transition (ET) energies of these EDA

complexes of chloranil and bromanil compare satisfactorily with the

reported pKa values.

The calculated values of ionisation energy of the donors of these

electron-transfer reactions (making use of eqn.(3)) as reported in Tables

3.18 and 3.19, show a clear trend of the electronic perturbation of the

substituent on the reaction center. It is evident from a comparison of

either ET or ID of m-substitued amines with similar para substituted

amines that the halogen and alkyl substituents at meta position require

more energy for this electron transfer reaction than at the para position.

Halogen substituents in general require more energy than alkyl

substituents since for halogen, opposing resonance and inductive effects

operate at p-position. At m-position the +R effect is weak and +I effect

is dominant for alkyl groups whereas -I effect is dominant for halogen

substituents. The order of ID fbr various amines are:

ID (ev) of 8.26 8.18 8.15 8.14 8.13 8.03 7.90 7.83

X-C6H4-hi2

Since the set of donors are identical for the two acceptor systems,

and in the absence of the literature values of a and b for bromanil

system, the ID values of donors are plotted against ET values of bromanil

system (Fig. 3.21; r = 0.998). From the slope and intercept of the

straight line obtained, the values of 'a' and 'b' for bromanil are found to

be 0.9047 and - 4.8560 respectively.

It can thus be concluded from the results obtained that the position

of the C.T absorption band varies with the ID of donor.

3.2.2. Influence of substituents on the magnitude of K & its

correlation with ET and ID:

Though different methods have been adopted for the

determination of K and E and all the estimates are reported, due to the

inadequacies present in all the linear methods of evaluation, the non-

linear methods are best suited for evaluation [71] and the Nelder Mead

method has been chosen. For the discussion of the results, the Nelder

Mead method estimate of K has been employed as it has been assessed

that, since both chloranil and bromanil have assays of about 98%, the

Ievel of errors in the acceptor concentration is much higher than that in

the donor concentration or in the absorbance readings due to

instrumental error.

It is seen from our results that the observed trends in the stability

constants of the series of amines remain the same in all the methods of

evaluation. It is possible to conclude that the changes in the values of K

are due to the changes in the electronic perturbations brought about by

the substituent's inductive effect at the meta position and mesomeric

effect at the para position. It will be meaningful to examine changes in

K,,, (= Kx / KH) for the series of anilines X-C6f&-NH2 with structural

changes in X and interpret the results

complex) K r e ~ (BA 0.401 0.484 0.585 0.641 1.000 1.400 1.830 3.673

complex)

The above order indicates that a very strong complex is formed by

p-methoxy aniline (3.5 times stronger than aniline) while a very weak

complex (3.34 times weaker than aniline) is formed by m-chloro aniline

for chloranil complexes. The same trend is observed in bromanil

complex systems also.

Bhattacharya and Basu [106], Chowdhury and Basu [104],

Dwivedi and Banga [I071 have reported from their works on EDA

complexes that the complex stability (K) order satisfies the condition

that the lower the ionization potential, higher will be the stability. From

our works, a linear variation is observed between the ionization energy

of the donors and the stability constant of the complexes formed as

shown below (rounded off to three decimals), the same being illustrated

in Fig. 3.24 and 3.25.

(M-') 0.374 0.481 0.544 0.598 1.250 1.607 2.056 4.331 CA complex

(M-l) 0.308 0.373 0.450 0.493 0.770 1.077 1.408 2.826 BA complex

One does infer that the reaction is facilitated by substituents which

are electron releasing and is retarded by electron withdrawing groups at

the reaction center. As observed and reported by several workers [107-

1091, fiom the data given in Table 3.18 and 3.19 it is apparent that there

is a good correlation between the measured CT energies and the

evaluated equilibrium constants in the logarithmic form as (I+ log K)

for both chloranil and bromanil systems (r = 0.95 and 0.97 respectively).

3.2.3 Establishment of LFER with Hammett's substituent constant

and evaluated equiiibrium constant & other spectral parameters :

In any discussion of substituent effects on reaction rates and

equilibria, three components must be borne in mind; the substituent or

"source" of the perturbation, X, the reaction site or "detector" of the

perturbation, Y and the molecular frame work (core) through which the

effect is transmitted. This is the benzene ring in the original Hamrnett

systems. The two basically distinct electronic effects namely inductive

and mesomeric effects that may be generated by a substituent on a

reaction site are responsible for the changes in the rate of a chemical

reaction or the position of chemical equilibrium. Model systems in

which reaction center and substituent are separated by an aromatic ring

as core, eg. the Hammett system, permit a mixture of resonance and

inductive interaction to be .transmitted, the former only weakly from a

meta- position.

The influence of polar substituents on reactivity concern the

potential energies of the reacting systems. Changes in log K are a good

measure of potential energy effects and it provides some explanation for

the success of the Hammett equation. The most common application of

the Harnmett equation is in connection with mechanistic studies. Many

authors have sought to show that the influence of meta- or para-

substituents in a given reaction of an aromatic system supports a

postulated mechanism or atleast does not disprove it. Relevant evidence

may be obtained variously from the p value, from the kind of o values

needed in the correlation, or from the linearity or non-linearity of a

particular Hammett plot.

A number of workers have examined the effect of substituents on

the donor-acceptor interactions of a related group of donors with an

acceptor. The equilibrium constants and the enthalpies of formation in

such system have been correlated with the substituent constant. Foster

and Goldstein [ I 101 in their work on spectroscopic studies of aryl

ketones- iodine complexes report that the first ionization potential of

some para substituted acetophenones show a correlation with the

Hammett constant for the substituent group and concluded the ionisation

potential refer to the removal of an electron fiom the carbonyl oxygen

rather than from the n- aromatic system. While the same two authors

[ I l l ] in their work on aryl ketones - TCNE complexes suggest, based

on the IR studies, that the complex formed should be of the n - 7c type.

Since our investigations too reveal a very good correlation between ID

and o (r = -0.95) and since the acceptors chloranil and bromanil in our

study are 7c acceptors like TCNE, we may conclude that the EDA

complex formed between aryl amines and the two acceptors may be both

n-n as well as n-n type which is supported by the earlier work [112].

The data in Tables 3.18 and 3.19 have been cast into a Harnmett

Plot of (1 + log K) vs o. The slope of the Hammett plot is p. Fig. 3.26

shows an excellent linear correlation (r=0.99) between the two

parameters with a slope of - 1.550 for chloranil system and Fig. 3.27

shows a straight line ( ~ 0 . 9 8 ) with a slope of - 1.3653 for bromanil

system. Thus the reaction constant p obtained for the two systems from

the slope of the straight lines is in accord with the theory of CT complex

reactions. The sign and magnitude of p indicates the spontaneity of the

reaction and a measure of the extent of electron demand at the reaction

center.

The p value obtained is in conformity with the work done by

Aloisi et al. [113] and by Srinivasan and Sindhu Jayarajan [ I 141. We

know that the parameter p measures the ability of the core to transmit the

electronic effects. The magnitude of the p gives a measure of the degree

to which the reaction is responding to substituents. Since this is on a

logarithmic scale, a change in p of 1, (i.e, 10' = 10) indicates a ten fold

change in rate / equilibrium. The negative sign indicates that the

reaction is facilitated by electron donation. From the value of p for the

EDA complexation reaction and fEom the perusal of Tables 3.18 and

3.19, we infer that for a change of substituent from m-chloro to

p-methoxy (Ao = 0.641) the reaction is sensitive to the substituents by 101.55 x 0.641 = 9.85 (or - lo), there is about 10 fold increase in the

equilibrium for chloranil system. The value gets reduced to 7.5 times

increase (10 1.3653 x 0.641 = 7.50) in bromanil systems. Thus the fitting of

an LFER in these reactions implies unchanging mechanism in spite of

changes being made to reaction conditions viz., changing the

substituents on the substrate.

When we utilise the Harnmett's [so] substituent constant (0) listed

in Tables 3.18 and 3.19 to correlate with the spectral property, namely

ET or ID of both chloranil and bromanil complexes, we find that a linear

free-energy relationship is established (Fig. 3.22 & 3.23; r = 0.95 & 0.96

for CA and BA systems respectively).

The positive and negative values of Gibb's free energy of

formation (AGO) (Tables 3.18 and 3.19) derived from the K values of

these donor-acceptor complexes indicate the retardation and

enhancement of these reactions by the substituents, electron donating

groups acccelerating the reaction and electron withdrawing groups

decelarating the same. A good correlation exists between AGO and ET

(r = 0.95 for CA system and 0.97 for BA system ; Fig. 3.28 & 3.29) in

both cases. As per the findings of Medina et al. [I 151, not only does the

correlation exist, but is also linear as is seen in Figs. 3.30 & 3 1, where

AGO values are plotted against pKa values of the bases (r = 0.98 and

0.97 respectively)

3.2.4 Comparison of the parameters of chloranil and

bromanil complexes:

Foster and Thomson [I 161, in their works on EDA complexes

reports that a plot of ET of tetracyano benzene complexes verses ET of

chloranil complexes of differently substituted same series of donor is

linear and reports that this relationship is generally observed for energies

of inter-molecular CT transitions. Venu Vanalingam and Subba Ratnam

[117] have reported a linear relationship between ET (1) and ET (2) for a

given donor with two different acceptors where (1) refers to the new

acceptor and ( 2 ) , the standard acceptor (chloranil), provided the standard

acceptor chosen has the same basic skeleton. Our studies of the effect of

changes in the substitution on the basic skeleton of p-benzoquinone is

very interesting. Since the substitutions are symmetrical, similarities

exist between the observed phenomena. Figure 3.32 shows a linear plot

(r = 0.998) between the observed ET values of chloranil and bromanil

complexes of ring substituted donors.

Figure 3.33 shows a linear relationship between the log - log plots

of formation constants of the two series of ring substituted donors with

chloranil and bromanil EDA complexes with a regression coefficient

value of 0.996. Establishment of the linear relationship is a clear

evidence for the changes in the structure producing proportional changes

in the AGO values and indicates an unchanging mechanism in spite of

changes brought out in the p-benzoquinone skeleton of the acceptor

moiety.

3.2.5 Correlation studies on N-substituted aniline complexes:

In the case of N-substituted aniline systems, it is found that the ACT

values of the EDA complexes gradually change as aniline (530 nm),

N-methyl (598 nm), N-phenyl(640 nrn), and N,N-dimethyl aniline (648

nm). The observed red shift on going from a primary amine to

s e c o n d a ~ amine and then to the tertiary amine is expected to be so

based on the electron donating strength of ArNR2 > ArhHAr > ArNHR

> ArhX2 predicted by the inductive and mesomeric effect of the alkyl

and aryl groups.

One of the most significant observations made in this investigation

is the smallest value of K obtained for N-phenyi aniline in its complex

formation with either chloranil or bromanil which may be attributed to

the polar and steric effects on the two bulky phenyl groups.

The factors that determine the stability of the EDA complexes

formed are : (a) Steric factor and (b) Electronic factor. Steric factor

include Van der waal's repulsive interaction (steric hindrance), strain

from bond angles, bond lengths, electrostatic interactions etc. Electronic

effect manifest from groups which can release electron to the reaction

site or withdraw electron from reaction site and it determines the

feasibility of the reaction. The release of electrons to the reaction site is

followed by a favourable equilibrium for the forward reaction and vice

versa.

The basic strength of the N-substituted anilines as reflected in their

pKa values (Table 3.26) is in the order: N-phenyl < aniline < N-methyl <

N,N-dimethyl. As the two phenyl groups attached to the nitrogen in

N-phenyl aniline can withdraw the lone pair electrons of the nitrogen by

the mesomeric effect, this arnine is a very weak base (weaker than

aniline). This property of N-phenyl aniline is reflected in the stability

constant values for its complexes with chioranil and bromanil.

A perusal of Tables 3.26 and 3.27 shows an observable anomaly in

the estimated values of K which follows the order:

DONOR N-Phenyl- Aniline N,N-Dimethyl- N-Methyl-

aniline aniline aniline

(M-') 0.7405 1.2499 (CA complex)

1.6623 1.9352

0.4516 0.7695 (BA complex)

0.9429 1.1084

As reported by Smith [118] and Tafi [119], the electron donor

strength of the N-substituted anilines decrease in the order ArNR2 >

ArNHR > ArNH2 predicted from inductive effect is expected to change

when bulky substituent groups exert steric hindrance in the formation of

EDA complex. The magnitude of the interaction is determined by the

factors, according to the works of Medina et al. [115], (a) one on the

electronic type and (b) the other being steric in nature.

Nogami et al. [81] considers that the values of the stability

constant reflect the donor strength of the arnine towards chloranil and

concluded that the same steric effects are present in the EDA complexes.

Hence the reason for the lower K values of the tertiary amines than that

of the secondary arnines may be purely due to the over whelming steric

factors. The stability order observed is N-phenyl < H < N-methy1 >

N,N-dimethyl. Positive AG" values have been obtained for diphenyl

arnine where as the other three amines gave negative AGO values in

chloranil systems. This trend is observed in bromanil complexes too,

though with a lesser magnitude, exhibiting a profound similarity

between the EDA complexes of these two acceptors.

CONCLUSION :

Several important conclusions may be drawn from these studies:

1. The results indicate that substitution in general affects the

electronic structure and charge distribution in the reaction center and that

stronger complexes are formed by electron-donating substituents and

weaker complexes are formed by electron-withdrawing substituents in

the substrate molecules.

2. Though several linear methods have been used for the

determination of stability constants, non-linear methods developed

recently (in particular, that of Nelder Mead Scheme) can be successfully

employed for obtaining the best estimates of stability constant, K.

3. Stability constants of the complexes with bromanil are lower

than those of chloranil.

4. The spectral shift and intensity changes observed are due to the

influence of the substituents.

5. Both elctronic and steric factors play a major role in stabilizing

the K-substituted arylarnine complexes.

6. It is seen that a good linearity exists between the Hamrnett

substituent constant (o) values and the log K values of our ring

substituted amine systems.

7. The magnitude of reaction constant p obtained for the two

(chloranil and bromanil) series suggest the extent to which the reaction is

responding to the substituents and its negative sign indicates that the

reaction is facilitated by electron donation.

8. A profound similarity has been found and demonstrated

between the EDA complexes formed by the two acceptors.

9. In order to understand the molecular properties of EDA

complex systems, one should take into account the electrostatic as well

as charge-transfer interactions.

10. The diversity of the sign and the magnitude of the overall

substituent effect is the result of different electronic perturbations caused

and cancellations between opposing effects.