Está en la página 1de 8

Elecmckimica Ada, Vol. 37, No. 12, pp. 2137-2144, 1992 00134686/92 55.00 + 0.

00
F’rinted in Great Britain. 0 1992. Pngamon Press Ltd.

ADSORPTION OF ORGANIC SUBSTANCES AT


ELECTRODES: RECENT ADVANCES

SERGIO
TRASATTI
Department of Physical Chemistry and Electrochemistry, University of Milan, Via Venezian 21,
20133 Milan, Italy

(Received 30 January 1992)

Abstract-Recent developments concerning four areas in the field of adsorption of organic compounds
at electrodes are critically examined and discussed. The topics are: (1) possibility of experimental
determination of the molecular size factor, n; (2) effect of the nature of the metal on the Gibbs energy
of adsorption; (3) effect of the crystallographic orientation of the metal surface; (4) effect of functional
groups in the adsorption of aromatic compounds from non-aqueous solvents. Results show: (1) the value
of n appears to be * 1 for all substances irrespective of the molecular size; (2) the nature of the metal
is mainly seen through the desorption of the preadsorbed water molecules; (3) the effect of the atomic
density of the metal surface cannot be assessed unambiguously from the few data available; (4) functional
groups can change the adsorbate-solvent interactions as well as the aromaticity of the benzene ring. The
combined effects can turn a physical interaction with the metal surface into a strong, specific chemical
interaction via the x-electron system, as a consequence of changes in the interfacial orientation.

Key worak: adsorption isotherm, single crystal face, metal-water interaction molecular size factor,
functional group.

INTRODUCTION that adsorption at solid/liquid interfaces is in fact a


solvent-replacement reaction[ 161 described by the
The study of the adsorption of organic substances at generalized equation:
electrodes, although it dates back to the early days
of interfacial electrochemistry with the pioneering B(sln) + nS(ads) c-) B(ads) + nS(sln), (1)
work of Gouy[l], still attracts wide interest because where B stands for adsorbing substance (adsorbate)
of its impact on our understanding of the structure of and S for solvent.
the electrical double layer[2,3], the kinetics of elec- Figure 1 provides a pictorial view of equation (1)
tron transfer[rl], the general field of corrosion (inhi- to emphasize the fact that under similar circum-
bition)[S], and the mechanism of electrode processes stances adsorption parameters are not representative
(role of intermediates)[6,7]. Recent reviews have of B only, but of both B and S. Therefore, they give
dealt with the adsorption of aliphatic[8] and aromatic information about the mode of adsorption of B and
compounds[9] at the Hg electrode, and with a at the same time they can be used to derive infor-
comparison of different molecular models of the mation about the state of the solvent at the interface.
adsorption layer[ lo]. In general, for a generic parameter Y, it is possible to
Although information on the molecular structure write:
of the adsorbate can be obtained in situ by means
of non-electrochemical techniques[ 111, the thermo- A Y(ads) = [ Ya(ads) + Y,(sln)] - [ Yi,(sln) + Ys(ads)]
dynamic analysis of surface concentrations (or sur- (2)
face excesses) is still the most used route to gain
insight into the mode of adsorption of organic com- or, in another form:
pounds[ 121. This entails the description of the exper- AY(ads) = [Ys(ads) - Ya(sln)] - [ Y,(ads) - Y,(sln)]
imental data by means of an adsorption isotherm,
and the construction of an adsorption model based (3)
on the derived adsorption parameters. Equation (3) emphasizes that the change of
In this context, how to proceed in the thermo- parameter Y upon adsorption also contains the
dynamic analysis and how to interpret the resulting change of Y for the solvent upon desorption. Typi-
adsorption parameters are still matters of discus- cal examples of parameters including contributions
sion[l315]. The aim of this paper is to analyse a few from the replaced solvent are the adsorption poten-
aspects which have recently emerged, and which can tial shift and the adsorption Gibbs energy. Rel-
contribute to gaining more insight into the molecular evant discussions have been presented recently
details of adsorption at electrodes. [14, 15, 17-201.

MODEL OF THE ADSORPTION LAYER SIZE FACTOR


There exists general consensus about the concept of The term n in equation (1) is the so-called size
the adsorption equilibrium: it is universally accepted factor. Physically, it is defined as the number of

2137
2138 S. TRASATTI

solvent molecules replaced by one molecule of adsor-


bate. Its quantitative value depends on the model of
the adsorption layer. 6 - .
t (4
A generalized definition of n is[21]:

n= k&lb&, (4)
where A, and A, are the area of interface occupied by
one molecule of B and S, respectively, and 1s and 1,
are the thicknesses of the adsorption layer expressed
in terms of unit of molecular thickness. Therefore, for
an adsorption layer whose thickness is the same as the .,,ll:::::::...*.............,..
thickness of the molecules of B, 1, = 1 and Is is equal 1’ -..
-.
to the number of monomolecular layers of S in the
adsorption layer.
The way n is expressed varies among different
authors. Some authors adopt a monomolecular layer
e
for the adsorption region[22,23]. Under similar cir-
Fig. 2. Simulation of isotherm tests. (a) Data conforming to
cumstances la = Is = 1 in equation (4) and n turns out
a Frumkin isotherm (1) plotted in terms of a Flory-Huggins
to coincide with the ratio of the projected areas of B isotherm (2) with n = 3; (b) data conforming to a
and S, respectively. Therefore, n is a model-dependent Flory-Huggins isotherm (2’) with n = 3 plotted in terms of
parameter whose value has been as a rule calculated, a Frumkin isotherm (1’).
since no direct experimental determination has been
envisaged for long.
A point of debate is whether or not n must test in terms of the Flory-Huggins isotherm
explicitly enter the adsorption equilibrium. Two pos- (0,/(1 - 8,)“xs us 6,). However, since experimental
itions have been taken. In Frumkin’s approach n is data, however accurate, are somewhat scattered, it
not introduced, which is tantamount to assuming is in practice hard to discriminate between the two
n = 1. The corresponding adsorption isotherm has models graphically. The only discriminating criterion
the form[24]: which can be adopted is a linear regression analysis
of the experimental tests, which does not provide any
KJa/(l- Wlev(-24d = I%, (3 physical certainty.
where t& is the fractional coverage of the interface There have only been two specific attempts to
with B, xa is the mole fraction of B in solution, u the determine the value of n. In one case, Damaskin[26]
particle-particle interaction parameter at the inter- concluded that the experimental data were better
face, and /I the adsorption constant. On the other represented by an equation with n = 1. Conversely,
hand, if n is introduced explicitly, the same mode1 on Lawrence and Parsons[22] reached the opposite con-
which the Frumkin isotherm is based leads to the clusion by fitting the data, for instance of sulpholane
isotherm bettem known as the Flory-Huggins adsorption, to a Flory-Huggins isotherm with
isotherm[25]: n = 2.27. On the other hand, the present author
observed[27] that if adsorption data are available
[e,/(i
-8,)“e”-i]exp(-2a’B,)=fl’x,. (6) only in a limited range of Be, it is impossible to reach
any conclusive evidence.
Proof that n must appear in the equations (ie that In the above context, a recent paper by Nikitas[28]
n # 1), might be obtained indirectly by analysing the
has apparently provided a clue to the solution of the
fitting of experimental data to adsorption isotherms. problem. Nikitas has claimed that n can be obtained
Figure 2 simulates adsorption data conforming thermodynamically from the raw experimental data.
either to a Frumkin isotherm or to a Flory-Huggins While the full derivation can be found in the original
isotherm. A linear test of the Frumkin isotherm paper, the fundamental equations are the following:
(t?,/( 1 - t&)x, us 0,) clearly produces a non-linear
w1 - 0
n = lim (7)
*d-1 c(iix, - e,)- X,/X,
where x is the mole fraction in solution, x0“ the mole

0 0
fraction at the interface, and C is given by the
expression:
+ C = exp{ln xa - @AJRT + K}. (8)

a 0 In equation (8) @ is the experimental surface pressure


due to adsorption (@ = y0 - y), and K is a constant
for a given system:

K = (YO- Y;)ABIRT. (9)

Fig. I. Model of adsorption as a solvent replacement In equation (9) y,, is the interfacial tension of the pure
reaction. solvent and 7; is that of the pure solute. Since the
Adsorption of organic substances at electrodes 2139

t b)

Fig. 4. Adsorption of dimethoxyethane at the free surface of


water: smoothed experimental data. Plot of equation (7) to
determine the value of n.
--.._._
8- -.-w-L
----____
1 is 0.8, the extrapolated K cannot be very accurate.
I
Once K is introduced into equation (8), and conse-
0 02 0.4 0.6 0.8 1 quently into equation (7), the resulting plot is as
e shown in Fig. 3b. Also in this case the extrapolation
Fig. 3. Adsorption of dimethoxyethane at the free surface of is not very accurate but a linear regression analysis in
water: raw experimental data[29]. (a) Plot of equation (8) to the range 0.3 c 0s < 0.8 provides a value close to 1
determine the constant term; (b) plot of equation (7) to for n.
determine the value of n. The same raw experimental data have then been
fitted by a Frumkin adsorption isotherm and the
adsorption parameters obtained. By means of the
value of 76 may not be available, K can be obtained corresponding state equation, smoothed e-0, and
from equation (8) by extrapolation: I&,-xs pairs of data have been recalculated. If the
smoothed data are introduced into equation (8), and
K = -iby, (In xs - @A,/RT). (10) then into equation (7), the resulting plot is as shown
in Fig. 4. Now the extrapolation turns out straight-
Thus, by means of two extrapolations it is poss- forward, and the resulting value of n is very close to 1.
ible, according to Nikitas, to determine the value It can be questioned that the use of an isotherm
of n. to fit the experimental data will produce smoothed
Nikitas has pointed out that the extrapolated value data reminiscent of the isotherm used. This is not the
of K is sensitive to the accuracy of the experimental case. Smoothing consists in fact in describing the
data, and that the value of n is extremely sensitive to @-log xs raw data by means of an analytical func-
the value of K. A smoothing procedure of the raw tion. It will be shown in a forthcoming paper that the
data is thus advisable. Unlike Nikitas, it is suggested choice of the function to describe the experimental
here that a very effective smoothing is obtained if the data is without any substantial effect on the derived
raw data are fitted by an isotherm and the isotherm value of n, provided the fitting is sound.
parameters are then used to recalculate the data. Data for other four organic compounds[29-311
While a more detailed discussion of this point will be have been processed in the same way. The results
given in a forthcoming paper, an example of data are summarized in Table 1. While the calculated value
processing is provided here. of n ranges from 4.5 to 7.8 in terms of molar
Figure 3a shows a plot of equation (8) using raw volume ratio (or from 2.8 to 3.5 in terms of ratio of
experimental data[29]. Since the highest value of 0s projected area), the values of n derived by extrapol-

Table 1. Theoretical and “experimental” sixe factor


n

Theoretical*
Adsorbate V A “Experimental”
1,4-Dioxane 4.7 2.9 Hg/sln 1.0*0.1
1,2-Dimethoxyethane 5.8 3.3 air/sin 1.0*0.1
Hg/sln 1.0*0.1
1,2_Diethoxyethane 7.8 3.5 air/sin 0.9 f 0.2
Hg/sln 1.2*0.2
Succinonitrile 4.5 2.8 air/sin 1.0 f 0.2
Hg/sln 0.8 f 0.2
Glutaronitrile 5.3 2.8 air/& 0.8 f 0.3
Hgjsln 0.8 f 0.3
2-Butanolt 5.0 2.8 air/& 1.6 f 0.3
Hg/sln 0.6 f 0.3
Trioctylposohine oxide? 10.9 8.7 Hg/MeGH 0.8 f 0.3
lV = Molar volume ratio; A = projected aa ratio.
tFrom Ref. [28].
2140 S. T~rri

ation are in the range 0.8-1.2, ie always approxi- If the organic compound is adsorbed in an orien-
mately 1. In Table 1 the previous data of Nikitas[28] tation allowing only weak, non-specific interactions
are also reported. Again, the value of n for two with the metal surface, AGO can also be con-
organic compounds is not far from 1, despite the fact sidered independent of the nature of the electrode
that in one case the area ratio is close to 9. surface. Under similar circumstances, equation (12)
The obtained results can suggest contrasting becomes:
conclusions.
A(A_,GO) = -A[AG”(S-M)], (13)
(1) The unit value of n indicates that the n
water molecules displaced by one molecule of adsor- ie the variation in Gibbs energy of adsorption be-
bate count (or behave) as a cluster from a statistical tween two metals equals approximately the difference
point of view. Therefore, n = 1 in equation (1). This in the work to desorb solvent molecules (n is the same
would support the earlier analysis of Damaskin[26] for different orientations of B in case the molar
and would corroborate the approach on which the volume ratio is used).
Frumkin isotherm is based. However, n = 1 does not The validity of equation (13) has been tested in
mean that solvent molecules desorb as clusters or that particular for sp metals[33,34]. As a measure of AC”
clusters of exactly the same size as the adsorbate (S-M), the enthalpy of formation of the oxide MO
pre-exist. More simply, from a statistical point of has been used. This is shown in the inset of Fig. 5
view the n solvent molecules must not be counted for the adsorption of n-pentanol. AadGo is seen to
separately. If this is the case, all existing analyses in decrease as ArHo increases, ie as the work to
which the value of n has been postulated as # 1 displace solvent molecules increases. In the plot the
should be reconsidered. position of Ag looks reasonable: with a low value of
(2) The derivation of n as suggested by Nikitas[ZS] ArHo( ie with a weak affinity for water mol-
is not a fully thermodynamic operation. In fact, in ecules, the AadGo is high. However, more recent
order to define n the size of the solvent must necess- results[35] give a much lower value of dadGo for Ag,
arily be introduced. The experimental data depend on as indicated by the arrow in the inset. This has been
the presence of the solvent but they do not contain its justified by the authors of Ref.[34] in terms of ion
parameters explicitly. Thus, in order to define n, co-adsorption (sulphates) in the case of their earlier
extra-thermodynamic assumptions or positions must results[36]. Thus, Ag turns out not to fit the plot in
be adopted. The situation is that, contrary to the the inset. The same is the case for Au, if recent results
possibility of thermodynamics, more information is are considered[37,38]. A,H”(MO) of Au is in fact not
looked for than the raw data can give. In other words, very different from that of Ag, and AB,Go is also
the raw experimental data do not contain any infor- small. Therefore, the plot in the inset of Fig. 5 does
mation on n, therefore they cannot provide any value not seem appropriate to include sd metals together
of n other than 1. with sp metals.
No firmer conclusions can be reached for the time This author has recently shown[39] that a more
being. Further analysis is needed and will be reported universal parameter than ArH”(M0) is obtained from
in a forthcoming paper. However, one may wonder the plot of the potential of zero charge vs. the electron
why the surface pressure, which is determined by the work function. This new parameter X, called the
surface concentration and depends on the bulk con- “interfacial parameter”, encompasses the various
centration, should be different depending on whether contributions to the interfacial potential drop:
or not solvent molecules are desorbed as monomers.
XM = 6~~ + g’(dip). (14)
In equation (14) 6~~ is the modification of the surface
NATURE OF THE ELECTRODE METAL
potential of the metal due to the contact with the
Equation (1) and Fig. 1 suggest that the Gibbs solution, and gs (dip) is the surface potential brought
energy of adsorption at electrodes must contain, in in by the liquid phase.
particular, a contribution from the desorbed solvent.
More specifically, it can be written as:
A,,G” = A,Go - nA,G”. (11) r 21-
E ..“Q
AsG” includes B-S and B-M(meta1) interactions at 3 #- -5.
the interface. Thus, to a first approximation, equation P
$ ,,_
(11) can be. rewritten as[32]:
AadGo = AG”(B-S) + AG”(B-M) - nAG”(S-M). * 15-
(12)
The first term on the right-hand side of equation (12) 0 -a1 -02 -0.3
allows for the different state of B in the bulk and on Ax/V
the surface in the absence of specific interactions with Fig. 5. Plot of the standard Gibbs energy change at zero
the metal, which is in turn accounted for by the coverage for the adsorption of n-pentanol on various
second term. Thus, AGo accounts for any metals, against the “interfacial parameter” AX derived from
squeezing out effect of B molecules from the solution. the work function-potential of zero charge relationship.
For a given organic compound AG”(B-S) is (to a Inset: the same data plotted against the standard enthalpy
first approximation) the same at different electrodes. of formation of the oxide M0[33,34].
Adsorption of organic substances at electrodes 2141

Since gs (dip) is related to any specific interactions although AadGo varied in the expected way from the
between the metal and solvent molecules[40], ArHo more open face (111) to the more compact face (2TT).
(MO), as used in the inset of Fig. 5, is likely to include The data existing for Ag at the time of the review also
only part of the factors governing AadGo. In fact, showed that this metal conformed to the same rule,
the modification of the electron distribution at the ie the A,d G” for the (111) face was higher than for the
surface of the metal (which determines the surface (100) face. However, those data were later shown[35]
potential of the metal) is also a consequence of to be vitiated by ion co-adsorption. Therefore, the
specific M-S interactions, although these may not agreement of the data for Ag with the predictions
contribute to the orientation of solvent molecules. above now appears rather fortuitous.
Therefore, XM is thought to measure the interaction Recently, work on single crystals of Au[43,44] and
between the metal and the solvent more globally. Ag[35] revealed an inversion in the expected order of
Figure 5 shows a plot of A,,G“ vs. AX, a relative crystal face specificity. This is shown in Fig. 6 where
value of XM taken with respect to Hg (absolute values the earlier data for Au and Ag are confronted with
cannot be obtained; on the other hand, they are not the more recent data on the same metals. In the case
needed in this correlation). The data points gather of Ag the authors have interpreted the new observed
around a straight line more closely than in the inset sequence as indicating that the (111) face is more
of the same Fig. 5. Moreover, the point for Ag now “hydrophilic” than the (100) face, a conclusion which
falls quite nicely on the common straight line. The contrasts with other evidence[41]. Also, for Au the
point of Au is low, but the interfacial parameters of authors have reached the conviction that the ob-
polycrystalhne surfaces (as it is the case for this served behaviour is related to specific interaction of
result) are not perfectly defined: the data for Ag on the OH group of alcohols with the metal surface.
the contrary refer to the (110) face. It is clear that the explanations provided by the
It can be concluded from Fig. 5 that the interfacial different authors for the same pattern observed with
parameter X simulates the strength of the interaction Au and Ag do not agree. In fact, the strong inter-
between the solvent and the metal surface better than action of RaH postulated for the (110) face of Au
ArH”(MO), and that the Gibbs energy of adsorption implies that this face interacts strongly with water as
of organic substances, provided no specific adsor- well, ie the (110) plane can be defined as more
bate-metal interaction takes place, varies with the “hydrophilic”. Opposite conclusions are in the end
nature of the metal in a way which can be closely reached for Ag and Au, two metals crystallizing in the
predicted by the value of X”: the larger the contri- same system and with almost the same atomic size.
butions (electronic and solvential) to the interfacial On the other hand, the data for Ag are taken in KF
potential drop in the absence of foreign adsorbates, solutions. Fluoride ions have been shown[45] to
the weaker the tendency of organic substances to be adsorb on Ag, especially at defects[46]. Thus, the
adsorbed. effect of ion co-adsorption cannot be ruled out for Ag
It is interesting that Ag and Au are poor adsor- even in fluoride solution, although this is excluded
bents despite the weak affinity for water (ie low value by the authors themselves on the basis of their
of A,HO). This confirms that, as stated earlier by this observation of constancy in E, =o with varying KF
author[33,41], in the case of sd metals electronic concentration[47].
contributions to the interfacial properties are more The above data indicate that some additional,
important than solvential contributions. not yet identified factors, should probably be taken
into account with single crystal faces of sd metals.
Other available data concern the adsorption on Ag
CRYSTAL FACE SPECIFICITY and Au[4&50] of pyridine, an aromatic compound

Work on adsorption of organic substances at metal


single crystal faces has been reviewed by this author
in a previous paper[42]. If adsorption at single crystal
faces is assumed to be governed by the same factors
as at heterogeneous (polycrystalline) metal surfaces, I l./'
/- //
Fig. 5 can be used to predict the order of crystal face
sequence. It has been shown previously[42] that the
more compact the crystal face, the more positive the
potential of zero charge. At the same time, XM
increases (linearly) as the potential of zero charge
becomes less positive[41] (ie XM is higher for open
faces than for close-packed structures). In this con-
text, it can be expected from Fig. 5 that AadG” should
decrease as the potential of zero charge becomes less
positive, ie on going from the more compact to the
more open faces.
The above prediction was confirmed in the pre-
Fig. 6. Plot of the standard Gibbs energy for the adsorption
vious review[42] on the basis of the very few data of various substances vs. the potential of zero charge of
available at that time. In particular, a linear corre- different crystallographic faces of metals (1,l’) Ag; (2,2’)
lation between AadGo and E,,, was found (with the Au. (1) n-Pentanol from KF solutions[35]; (1’) n-hexanol
expected trend) for Pb, Zn and Au. In the case of from Na$O, solutions[36]; (2) tqxntanol[43,44]; (2’) di-
Bi, there appeared to be a non-monotonic variation ethylether[S9].
2142 S. Trww-r~

whose ascertained specific interaction with the metal


surface[Sl, 521 prevents discussion of the data in the
context of the above approach. Further studies are
thus needed to gain decisive insight into these factors.

ADSORPTION FROM NON-AQUEOUS


SOLVENTS

Equation (12) clearly shows that the energetics of


adsorption can be understood in terms of a number
of pairwise interactions: adsorbate-solvent, adsor-
bate-metal, metal-solvent, and also solvent-solvent
and adsorbate+adsorbate which are included but not
explicitly expressed. In order to single out (to a first
approximation) one contribution at a time, some of
the factors must be made constant and the others
varied. Thus, metal-water interactions can be seen (as
shown above) by working “at constant solvent and
adsorbate” and varying the metal. Further, metal-
adsorbate interactions can be detected by working “at
constant adsorbate and solvent” and by confronting Fig. 7. Plot of the relative variation of the standard Gibbs
the situation at the air/solution interface[29,30]. Ac- energy of adsorption for various substituted benzene com-
cordingly, adsorbate-solvent interactions can be in- pounds on Hg from ethylene glycol solutions us the elec-
vestigated by working “at constant metal and trode potential in the rational male. (1) Benzene; (2) benzoic
adsorbate” and by varying the solvent. acid, (3) dimethyl-o-toluidine; (4) dimethylaniline; (5) phe-
Although the above reasoning is only a simplifica- nol; (6) aniline.
tion, it illustrates the general strategy which can be
adopted depending on the aim of the research work. which fundamental interfacial studies have already
A topic which falls in the above context is the effect been performed[54].
of substituents and functional groups on the inter- Table 2 summarizes the substances investigated.
facial behaviour of organic substances. Investigations The basic molecular unit is the aromatic nucleus to
of the adsorption of classes of homologous com- which the following substituents have been attached:
pounds have been usually carried out from aqueous -CH3, -OH, -NH2, -COOH, alternatively further
solution[8]. Due to the hydrophobic interactions in substituted with CH, groups. The aim was to change
aqueous solution, organic substances are as a rule the aromaticity of the ring and at the same time
adsorbed by squeezing out effects so that they nor- the adsorbat-olvent interaction. Some of these
mally interact with the electrode surface through the substances were studied in aqueous solution by
inert hydrocarbon tail. This leads to A,,G” being Blomgren and Bockris[55] long ago with a different
correlated with the solubility of the adsorbing sub- aim. While detailed results are reported in another
stance, and varying approximately quadratically with paper[56], Fig. 7 shows that the electric field depen-
the electric field at the interface[53]. dence of Asd G” (among other parameters) splits these
In non-aqueous solvents the nature of the adsor- substances into two groups. For some of them a
bate-solvent interaction changes substantially and quadratic dependence of AndGo on the electric field
this may lead to interfacial orientations quite differ- is observed, while for the others the dependence is
ent from those in aqueous solution, and therefore to linear and AndGo increases as the field becomes more
new and unexplored interfacial situations. In this positive.
context, research work on the interfacial properties of The picture in Fig. 7 suggests that competition
differently substituted aromatic compounds has been exists between the effect of the substituents on the
carried out in ethylene glycol solutions, a solvent in solubility and their effect on the aromaticity of the
ring. It is clear that a quadratic dependence of AadGO
entails strong adsorbate-solvent interaction so that
Table 2. Organic compounds only squeezing out effects operate. This is the usual
Benzene
pattern obtained in the adsorption of these substance
C,HrH
Toluene C,H+H,
from aqueous solution. A linear dependence of AadG”
Phenol C,H&H on the electric field simulates the adsorption of
Aniline C,HrNH, anions, ie increasing interaction with positive charges.
Benzoic acid C,HdOOH Similar behaviour has been observed with thiourea,
Dimethylaniline C,H+(CH,), known to interact chemically with Hg via the sulphur
NH, atom[57]. Therefore, the linear dependence of Asd G”
o -Toluidine C,H/ suggests that a chemical interaction takes place via
\CH, the n-electron system of the aromatic ring with
,N(CH,), partial charge transfer to the metal. This model is
Dimethyl-o-toluidine C,H, supported by the observed shift of the potential of
\CH, zero charge towards more negative values[S 581.
Solvent: ethylene glycol. Table 3 summarizes the above conclusions.
Adsorption of organic substances at electrodes 2143

Table 3. Subdivision of substituted benzene well as the aromaticity of the benzene ring in such a
compounds adsorbed on Hg from EG way that it is possible to obtain either physical or
Group A Group B strong chemical adsorption. Working in non-aqueous
(non-interacting) (interacting) solvents is essential to observe these effects, since
Benzene Phenol squeezing out effects as a rule prevail in aqueous
Toluene Aniline solution thus smoothing down any further details.
Benzoic acid DM-aniline
DM-o-toluidine o-Toluidine Acknowledgements-This work was supported by the
MURST (40% funds) and the CNR.

The interfacial behaviour of organic compounds


can thus be modulated by modifying substituents REFERENCES
and functional groups. Although this is a well-known 1 G. Gouv. Ann. chim. Phvs. 8. 294 11908).
s.
strategy, it is essential to modulate at the same 2. P. Nikit&, J. electroanai Chem. &l, 235 (1988).
time the adsorbat+solvent interaction for the 3. S. Trasatti, in Trenris in Interfacial Electrochemistry
adsorption behaviour to be modified completely. (Edited by A. F. Silva), p. 1. Reidel, Dordrecht (1986).
Thus, molecules which are otherwise only physically 4. J. Lipkowski, C. Buess-Herman, J. P. Lambert and
held at the metal surface can be changed into chemi- L. Gierst, J. electroanal. Chem. 202, 169 (1986).
cally adsorbed entities. This is especially important 5. L. A. Avaca, E. R. Gon&lez and A. Ruvolo Filho,
in the field of corrosion inhibitors, and the results J. appl. Electrochem. 12, 405 (1982).
illustrated here provide an example of the strategy 6. V. Solis, T. Iwasita, A. Pavese and W. Vielstich,
J. electroanal. Ckem. 255, 155 (1988).
which can be adopted to obtain the desired effects. 7. S. G. Sun and J. Clavilier, J. electroanal. Chem. 236,
95 (1987).
8. F. D. Koppitz, J. W. Schultze and D. Rolle,
CONCLUSIONS J. electroanal. Chem. 170, 5 (1984).
9. D. Rolle and J. W. Schultze, J. electroanal. Chem. 229,
Four recently developed topics in the field of 141 (1987).
adsorption of organic substances have been examined 10. P. Nikitas, Electrochim. Acra 32, 205 (1987).
and discussed. The following conclusions can be 11. K. Ashley and S. Pans, Chem. Rev. 88, 673 (1988).
12. L. Stolberg, S. Morin, J. Lipkowski and D. E. Irish,
drawn. J. electroanal. Chem. 307, 241 (1991).
(1) Nikitas’ approach to the determination of the 13. P. Nikitas, J. electroanal. Chem. 309, 1 (1991).
size factor gives n N 1 for all substances examined. 14. J. Jastrzebska. M. Jurkiewicz-Herbich and S. Trasatti,
This outcome warns against introducing a calculated J. electro&al.’ Chem. 216, 21 (1987).
15. J. Torrent and F. Sanz, J. electroanal. Chem. 303,
size factor as a postulate, but at the same time hints 45 (1991).
of the possibility that the true value of n, if different 16. J. O’M. Bockris, M. A. V. Devanathan and K. Miiller,
from 1, cannot in fact be obtained thermodynami- Proc. R. Sot. 274A, 55 (1963).
cally from the experimental data. Further studies are 17. P. Nikitas, J. electroanal. Chem. 170, 333 (1984).
needed to assess the situation. 18. M. Karolczak, Electrochim. Acta 34, 511 (1989).
(2) The adsorption Gibbs energy of organic com- 19. D. Schuhmann, J. electroanal. Chem. 201, 247 (1986).
pounds varies with the nature of the electrode metal 20. A. Daghetti, S. Trasatti, I. Zagorska and Z.
essentially as a result of different metal-solvent inter- Koczorowski, CON. SurJ 51, 29 (1990).
action strengths. Data for different metals turn out 21. H. Nakadomari, D. M. Mohilner and P. R. Mohilner,
J. phys. Chem. 80, 1761 (1976).
to be in a closely linear correlation with the “inter- 22. J. Lawrence and R. Parsons, J. phys. Chem. 73,
facial parameter” X obtained from work func- 3577 (1969).
tion-potential of zero charge plots. This result 23. B. E. Conway and L. G. M. Gordon, J. phys. Chem. 73,
confirms the strict consistency of different parameters 3609 (1969).
and of their analysis as carried out by this author in 24. A. N. Frumkin, Z. phys. Ckem. 116, 466 (1925).
previous papers. 25. H. P. Dhar, B. E. Conway and K. M. Joshi,
(3) Consistently with the logic of point (2), the Electrochim. Acta 18, 789 (1973).
adsorption Gibbs energy is found to vary linearly 26. B. B. Damaskin, EIektrokhimiya 1, 63 (1965).
with the potential of zero charge of the different 27. S. Trasatti, J. electroanal. Chem. 2% 257 (1970).
28. P. Nikitas, J. electroanal. Chem. 263, 147 (1988).
crystallographic faces of a given metal. However, 29. A. Daahetti. S. Trasatti. I. Zadrska and Z.
recent results for Ag and Au have shown an inversion Koczoriwski; Electrochim. Acta 33, 1705 (1988).
in the expected sequence of crystal faces. While 30. S. Minjozzi, A. Daghetti, S. Trasatti, I. Zagorska and
possible reasons of experimental interference have Z. Koaorowski, J. ckem. Sot. Faraday Trans. 1 79,
been suggested, the results point to the possible 2801 (1983).
existence of still unidentified factors in the adsorption 31. A. Daghetti, C. Gatti and S. Trasatti, J. electroanal.
of organic substances on well-defined metal crystal Chem. I%, 179 (1985).
faces. Further studies with different adsorbates are 32. S. Trasatti, J. electroanal. Chem. 6S, 815 (1975).
33. S. Trasatti, Croat. Chem. Acta 60, 357 (1987).
necessary.
34. S. Trasatti, Electrochim. Acta 28, 1083 (1983).
(4) Dramatic changes in adsorbat+metal inter- 35. A. Popov. 0. Velev and T. Vltanov, J. electroanal.
action are seen as aromatic compounds with various Chem. 256,405 (1988).
substituents and functional groups are adsorbed on 36. T. Vitanov and A. Popov, Elektrokhimiya lo,1373 (1974).
Hg from non-aqueous solvents. Different substituents 37. R. Holze and M. Beltowska-Brzezinska. Electrochim.
can modulate the adsorbate-solvent interaction as Acta 30, 937 (1985).
2144 S. TrusATn

38. R. I. Tucceri and D. Posadas, Electrochim. Acta 32, 50. L. Stolberg, S. Morin, J. Lipkowski and D. E. Irish,
27 (1987). J. electroad. Chem. 307, 241 (1991).
39. S. Trasatti, Electrochim. Acta 36, 1659 (1991). 51. B. E. Conway, J. G. Mathieson and H. P. Dhar,
40. S. Trasatti, .I. electroanal. Chem. 54, 437 (1974). J. phys. Chem. 78, 1226 (1974).
41. S. Trasatti, J. electroanal. Chem. 172, 27 (1984). 52. R. I. Kaganovich, B. B. Damaskin and M. A. Suranova,
42. S. Trasatti, Mater. Chem. Phys. 12, 507 (1985). Elektrokhimiya 7, 1158 (1971).
43. J. Richer and J. Lipkowski, J. electroanaf. Chem. 251, 53. S. Trasatti, J. electroanal. Chem. 53, 335 (1974).
217 (1988). 54. J. I. Japaridze, N. A. Abuladze, Sh. S. Japaridze,
44. J. Richer, L. Stolbcrg and J. Lipkowski, Lungmuir 2, A. De Battisti and S. Trasatti, Electrochim. Acta 31,
630 (1986). 621 (1986).
45. M. Bacchetta, S. Trasatti, L. Doubova and A. Hamelin, 55. E. Blomgrem and J. GM. Bockris, J. phys. Chem. 63,
J. electroanal. Chem. 255, 237 (1988). 1475 (1959).
46. G. Valette, J. electroanaf. Chem. 122, 285 (1981). 56. J. I. Japaridze, Sh.S. Japaridze, N. A. Abuladze,
47. T. Vitanov, A. Popov and E. S. Sevastyanov, J. A. De Battisti and S. Trasatti, J. them. Sot. Faraday
efectroanal. Chem. 142, 289 (1982). Trans. 87, 3399 (1991).
48. A. Hamelin, S. Morin, J. Richer and J. Lipkowski, 57. R. Parsons, Proc. R. Sot. MlA, 79 (1961).
J. electroad. Chem. 272, 241 (1989); 304, 195 58. V. N. Gerovich, E. N. Protskaya and B. B. Damaskin,
(1991). EIektrokhimiya 16, 1257 (1980).
49. L. Stoberg, J. Lipkowski and D. E. Irish, J. efectroanal. 59. J. Lipkowski, C. Nguyen van Huong, R. Parsons and
Chem. 238, 333 (1987). J. Chevalet, J. electroanal. Chem. 143, 375 (1983).

También podría gustarte