The relative standard deviation at the 1.5 pg ml-1 level is. 3.5%. The method was applied satisfactorily to the determination of warfarin in irrigation water.
1003
ANALYST, JULY 1990. VOL. 115
Enhanced Spectrofluorimetric Determination of the Pesticide Warfarin by Means of the Inclusion Complex With P-Cyclodextrin J. C. Marquez, M. Hernandez and F. Garcia Sanchez Department of Analytical Chemistry, Faculty of Sciences, University of Malaga, E-29071 Malaga, Spain
The warfarin - p-cyclodextrin inclusion complex alters the luminescent properties o f warfarin. This feature allows the use of a spectrofluorimetric method for its determination that gives improved analytical performance. The spectral changes associated w ith the inclusion process allowed a value of 160 I mol-1 t o be obtained for the complexation constant. Fluorescence intensity is linearly related t o warfarin concentration for a quantification limit of from 0.2 t o 4 pg ml-1. The relative standard deviation at the 1.5 pg ml-1 level is 3.5%. The method was applied satisfactorily t o the determination of warfarin in irrigation water. Keywords: Warfarin determination; spectrofluorimetry; (3-cyclodextrin inclusion
Warfarin [3-(cx-4-acetonylbenzyl)-4-hydroxycoumarin]has found extensive use both as a rodenticide to help control rat populations and, given its anticoagulant character, in the treatment of thromboembolic disorders. The use of warfarin as a pesticide has led to a requirement for analytical methods for its determination in cases of suspected poisoning as a result of direct ingestion. Spectrophotometric,'.' spectrofluorimetric3-4 and phosphorimetrics-0methods have been used for the determination of warfarin, but these methods usually suffer from a lack of specificity and sensitivity and are not suitable for determining low levels of pesticide. The detection limits were in the range 1-10 vg ml-1. Best results have been obtained by using liquid chromatographic methods (0.02 mg kg-1,7 0.01 mg kg-1,g 0.3 ng ml-19 and 0.1 pg10). Cyclodextrins are cyclic compounds, produced from starch by the cyclodextrintransglycosylase enzyme, containing six, seven or eight glucose units fused tooform a ring with an internal diameter of 5.7, 7.8 or 9.5 A, namely a-,b- and y-cyclodextrin. Cyclodextrins are capable of forming inclusion complexes with compounds having a size compatible with the dimensions of the cavity. The extent of complex formation, quantified by the association constant, depends on the polarity of both the host and guest molecules. However, geometric, rather than chemical, factors are decisive in determining the type of guest molecules which can penetrate into the cyclodextrin cavity. 1 1 Because of the ability of cyclodextrin media to form complexes, luminescence phenomena are enhanced, the molecules introduced into the internal cavity are isolated from the surrounding environment and their excited states shielded from quenching processes. Cyclodextrins in aqueous media have received considerable attention in recent years and their ability to enhance spcetrofluorimetric or room temperature phosphorimetric methods has been exploited in the determination of a wide variety of organic compounds, mainly pesticides and drugs. 12-14 Recently, cyclodextrins have been considered as confined spaces in which the ligand acts as a shielded functionalised assembly which is a selective spectrofluorimetric reagent for the determination of metal ions, or generally, for the determination of a particular analyte. 15-17 In the present study the enhanced fluorescence of the rodenticide warfarin in (3-cyclodextrin media was used to test the feasibility of determining this compound. The proposed method was applied to the determination of warfarin in irrigation water; the limit of detection is 0.059 pg ml-1 and the relative standard deviation (RSD) at the 1.5 pg ml-1 level is 3.5%.
Experimental Apparatus Emission measurements were carried out with a luminescence spectrometer, Perkin-Elmer LS-5, (Perkin-Elmer, Beaconsfield, Buckinghamshire, UK) equipped with a xenon discharge lamp pulsed at line frequency (9.9 W), monochromators, j.73 Monk - Gilleon type, and 1 x 1 cm quartz cells. The spectrofluorimeter was operated in the computer-controlled mode via the data station microcomputer. Instrumental control and data collection were achieved by using the commercially available Perkin-Elmer computerised luminescence software (PECLS-11) . T o compare all measurements and ensure the reproducibility of the experiments, the LS-5 spectrometer was checked daily. A fluorescent polymer sample of p-terphenyl (10-7 M) gave a relative fluorescence intensity (RFI) of 90% at kern,= 340 nm with a kex.of 295 nm, a slit-width of 2.5 nm and a sensitivity factor of 0.5973. For graphical recording, an Epson FX-85 printer - plotter was connected to the spectrofluorimeter . All fluorescence spectra are uncorrected because no significant wavelength shifts were observed when these spectra were compared with corrected spectra. Reagents Stock solutions of warfarin (Aldrich, Milwaukee, MI, USA) were prepared in ethanol at concentrations of 1.0 mg ml-1; aand (3-cyclodextrin (Sigma, St Louis, MO, USA) were recrystallised once from boiling water, 1 X 10-ZM aqueous solutions being used. A buffer solution (pH = 9) was prepared from 0.1 M boric acid and 0.1 M sodium hydroxide solution. Procedures
Study of inclusion phenomena To aliquots of ethanolic warfarin solution, gently evaporated to dryness on a hot-plate, were added increasing volumes of 1 x 10-2 M (3- o r a-cyclodextrin solutions, 2 ml of buffer solution (pH = 9) and de-ionised water to give a final volume of 10 ml. These samples were sonicated for 15 min and their fluorescence spectra measured. P-Cyclodextrin concentration was chosen as the highest that could be obtained within the limitation of its solubility in water (1.2 X M) and the physical volume of the calibrated flask used (10 ml).
ANALYST, JULY 1990, VOL. I15
L
a, c
.-c
a, 150 0
::100 a,
2
2
50
a,
.-r s
-
$
-
o
> B
A
360
410
460
500
Wavelengthhrn
Fig. 1. Fluorescence emission spectra of warfarin in A, water; B , a-CD; and C. (3-CD. [Warfarin] = 1 pg ml-l; [a-CD] and [P-CD] = 8 x 1 0 - 3 ~ pH . = 9.0, Lex, = 310 nm
.-i r
160 c,
.-
a, a, c v)
2 2 L
O
300
350
400
450
500
Wavelengthhrn Fig. 3. 1, Excitation and 2, emission spectra of warfarin - (3-CD complex. [Warfarin] = 1 pg ml-l; and [p-CD] = 8 X 1 0 - 3 ~
.-t v)
4-
:
2
*O
r .a
k
.-rs
.-I
c v)
c)
+? 250
2
.-
s
44
t
$ 20
a, c
v)
c)
150
L
0
-3
c 0
3 100
LL
-m
o)
a
4
6
8
1
0
1
2
Fig. 4. Effect of p H on fluorcscence intensity of warfarin complex. [Warfarin] = 0.7 pg ml-1; and [P-CD] = lo-‘ M
50
a >,
.r
2
PH
2
2
40
a,
‘v, 200
;.
60
O 350
400
450
500
Wavelengthhm
Fig. 2. Influence of the p-CD concentration on thc formation of the inclusion complex. [Warfarin] = 1Fig ml-1; [(3-CD]: A, 0; B, 1 x 10-3; C. 3 X 10-3; D, 5 X 10-3; and E , 8 x lo-- M
Spectrofluorimetric determination of warfarin Ethanolic aliquots of samples, containing 2, 5 , 10,20 or 40 pg of warfarin, were placed in 10-ml calibrated flasks. The samples were slowly evaporated and then 8 ml of 1 x 10-2 M 6-cyclodextrin solution and 2 ml of the buffer solution were added. The fluorescence intensity was measured at 386 nm, with excitation at 310 nm, against a reagent blank. Determination of warfarin in irrigation water The water samples were filtered through Whatman No. 5 paper (pore size 2.5 pm) in order to eliminate the suspended organic matter. T o aliquots of samples, a mass of solid P-cyclodextrin (final concentration 8 X 10-3 M) and 2 ml of a buffer solution (pH = 9) were added. These samples were sonicated for 15 min and their fluorescence intensities measured as previously described.
Results and Discussion Study of Inclusion Processes Warfarin was found to form an inclusion complex with cyclodextrin. In Fig. 1 the fluorescence emission spectra of warfarin in water, a-, and 6-cyclodextrin solutions are presented; the relative fluorescence intensity is more significant in (J-cyclodextrin solutions because the enhanced fluorescence emission in (J-cyclodextrin media involves the forma-
-
(3-CD
tion of an inclusion complex between the cyclodextrin and warfarin. Quantitative data can be obtained from the enhanced fluorescence emission which occurs as the fl-cyclodextrin concentration increases (Fig. 2). We attribute this effect to a diminution in rotational freedom in the medium of the hydrophobic cavity, where one part of the molecule of warfarin is protected from collisions, consequent on insertion of the guest into the host. The binding constant of warfarin with fLcyclodextrin was obtained by the expression described previously18
A P = (a[~arfarin](~k[(J-CD])-~ + (a[~arfarin](~)-l where CD is cyclodextrin, AF is the change of fluorescence intensity upon addition of 6-CD, [warfarin],, is the initial concentration of warfarin molecules and a is the proportionality constant. The linear relationship between AF-1 and [f3-CD]-l gives k , the host - guest association constant. The value obtained is 160 I mol-1 (at 25°C). This apparent stability constant is in agreement with the result (148.9 1 mol-1) obtained by Lin19 from a solubility data and membrane permeation study. Effect of Experimental Variables The fluorescence excitation and emission spectra of warfarin (1 pg ml-1) in a p-cyclodextrin medium, at p H 9.0, are shown in Fig. 3. Maxima are found at 310 and 386 nm for excitation and emission, respectively. Studies on the effect of p H on the fluorescence intensity show that the fluorescence is constant for p H values between 6.8 and 11 (Fig. 4). The p H can be maintained at 9.0 by the addition of 2 ml of the buffer solution. The complex prepared as described was stable for at least 2 h. The order of addition of the reagents is immaterial.
1005
ANALYST, JULY 1990, VOL. 115
Application to Water Samples
Table 1. Analytical parameters
Analytical sensitivity, SA = s,/m (vg ml-l) cL(k = 3)/pg ml.. .. . . .. ~ ( k 10)/pgml-l = .. .. .. .. Linear dynamic range/pg ml-1 .. .. RSD,*% . . . . . . . . . . . .
* At the 1.5 pg ml-l
.. ..
.. ..
..
0.052 0.059 0.2 0.2-4 3.5
level.
Table 2. Determination of warfarin in irrigation water
Irrigation water sample Synthetic* . . It .. .. 2 .. .. .. 3 . . .. ..
.. .. .. ..
Warfarin spiked pg ml-1 2 1 2 4
Warfarin found (SD)/ pg ml-1 1.91 (0.10) 0.95 (0.07) 2.04 (0.02) 4.20 (0.11)
* Nine determinations. t Samples 1, 2 and 3 (three determinations). Analytical Parameters The calibration graph was linear over the range 0.2-4 pg ml-1. The calibration graph obtained by the least-squares treatment is described by:
If = 97.5 [warfarin]
+ 18, r = 1
where If is the fluorescence intensity, r the correlation coefficient and [warfarin] is in pg ml-1. The sensitivity of the method is reported as the analytical sensitivity, SA = s,/m [s, is the standard deviation of the analytical signal (seven measurements) and m is the slope of the calibration graph]. The limit of detection, cL ( k = 3) and the limit of quantification, CQ ( k = 10) are reported as defined by IUPAC,20 where k is a numerical factor chosen in accordance with the confidence level desired. The limit of quantification, cQ, is employed to establish the lower limit of the linear dynamic range. These analytical parameters are given in Table 1. The method shows good reproducibility with an RSD of 3.5% at the 1.5 pg ml- 1 level (seven determinations). The proposed method compares favourably with spectrophotometric methods in terms of sensitivity.21.22 Detection limits of spectrofluorimetric3 and phosphorimetrics methods are at about the 1 pg ml-1 level. An additional advantage of the proposed method is its ability to determine warfarin in aqueous media without previous extraction processes, owing to the solubility of (3-cyclodextrin compared with moderately apolar compounds.15 Interference arising from warfarin isomers ( R and S) and metabolites has not been checked.
The (3-cyclodextrin procedure was applied to the determination of warfarin added to irrigation water samples. Table 2 summarises the results, from which it can be concluded that acceptable accuracy and precision can be obtained in the analysis of samples containing 2 pg ml-1 of warfarin. The authors thank The Direccion General de Investigacion Cientifica y Tkcnica for supporting this study (project number PB086-0247).
References 1. Fishwick, F. B., and Taylor, A., Analyst, 1967, 92, 192. 2. Welling, P. G., Lee, K. P., Khanna, V., and Wagner, J. C., J. Pharm. Sci., 1970, 59, 162. 3. Corn, M., and Berberich, R., Clin. Chem. (Winston-Salem, N C ) , 1967, 13, 126. 4. Hollifield, H. C., and Winefordner, J. D., Talanta, 1967, 14, 103. 5. Su, S. Y., Asafu-Adjaye, E., and Ocak, S., Analyst, 1984,109, 1019. 6. Vanelli, J. J., and Schulman, E. M., Anal. Chem., 1984, 56, 1030. 7. Hunter, K., J. Chromatogr., 1983, 270, 255. 8. Hunter, F. S., J. Chromatogr., 1985,321, 255. 9. Steyn, J. M., and van der Merwe, H. M., J. Chromatogr., 1986, 378, 258. 10. Banfield, C., and Rowland, M., J. Pharm. Sci., 1983,72,921. 11. Szejtli, J., “Cyclodextrins and Their Inclusion Complexes,” Akademiai Kiado, Budapest, 1982. 12. Scypinski, S., and Cline Love, L. J., Anal. Chem., 1984, 56, 322. 13. Cline Love, L. J., Grayeski, M. L., and Noroski, J., Anal. Chim. Acta, 1985, 170, 3. 14. Alak, A. M., and Vo-Dinh, T., Anal. Chem., 1988, 60,596. 15. Garcia Sanchez, F., Hernandez Lopez, M., and Heredia, A., Anal. Chim. Acta, 1986, 187, 147. 16. Garcia Sanchez, F., Hernandez Lopez, M., and Marquez Gomez, J. C., Analyst, 1987, 112, 1037. 17. Garcia Sanchez, F., Hernandez Lopez, M., and Marquez Gomez, J. C., Fresenius 2. Anal. Chern., 1987,328, 499. 18. Hoshino, M., Imamura, M., Ikehara, K., and Hama, Y., J. Phys. Chem., 1981, 85, 1820. 19. Lin, S. Y., and Yang, J. C., Pharrn. Weekbl., Sci. Ed., 1986,8, 223. 20. Nomenclature, Symbols, Units and Their Usage in Spectrochemical Analysis 11, Spectrochirn. Acta, Part B , 1978,33,242. 21. Koget, T. O., apd Tsisar, J., Farm. Zh. (Kiev), 1972, 27,80. 22. Giuran, V., Rom. 62197 (Cl. GOlN21/00), 20 May 1977, Appl. 74590, 24 April 1973.
Paper 9/01816F Received May 2nd, 1989 Accepted February 26th, I990
