ANALYST, FEBRUARY 1990, VOL. 115 217

Flow Injection Determination of Thiamine Based on Thiochrome by Mercury(l1)

Carmen Martinez-Lozano, Tomas Perez-Ruiz, Virginia Tomas and Concepcion Deparment of Analytical Chemistry, Faculty of Sciences, University of Murcia,

its Oxidation to

Abellan Murcia, Spain

The reaction involving the oxidation of thiamine to fluorescent thiochrome by mercury(l1) in basic solutions has been adapted to flow injection. Linear calibration graphs were obtained between 2 x 10-7 and 7 x 10-6 M,

with a sampling rate of 22 samples h-1 and a relative standard deviation of 0.42-0.17%. The applicability of the method to the determination of thiamine in vitamin - mineral preparations was demonstrated by investigating the effect of potential interferences and by the analysis of commercial preparations.

Keywords : Th jam in e determination; flow injection; p ha rmaceu tica I application

Thiamine (vitamin €3,) is a natural nutrient in many foods and is also added as an essential nutrient. Since its discovery and isolation, there have been numerous reports in the literature on the determination of thiamine by microbial, chromato- graphic, direct molecular absorption, spectrophotometric, spectrofluorirnetric, polarographic and reaction-rate met hods. 1-3

The most widely used method for the assay of thiamine is the so-called thiochrome method which involves the reaction between the vitamin and hexacyanoferrate( 111) in alkaline solution and the extraction of thiochrome (TC) from the aqueous phase into an organic phase. The amount of potassium hexacyanoferrate(l1T) must be sufficient to oxidise the thiamine; however, a large excess is undesirable as it may result in the decomposition of TC.2 The yield of the reaction is about 67%.-' The extraction of TC is necessary in order to separate it from other fluorescent substances and from potassium hexacyanoferrate(Tl1) which quenches the TC fluorescence .5

The oxidation of thiamine to fluorescent TC is always accompanied by the simultaneous formation of thiamine disulphide (TDS), a condensation product of two thiamine molecules. The ratio of TC to TDS is affected by the pH," solvent7 and oxidising agent .X Apart from potassium hexa- cyanoferrate(III), many other oxidising agents have been used to oxidise thiamine, such as KMn04 and Mn02,Y cyanogen bromide,1° and H202, I2 and HgII."-ls Oxidising agents such as KMn04, H 2 0 2 and I2 favour the production of non- fluorescent TDS.8 Hexacyanoferrate(III), in the absence of solvent extraction, is not an appropriate oxidising agent [3 X 10-3 M K3Fe(CN), reduced the TC fluorescence signal by more than 7.5%15]. O n the other hand, HglI has been used to oxidise thiamine to TCll.12,15 and it has been found that the quenching of the T C fluorescence does not occur and that the reaction provides linear results over a much wider thiamine concentration range.

The reaction time is an important parameter in the TC fluorescence method. The adaptation of this method to a continuous flow system appears, therefore, feasible as the reaction time could then be kcpt constant. Consequently, automated methods have been proposedl"17 and have gained acceptance in some laboratories.

The TC fluorescence method has also been applied to flow injection (Fl). The procedure involves the reaction between thiamine and hcxacyanoferrate(ll1) in alkaline solution; the aqueous phase is segmented by an organic phase introduced via a specially constructed segmentor. 18

The electrochemical oxidation of thiamine to T C has also been applied to FI. A flow-through electrolytic cell, consist- ing of a vitreous-carbon tube as the working electrode and a vitreous-carbon rod as the counter electrode, was used. l 9

This paper describes a procedure for the determination of

0.87 ml rnin-'

Thiamine

300 cm Detector HSC12

I Buffer

Fig. 1. values of the variables. V, = Sample volume; 1 = coil length

FI manifold for the determination of thiamine with optimum

thiamine, involving a simple FI technique and no solvent extraction; it is based on the oxidation of the vitamin to TC by Hg". The results of an extensive interference study and the application of the method to the determination of thiamine in multi-vitamin - mineral preparations are reported.

Experimental Reagents

All chemicals used were of analytical-reagent grade and doubly distilled water was used throughout.

Thiamine stock solution, 10-3 M. Dissolve thiamine, previ- ously dried, in water and adjust to pH 4 with HCI; this solution was stable for 3 months if kept refrigerated. Working standard solutions were prepared daily from the stock solution by appropriate dilution with water.

Mercury(Z1) solutions. Prepared from HgClz and adjusted to pH 4 with HCI.

Phosphate buffers. Prepared from 0.2 M disodium or dipotassium hydrogen phosphate and sufficient 5 M potassium hydroxide to give the desired pH.

The substances for the interference study were dissolved in water adjusted to pH 4 with HCI.

Apparatus

A Perkin-Elmer Model 3000 spectrofluorimeter equipped with an Omniscribe recorder, a Perkin-Elmer chromato- graphic flow cell, a Gilson Minipuls HP4 peristaltic pump and an Omnifit injection valve was used. P'TFE tubing of 0.5 mm i.d. was used for the mixing coil and for all connections.

Manifold The manifold employed is shown in Fig. 1. The sample solution is introduced with the aid of an injection valve having a 17.5-p.1 loop. The Hg" and sample solutions are mixed with the 0.2 M phosphate buffer solution. The reaction coil, R, is submerged in a water-bath, the tempcrature of which is

Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by Y

ork

Uni

vers

ity o

n 08

/10/

2013

22:

44:4

6.

View Article Online / Journal Homepage / Table of Contents for this issue

21 8

A

ANALYST, FEBRUARY 1990, VOL. 115

2.2

adjustable. A timer, synchronised to the injection system, allows the flow to be stopped for a given period of time.

Recommended Procedure

A 5 x 10-3 M HglI solution and 0.2 M phosphate buffer (pH 12.5) are each pumped at a rate of 0.87 ml min-1. A 175-pl sample containing between 2 x 10-7 and 7 x 10-6 M thiamine is injected into the FI manifold. The timer is programmed so that after 30 s the flow stops for 1 min and then the pump starts again. The TC fluorescence (hey. = 370 nm, kern. = 465 nm) is measured and recorded.

Preparation of Assay Solutions

Ta h lets Ten tablets are powdered and an amount equivalent to 1 mg of thiamine is weighed accurately and dissolved in ca. 100 ml of water. The solution is filtered through Whatman No. 1 filter-paper and the filtrate and two washings, each of 20 ml, are collected in a 1000-ml calibrated flask and diluted to volume with water.

Injections An injection vial solution is shaken thoroughly and a volume equivalent t o 1 mg of thiamine is transferred into a 1000-mi calibrated flask and diluted to volume with water.

Results and Discussion When thiamine is oxidised by HgJI in a FI system the largest yield of T C is obtained at a high Hg" concentration and in an alkaline medium. The time taken to reach equilibrium decreases with increasing Hgtl and OH-- concentrations. As the reaction rate is pH dependent, the reaction mixture must be well buffered (Table 1). The phosphate buffer was selected because it provided a good buffering capacity at pH 12-13 and good precision in the measurements.

The formation of a precipitate in an FI system with speetrofluorimetric detection is unacceptable because of the lack of reproducibility of thc measurements. Therefore, it was necessary to study the solubility of HglI at alkaline pH. It was found that with 5 x 10-3 M Hg" and 0.2 M phosphate buffer, precipitation did not occur after 8 min in the pH range 11-13.5.

Table 1. Influence of HgI' concentration and p H on peak height

PH Peak heightkm 12.5 12.5 12.5 12.5 11.0 I L.5 12.0 12.5 13.0

5.6 9.2

11.4 13.0 3.3 6.2 7.1

11.4 15.2

The order of addition of the reagents is critical. If the thiamine is mixed with the base first, a low yield of T C is obtained on addition of Hg". The addition of a large excess of Hg" to the thiamine solution apparently stabilises the latter by the formation of a complex which can then be oxidised to TC when the basc is added.

The peak height increases with increasing temperature. A temperature of 30 "C was found to be suitable.

Flow Injection Variables

Volume of sample The intensity of the fluorescence signal is a function of the volume of sample injected. The graph is almost linear up to a 125-pl sample volume, after which it deviates so that above 175 pi the readings are almost independent of the sample volume. A 175-u1 sample volume was selected as the optimum in terms of sensitivity and reproducibility.

Influence of length of reactor coil and flow-rate It has been shown that, to a first approximation, for a first- or pseudo-first-order chemical reaction, the reaction and disper- sion processes are independent of each other and that an

Time - Fig. 2. Calibration of the FI system for duplicate injections o f 175 pl of thiamine standard solutions. The values above the peaks are concentrations in pg ml-1

Table 3. Tolerance of the proposed procedure to other substances

Substance Tolerable mole ratio"

Lactose, sucrose, fructose, starch, acetate, sulphate, nitrate, perchlorate, Mg", Call, Zn", Co". Cu", Ni". Ball, Sr1I . . . . . . . . . . . . . . . .

Biotin (vitamin H ) , cyanocobalamin (vitamin BIZ), calcium pantothenate, pyridoxine (vitamin Bh) . .

Mn'1 . . . . . . . . . . . . . . . . . . Nicotinamidc . . . . . . . . . . . . . . . . Ascorbic acid (vitamin C), Fe". Vv . . . . . . . . Fe"1 . . . . . . . . . . . . . . . . . . MoV1 . . . . . . . . . . . . . . . . . . Bromide . . . . . . . . . . . . . . . . . . Iodide . . . . . . . . . . . . . . . . . .

lO00-l

100 100 50 10 5 I 0.1 0.05

* Smallest amount that caused an error of 2% in the determination

i Maximum tested. of 2.5 X M thiamine.

Table 2. Peak height and sampling rate as a function of rcaction coil length, flow-rate and stop-time (thiamine concentration. 2 x 10-" M )

Reaction coil

3 x

1 1 1 3 1 5 2 3 3 3 4 3

Stop-timelmin length/m -

-

Flow-rate/ ml min-I

0.8 1.2 0.8 0.8 1.2 0.8 0.8 0.8

Sampling rate/ Peak height/cm samples h-'

6.2 34 8.7 26 9.0 26

11.4 22 10.9 21 14.2 16 18.0 12 19.2 10

Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by Y

ork

Uni

vers

ity o

n 08

/10/

2013

22:

44:4

6.

View Article Online

ANALYST, FEBRUARY 1990, VOL. 115 219

Table 4. Determination of thiamine in pharmaceutical preparations

Sample*

Thiamine foundrng

Thiamine A O A C manual Source eontent/mg FI methodt

Betabion . . . . . . Merck 100 98.8 IfI 0.04 98.9 k 0.05

Renerva . . . . . . Roche 100 98.2 k 0.05 98.5 k 0.06 Hepacomplet . . . . Reig Jofre 10 9.8 k 0.04 9.7 k 0.03 Vitallon3B . . . . . . Merck 300 302 k 0.02 301 k 0.03

Becozyme . . . . . . Roche 10 9.9 k 0.06 9.9 k 0.04

Compositon of samples. Betabion: thiamine. 100 mg; and water, 2 g. Becozyme: thiamine, 10 mg; riboflavine. 4 mg; pyridoxine, 4 mg; nicotinamide. 40 mg; panthenol, 6 mg; cyanocobalamin, X pg; biotin, 0.5 mg; and water, 2 g. Benerva: thiamine, 100 mg; and water, I g. Hcpacomplct: thiamine, 10 mg; riboflavine, 0.2 mg; pyridoxine, 10 mg; nicotinamide. 1 0 mg; adenine, 10 mg; cobalamin, I mg; hepatic extract. 125 mg: and mater, 2 g. Vitallon 3R: thiamine, 300 mg; pyridoxine, 150 mg; cobalamin. 15 mg; and L-glutamine, 100 mg.

+ Average o f four determinations. $ Avcrage of two determinations.

Table 5. Recovery experiments for thiamine added to samplc solutions of commercial formulations

Thiamindpg ml-1

Initially Recovery. Formulation present Added Recovered %

Bcco7yme . , 1 1 1.96 98.0 2 2.95 98.3 3 4.06 101 .5

Renerva . . . . 1 1 2.01 100.5 2 2.98 99.3 3 3.95 98.7

Hepacomplet . . 2 2 4.03 100.7 3 5.05 101.0 4 5.92 98.6

optimum relationship between flow-rate and reaction coil length can be established in order to obtain the maximum analytical signal.”).21 For selection of the working conditions, 175 pi of a 2 X 10-6 M thiamine solution were injected into reaction coils of different lengths. The flow-rate was varied and the peak height and residence time were measured. The peak height increased with increasing residence time. The best results were obtained if the pump was stopped when the sample plug was located in the reaction coil, R. After a given period of time the pump was triggered by a timer and the sample zone was directed towards the detector. Table 2 summarises the results. A coil length of3 m, a flow-rate of 0.87 ml min-1 and a stop-time of 1 min were chosen as a compromise between analytical signal and sampling rate.

Determination of Thiamine

A series of standard solutions was injected in duplicate to test the linearity of the calibration graph. The results obtained by the injection of 175 pl of 2 X 10-7-7 X 1 0 k h M thiamine solutions are shown in Fig. 2 (conditions as in Fig. 1). The calibration graph is linear from 2 x 10-7 to 7 x 10-6 M with a correlation coefficient of 0.9993. The sampling rate is about 22 samples h-1. The repeatability o f the method is good; a 4.5 x 1 0 - 6 M standard solution of thiamine was injected 11 times (eight injections are shown in Fig. 2) and a relative standard deviation o f 0.2% was obtained. The limit of detection, defined as the thiamine concentration for which the fluores- cence intensity was three times the standard deviation of the blank,” is 3.4 x 10-8 M and the quantification limit is 1.1 x 10- M .

Interferences An extensive interference study, aimed at the determination o f thiamine in vitamin - mineral preparations, was performed.

Samples containing a fixed concentration of thiamine (2.5 x lo-” M) and various concentrations of the foreign substance were injected into the FI system. A substance was considered not to interfere if the variation in the peak height of thiamine was less than 2% in its presence. The results are shown in Table 3.

Applications

The method was applied to the determination of thiamine in multi-vitamin preparations. Table 4 presents results from the routine analysis of several different vitamin preparations. The results are in good agreement with those obtained with the manual Association of Official Analytical Chemists (AOAC) method,23 which involves the oxidation of the vitamin to T C by hexacyanoferrate(II1) in alkaline solution, the extraction of T C from the aqueous phase into butan-2-01 and the measure- ment of the fluorescence of the organic phase.

The accuracy of the proposed method was tested by performing recovery experiments on solutions prepared from thiamine formulations. A mean recovery of 99.6% was obtained (range 98-101.5%) (Table 5 ) .

Conclusions The oxidation of thiamine to T C in a flow system appears to be simpler than the electrochemical oxidation because it is not necessary to use a special electrolytic flow cell. However, the main advantage of the oxidation with Hg” is that it is not necessary to use a segmentor to extract the TC; this is cssential when hexacyanoferrate(TT1) is used. 18

A comparison between the proposed F1 method and the manual AOAC method indicates that the former has the advantages of a good sampling rate, simplicity and automa- tion. The FI method also gives results that are in good agreement with the expected values when applied to the determination of thiamine in multi-vitamin preparations.

We acknowledge the financial support given by DGICYT, project PB87-0053.

1.

2.

3 .

4.

5 .

References Strochccker, R.. and Henning, H . M. , “Vitamin Assay Tested Methods.” Verlag Chemic. Weinheim, 1965. “Methods of Vitamin Assay,” Association of Vitamin Chem- ists. Interscience, New York, 1066. Davidson, 1. E . . in Smyth, W. F., Editor, “Polarography of Molecules of Biological Significance,” Academic Press, London. 1979. Nichelson, 0.. and Yamamoto. R . S., in Glick, D., Editor, “Methods o f Biochemical Analysis,” Volume 6, Interscience, New York, 1958. Ellinger, P.. and Holdcn, M., Biochrm. J., 1944, 38, 147.

Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by Y

ork

Uni

vers

ity o

n 08

/10/

2013

22:

44:4

6.

View Article Online

220 ANALYST, FEBRUARY 1990, VOL. 115

6.

7. 8. 9.

10. 11. 12.

13.

14. IS.

16.

17.

Kawasaki. C., in Harris, R. S. , Editor, “Vitamins and Hormones,” Academic Press, New York, 1963. Wostman. B. S . , and Knight, P. L . . Expericxtia, 1960, 16,500. Syke5, P.. and Todd. A . R., J . Chem. SOC. , 1951, 534. Dyke, S. F . , “The Chemistry o f the Vitamins,” Intcrscicncc, New York, 1965. Fujiwara, M., and Matsui, K. , Anal. Chem., 195.3, 25, 810. Holman, W., Riocliem. J . , 1944, 38. 388. Edwin, E. E., Jackman, R., and Herbert, N., Analyyt, 1975. 100, 689. Morita, M., Kanaya, T . , and Minesita, T., J . Viturninol., 1969. 15, 116. Prokhovnik, S. J . , Analyst, 1952. 77, 257. Ryan, M. A.. and Ingle, J . D.. Jr., Anal. Chem., 1980, 52, 2177. Khoury. A . J . , “Automation in Analytical Chemistry,” Techni- con Symposia 1966, Mediad, White Plains, New York, 1967. Kirk, J . R., J . Assoc. Off. Anal. Chem.. 1974, 57. 1081.

18. Karlberg, €3.. and Thclander, S., Anal. C’him. A d a , 1980, 114, 129.

19. Kusube, A . , Hiroshima, O., lshiguro, Y., Ishikawa. S., and Hoshida, H.. Chern. P h r r n . Bidl.. 1983, 31, 3589.

20. Painton, C. C . , and Mottola, H. A , , Anal. Chim. Acta, 1984, 158, 67.

21. Shih, J . T., and Carr, P. W.. Anul. Chirn A m , 19x5. 167, 137. 22. ACS Committee on Environmental Improvement, Anal.

Clierrz.. 1980. 52, 2242. 23. “Official Method of Analysis of the Association of Official

Analytical Chemists.” Thirteenth Edition. Association of Official Analytical Chemists. Washington, DC, 1980.

Paper 9l01693G Received April 24th, I989

Accepted September lath, 1989

Publ

ishe

d on

01

Janu

ary

1990

. Dow

nloa

ded

by Y

ork

Uni

vers

ity o

n 08

/10/

2013

22:

44:4

6.

View Article Online