You are on page 1of 8


ELSEVIER Analytica Chimica Acta 321 (1996) 53-60

Determination of gallic acid by an oscillating chemical reaction

using the analyte pulse perturbation technique
Rafael Jim~nez-Prieto, Manuel Silva, Dolores P~rez-Bendito *

Deparfraent of Analyai¢@lChemistry, Fc:cultyof Science~, Universi~ of Cdrdoba, E-14094 C6rdob~ Spare

Recei'~ed31 Jury 1995; revised 23 October 1995; accel~ed 30 Oclobcr 1995


A new analytical method for the determination of gallic acid by the sequential perturbation caused by different amounts
of acid on the oscillating chemical system involving the Cu(iI)-catalysed reaction between hydrogen peroxide and sodium
thiocyanate in an alkaline medium is proposed. The method relies on the lineal" relationship between the changes in the
oscil[atien amplitude of the chemical system and the concentration of gallic acid. which is in this work exposed for the first
time. The t:alibmtion curve is linear over the range 0.075-2.00/~mol of gallic acid. The method features good ptecisiou (rsd
G.67%) and excellent throughput (14 samples h - t) thanks to the expeditiousness with which the steady state is regained after
each perturbation. Some aspects of the interaction of gallic acid with the oscillating chemical system are discussed. This is
the first reported determination of gallic acid based on an oscillating reaction.

Keywords: Oscillating chemical reaclion; Analyt¢ pulse perturbation; Gallic acid

1. I n t r o d u c t i o n have been primarily studied from a physico-chemical

point of view in order to elucidate the complex
Some far-from-equilibrium chemical s y s t e m s ex- reaction m e c h a n i s m s behind oscillations, and to de-
hibit an oscillating behaviour as a result o f their velop theoretical models ac<:ounting for the experi-
complex m e c h a n i s m s including and autocatalytic step mental results [7]. The most widely k n o w n and
[1]; such s y s t e m s are usually referred |o as 'oscillat- studied oscillating chemical s y s t e m s are those based
ing reactions', Oscillating reactions can take place on the Belousov-Zhabotinskii [ 8 - 1 0 ] and B r a y -
both in the liquid phase [ 2 - 4 ] and in the g a s e o u s Liebhafsky reactions [11-13]. More Tccently, oscil-
phase from heterogeneous catalytic systems [5]. Ther,~ lating s y s t e m s involving no halogen c o m p o u n d (pat-
are also several references to naturally occurring ticulafly the CuSO4-catalysed reaction o f hydrogen
oscilqating reactions such as those resulting in peri- peroxide with thiocyanate ion [14]) have aroused
odic changes in the calcium concentration in a vari- increasing attention. For oscillations to last long
ety o f cell types, which are highly significant to enough, the reagent inflow ~ product outflow m u s t
protein phosphorylation [6]. Most o f such s y s t e m s be carefully maintained constant in order to ensure
that the reactor is kept under steady concentration
conditions [15,16]; this is usually accomplished by
• Corresponding author. using a continuous-flow stined-tank reactor (CSTR).

0003-2670/96/$15.00 ~, 1996 El~cvicr Scicnce B,V. All rights t~rved

SSDI 0003-2670(95)00563-3
54 R. Jim~nez-Prieto et aL /Ana.rytica Chimica Acre 321 (1996) 5 3 - 6 0

Oscillating reactions have scarcely been explored continuous manner; the system regains the steady
for analytical purposes with the exception of ~ome state after each perturbation, which allows several
discrete determinations including those of hexa- determinations to he carried out simultaneously on
cyanoferrates [17] and chloride in human serum [18] the same system [21].
by use o f the Bclousov-Zhabotinskii reaction, or The proposed method widens the scope of oscil-
that of Cu(I1) at the microgram-per-millilitre level lating chemical reactions in quantitative analytical
with the H~O2/malonie acid/Mn(II) system [19]. determinations. It relies on the analytical potential of
Also, the Belousov-Zhabotinskii reaction has been the copper snlphate-catalysed reaction of hydrogen
implemented in a flow injection system [20]. Re- peroxide with thiocyanate ion in an alkaline medium
cently, our group developed the so-called 'analyte to determine gallic acid with the aid of a CSTR. The
perturbation technique' as a tool for analytical deter- chemical system is perturbed with variable amounts
minations in far-from-equilibrium dynamic systems, of gallic acid (3,4,5-trihydroxybenzoic acid), which
where the oscillating system is sequentially per- results in substantial changes in the oscillation ampli-
turbed by the analyte, which is determined in a tude that are proportional to the acid concentration.

,','l,','~,~;'; I ~ ' ',b~,~


Pt e t e c t r o d e

Satured calomel electrode f ~ .
N a g C N / NaOH
Sample tM'cropipette) C u S O 4 / NaCI

CS;R~' "~


M a g n e t i c stirrer Peristaltic p u m p
Fig. 1. Experimentalset-upfor the implementationof oscillatingleactionsusingthe anaiytepulsepenmbationtechnique.
R. Ji~nez-PrieW et aL/Analytica Ckimica Acre 321 (1996) 53-60 55

In addition, the oscillating system regains the steady wet,: injected in the analyte pulse pertm4bation m o d e
state shortly after each perturbation. by m e a n s o f a Nichiryo 5000 m i o ' o p i l x ~ . T h e
reactor w a s kept under steady state conditions by
using a Gilson Minipuls-3 four-channel peristaltic
2. Experimental pump; three o f the channels were used to deliver the
reactants and the fourth to keep the v o l u m e o f the
2.1. Reagents reaction mixture in the C S T R constanL Poly(vinyl
chloride) p u m p i n g tubes and PTFE tubing for the
All chemicals used were of analytical reagent manifold were used throughout.
grade (Merck). Gallic acid, sodium thiocyanate, cop-
per sulphate pentahydrate, hydrogen peroxide, 2.3. Procedure
sodium hydroxide and sodium chloride solutions were
all prepared in bidistilled water. To the CSTR~ thermostatcd at 30°C, were added
in the following sequence: 2.0 ml o f 0.7 M sodium
2.2. Apparatus chloride, 1.0 ml o f 0.245 M s o d i u m thiocyanat¢, L 0
ml o f 0.42 M sodium hydroxide, 0.23 ml o f 33%
The experimental set-up used to implement the ( m / v ) hydrogen peroxide and bidistilled water up to
oscillating reaction (Fig. 1) consisted o f a 10 ml a final v o l u m e o f 6.0 ml, the mixture being h o m o g e -
glass reaction vessel (CSTR) furnished with a ther- nized by magnetic stirring. The oscillating reaction
mostated jacket connected to a Selecta 6000383 ther- w a s started by adding 1.0 ml o f 1.75 × 10 -3 M
mostat and an Eyela RC-2 magnetic stirrer for ho- Cu(ll) and, without delay, the peristaltic p u m p deliv-
mogenization. Potential changes were recorded by ered the tlwee reactant streams B the overall feed
m e a n s o f a M e t r o h m E A 2 0 2 Pt electrode and an stream w a s 0.64 M in H 2 0 2 , 7 × 10 -~- M in NaSCN,
A m e l 303 SCE interfaced to a P C - A T 12 M H z 0.120 M in NaOH, 2.5 × 10 -~ M in C u S O 4 and 0.2
compatible computer via a PC-Multilab PCL-812PG M in NaC! - - at a constant f l o w - a t e o f 7 nfl r a i n - 1.
12-bit analog-to-digital converter (ADC). Samples The system evolved gradually to constant oscillating

125 - -

ZoneB i i


i 75

50 I I I I |
0 ~ logo 1~00 200o 25o0
Time, s

Fig, 2, Typical oscillation p~ofilcsobtaincd in the absence and p~SCn~ of vafiabl© amv~m~ of gallic acid tmdct sclc~ctl e.xpcrirncnlal
conditions. The inset illustrates Ihe selection of the injection poinL The meani,zg of zones A and B, and [mint 1, is explailled in the [~XL
56 R. Jim~nez-Prieto et aL / Analytica Ch[m[ca Acta 321 (1996) 53-60

amplitude and period (a steady state). As soon as the containing the analyte should be injected once recov-
signal and time increments for the oscillations lev- ery has started (point 1) since this is the zone where
elled off, variable volumes (a few microlitres) o f the system will respond maximally to the gallic acid
sample or standard containing also variable amounts perturbation. Injections result in a sharper rise in the
o f gallic acid from 75 to 2000 nmol were sequen- potential, which reaches a maximum and then falls to
tially injected. Data were acquired at a rate of t.35 normal oscillation levels. The system regains the
p o i n t / s using software developed by the authors in oscillating conditions prevailing before the perturba-
Microsoft QuickBasic v. 4.0; the oscillating ampl~- tion is introduced within 4 - 5 rain after injection.
rude obtained in the presence o f gallic acid was used Once that time has elapsed, a new determination can
a s the measured parameter to construct the calibra- be started.
tion plot.
3.1. Influence of experimental variables
3. Resnlts a n d discussion The effect o f experimental variables on the pro-
posed oscillating reaction was studied in order to
Oscillating chemical systems exhibit periodic establish the optimal working conditions for deter-
changes in the concentration of some species that mining the species vi.', which the perturbation was
reflect in cyclic colour, pH or potential changes, introduced: gallic acid. In order to ensure the maxi-
depending on the particular system [22,23]. As a mum possible sensitivity and precision in the deter-
rule, oscillating systems arc highly vulnerable to mination, as well as the ability to perform large
external perturbations such as temperature changes series of analyses, conditions were optimized with
[24] or variations o f the 0 2 concentration in the two primary objectives in mind, namely: (a) to
medium [25]. The Cu(ll)-catalysed oxidation of achieve the maximum possible stability in the oscil-
S C N - ion by H2O 2 in a strongly alkaline medium
involves successive oscillations in the concentrations

of the oxidized and reduced form o f the metal ion 7 3.0 B

which give rise to the formation and subsequent 5 2.S

disappearance of a yellow superoxide-cuprous com-
plex (HaO-Cu(I)) that governs the rate of the pro-
cess and is responsible for the colour changes ob-
served [26]. On interaction with this oscillating sys-
tem, gallic acid causes ch'mges in the oscillation
" i
o3 o.~
[ Hydrogen [0qlo)d(~ l, M
o.~ o,~
[ Sodium Ulloc~nMm I. M

amplitude and period, thereby perturbing the steady

state; however, the system gradually regains the
steady state after the perturbation ceases. Because - 3,0 -I 2.0
the above-described changes are proportional to low 25 L5 *

gallic acid concentrations and the system recovers

from the perturbation quite rapidly, we explored the 15 "1 0.5
10 [ i : i 0.0 i ~ i
possibility o f exploiting this behaviour for the deter- O.OO 0.02 004 O~OIS OOO 0 1 0 10 1~ 20 2.§ 3.0 3.5 .
mination o f gallic acid. [~ m I'o~gc~dde ], M (C~Qge~ll) ] x 10"4 M
Fig. 2 shows typical oscillation profiles for the Fig. 3. I n f l u e n c e o f ~h¢ concentration o f ( A ) h y d r o g e n p e r o x i d e ,
proposed system in the presence and absence of (B) sodium thiocyanate,(C) sodium hydroxide and (D) copper(ll)
gallic acid under the above-described working condi- on the gallic acid-perturbedoscillating reaction. Conditiolts:(A)
tions. The inset shows the oscillating region where 560 nmol of gallic acid, 2.SxlO -2 M NaSCIq, 2.5:<10-2 M
the perturbation must be introduced. Thus, if a nor- NaOH and 1.5 × 10-4 M Cu(ll); (B) 280 nmol of gallic acid, 0.32
mal oscillation cycle is assumed to raise the potential M H202 , 2.5 x 10-2 M NaOH and t.5 × 10-'~ M Cut[l); (C) 225
nmol of gallic acid, 0.32 M H 2 O 2 , 3.5:'(10 -'~ M NaSCN and
(zone A) to a maximum, from which it subsequently 1.5× 10-4 M Cu(i]}; (D) 168 nmo| of gal|ic acid, 0.32 M ] ' I 2 0 2 ,
drops (zone B) to the starting conditions, the sample 3 . 5 X 10 - 2 M N a S C N a n d 6 . 0 : < l 0 ~2 N'aOH.
R. Jin~nez.Prleto ¢t al. /Aaalytica Citimica Acla 321 (19v~) 53-00 57

luting system over time; and (b) to ensure that the finally chosen in order to maximize ~ system re-
oscillation amplitude allowed the effect of gallic acid sponse to the gallic acid perturbation.
on it to be accurately determined. Thus, the influence "['he effect of the Cu(ll) concentration in the
of selected experimental variables was studied in the reactor was studied over the range (1.5-3.5) × 10 -4
presence and absence of gallic acid using the ampli- M; the oscillation period w ts found to decrease with
tude ratio obtained under both types of conditions as increasing copper concenW~tion without s ~ t l y
the measured parameter. altering the oscillation amplitude. As can be seen in
The effect of the hydrogen peroxide concentration Fig. 3D, the effect of the gallic acid perturbation was
was studied over the range from 0.1 to 0.4 M. Low similar at every Cu(ID concentration tested_ A cop-
H202 concentrations were found to adversely affect per(II) concentration was thus chosen as optimal
the oscillating system (induction periods exceeding since it caused the system to oscillate higllly eai-
10 m ~ were observed at the beginning of oscilla- fonnly prior to the perturbation and rest'ended qeite
tions that resulted in the loss of the oscillating state predictably to it.
by effect of minimal perturbations and eventually led Changes in the ionic strength of the reaction
to a chaotic state). As the hydrogen peroxide concen- medium were also found to affect the determination;
tration was increased, both the oscillation amplitude raising ionic strength by adding increasing coacen-
and period decreased, and the induction period even- tratious of NaCl up to 1.0 M increased the response
tually disappeared altogether. On the other hand, the to the gallic acid perturbation, with no appreciable
perturbation with gallic acid elicited a response that change in the oscillation period or ampliUtde. Also,
fitted a curve such as that of Fig. 3A. As can be in q ~ i t a t i v e terms, an in,:rease in the NaCI concen-
seen, the curve fell as the H202 concentration was tration was found to ialfibit the passivation of the Pt
increased; it reached a minimum and then rose up to electrode by precipitated copper hydroxide, so the
a maximum at a concentration of 0.32 M, which was measured potential was subject to no increasing
adopted as optimal - - while the response to the background noise as time elapsed. An NaCI concew
perturbation was somewhat small relative to lower lration of 0.2 M was thu:~ adopted in order to exploit
H202 concentrations, the system oscillated more all the above-mentione,a advantages.
uniformly and was subject to no induction period. The temperature also had a strong influence of the
The sodium thiocyanate concentration in the oscillating system: raising it from 2 5 ~ 45"C dramat-
CSTR was changed between 2.0 and 4.5 X 10 -2 M. ically decreased the oscillation pc~od (by about
As the concentration was increased, the oscillating 70%), with virtually no change in the oscillation
period for the chemical system lengthened while the amplitude. However. the system resfmnse to the
amplitude remained fairly constant. The effect of the gallic acid perturbation was not altered by tempera-
gallic acid perturbation was such that it increased the ture changes. A temperature of 30"C was therefore
response with increase in the sodium thiocyanate chosen as optimal as it ensured correct, reproducible
concentration (Fig. 313). An NaSCN concentration of oscillations.
3.5 x 10 -2 M was chosen as a compromise between Finally, the effect of the overall reactant flow-rate
maximum sensitivity (amplitude) and minimum anal- fed to the CSTR was studied; this variable influ-
ysis time (period of the oscillating system). enced the ease with which the steady state was
The influence of the NaOH concentration was regained after each perturbation and hence the repro-
investigated over the range (1.5-9.0)× 10 -2 M; ducibility and the system resistance to the pertmbao
concentrations below 2.5 × 10 -2 M resulted in a tions. Decreasing the ratio of the system volume (7
marked drift in the potential zone where oscillations ml) to the overall reactant flow-rate - - which was
took place. Raising the pH suppressed the drift and changed between ! ~nd 7 ml min -z - - was found to
caused the system to oscillate more uniformly with- result in an increasingly reproducible response to the
out significantly affecting the oscillation amplitude gallic acid perturbation. A unity ratio (i.e. an overall
or period. The effec~ of the NaOH concentration in reactant flow-rate of 7 ml rain -1, obtained with one
the presence of the gallic acid is illustrated in Fig. of ca. 2.3 ml min-J for each reactant) was adopted
3C. An NaOH concentration of 6.0 × 10 -2 M was as optimal.
58 R. Jimdnez.Priezo et al./Analytica Chimica Acta 321 (19~) 53-60

3.Z Determination of gallic acid Table 1

Regression data and analytical figures of merit for the determina-
The oscillating system was perturbed with vari- tion of gallic acid
able amounts o f gallic acid under the above-de- Dclerminationmethod
scribed optimum conditions and the response ob- Paramelcr Amplitude Maximum
tained was analysed in two ways, namely: (a) by potcnlial
measuring the amplitude on each perturbation; and Linear range, p.mol 0.075 - -2.0 0.075---2..0
(b) from the potential measured at the peak o f each Slope, mV p.mol- i 28.80 4-0.73 35.624-0.68
perturbation. Fig. 4 shows the maximum potentials Intercept, mV 8.50 4-0.50 63.1 4-1.0
obtained for all the oscillations (both perturbed and S~andard error of estimate 1,28 1,75
Regressioncoefficient (n = 9) 0,9977 0.9973
unperturbed) in a typical determination sequence. As Dot¢ctionlimil a nmoI 22 20
can be seen, there is a w e l l - d e f i n e d baseline defined Precision (rsd) b % 2,05 0.67
by the potentials for the system oscillations in the
steady state and a series of points lying off the line a Calculated ~ ' n r d i n g to I U P A C ' s recommendations [27].
b Fromeleven determinations of 0.3 #tool gallic acid.
that represent the perturbations. A similar recording
can be obtained by using oscillation amplitudes rather
than potentials. method was somewhat better on account o f its higher
A plot of amplitude ~r maximum potential for precision. We finally detcrminod the throughput; un-
each perturbation as a function of the injected ana- der the above-described experimental conditions, the
lyte conccn~ation was I~near in both cases. This chemical system was allowed to oscillate for at least
allows one to address the determination of gallic acid 4 h, time during which the oscillating reaction acted
by using an amplitude or a maximum potential by way of a detection system (it produced signals on
method. Table 1 summarizes the figures of merit of injection of micro amounts of gallic acid). From the
the calibration graphs. As can be seen, the precision time taken by the system to regain the steady state
and sensitivity achieved in both types of determina- after each perturbation, the throughput was estimated
tion were quite good; in any case, the potential to be about 14 samples h - I .






60 t i i l I
0 50 100 150 200 250

Number of oSCillations
Fig. 4. Effect of variable perturbationsof gul]ic acid on the maximumpotential for the oscillating reaction. For gallic acid concentrations,
linear dynamic ranges and other conditions, see under Experimental.
R. 1imdnez.F, ~c:oet aL /Analylica Chimtca Acta 321 (199~) 53-60 59

Table 2 feedback loops by which a yellow s u p e r o x i d e -

Influence of chemically related specieson the determinationof 0,4
#tool ganie acid copper(I) complex is formed and destroyed, respec-
tively, and on which the autocatalytic process relies.
Foreign species Tolerated ratio
In order to elucidate the interaction o f gallic acid, the
Benzoic acid, phenylacetic acid, 61) oscillating system w a s monitored photometrically;
salicylic acid, 2-chlorophenol
Resorcinol 25 no effect w a s observed as a result o f the gallic acid
Phenol, 4-acetamidophenol, 10 perturbation, however. This allows one to infer that,
4-hydroxybenzaldehyde at least, gallic acid does not take part in the appear-
4-Hydroxypheaylaeeticacid, vanillin 6 anee or disappearance of the yellow s u p e r o x i d e -
3-Dimethylaminophcnol 2 Cu(I) complex, whether or not it is involved in any
4-Aminophcnol 0.6
Pyrogalloi 0.2 other step o f the oscillating mechanism. Further batch
assays in which binary, ternary and quaternary mix-
tures o f the ingredients o f the oscillating reaction
where supplied with a small a m o e n t o f gallic acid,
Oscillating s y s t e m s are reportedly very easily al- revealed a significant increase in the potential in the
tered by the presence of foreign species in the reac- presence o f Cu(ll), whether in the aqueous or in the
tion medium. Consequently, we investigated the ef- alkaline m e d i u m , but always in the absence o f hy-
fect o f some potential interferences from substances drogen peroxide, From these results, the ~allic acid
with chemical structures resembling that of gallic m a y be a s s u m e d to interact with copper owi;,g to its
acid. For this purpose, the oscillating system w a s reducing character.
perturbed with a sample containing a fixed a m o u n t
o f gallic acid (0.4 /xmol) and variable a m o u n t s of
interferents (less than 24 /.tmol in order 1o avoid
solubility problems). The results obtained are s h o w n
in Table 2. A s can be seen, the proposed method is
The authors gratefully acknowledge the financial
acceptably selective, which can be attributed to the
support o f the Spanish 'Direcei6n General Intermin-
special features of the analyte pulse perturbation
technique used to monitor the oscillating reaction. isterial de Ciencia y Tecnologfa (DIGICyT)" for the
We should note the strong interference o f pyrogallol, realization of this work as part of Project P B 9 1 4 ~ 4 0 .
which is structurally highly similar to gallic acid, and
that o f 4-aminophenol - - the oscillating system is
readily altered by the presence of primary amino References
groups in the reaction m e d i u m [14]. A s a rule, the
reducing character of the species studied had a strong [i] A.S. Toralin, Anal. Proc., 30 (I.tY~3)307.
influence on the extent to which they were tolerated: [2] M. Orban and I.R. Eps~in, J. Am. Chem. Sou., 111 (1989)
the more reductant, the more interfering.
[3] G. Schmitz, J. Chim. Phys. Phys.-Chim. Biol., 84 (1987)
3.3. Mechanism o f action o f gallic acid on the [4] J.D. Druliner and E. Wasserman, J. Am. Chem. Soc., IlO
oscillating system (1988) 5270.
[5] R.F. Melka, G. Olsen, L. Beavers and J.A. Draeger, i. Chem.
Educ., 69 (1992) 596.
The m e c h a n i s m s of oscillating reactions are usu-
[6] G. Dupont and A. Goldheter, Biophys. Chem, 42 (1992)
ally rather complex; while s o m e authors have at- 257.
tempted to elucidate them, the inferences are mere [7] R.M. Noyes, J. Chem. Edae., 66 (1989) 190.
approximations that necessitate s o m e refining. Ac- [8] RJ. Field and F.W. Schneider, J. Chem. Edu¢., 66 (1989)
cording to Epstein et al. [26], the oscillating reaction 195.
used in this work consists of 30 kinetic steps that [9] A.I. Zhuravlev and V.M. Trdinin, J. Biolumin. Cl~milumin.,
5 (1990) 227.
involve 26 independent variables; also, the centre- it0] V.P. Kazakov. A.D. Kaxavayevand S.R. Vakhidova, React.
pieces o f the m e c h a n i s m are positive and negative Fillet. CataL Lea., 45 (1991) 99.
60 R. Jim$nez-Prc'esoet al. / Analytica Chimica Acea 321 (1996) ~3-60

[11] S. Ani¢, D. Miti¢ and M. Cur~-ija, J. Scrb. Chem. Sot., 52 [20] R,T. Echols, M.K. Carroll and J.F. Tyson, Anal. Proc., 32
(1987) 575. ( 1 ~ 5 ) 3.
[12] S. Anic and L.Z. Kolar-Anic, J. Chem. Soc. Faraday Trans., [21] R. Jin~nez-Prieto, M. Silva and D. P~rez-lkndito, Anal.
84 (1988) 3413. Chem.. 67 (1995) 729.
(13] L. Kolar-Anlc and G. Schmitz, J. Chem. Soc. Faraday. [22] Z. Xu, S. Ni, J. Xu and Y. Tian, Huaxue Xuebao, 50 qL1992)
Tram., 88 (1992) 2343. 1085.
[14.] M. O r b S , J. Am. Chcm. Soc., 108 (1986) 6893. [23] M.T. Beck, l.P. Nagy and O. Szekely, int. J. Chem. Kinet.,
[MS] H,R. Weigt, Angew. Chem. Int. Ed. Engt, 31 (1992) 355, ( 1 ~ ! ) 881.
[16] R. Gyula and LE Epstein, J. Am. Chem. Sot., 114 (1992) [24] s. At'tic, L. Kolar-Anie, D. Slanisavijev, N, Begovic and D.
1529. Mitic, React. Kiaet. Calal. Lclt., 43 (1991) 155.
[17] M. Jiang, Y. Li, X. Zhou, Z. Zhao, H, Wang and J, Mo, [25] K. Saignsa, Chem. Phys. Lelt., 157 (1989) 251.
Anal. Chime. Acla, 236 (1990) 411. [25] Y. Luo, M. Orblln, K. Kuslin and I.R. Epstein, J. Am. Chem.
[18] O. Zhaag and J. Chen, Fenxl Shiyanshi, 7 (1988) 4. Soc., 111 (1989) 4541.
[19] T. Sekimoto, IC Hirayama and N. Unohera, Nihon Daigaku [27] G.L. Long and J.D. Winefordner, Anal. Chem., 55 (1983)
Kogakubu Kiyo Bunrui A, 31 (1990) 105. 712A.