- Email: [email protected]
Periodate determination by fia with chemiluminescence emission detection, and its application to ethylene glycol
~39-91~~9 f3.00 f0.00 Copyright Q 1989 F%gamon Press pk
Talmta. Vol. 36, No. 3, pp. 357-362, 1989 Printed in Great Britain. AI1 rights reserved
PERIODATE DETERMINATION BY FIA WITH CHEMILUMINESCENCE EMISSION DETECTION, AND ITS APPLICATION TO ETHYLENE GLYCOL N. P. EVMIRIDIS* Laboratory of Analytical Chemistry, Department of Chemistry, The University, Hull, England (Received 1 April 1987. Revised 30 September 1988. Accepted 27 October 1988) Summary-A method for periodate determination is given which combines the rapidity of flow-injection analysis and the sensitivity of chemiluminescence (CL) detection. It is based on the CL emission generated during oxidation of pyrogallol by periodate, and gives a relative standard deviation of 3% and a detection limit of 350 ng with the instrumentation used. The method has been applied to determination of ethylene glycol, with a detection limit of 0.5 gmole. The accuracy of the method is quite good when the ethylene glycol is oxidized in unbuffered solutions, and the interference- due to fo~aldehyde produced can be halved by prior addition of an appropriate amount of iodate. A throughput of I5 sample@& is possible, and the method is suitable for automation and remote control.
of periodate is almost exclusively related to analysis of organic compounds that have vi&al -CH@H, -CHO, >CO, or -COOH groups, especially polyhydroxy compounds. The consumption of periodate is determined by measurement of the periodate concentration before and after oxidation of the organic compound; titrimetric methodsle3 are usually used. The Miller-F~~~rger method’ is based on addition of iodide and titration of the iodine produced, but has the disadvantage that the iodine may react further with some of the products of oxidation. Some workersc6 have reported doubts of the reliability of the Fleury-Lange method2 for determination of excess of periodate in analysis of carbohydrates. Other workers,“‘O dealing with specific carbohydrate compounds, have found errors due to iodination or further oxidation by iodine formed during the determination procedure, or to incomplete reaction between iodine and arsenite. In addition, all the procedures above are time~ons~ng and not applicable in flow systems. A spectrophotometric method” is also used for the determination of very small quantities of periodate by monitoring of the absorbance at about 222.5 nm, but it suffers serious interference from iodate and the probability of oxidation of HCHO and HCOOH in the light-beam. Ponsi2 employed H-acid as a spectrophotometric reagent for periodate but the method is time-consuming and needs further statistical validation. It has recently been reported*3v’4 that chemiluminescence (CL) is generated during the oxidation Determination
*Present address: Laboratory of Analytical Chemistry, Department of Chemistry, University of Ioannina, Ioannina, Greece.
of pyrogallol with periodate, and this forms the basis of the method reported here. EXPERIMENTAL Reagents All chemicals were of analytical-reagent grade and demineralized distilled water was used throughout. The reagent solutions were prepared as described earlier.” Ethylen~Iyco~ (EG) aqueous solution, 0.5&f. Dissolve 31.04 g of ethylene glycol in 1 litre of water, and dilute further as required for lower concentration solutions. Phosphate buffer, pH 8.0. Dissolve 13.6 g of potassium dihydrogen phosphate in water, adjust the pH with 1M sodium hydroxide and dilute to 1 litre. Apparatus The flow manifold (Fig. 1) and deteztor’5 used have already been described.13 Method A sample mixture of injected into mixed with
(40~1) of periodate excess of periodate the carrier stream of tbe reagent (either
solution or the reaction with ethylene glycol is buffer solution, and then pyrogallol solution or
pyrog~loI-hydroxyla~ne solution). For periodate determination the reagent stream is [pg], = mx], = 1.0 x 10e3M in pH 8.0 phosphate btier. RESULTS AND DISCUSSION
The CL emission generated during oxidation of pyrogallol with periodate is rather weak and relatively insensitive to changes in periodate conoentration, but can be enhanced by the addition of hydroxylamine to the system.13 Optimization
of conditions
The reagent is a btiered solution of pyrogallol (pg) and hydroxylamine (Hx), which are considered to form a condensation product Pg + Hg z$ Pg-Hx 357
358
N. P. &SIIRIDts 1 ml /min
V
1 ml/min
Fig. I. Manifold and instruments used for the mixing and for monitoring the generated CL emission. (SI) carrier stream; (SZ) reagent stream; (V) sampie valve; OK) waste; (C) cell; (A) pump; (3’) photom~tiplier; (R) recorder.
that acts as a sensitizer in the oxidation of Pg with periodate.” However, any free hydroxylamine acts as an inhibitor by reduction of periodate. The effect of the concentration of each reactant on the overall CL emission was studied by varying the concentration of either Pg or Hx (keeping the concentration of the other fixed) at two levels of periodate concentration, 1.0 x 10A3M (low) and 1.0 x IO-*&# (high). Figure 2 shows that for [IO;] = 10W3Mand [Hx] = 10-3M the CL is maximal at pg] = 5 x 10T4A4, and that for [IO,-] = 10V2M and [Hx] = 10W3Mthe maximum is at [pg] = lo-‘M. The decrease in CL with
increase in f_Pg]after the maximum is presumably consequence of an inner filter effect of the Pg.
a
Efict of ~ydroxyiam~~ conee~tratio~ This was examined at two concentration levels of Pg, 5.0 x 10V4 and 1.0 x 10e3M and two periodate concentrations, 1.0 x lo-’ and 1.0 x 10e2M, Fig. 3. At the lower periodate level (lo-‘M) the CL obtained from 5 x 10-4M Pg is maximal at 5 x 10e4M Hx concentration, and is higher than that for 1.0 x 10-‘M Pg, which occurs at 2 x 10-3M I-Ix. This suggests competition between Nx and the Pg-Hx complex for oxidation by periodate, and this is substantiated by the results for the 1.0 x IOmZM periodate system, where the larger excess of periodate can accommodate the competitive reduction by Hx.
500
1500
Analysis of effects and interactions
ooo
500
-
1
2
3
t
I:Pgl x IO3 Fig. 2. Effect of pyrogallol concentration at [Hx]@= 1.0 x IO-‘M. Conditions: pH 8.0; [IO,-], 1.0 x 10-3&I (I), 1.0 x lo-*M (2).
It was obvious that the system was complicated, so a 2j factorial experiment was performed at two levels of each reactant, 1.0 x IO-* and 1 x 10m3A4 periodate, 5.0 x 10W3 and 5.0 x 10-*M Hx and 2.5 x 10m3and 5 x IOm4M Pg. The residual variance found for 12 measurements under identical conditions was S; = 379.5. Table 1 clearly demonstrates that all the effects and interactions of the analytical concentration levels are significant at the 95% level, especially the interaction between Pg and Hx. The next most significant are the effect of periodate and the ternary interaction between all three reactants. Effect of reagent concentrations on periodute calibration curues The concentrations of the reactants at equilibrium were calculated for an analytical Hx concentration of 1.0 x 10e3A4 and analytical Pg concentrations of 5.0 x 10m4, 1.0 x 10V3 and 2.5 x 10W3M, on the basis of reaction (1) and the assumption that Ic; = 1 x l@, and are given in Table 2 together with ~g-Hx~/[Pg~ which represents the ratio of [sensitizer]/[CL generator]. Calibration curves for periodate over the range O-O.OlM, for each of the [pg], [Hx] combinations, are shown in Fig. 4. It is clear that the equimolar mixture
Periodate determination by PIA 1500
4500
(a)
359
(b)
1000 s z c 0 1 2 E al d 500
I 1
u
234567
6
9
101
I
I
I
I
I
I
I
I
2
3
4
5
6
7
6
910
I
CHx3,mM
Fig. 3. Effect of hydroxylamine concentration at (a) [IO; ]o = 1.O x lo-‘M and (b) PO; ]o = 1.O x lo-*M. Conditions: pH 8.0; pg],, 5.0 x 10m4M (l), 1.0 x lo-‘M (2).
of Pg and Hx gives the greatest sensitivity and widest useful range. Optimization
of the sensitivity
The shape of the calibration curve is dependent on both [CL generator] and the ratio [sensitizer]/[CL generator]. Since it is not possible to adjust [pg-Hx], independently of PgL, it is necessary to find the analytical concentrations [Pg], and [Hx], which optimize the CL emission. As an equimolar mixture seemed best (Fig. 4), the effect of varying [pgb (= [Hx],) was examined. Figure 5 shows that
[Hx], = Pg10 = 1.0 x lo-‘A4 gives the best result overall. The crossing of curves 2 and 3 at low periodate concentration is due to the competitive oxidation of Hx by periodate when the [Hx]/[IO;] ratio is high, resulting in less CL. For low periodate concentration (
Table 1. Analysis of significance of effects and interactions (F-test)‘%*’ Effect on CL signal height, mV
Effects and interactions
P’gb
121
Fi]$xl0 [PgkCr
432 335.5 112.5 216 210 300.5
b Wxb[IO,_ b P~bFWo[IOilo
B = mean* square
95% C = S: x F;,
29,282
225,120 373,248 25,312 93,312 88,200 180,600
379.5 x 4.84 = 1836
B/C
Significance
>l
Yes
>l >1 >l >I >1
Yes Yes Yes Yes
*Mean square of each effect or interaction.
Table 2. Calculated equilibrium concentrations Dw& 1.0 1.0 1.0
Er 0.5 1.0 2.5
Fg;--Ileq*
lPg 8 mb
0.22 0.38 0.65
0.28 0.62 1.85
of Pg-Hx, Pg and Hx
y-lg 0.18 0.62 0.35
PHW,,Rpgl, 0.78 0.62 0.35
N. P. EVWRIDIS 1500
r
.9---”
~-1 ;;;pz .
2 1000
5 d
.c u)
.
/
/
d 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0
1.0
0.1
0.2
0.3
0.4
0.5
Fwiodatel
[Periodate] x10* (M)
0.6
0.7
0.9
0.9
1.0
x102(M)
Fig. 4. Periodate calibration curves at constant [Hx]. Conditions: [Hxl, 1.0 x IO-‘M; pH 8.0; [pg], 2.5 x lo-‘M (I), 1.0 x 10-‘&f (2) 5.0 x 10-4M (3).
Fig. 5. Periodate calibration curves at a ratio of [Pgb/[Hx], = 1. Conditions: pH 8.0; [Hxl, 1.0 x 10-4M (l), 5.0 x 10-4M (2), 1.0 x 10-3&f (3).
Application to determination of organic compounds
quantities of ethylene glycol, but lower consumption than expected at the higher quantities. This may be explained by the fact that with high excess of periodate there is oxidation of the formaldehyde formed’61g but with low excess of periodate the oxidation of ethylene glycol reaches completion very slowly.
used in analysis of organic compounds that have vicinal hydroxyl, carbonyl or carboxylic acid groups, which are oxidized to aldehyde, carboxylate or carbon dioxide respectively. The most important group determined by periodate oxidation is that of polyalcohols. Excess of periodate is added and the surplus determined. This can be done by the method proposed here. The application has been tested with ethylene glycol. Periodate
is almost
exclusively
Ethylene glycol oxidation in bu$ered solution (pH 8.0). Ethylene glycol was treated with excess of
Ethylene glycol oxidation in unbuffered solution. The results in Table 4 for oxidation with 5 ml of
2.5 x lo-‘A4 periodate mixed with ethylene glycol show an improvement, but the reaction is still slow to reach completion when the excess of periodate is small. The recommended method is addition of a fixed volume of sample solution, (e.g., 0.05 ml) containing not more than 10 pmole of ethylene glycol, to 5.0 ml of 2.5 x 10-‘&f periodate, and measurement of the periodate excess as already described. A linear calibration graph for the amount of ethylene glycol added is obtained over the range O-10 pmole.
periodate in aqueous solution buffered at pH 8.0 and after the end of the reaction a small volume of the reaction mixture was injected into the FIA manifold for periodate determination. Table 3 gives the results for a series of 5.0-ml volumes of 1.0 x IO-*&f periodate to which 0.05 ml of 0.5M ethylene glycol plus O-0.30 ml of 0.05M ethylene glycol had been added. The residual periodate concentration in each Selectivity sample was calculated and compared with the Oxidation of ethylene glycol by periodate leads value expected from the reaction stoichiometry. to the formation of formaldehyde and iodate: Comparison of columns 3, 5 and 7 of Table 3 shows there is over-consumption of periodate by low HOCH,CH,OH + IO; -, ZHCHO + IO; + H,O Table 3. Excess of periodate, calculated from CL emission after oxidation of ethylene glycol with periodate in pH 8.0 buffer solution Experimental signal peak height 30min after mixing Ethylene glycol added, mole
Calculated periodate excess, mole
[Periodate] excess, mM
CL,
[Periodate] excess,
mV
mM
25 30 35 40 45 50 55
25 20 15 10 5 0 0
5.0 4.0 3.0 2.0 1.0 0 0
850 820 720 630 500 350 200
3.10 3.05 2.70 2.50 1.95 1.45 0.95
60 min after mixing
CL, mV
[Periodate] excess, mM
780 750 670 620 480 325 200
2.80 2.70 2.50 2.30 1.85 1.35 0.95
Periodate determination
by FIA
361
Table 4. Excess of periodate calculated from CL emission after oxidation of ethylene glycol with periodate
in unbuffered aaueous solution Experimental signal peak height Immediately after mixing Ethylene glycol added,
Calculated periodate excess, mole
mole
0.0 2.5 5.0 7.5 10.0 12.5 15.0
[Periodate] excess, FnM
12.5 10.0 7.5
2.5 2.0 1.5 1.0 0.5
:3
8
is a reductant and the iodate an oxidant, so both may interfere in determination of the periodate excess. Table 5 shows the effect of some products from periodate oxidation of polyalcohols. The interference from fo~aldehyde and formic acid is rather high, but can be decreased by the presence of an equimolar amount of iodate.
The formaldehyde
Application
90min after mixing [Periodate]
CL, mV
[Periodate] excess, mM
CL, mY
CXCZSS,
690 555 410 290 175 165 150
2.55 2.10 1.55 1.25 0.85 0.82 0.77
680 530 350 210 60 25 0
2.50 2.02 1.40 1.00 0.52 0.25 0
fluorescent dyestuff) in water, or by diluting the water/antifreeze mixture from a car radiator. The samples were analysed by the Fleury-Lange and CL-emission methods. The results are shown in Table 6. There was good agreement for the pure ethylene glycol solutions, and fairly good agreement for the other samples. No problems were caused by the fluorescent dyestuffs.
of the method to aqueous samples
Sample solutions were prepared by dissolving a known volume of ethylene glycol (analyticalreagent grade) or commercial antifreeze (containing
Conchisions
The method is rapid, and reasonably accurate and reproducible. When the oxidation reaction used in
Table 5. Selectivity
Added, vole Ethylene glycol, 2.75 Formaldehyde, 2.75 Formic acid, 2.75 Acetic acid, 2.75 Formaldehyde + iodate (2.75 + 2.75) Formic acid + iodate (2.75 + 2.75) Table 6. Dete~nation
Relative decrease of CL emission, %
Equivalent amount of ethylene glycol, lunofe
100 13.3 27.7 3.2 6.9 8.2
2.75 0.37 0.76 0.09 0.19 0.23
of ethylene glycol content of aqueous samples Ethylene glycol content, mg Iml
Sample
Fleury-Lange* method
Ethylene glycoi solutiona A B C D Antifreeze solutions BP Glycoshell Cur radiator water One-year old car Six-year old car# Seventeen-year old car$ lr.s.d. 2%. 7r.s.d. 3%. gRadiator content renewed in previous year. #Car radiator with tiny leaks. TAL M/3-B
mM
CL detectiont method
2.79 3.26 3.72 4.06
2.79 3.26 3.63 4.03
3.30 3.35
3.40 3.53
0.336 0.458 0.031
0.372 0.418 0.062
N. P. E-IS
362
the application is fast, the whole determination can be performed in the flow-injection system by merged flow of separate solutions of the sample and the periodate, followed by passage of the mixture through the sampling valve, and injection of a sample into the carrier stream. The system could even be used for continuous monitoring of a process stream. The kinetics of oxidation of a specific polyhydroxycompound with periodate can also be followed fairly easily with approp~ate a~~gement of the stream which leads to the sample loop. Such systems can be automated and are suitable for remote control. Acknowledgements-I thank Professor Alan Townshend for giving me the opportunity to work in the Analytical Laboratory of the Chemistry Department of the University of Hull, and my students N. ~an~~ias and N. Santiris for helping in some of the experimental work.
REFERENCES E,
Milller and 0. Friedberger, Ber., 1902, 35, 2652. P. F. Flemy and J. Lange, J. Pharm. Chim., 1933,17, 107, 196. L. Malaprade, Bull. Sot. Chitn. France, 1928, 43,683: Compt. Rend., 1928, 186, 392. G. Hughes and T, P. Nevell, Trans. Faraday Sot., 1948, 44, 941.
5. P. Fleury, J. E. Courtois and A. Bieder, Bull. Sot. Chim. France, 1952, 118. 6. J. fi. Taylor, J. Am. Chem. Sot., 1953, 75, 3912. 7. J. C. P. -!jchwarz, Chem. Ind. London, 1954, 1000. 8. W. A. Bonner and R. W. Drisko. J. Am. Chem. SW., 1951,?3, 3699. 9. P, Fleury and M. Fatome, J. Phurm. C&m., 1935, 21, 247. 10. G. Lundblad, Ark& Kemi, Min. Geol., 1947,24, No. 25. 11. J. S. Dixon and D. Lipkin, Anal. Chem., 1954,26, 1092. 12. J. A. Pons, Pubt. Inst. Invest. Microquim., Univ. Na Litor& (Rosuriv. Arg.), 1965, 26, 175, 199, 215. 13. N. P. Evmiridis, Art&~, 1987, 112, 825. 14. I&m, ibid., 1988, 113, 1051. 15. J. L. Burguera, A. Townshend and S. Greenfield, Anal. Chim. Actu, 1980, 114,209. 16. L. HOI@ and B. M. Woods, Chem. Ind. London, 1957, 1421. 17. G. Lindstedt, Nature, 1945, 156, 448. 18. D. J. Bell, A. Palmer and A. T. Johns, J. Chem. Sac., 1949, 1536. 19. G. Neumiiller and E. Vasseur, Ark& Kern& 1953,5,235. 20. A. B. Calder, I. M. Calus, J. D. Chamberlain, J. Dwyer, A, L. Glenn and A. F. H. Ward, Course Manual Postgraduate School in Quantitative Treatment of Experimental Data in Chemistry, U.M.I.S.T., Chemical Society, London, 1973. 21. F. Yates, Design and Analysis ofFactorial Experbnents, Imperial Bnre& of Soil &en&, London, i937. 22. M. A. Sharaf, D. L. Ilhnan and B. R. Kowalski. Che~metrics, Wiley, New York, 1986.
Talmta. Vol. 36, No. 3, pp. 357-362, 1989 Printed in Great Britain. AI1 rights reserved
PERIODATE DETERMINATION BY FIA WITH CHEMILUMINESCENCE EMISSION DETECTION, AND ITS APPLICATION TO ETHYLENE GLYCOL N. P. EVMIRIDIS* Laboratory of Analytical Chemistry, Department of Chemistry, The University, Hull, England (Received 1 April 1987. Revised 30 September 1988. Accepted 27 October 1988) Summary-A method for periodate determination is given which combines the rapidity of flow-injection analysis and the sensitivity of chemiluminescence (CL) detection. It is based on the CL emission generated during oxidation of pyrogallol by periodate, and gives a relative standard deviation of 3% and a detection limit of 350 ng with the instrumentation used. The method has been applied to determination of ethylene glycol, with a detection limit of 0.5 gmole. The accuracy of the method is quite good when the ethylene glycol is oxidized in unbuffered solutions, and the interference- due to fo~aldehyde produced can be halved by prior addition of an appropriate amount of iodate. A throughput of I5 sample@& is possible, and the method is suitable for automation and remote control.
of periodate is almost exclusively related to analysis of organic compounds that have vi&al -CH@H, -CHO, >CO, or -COOH groups, especially polyhydroxy compounds. The consumption of periodate is determined by measurement of the periodate concentration before and after oxidation of the organic compound; titrimetric methodsle3 are usually used. The Miller-F~~~rger method’ is based on addition of iodide and titration of the iodine produced, but has the disadvantage that the iodine may react further with some of the products of oxidation. Some workersc6 have reported doubts of the reliability of the Fleury-Lange method2 for determination of excess of periodate in analysis of carbohydrates. Other workers,“‘O dealing with specific carbohydrate compounds, have found errors due to iodination or further oxidation by iodine formed during the determination procedure, or to incomplete reaction between iodine and arsenite. In addition, all the procedures above are time~ons~ng and not applicable in flow systems. A spectrophotometric method” is also used for the determination of very small quantities of periodate by monitoring of the absorbance at about 222.5 nm, but it suffers serious interference from iodate and the probability of oxidation of HCHO and HCOOH in the light-beam. Ponsi2 employed H-acid as a spectrophotometric reagent for periodate but the method is time-consuming and needs further statistical validation. It has recently been reported*3v’4 that chemiluminescence (CL) is generated during the oxidation Determination
*Present address: Laboratory of Analytical Chemistry, Department of Chemistry, University of Ioannina, Ioannina, Greece.
of pyrogallol with periodate, and this forms the basis of the method reported here. EXPERIMENTAL Reagents All chemicals were of analytical-reagent grade and demineralized distilled water was used throughout. The reagent solutions were prepared as described earlier.” Ethylen~Iyco~ (EG) aqueous solution, 0.5&f. Dissolve 31.04 g of ethylene glycol in 1 litre of water, and dilute further as required for lower concentration solutions. Phosphate buffer, pH 8.0. Dissolve 13.6 g of potassium dihydrogen phosphate in water, adjust the pH with 1M sodium hydroxide and dilute to 1 litre. Apparatus The flow manifold (Fig. 1) and deteztor’5 used have already been described.13 Method A sample mixture of injected into mixed with
(40~1) of periodate excess of periodate the carrier stream of tbe reagent (either
solution or the reaction with ethylene glycol is buffer solution, and then pyrogallol solution or
pyrog~loI-hydroxyla~ne solution). For periodate determination the reagent stream is [pg], = mx], = 1.0 x 10e3M in pH 8.0 phosphate btier. RESULTS AND DISCUSSION
The CL emission generated during oxidation of pyrogallol with periodate is rather weak and relatively insensitive to changes in periodate conoentration, but can be enhanced by the addition of hydroxylamine to the system.13 Optimization
of conditions
The reagent is a btiered solution of pyrogallol (pg) and hydroxylamine (Hx), which are considered to form a condensation product Pg + Hg z$ Pg-Hx 357
358
N. P. &SIIRIDts 1 ml /min
V
1 ml/min
Fig. I. Manifold and instruments used for the mixing and for monitoring the generated CL emission. (SI) carrier stream; (SZ) reagent stream; (V) sampie valve; OK) waste; (C) cell; (A) pump; (3’) photom~tiplier; (R) recorder.
that acts as a sensitizer in the oxidation of Pg with periodate.” However, any free hydroxylamine acts as an inhibitor by reduction of periodate. The effect of the concentration of each reactant on the overall CL emission was studied by varying the concentration of either Pg or Hx (keeping the concentration of the other fixed) at two levels of periodate concentration, 1.0 x 10A3M (low) and 1.0 x IO-*&# (high). Figure 2 shows that for [IO;] = 10W3Mand [Hx] = 10-3M the CL is maximal at pg] = 5 x 10T4A4, and that for [IO,-] = 10V2M and [Hx] = 10W3Mthe maximum is at [pg] = lo-‘M. The decrease in CL with
increase in f_Pg]after the maximum is presumably consequence of an inner filter effect of the Pg.
a
Efict of ~ydroxyiam~~ conee~tratio~ This was examined at two concentration levels of Pg, 5.0 x 10V4 and 1.0 x 10e3M and two periodate concentrations, 1.0 x lo-’ and 1.0 x 10e2M, Fig. 3. At the lower periodate level (lo-‘M) the CL obtained from 5 x 10-4M Pg is maximal at 5 x 10e4M Hx concentration, and is higher than that for 1.0 x 10-‘M Pg, which occurs at 2 x 10-3M I-Ix. This suggests competition between Nx and the Pg-Hx complex for oxidation by periodate, and this is substantiated by the results for the 1.0 x IOmZM periodate system, where the larger excess of periodate can accommodate the competitive reduction by Hx.
500
1500
Analysis of effects and interactions
ooo
500
-
1
2
3
t
I:Pgl x IO3 Fig. 2. Effect of pyrogallol concentration at [Hx]@= 1.0 x IO-‘M. Conditions: pH 8.0; [IO,-], 1.0 x 10-3&I (I), 1.0 x lo-*M (2).
It was obvious that the system was complicated, so a 2j factorial experiment was performed at two levels of each reactant, 1.0 x IO-* and 1 x 10m3A4 periodate, 5.0 x 10W3 and 5.0 x 10-*M Hx and 2.5 x 10m3and 5 x IOm4M Pg. The residual variance found for 12 measurements under identical conditions was S; = 379.5. Table 1 clearly demonstrates that all the effects and interactions of the analytical concentration levels are significant at the 95% level, especially the interaction between Pg and Hx. The next most significant are the effect of periodate and the ternary interaction between all three reactants. Effect of reagent concentrations on periodute calibration curues The concentrations of the reactants at equilibrium were calculated for an analytical Hx concentration of 1.0 x 10e3A4 and analytical Pg concentrations of 5.0 x 10m4, 1.0 x 10V3 and 2.5 x 10W3M, on the basis of reaction (1) and the assumption that Ic; = 1 x l@, and are given in Table 2 together with ~g-Hx~/[Pg~ which represents the ratio of [sensitizer]/[CL generator]. Calibration curves for periodate over the range O-O.OlM, for each of the [pg], [Hx] combinations, are shown in Fig. 4. It is clear that the equimolar mixture
Periodate determination by PIA 1500
4500
(a)
359
(b)
1000 s z c 0 1 2 E al d 500
I 1
u
234567
6
9
101
I
I
I
I
I
I
I
I
2
3
4
5
6
7
6
910
I
CHx3,mM
Fig. 3. Effect of hydroxylamine concentration at (a) [IO; ]o = 1.O x lo-‘M and (b) PO; ]o = 1.O x lo-*M. Conditions: pH 8.0; pg],, 5.0 x 10m4M (l), 1.0 x lo-‘M (2).
of Pg and Hx gives the greatest sensitivity and widest useful range. Optimization
of the sensitivity
The shape of the calibration curve is dependent on both [CL generator] and the ratio [sensitizer]/[CL generator]. Since it is not possible to adjust [pg-Hx], independently of PgL, it is necessary to find the analytical concentrations [Pg], and [Hx], which optimize the CL emission. As an equimolar mixture seemed best (Fig. 4), the effect of varying [pgb (= [Hx],) was examined. Figure 5 shows that
[Hx], = Pg10 = 1.0 x lo-‘A4 gives the best result overall. The crossing of curves 2 and 3 at low periodate concentration is due to the competitive oxidation of Hx by periodate when the [Hx]/[IO;] ratio is high, resulting in less CL. For low periodate concentration (
Table 1. Analysis of significance of effects and interactions (F-test)‘%*’ Effect on CL signal height, mV
Effects and interactions
P’gb
121
Fi]$xl0 [PgkCr
432 335.5 112.5 216 210 300.5
b Wxb[IO,_ b P~bFWo[IOilo
B = mean* square
95% C = S: x F;,
29,282
225,120 373,248 25,312 93,312 88,200 180,600
379.5 x 4.84 = 1836
B/C
Significance
>l
Yes
>l >1 >l >I >1
Yes Yes Yes Yes
*Mean square of each effect or interaction.
Table 2. Calculated equilibrium concentrations Dw& 1.0 1.0 1.0
Er 0.5 1.0 2.5
Fg;--Ileq*
lPg 8 mb
0.22 0.38 0.65
0.28 0.62 1.85
of Pg-Hx, Pg and Hx
y-lg 0.18 0.62 0.35
PHW,,Rpgl, 0.78 0.62 0.35
N. P. EVWRIDIS 1500
r
.9---”
~-1 ;;;pz .
2 1000
5 d
.c u)
.
/
/
d 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0
1.0
0.1
0.2
0.3
0.4
0.5
Fwiodatel
[Periodate] x10* (M)
0.6
0.7
0.9
0.9
1.0
x102(M)
Fig. 4. Periodate calibration curves at constant [Hx]. Conditions: [Hxl, 1.0 x IO-‘M; pH 8.0; [pg], 2.5 x lo-‘M (I), 1.0 x 10-‘&f (2) 5.0 x 10-4M (3).
Fig. 5. Periodate calibration curves at a ratio of [Pgb/[Hx], = 1. Conditions: pH 8.0; [Hxl, 1.0 x 10-4M (l), 5.0 x 10-4M (2), 1.0 x 10-3&f (3).
Application to determination of organic compounds
quantities of ethylene glycol, but lower consumption than expected at the higher quantities. This may be explained by the fact that with high excess of periodate there is oxidation of the formaldehyde formed’61g but with low excess of periodate the oxidation of ethylene glycol reaches completion very slowly.
used in analysis of organic compounds that have vicinal hydroxyl, carbonyl or carboxylic acid groups, which are oxidized to aldehyde, carboxylate or carbon dioxide respectively. The most important group determined by periodate oxidation is that of polyalcohols. Excess of periodate is added and the surplus determined. This can be done by the method proposed here. The application has been tested with ethylene glycol. Periodate
is almost
exclusively
Ethylene glycol oxidation in bu$ered solution (pH 8.0). Ethylene glycol was treated with excess of
Ethylene glycol oxidation in unbuffered solution. The results in Table 4 for oxidation with 5 ml of
2.5 x lo-‘A4 periodate mixed with ethylene glycol show an improvement, but the reaction is still slow to reach completion when the excess of periodate is small. The recommended method is addition of a fixed volume of sample solution, (e.g., 0.05 ml) containing not more than 10 pmole of ethylene glycol, to 5.0 ml of 2.5 x 10-‘&f periodate, and measurement of the periodate excess as already described. A linear calibration graph for the amount of ethylene glycol added is obtained over the range O-10 pmole.
periodate in aqueous solution buffered at pH 8.0 and after the end of the reaction a small volume of the reaction mixture was injected into the FIA manifold for periodate determination. Table 3 gives the results for a series of 5.0-ml volumes of 1.0 x IO-*&f periodate to which 0.05 ml of 0.5M ethylene glycol plus O-0.30 ml of 0.05M ethylene glycol had been added. The residual periodate concentration in each Selectivity sample was calculated and compared with the Oxidation of ethylene glycol by periodate leads value expected from the reaction stoichiometry. to the formation of formaldehyde and iodate: Comparison of columns 3, 5 and 7 of Table 3 shows there is over-consumption of periodate by low HOCH,CH,OH + IO; -, ZHCHO + IO; + H,O Table 3. Excess of periodate, calculated from CL emission after oxidation of ethylene glycol with periodate in pH 8.0 buffer solution Experimental signal peak height 30min after mixing Ethylene glycol added, mole
Calculated periodate excess, mole
[Periodate] excess, mM
CL,
[Periodate] excess,
mV
mM
25 30 35 40 45 50 55
25 20 15 10 5 0 0
5.0 4.0 3.0 2.0 1.0 0 0
850 820 720 630 500 350 200
3.10 3.05 2.70 2.50 1.95 1.45 0.95
60 min after mixing
CL, mV
[Periodate] excess, mM
780 750 670 620 480 325 200
2.80 2.70 2.50 2.30 1.85 1.35 0.95
Periodate determination
by FIA
361
Table 4. Excess of periodate calculated from CL emission after oxidation of ethylene glycol with periodate
in unbuffered aaueous solution Experimental signal peak height Immediately after mixing Ethylene glycol added,
Calculated periodate excess, mole
mole
0.0 2.5 5.0 7.5 10.0 12.5 15.0
[Periodate] excess, FnM
12.5 10.0 7.5
2.5 2.0 1.5 1.0 0.5
:3
8
is a reductant and the iodate an oxidant, so both may interfere in determination of the periodate excess. Table 5 shows the effect of some products from periodate oxidation of polyalcohols. The interference from fo~aldehyde and formic acid is rather high, but can be decreased by the presence of an equimolar amount of iodate.
The formaldehyde
Application
90min after mixing [Periodate]
CL, mV
[Periodate] excess, mM
CL, mY
CXCZSS,
690 555 410 290 175 165 150
2.55 2.10 1.55 1.25 0.85 0.82 0.77
680 530 350 210 60 25 0
2.50 2.02 1.40 1.00 0.52 0.25 0
fluorescent dyestuff) in water, or by diluting the water/antifreeze mixture from a car radiator. The samples were analysed by the Fleury-Lange and CL-emission methods. The results are shown in Table 6. There was good agreement for the pure ethylene glycol solutions, and fairly good agreement for the other samples. No problems were caused by the fluorescent dyestuffs.
of the method to aqueous samples
Sample solutions were prepared by dissolving a known volume of ethylene glycol (analyticalreagent grade) or commercial antifreeze (containing
Conchisions
The method is rapid, and reasonably accurate and reproducible. When the oxidation reaction used in
Table 5. Selectivity
Added, vole Ethylene glycol, 2.75 Formaldehyde, 2.75 Formic acid, 2.75 Acetic acid, 2.75 Formaldehyde + iodate (2.75 + 2.75) Formic acid + iodate (2.75 + 2.75) Table 6. Dete~nation
Relative decrease of CL emission, %
Equivalent amount of ethylene glycol, lunofe
100 13.3 27.7 3.2 6.9 8.2
2.75 0.37 0.76 0.09 0.19 0.23
of ethylene glycol content of aqueous samples Ethylene glycol content, mg Iml
Sample
Fleury-Lange* method
Ethylene glycoi solutiona A B C D Antifreeze solutions BP Glycoshell Cur radiator water One-year old car Six-year old car# Seventeen-year old car$ lr.s.d. 2%. 7r.s.d. 3%. gRadiator content renewed in previous year. #Car radiator with tiny leaks. TAL M/3-B
mM
CL detectiont method
2.79 3.26 3.72 4.06
2.79 3.26 3.63 4.03
3.30 3.35
3.40 3.53
0.336 0.458 0.031
0.372 0.418 0.062
N. P. E-IS
362
the application is fast, the whole determination can be performed in the flow-injection system by merged flow of separate solutions of the sample and the periodate, followed by passage of the mixture through the sampling valve, and injection of a sample into the carrier stream. The system could even be used for continuous monitoring of a process stream. The kinetics of oxidation of a specific polyhydroxycompound with periodate can also be followed fairly easily with approp~ate a~~gement of the stream which leads to the sample loop. Such systems can be automated and are suitable for remote control. Acknowledgements-I thank Professor Alan Townshend for giving me the opportunity to work in the Analytical Laboratory of the Chemistry Department of the University of Hull, and my students N. ~an~~ias and N. Santiris for helping in some of the experimental work.
REFERENCES E,
Milller and 0. Friedberger, Ber., 1902, 35, 2652. P. F. Flemy and J. Lange, J. Pharm. Chim., 1933,17, 107, 196. L. Malaprade, Bull. Sot. Chitn. France, 1928, 43,683: Compt. Rend., 1928, 186, 392. G. Hughes and T, P. Nevell, Trans. Faraday Sot., 1948, 44, 941.
5. P. Fleury, J. E. Courtois and A. Bieder, Bull. Sot. Chim. France, 1952, 118. 6. J. fi. Taylor, J. Am. Chem. Sot., 1953, 75, 3912. 7. J. C. P. -!jchwarz, Chem. Ind. London, 1954, 1000. 8. W. A. Bonner and R. W. Drisko. J. Am. Chem. SW., 1951,?3, 3699. 9. P, Fleury and M. Fatome, J. Phurm. C&m., 1935, 21, 247. 10. G. Lundblad, Ark& Kemi, Min. Geol., 1947,24, No. 25. 11. J. S. Dixon and D. Lipkin, Anal. Chem., 1954,26, 1092. 12. J. A. Pons, Pubt. Inst. Invest. Microquim., Univ. Na Litor& (Rosuriv. Arg.), 1965, 26, 175, 199, 215. 13. N. P. Evmiridis, Art&~, 1987, 112, 825. 14. I&m, ibid., 1988, 113, 1051. 15. J. L. Burguera, A. Townshend and S. Greenfield, Anal. Chim. Actu, 1980, 114,209. 16. L. HOI@ and B. M. Woods, Chem. Ind. London, 1957, 1421. 17. G. Lindstedt, Nature, 1945, 156, 448. 18. D. J. Bell, A. Palmer and A. T. Johns, J. Chem. Sac., 1949, 1536. 19. G. Neumiiller and E. Vasseur, Ark& Kern& 1953,5,235. 20. A. B. Calder, I. M. Calus, J. D. Chamberlain, J. Dwyer, A, L. Glenn and A. F. H. Ward, Course Manual Postgraduate School in Quantitative Treatment of Experimental Data in Chemistry, U.M.I.S.T., Chemical Society, London, 1973. 21. F. Yates, Design and Analysis ofFactorial Experbnents, Imperial Bnre& of Soil &en&, London, i937. 22. M. A. Sharaf, D. L. Ilhnan and B. R. Kowalski. Che~metrics, Wiley, New York, 1986.