摘要:

Analyst, March, 1979, Vol. 104, pp. 241-247 241 Investigations on Reaction Mechanisms in the Determination of Non-ionic Surfactants in Waters as Potassium Picrate Active Substances L. Favretto, B. Stancher and F. Tunis Istituto d i Merceologia, UniversitiE d i Trieste, 34100 Trieste, Italy The two-phase extraction and spectrophotometric determination of poly- oxyethylene non-ionic surfactants in water at trace levels is examined in detail by considering both monodisperse and polydisperse surfactants of the type RO(CH,CH,O),H, where R = p-tert-nonylphenyl and n is the degree of polymerisation. Potassium picrate is used as a reagent for the polyoxyethylene chain and 1,2-dichloroethane as an extracting phase. Monodisperse surfactants with n from 4 to 15 were isolated by liquid - solid absorption chromatography.Their purity was checked by temperature- programmed gas - liquid chromatography. Their reactivity to the reagent is explained qualitatively by considering the equilibria involved in the extrac- tion. Polydisperse surfactants with ;ii (number-average degree of polymerisation) ranging from 3.3 to 21.5 are also considered and compared with other poly- disperse surfactants in which R = dodecyl. The concentration of these non- ionics in waters is conveniently expressed as potassium picrate active sub- stances (PPAS). It can be referred to the standard synthetic monodisperse surfactant RO(CH,CH,O),H, where R = dodecyl, which gives a spectrophoto- metric response acceptably near to that of the examined series of commercial surfactants. Keywords : Polyoxyethylene alkylphenyl ether non-ionic surfactant trace determination ; water analysis ; spectvophotometry ; potassium picrate ; reaction mechanism Potassium picrate has recently been suggested as a sensitive reagent for the extraction and spectrophotometric determination of polyoxyethylene non-ionic surfactants in water a t trace levels (less than 1 mg kgl).l The method is based on the coordination reaction between the polyether chain and the potassium ati ion.^,^ The cation complex is extracted into 1,2-dichloroethane as an ion pair with picrate, which is the anionic chromophore measured in the spectrophotometric determination at 378 nm.4 Distinct advantages of the picrate anion over the classical tetrathioisocyanatocobaltate(I1) at 620 nm are the higher molecular absorbance and the greater stability with time and towards the extraction solvent .5 Although the maximum at 321 nm in 1,2-dichloroethane at 20 "C of the tetrathioiso- cyanatocobaltate(I1) (log E = 4.06) is higher than that at 620 nm (log E = 3.25), absorbance measurements at 321 nm are less accurate than those at 620nm, because ethylene thio- cyanate is also extracted into the organic phase and it has a broad absorbance maximum at 240 nm, which causes appreciable blanks in the ultraviolet region.Pre-extraction of non-ionic surfactants from water with 1,2-dichloroethane and purifica- tion of the organic extract are essential steps in the trace analysis of actual samp1es.l The polyoxyethylene non-ionic surfactants are then determined on an aqueous solution of the residue of the extract with the potassium picrate reagent.The most common commercial polyoxyethylene non-ionics are polydisperse mixtures having the formula RO(CH,CH,O)i;H, where R is an alkyl or alkylaryl group and $i the number-average degree of polperisation. Only if R is known and iz is directly evaluable in some way (for instance, by gas-liquid Chromatography, at least up to a certain value of .ti), is an absolute determination of the surfact ant concentration in the aqueous phase possible. Otherwise the concentration of potassium picrate active substances (PPAS) is conveniently referred to a standard mono- disperse surfactant .l In this paper, the determination of monodisperse (4<rt< 15) and polydisperse (3.3GFiig21.5) surfactants in which R = +-tert-nonylphenyl is discussed, with the aim of determining the reactivity toward potassium picrate and the factors controlling this effect.242 Apparatus Unicam SP500 and Perkin-Elmer 402 (with hix scale expansion) ultraviolet - visible light spectrophotometers were used with matched silica cells of various path lengths for the measurement of the absorbance of the solutions.The pH values of aqueous solutions were determined with a Beckman 4500 digital pH meter using Fisher ACS certified buffer solutions. Monodisperse surfactants (4 <n< 15) were isolatqd by preparative gradient-elution adsorption chromatography with an LKB Ultrorac 7000 fraction collector at 20 "C. By means of a glass column of 2 cm id., 50-100 mg of surfactant at 70-80y0 purity was obtained and this was further purified by repeated chromatography in a 1-cm i.d.column. Silica gel (AR grade, Mallinckrodt, St. Louis, Mo.), 30-'50 pm, activated at 190 "C, was used as an adsorbent, packed to a length of 15 cm in the first column and 30 cm in the second. Elution was carried out with mixtures of propan-2-one in dichloromethane of composition given by the expressions P = 0.083V (first column) and P = 0.040V (second column), where P is the concentration of propan-2-one (yo V/V at 20 "C before mixing) and V ml is the volume of the eluting agent at the top of the column. The composition of 5-ml fractions was systematically checked by gas - liquid chromatography after evaporation of solvent under nitrogen. A Pye Unicam 104 double-column chromatograph with flame-ionisation detectors was used.Pyrex glass columns (35 cm x 1.7 mm i.d.) were packed with 80-100-mesh Gas- Chrom Q coated with 3% m/m OV-1 (Applield Science Laboratories, State College, Pa.). At temperatures above 300 O C , this phase gives a better performance than GE SE-30, which is extensively used for high-temperature gas - liquid chromatographic analyses.6 Fraction- ation was carried out with a linear temperature programme of 10 "C min-1 from 100 to 370 "C and with injector and detector temperatures of 400 "C. Flow-rates of the gases were nitrogen 45, hydrogen 45 and air 400 ml min-l. Peak areas were determined by means of a Hewlett-Packard 3380A integrator. FAVRETTO et al. : REACTION MECHANISMS IN THE Experimental Analyst, Vol.104 Reagents 1,2-Dichloroethane. E. Merck, extra-pure grade, freshly distilled. Potassium nitrate solution, 2 . 5 0 ~ . Potassium +hate solution, 0.020 M. In a 100-ml calibrated flask, dissolve 25.28 g of potassium nitrate (AnalaR grade) in water and dilute to the mark with water. Dissolve 0.534 g of potassium picrate [re-crystallised from an aqueous solution and dried with phosphorus(V) oxide] in water in a 100-ml calibrated flask. Monodisperse surfactants Monodisperse surfactants with 4 <n < 15 were separated from commercial polydisperse materials having a number-average degree of polymerisation of Ei = 5.4, 6.5, 7.5 and 8.6. The gas - liquid chromatographic purities of the surfactants are shown in Table I. Aqueous stock solutions containing 1-20 mg I-1 (depending on solubility of surfactants) were prepared weekly by weighing the compounds [dried with phosphorus(V) oxide] on a Mettler M5 microbalance.The 1 mg 1-1 solution was prepared daily and was used immedi- ately. Polydis9erse surfactants Stock solutions of polydisperse surfactants were prepared by following the above pro- cedure from commercial products (Chemische Werke Huls) with f i = 3.3, 5.4, 6.5, 7.5, 8.6, 9.7, 10.8, 12.9, 15.0 and 21.5. All of the surfactants were previously dried at 80 "C under vacuum (1 mmHg) for 4 h. The value of fi was checked by vapour-pressure osmometry at 37 "C in 1,2-dichloroethane with a Hewlett-Packard Mechrolab vapour pressure osmometer. Procedures Spectrophotometric calibration graphs In a 50-ml calibrated flask 20.00 ml of potassium nitrate solution were added to an aliquot of the standard solution of surfactant.The aliquots were chosen in order to have 0.10-1.00 mg 1-1 Calibration graphs were obtained at 20 "C by means of the following procedure.March, 1979 DETERMINATION OF NON-IONIC SURFACTANTS IN WATERS TABLE I GAS - LIQUID CHROMATOGRAPHIC PURITY OF THE MONODISPERSE SURFACTANTS RO(CHZCH2O)nH (R = P-tert-NONYLPHENYL) The purity of the surfactants was calculated from measurements of the peak areas. Amounts of surfactants* n Purity, % present as impurities, yo 4 98.05 1.35 (3), 0.60 (6) 5 98.75 1.05 (4), 0.20 (6) 6 99.82 0.18 (5) 7 97.50 0.98 (6), 1.52 (8) 8 98.99 0.77 (7), 0.24 (9) 9 98.24 0.91 (8), 0.85 (10) 10 96.32 2.81 (9), 0.87 (11) 11 98.38 1.62 (10) 12 97.0 2.8 (ll), 0.2 (13) 13 98.6 0.5 (ll), 0.9 (12) 14 97.8 1.6 (13), 0.6 (15) 15 98.5 0.3 (13), 0.4 (14), 0.8 (16) * n values of impurities are given in parentheses. 243 of surfactant in the final 50-ml volume of aqueous phase.After mixing and allowing to stand for 1 h, 5.00 ml of the potassium picrate solution were added and the volume was made up to the mark with water. The mixed solution (pH approximately 7) was allowed to stand for 15 min and then transferred into a separating funnel with a PTFE stop-cock. A 5.00-ml volume of 1,2-dichloroethane was added and the mixture was shaken for 5min. The organic layer was transferred dropwise into a conical centrifuge tube fitted with a polyethylene stop-cock and centrifuged at 1000 g for 5 min. The absorbance ( A ) of the clear organic extract was measured at 378 nm in a l-cm (or 2-cm) cell against the reagent blank.Calibration graphs were also obtained at pH values from 7 to 11 by the addition of a few microlitres of 6 M potassium hydroxide solution to the potassium nitrate solution. Dnta processing Statistical analysis of the regression was performed on an Olivetti P 6040 desk calculator. Linearity was tested by the analysis of variance for the regression, using R2 and F ratios as criteria of adequacy,' where R2 is the ratio between the sum of squares attributable to regression and the total sum of squares and F is the ratio between the variance attributable to regression and the variance attributable to the deviation from the regression. Results and Discussion Monodisperse Surfactants When the absorbance ( A ) of the 5 ml of organic extract is plotted against the surfactant concentration (c') existing in 50ml of aqueous phase before the extraction, straight lines passing through the origin are obtained ( A = abc'), at least in the c' range 1 4 .1 mg l-l. Table I1 indicates the values of the slope (a) of the calibration lines obtained at 20 "C. ' The slope represents the absorbance [cell length (b) = 1 cm] of the 5.00ml of extract from 50.0 ml of aqueous phase containing 1.00 mg 1-1 of monodisperse surfactant. At least in the range 5<n<15 (the surfactant with n = 4 does not respond appreciably to the picrate reagent under the proposed experimental conditions), the linearity of the calibration lines is demonstrated by the analysis of variance for the linear regression.The value of R2 is higher than 0.999 and that of F is higher than 6000 except for the surfactants with n = 5 (R2 = 0.9637, F = 160) and n = 6 (R2 = 0.9970, F = 2171). However, the latter surfactants have too low a response to the reagent to allow statistical calculations to be performed comparable to those on results obtained with the others. If the polyether surfactant is detennined at trace-level concentration, the linearity of the relationship between the absorbance of the organic extract and the concentration in the aqueous phase can be derived simply from the equilibrium expressions. For instance, the basic equilibria involved in the extraction of a monodisperse surfactant ligand (L) reacting244 FAVRETTO et aE. : REACTION MECHANISMS IN THE Analyst, VoZ.104 TABLE: I1 SLOPES OF THE CALIBRATION STRAIG:HT LINES FOR MONODISPERSE AND POLYDISPERSE POLYOXYETHYLENE e-tef’t-NONYLPHENYL ETHER SURFACTANTS a is the slope (a = A/bc’) obtained by linear regression from a monodisperse surfactant with degree of polymerisation n ; rZ is the slope (Z = A/bc’) obtained from a polydispersed surfactant with number-average degree of polymerisation 1z; RB is the ratio between the sum of squares attributable to regression and the total sum of squares; F is the ratio between the variance attributable to regression and the variance attributable to the deviation from the regression. In .all instances eight observations were processed by regression. 7 % 5 6 7 8 9 10 11 12 13 14 15 Monodisperse surfactants h 7 a R2 F 0.030 0.9637 160 0.081 0.9970 2 171 0.224 0.9994 9 808 0.287 0.9991 6 590 0.293 0.9994 11698 0.289 0.9997 22030 0.280 0.9997 19807 0.264 0.9991 6 647 0.251 0.9993 8 636 0.248 0.9993 9 241 0.247 0.9996 17357 r- 9 i 31.3 51.4 61.5 7.5 8.6 91.7 121.9 15.0 21.5 10.8 Polydisperse surfactants L i Ra 0.036 0.9870 0.121 0.9974 0.191 0.9985 0.211 0.9988 0.228 0.9992 0.236 0.9988 0.240 0.9986 0.243 0.9992 0.244 0.9993 0.245 0.9987 A 1 F 475 2 402 4078 5 187 7 096 5 129 4 168 8 074 8813 4521 1 + 1 with picrate (A-) are the following (a ratio of organic to aqueous phase of 1: 1 is assumed for sake of simplicity) : KKLS.KKLA Aqueous phase L + K+ KL+ KL+ + A- + KLA (pH 7-11) Organic phase L KLA PL 11 11 BKLA Assuming that the extraction is performed at a constant alkaline pH, the A- is also constant and the side AH - A equilibrium can be neglected.It is also assumed that the ionisation of KLA in the organic phase is negligible. In the aqueous phase activity coefficients are unity for non-ionic compounds at trace concentrations. Therefore, the distribution of the surfactant between the organic (org) and aqueous (as) phases is ISL = [L]org/[L]aq and that of the chromophore compound is PmA = [KLA],,,/ [KLAIaq, where equilibrium concentrations (moll-1) are indicated in brackets for every species considered ( [KLAIorg is proportional to the absorbance). As the ionic strength is constant (p M l), the activity coefficients of non-ionic species are also constant. Therefore, K,,, = f [KLA]aq/((IKL+]aq[A-],q), where f = I/( fKL + fA-) is a constant and KKL+ = f.’[KL+]aq/([K+]aq[L]aq), wheref’ = fKL+lfrr+ is also a constant.If the above expressions are introduced into the mass-balance equation of the surfactant (cL = [Llorg + [Llaq + [KL+]aq + [KLA],, -+ [KLAIoTg, where c, is total surfactant concentration in the aqueous phase), the following expression is obtained by resolving for [KLAlorg : At trace levels of surfactant and with an excess of potassium picrate, [K+]aq m cK+ (total concentration of the co-ordinating cation in the aqueous phase) and [A-laq m c,- (total concentration of picrate in the aqueous phase at pH > 7). As c,+ and cA- are constant, the above expression is reduced to an equation that represents a straight-line graph passing through the origin and with a constant slope.Linearity of the calibration graph is always lobserved in the determination of both mono- disperse and polydisperse polyoxyethylene compounds at trace levels with various extraction reagents, for instance, sodium picrate5 and ammonium tetrathioisocyanatocobaltate( 11) .* With the monodisperse surfactant considered in this work, a single extraction was adopted, as a negligible absorbance ( A (0.003) occurred in the organic phase when an aliquot of theMarch, 1979 DETERMINATION OF NON-IONIC SURFACTANTS IN WATERS 245 centrifuged aqueous phase was further extracted with 1,2-dichloroethane in the same 1 : 10 organic to aqueous phase ratio. Starting with the surfactant with n = 5, the slope a increases with n up to a maximum at n = 9, then decreases to a constant limiting value (a = 0.247), which is also observed in a binary mixture consisting of .n = 16 (60%) and n = 17 (40%) (the composition of the mixture is approximate, as the gas- chromatographic analysis is beyond the limit of complete resolution of the peaks).The variation of a as a function of n is shown in Fig. 1. 0.30 L. I(0 0 :. 0.20 al .- - C 0 .- r 2 s ; 0.10 m - a, Q m - 0 A- &---A, I 'A, A 1 'A, .,.x .,a<*2A'.p' .. . . .. x.. . . . . . . . * .. . . . . . . . . . .. . . . . . . . . . . . . - 1 - i) " " - - / / i ' ' ' ~ ~ ' ' ' ~ ' " ' ' ~ I t " I ' 1 1 I I 1 1 5 10 15 20 25 Fig. 1. Graphs of the slopes of the calibration straight lines versus degree of polymerisation of the surfactant. A, Slope (a) of the calibration straight lines veYsus the degree of polymerisation (.n) for monodisperse surfactants RO (CH,CH,O) .H where R =p-tert-nonylphenyl ; A indicates the mixture TZ = 16 + 17.B, Slope ( H ) of the calibration straight lines Venus the number-average degree of polymeris- ation (d) for the polydisperse surfactants RO(CH,CH,O),H, where R=p-tert-nonylphenyl. C , Slope (Z) of the calibration straight lines versus the number-degree of polymerisation ( f i ) for the polydisperse surfactants RO(CH,CH,O),H, where R= dodecyl. At least in the range of n considered, the variation of a as a function of n can be explained qualitatively by considering the equilibria existing in the two-phase extraction of the poly- oxyethylene non-ionic surfactant with potassium picrate. The analytical expression for [KLA],,, shows that the increasing reactivity with n up to 9 is dominated by (i) the distribution equilibrium of the surfactant between the phases and (ii) the co-ordinating equilibrium of the surfact ant with the potassium cation. Hydrophobic polyethers with n<5 are almost completely extracted into the organic phase (/3,>20 at 25 oC)6 and in aqueous solution they do not co-ordinate appreciably with potassium.Although the surfactant with n = 5 has /3,>20, it shows an appreciable reactivity, as the ligand has at least five polyether oxygen atoms for the co-ordination. The same limiting value is also observed with ammonium as the co-ordinating cation.9 The co-ordination equilibrium reasonably assumes a 1 + 1 reaction between the polyether ligand and the potassium cation (the possible side equilibrium 2L + K+ + KL2+ is not considered, because the molar ratio L/K+ M 1/1000 is unfavourable to the formation of KL2+). There are no direct equilibrium data available on this ligand in an aqueous phase.However, a 1 + 1 stoicheiometry was observed for L = CH,O(CH,CH,O),CH, reacting with K+ in methanol; in this solvent the value of the concentration stability constant (K' = [KL+],,/[K+].,[L],, at 25 "C) is 158 1 mol-l but in water a lower value is expected. A 1 + 1 reversible reaction was also found potentiometrically for the monodisperse polyether RO(CH2CH20),R (R = phenyl, 6<n<lO) reacting in methanol2 with Na+, K+ and Rb+; with these cations, log K' increases almost linearly with ~ 2 . ~ This fact suggests that in polyoxyethylene 9-tert-nonylphenyl ethers from n = 6 to 9 the increment in the sensitivity mainly reflects the increase in ITKL+.246 FAVRETTO et al.: REACTION MECHANISMS IN THE Analyst, VoZ. 104 The maximum at n = 9 is probably related to a maximum in the stability of KL+, which probably corresponds to the complete wrapping of the K+ by the polyether chain. In order to evaluate the amount of surfactant extracted from an aqueous into an organic phase by means of potassium picrate, the molar absorbance coefficient at 378nm of the picrate in 1,2-dichloroethane was first determined, using re-crystallised dodecyltrimethyl- ammonium picrate as a standard. In the 300-700-nm region, this ion-pair compound has the same absorbance spectrum as that of the KLA compound, and it is assumed that its picrate chromophore has the same molar absorbance coefficient as that of the KLA ion pair.By use of this coefficient the theoretical absor'bance at 378 nm (b = 1 cm) that 1.00 mg 1-1 of surfactant in aqueous phase would give if the surfactant was totally extracted into the organic phase was calculated. The further assumption was made that surfactant and picrate were in a 1 + 1 stoicheiometric ratio in the extracted compound. The comparison with the observed adsorbance gives the propor tion of surfactant extracted. For n = 5 the extraction efficiency is poor (7.3y0), but it increases up to n = 9, when it reaches 100~o. In the range 10<n<13 the efficiency gradually increases to a constant value of llOyo, but for longer chain mono- disperse surfactants (n = 14 and 15) the rise is gradual; this effect is also confirmed by the behaviour of the binary mixture n = 16 $- 17 (extraction efficiency = 128%).This behaviour showing more than lOOyo recovery can be explained only by assuming the forma- tion of the K,L2+ bivalent cation with the surfactants with n>9. The progressive formation of a polycation with the increase in the length of the poly- oxyethylene chain seems to be reasonable. In the structure of KL+ the potassium cation is progressively wrapped by the polyether chatin with the oxygen atoms stretched inwards and the ethylene groups outwards. At n = 9 just one K+ is co-ordinated by the polyether, whereas at higher values of n a second K+ begins to be progressively co-ordinated. There is no direct information regarding the n value at which the bication K2L2+ reaches maximum stability.Assuming that the limiting slope value (a = 0.247) is constant, an extraction efficiency of 200% would be reached at about n = 27. As the minimum chain length for complete wrapping of the first Kf is only 9 polyoxyethylene groups, about 18 would be required to wrap the second K+. Eighteen oxyethylene groups may be the number required for co-ordinating every other K+ after the first in the K,Lm+ polycation. The results are shown in Fig. 2 as a function, of n. I I 0 5 10 15 Degree of polymerisation, n Fig. 2. Graph of extraction efficiency versus degree of polymerisation (n) for the mono- disperse surfactants R(3(CHzCH20)nH, where R=$-tert-nonylphenyl. Indicates the mix- turen = 16 + 17.Polydisperse Surfactants Table I1 indicates the a' values observed for these surfactants. The calibration graphs are ays straight lines through the origin with F>2000, except for +i = 3.3 (F = 475); inMarch, 1979 DETERMINATION OF NON-IONIC SURFACTANTS IN WATERS 247 this surfactant the absorbance response is so low that the experimental error makes a signifi- cant contribution to the variance attributable to the deviation from the regression, thus lowering the F value. In Fig. 1 the slope of the calibration lines is shown as a function of G. Increasing the pH of the aqueous phase from 7 to 11 does not change either the zero intercept or the slope of the calibration graphs. In comparison with the behaviour of the monodisperse surfactants, the trend of ii versus +i is slightly different.At low values of G, the response of the polydisperse surfactants is higher than that of corresponding monodisperse materials, but with increasing f i this feature inverts and no maximum is observed. The limiting value for 13GfiG22.5 appears to approach that observed with the monodisperse surfactants. The slope of the calibration lines observed in the polyoxyethylene dodecyl ether non-ionic surfactants under the same experimental conditions1 is also shown in Fig. 1 as a function of pi. The trend is similar to that found in the polyoxyethylene 9-tert-nonylphenyl ethers and the limiting value is nearly the same for surfactants. An absolute determination of these commercial surfactants is thus possible for +i>8, using a mean slope as a calibration graph. However, as the evaluation of +i is possible only in special instances, and no information can be obtained on the R group at trace levels, it is convenient to express the concentration of non-ionics as “potassium picrate active substances” (PPAS) -1 It has been proposed to refer the concentration of PPAS to the standard monodisperse surfactant hexaoxyethylene dodecyl ether.The slope (a = 0.268) of this compound is acceptably near to those observed in both surfactant series at %>lo. The slope for a dodecyl ether is always higher than that for the corresponding 9-tert-nonyl- phenyl ether, but this effect becomes more pronounced at G<9. At least two facts can possibly contribute to explain this difference. If two monodisperse potassium picrate reactive surfactants RO(CH,CH,O),H with same n are compared it will be found that over the range 5<n<9 the pL constant with R = dodecyl is systematically lower than that with R = p-tert-nonylphenyl, that is the former is slightly more hydrophilic (more reactive) than the‘latter. Moreover, the KKL+ constant with R = dodecyl is systematically higher than that with R = 9-tert-nonylphenyl, as a consequence of the possible steric hindrance of the P-tert-nonylphenyl group to the co-ordination of Kf by means of the phenoxy oxygen atom. When R = dodecyl, the ether oxygen atom bridging this group can also participate in the co-ordination, and therefore the surfactant probably reacts as an n + 1 polyoxyethylene compound. Further studies are still in progress on monodisperse n-alkyl ethers in order to confirm these speculations. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Favretto, L., Stancher, B., and Tunis, F., Analyst, 1978, 103, 955. Calzolari, C., Favretto, L., Pertoldi Marletta, G., and Favretto Gabrielli, L., Annuli Chim., 1974, Favretto, L., Annuli Chim., 1976, 66, 621. Calzolari, C., Favretto, L., and Tunis, F., Analyst, 1974, 99, 171. Favretto, L., and Tunis, F., A~talyst, 1976, 101, 198. Stancher, B., Tunis, F., and Favretto, L., J . Chronzat., 1977, 131, 309. Draper, N. R., and Smith, H., “Applied Regression Analysis,” John Wiley, New York, 1966, Favretto, L., and Bruni, G., “Proceedings of the 6th Congress on Quality,’’ Genoa, 11-13 September Crabb, N. T., and Persinger, H. F., J . Am. Oil Chenz. Soc., 1968, 45, 611. Frensdorff, H. K., J . Am. Chew. Soc., 1971, 93, 600. 64, 463. pp. 7-26. 1967, pp. 259-274. Received September llth, 1978 Accepted September 20th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400241

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

248 Analyst, March, 1979, Vol. 104, pp. 248-257 Limit of Detection in Analysis with I on -selective Electrodes Derek Midgley Centyal Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey KT22 7SE The limit of detection in analysis with ion-selective electrodes is discussed and definitions that are based only on the deviation of an electrode’s Cali- bration from the theoretical, and take no account of the random errors of measurement, are shown to be inadequate. Equations are derived that express the limit of detection in terms of the random error of measurement and the factors determining the deviation of the electrode response from the Nernstian value, i.e., reagent blanks, solubility products and inter- ferences. The equations enable one to predict (a) the degree of precision with which the e.m.f.has to be measured if an electrode is to attain a desired limit of detection in specified conditions or (b) whether changing the conditions might bring the desired limit of detection within reach of a given precision of measurement. Practical examples with ion-selective electrodes justify the proposed statistical treatment of limit of detection and demonstrate that the errors for electrodes operating in the non-Nernstian region are normally distributed. Keywords : Ion-selective electrodes ; potentiowetry ; limit of detection The limit of detection of an ion-selective electrode has been defined in terms of the extent of the deviation of its calibration from the theoretical Nernstian response, but in such In the IUPAC definition,l the limit of detection is taken as the concentration at the point of intersection of the extrapolated linear segments of a graph of e.m.f.against the logarithm of the concentration. In an earlier recommendation,2 the limit of detection was the con- centration at which the calibration deviated by klog2/z mV from the extrapolated Nernstian response ( K being the Nernst slope factor for a univalent electrode and z being the ionic charge). Although such definitions may be useful as rule-of-thumb characteristics of electrode performance, they are not rigorously related to analytical performance, in contrast to the statistically based definitions of limit of detection that have been applied to other techniques of chemical analy~is.~ The limit of detection should not be arbitrarily deter- mined from the calibration graph without consideration of the random errors associated with the measurement but should be decided on a statistical basis that allows a solution containing a given concentration of determinand to be discriminated, with a specified degree of confidence, from a blank solution.A rigorous definition is needed for ion-selective electrode methods, not only for consistency with other methods of analysis, but on the practical grounds that ion-selective electrodes are routinely used to determine chloride4-6 at levels below the IUPAC limit of detection for the electrode concerned and have been demonstrated to be of use for determining othler species at comparable levels.’ This paper shows that when a statistical approach is adopted the limit of detection can be expressed as a function of the factors determining the deviation of the electrode response from the Nernstian valbe, i.e., reagent blanks, solubility products and interferences.Such equations enable one to predict (a) the degree of precision with which the e.m.f. has to be measured if an electrode is to attain a desired limit of detection in specified conditions or (b) whether changing the conditions might bring the desired limit of detection within reach of the precision of measurement, e.g., by working at a low temperature in order to reduce a solubility product. the critical deviation has been assigned in an arbitrary way. Theory The approach adopted is that of ROOS,~ combined with the conventional choice of 5% probability.The criterion of detection is then defined as the level of determinand corre- sponding to the analytical result that will not be exceeded on more than one occasion inMIDGLEY 249 20 unless determinand is present in the sample, i.e., there is a 5% chance of claiming that a determinand has been detected when none is present. The limit of detection is defined as the concentration for which there is a 5% probability that the analytical result will be less than the criterion of detection, i.e., at which there is a 5% chance of failing to detect the determinand. The criterion of detection, Q, for 95% confidence is where uQ is the within-batch standard deviation of the result, i.e., the difference between the sample reading, A , and the blank reading, B.Thus, When Q is estimated empirically, we do not know beforehand the value of A at which measurements should be made. It is convenient, therefore, to make the assumption and so to write oQ = d%o, and Q = 2.330, . . .. .. .. * . (3) Once Q has been estimated from equation (3), a first estimate of A is obtained and the validity of the assumed equality (2) tested with standard deviations for results in the vicinity of A . By a similar process and assuming the approximate equality ( 2 ) is still valid, it can be shown that the limit of detection, L, for 95% confidence is L = 1.6450, = 4.650, If the concentration is linearly related to the experimentally observed quantity, as in titrimetric or photometric methods, L and o, can be expressed equally well in either con- centration units or experimental units.When this linear relationship does not hold, as in potentiometry, L and oL should be expressed in the units in which the errors have a normal distribution. I t is shown below that in the instances where data are available the distri- bution can be assumed to be normal irrespective of the units chosen, but for the following reasons it is better to use e.m.f. units in the absence of experimental proof of the nature of the distribution. 1. The blank reading cannot be plotted as a point on a graph of e.m.f. against the loga- rithm of the concentration, which is the usual representation of the calibration of an ion- selective electrode, and hence the standard deviation in concentration terms cannot be obtained from such a graph.By using the standard deviation in e.m.f. units and drawing the lines corresponding to 2.330, and 4.650, on the calibration graph (as in Fig. l), the criterion and limit of detection can be obtained in Concentration terms. Although it is possible to plot the blank reading as a point in other forms of calibration and so to use the standard deviations in concentration units, standard deviations in e.m.f. have the advantage of being applicable in all instances. 2. On the basis of the small amount of data available for ion-selective electrodes operating near their limits of detection, the assumed equality (2) is more readily satisfied by taking standard deviations in e.m.f. units. Applying the F-test to the within-batch standard deviations in Table I shows that at the P = 0.05 level the hypothesis that the variances are equal throughout the range covered would be rejected only for the sodium electrode for measurements in e.m.f.units and for the sodium, fluoride and silver - silver chloride electrodes for measurements in concentration units. The use of standard deviations in e.m.f. units is consistent with practice at concentra- tions in the Nernstian response range, e.g., in testing the significance of recovery tests,14 where the standard deviations in concentration units increase so much with concentration (even 3.250 MIDGLEY: LIMIT OF DETECTION IN TABLE I WITHIN-BATCH STANDARD DEVIATIONS (s.d.) FOR ION-SELECTIVE ELECTRODES OPERATING IN THE NON-NERNSTIAN RANGE Analyst, Vol. 104 Concentrationlpg 1-1 . . . .20 5 S.d./vgl-1 . . . . . . 0.35 0.37 S.d./mV . . . . . . 0.19 0.20 Concentration/mg 1-1 . . . . 10 3.38 S.d./mgl-l . . . . . . 0.23 0.10 S.d./mV . . . . . . 0.48 0.67 Concentration/wg 1-1 . . . . S.d./pg 1-l . . . . . . S.d./mV . . . . . . 25 1.6 1.5 Concentration/yg 1-1 . . . . 20 S.d./pg 1-1 . . . . . . 2.6 Concentration/mg 1-1 . . . . 1.0 S.d./mV . . . . . . . . 1.73 S.d./mg 1-1 . . . . . . 0.057 S.d./mV . . . . . . 1.23 Gmcentrationlyg 1-1 . . . . S.d./pg 1-1 . . . . . . S.d./mV . . . . . . 45.6 3.2 0.9 1 0.62 0.29 0.74 0.02 0.38 2.5 0.6 2.7 1.2 1.00 0.1 0.011 2.35 10 18.2 1.8 1.8 Degrees of freedom Determinand Electrode material Reference 0 5 c1- HgSCl, - HgS 9 0.75 0.41 0.1 5 c1- Ag - AgCl 4 0.01 0.45 0.25 11* Na+ Glass 10 0.5 3 .O 1.4 1.65 0.05 9 F- LaF, 12 0.008 2.18 0 4 NH,+ Glass 11 3/5 9- A&S 13 * These results include between-batch variation.though the relative standard deviations decrease) that combination or comparison of errors expressed in concentration units would be misleading. When e.m.f. units are adopted, the criterion of detection, Q, is the positive value of the difference between e.m.f. E,, given by the blank solution, and e.m.f. E,, given by a sample containing such a concentration of determinand, C,, that E, = E, + iQ where i = +1 for a cation-selective electrode and -1 for an anion-selective electrode. The limit of detection, L, is defined in an analogous way. Thus, E, = E, + iL at a concentration of determinand C,. Normal Distribution of Error in E.m.f. Measurements Within-batch errors of e.m.f.measurements made with various electrodes near their limits of detection have been analysed by means of the Lilliefors test for normality15J6 (using the corrected table given by Conoverl'). The electrodes used and the conditions of measure- ment are given in Table 11. The only instances in which the hypothesis that the distri- bution was normal was rejected at the 576 level were for the sodium-responsive glass electrode at 25 pg 1-1 and the mercury( I) chloride - mercury( 11) sulphide chloride-selective electrode at 10 pg 1-1 and 25 "C. In no instance did acceptance or rejection depend on whether the errors were expressed in the original e.m.f. units or the derived concentration units. The use of parametric statistics in the calculations below is, therefore, justified, as are the definitions of limit and criterion of detection in e.m.f. terms.TAELE I1 ION-SELECTIVE ELECTRODES TESTED FOR DISTRIBUTION OF ERRORS IN THE NON-NERNSTIAN RESPONSE RANGE Determinand Electrode material vg I-' results Conditions Reference c1- Hg C1 - HgS 1, 2) 5,lO 10 25 and 4 "C, 0.01 moll-' HN08, flow cell 9 c1- Aga- &Cl 100,1760,5940 10 25 "C, pH 4.7 (CH,COONH,, buffer) 4 18.2, 46.5 4/; 0.6 mbfl- NaOH - 0.1 m o l V Na,H,EDTA 13 19 SZ- &as cup+ CUS - Ag,S 6.3 Concentrations tested/ NO. of c1- HgSCl, - HgS 60 10 25 "C, 0.01 HNOJ 18 NH + Glass 0, 10, 20 6 pH 8.2 (triethanolamine buffer) 11 Naf Glass 0.25, 2.5, 25 12 29OC H 11 (NH,) 10 22 "C, no added saltsMarch, 1979 ANALYSIS WITH ION-SELECTIVE ELECTRODES 251 Calculation of Limits and Criteria of Detection for Ion-selective Electrodes response, where the e.m.f.can be expressed by the equation Measurements with ion-selective electrodes are usually made in the region of Nernstian E = E" + kl0gC . . .. .. .. - (4) where E" is the standard potential, k == RTlnlOIzF is the slope factor and C is the con- centration of determinand from the sample. Near the limit of detection, however, other factors influence the e.m.f., which can be expressed by the equation E = E" + klog(C + s + b, + Cbi) . . .. - (5) where s is the contribution of determinand from dissolution of the materials of the electrode itself, b, is the concentration of determinand in the reagent blank and b, is the interference effect of the ith interfering species. The criterion (C,) and limit (C,) of detection in concentration terms can be calculated from b,, b,, s and K , the solubility product governing s, together with the values of Q or L.If, as is usual in potentiometric analysis, the ionic strength of the samples is adjusted to a constant value, the activity coefficients are constant and can be included in the E" and K terms of the equations derived below ; concentrations have, therefore, been used throughout instead of activities. The derivation of equations (7), (9), ( l l ) , (13) and (15) for electrodes with non-ideal responses has been described by Midgley.20 The equations are shown for cation-selective electrodes only; those for anion-selective electrodes are identical except for the signs before the logarithmic terms in the equations for e.m.f.Respon.se not limited by solubility product In this instance b, + Cbi = b > s. and s = 0: The e.m.f., E,, is given by equation (5) with C = 0 EB = E" + klogb At the criterion of detection C = CQ and E = E,, given by By definition, Q = I E, - E, I and, therefore, Hence, Similarly, c, = ( l O L / * - 1)b .. .. .. . . (6') Response limited by sohbility product In this instance b, = Cbi = 0. For electrodes whose response is governed by the solubility product of a crystalline phase with 1 : 1 stoicheiometry, the e.m.f. can be expressed as follows~0: .. * . (7) At C = 0, E, = E" + klogK3252 At C = C,, MIDGLEY: LIMIT OF DETECTION IN Analyst, Vol. 104 E, = E" + klog[;; + (7 + K ) t ] Hence, Expanding by the binomial theorem for C, < ,K*, we obtain Exponentiating and solving the resultant quadratic, we obtain ...' (8) c, = 2K*[(2 x lOQlk - 1p - 11 .. Similar1 y , c, = 2K&[(2 x 10Llk - 1)* - 11 .. .. . . (8') Response limited simultaneously by solubility product and interference When b, = 0 and Xbi # 0, we have,20 with the same conditions as for equation (7) .. .. (9) Proceeding as from equation (7) to equation (8), we obtain and . . (10') C L = 2K*([1 - 2(1 - 10L/k)(1 + b/K*)]* - 1) .. Response limited simultaneously by s o h bility piloduct and reagent blank determinand When b, # 0 and Xbi = 0 and with the saime conditions as for equation (7), we obtain14 .. . . (11) E = E" + Klog {s: b + C + "b" + c)2 + K ] ' ) 4 Proceeding as from equation (7) to equation (8), we obtain (10QIk - 1) 4K * c, = (2K* + b) lo,/* -t .. (12) {[ ( 2 K * + b)2 ] '} * * Response limited by solubility product with 2 : 1 stoicheiometry as f ollows20 : Univalent electrodes. With b, = Zbi = 0 and (c + s ) ~ s / ~ = K , the e.m.f. can be written E = E" + klog[(2K/s)q . . .. .. . . (13)March, 1979 At C = 0, ANALYSIS WITH ION-SELECTIVE ELECTRODES 263 E, = E" + klog[(2K)*] Hence, s, = (2K)$ x 10-2Qlk and CQ = - sQ + (2K/sQ)' . . .. .. . . (14) The calculation of s, and C, is exactly analogous. as follows14: Divalent electrodes. With b, = Zbi = 0 and (C + s)(2s)2 = K, the e.m.f. can be written . . (15) E = E" + K1og(K/4s2) . . .. .. A t C = 0, At C = C,, Hence, and The equation for C, is exactly analogous. Applicability to Different Types of Electrodes Equation (6) is valid for all electrodes, regardless of mechanism, provided that b 3- s, i.e., the presence of the reagent blank or interferences is the only cause of deviation from the Nernstian response at the observed criterion of detection.This can be checked by the procedure described by Midgley.20 Equations (8)-(16) are valid for electrodes with crystalline phases that participate in solubility equilibria with the adjacent solution but not for electrodes in which the active component dissolves as a result of an irreversible chemical reaction, such as the oxidation of sulphide in the copper(I1) sulphide - silver sulphide membranes of copper-selective electrodes.21 The validity of equations (8)-(16) does not depend on how the crystalline phase is incorporated into an electrode, e.g., for electrodes based on silver chloride, the equations describing the e.m.f.are equally valid for silver - silver chloride electrodes, hetero- geneous (Pungor-type} membrane electrodes and mixed-crystal (silver sulphide - silver chloride) membrane electrodes.20 With mixed-crystal membrane electrodes the more soluble component determines the choice of equation, e.g., for silver sulphide - silver chloride electrodes, the greater solubility of silver chloride means that equation (8) is used, not equation (14).254 MIDGLEY: LIMIT OF DETECTION IN Analyst, Vol. 104 In theory,20 liquid ion-exchange electrodes can be accommodated to equations (7)-( 16) by replacing K by K,, where K, is equal to (C*)2/KD for equations (7)-(12), 0.5(C*)3/K, for equation (14) and 4(C*)3/KD for equation (16).In each instance C* is the concentra- tion of exchanger in the membrane phase and K , is the appropriate distribution coefficient between the aqueous and membrane phases. This model has not been tested experimentally. Influence of Blanks, Interferences and Sollubility Products on Limits and Criteria of Detection Fig. 1 shows the calibration graphs for different types of hypothetical ion-selective electrodes. All of the electrodes respond to univalent cations, have standard potentials of 337.15 mV and calibration slopes of 59.16 mV per decade at concentrations well above the limit of detection. Line N represents the extrapolation of the Nernstian portions of the calibrations of all the electrodes. To show the comparative effects of the various causes of non-Nernstian behaviour, the solubility prodlucts, interferences and reagent blanks were chosen so that the potential of each electrode at zero nominal determinand concentration was 0.0 mV, corresponding to an apparent concentration of 2 x moll-1 on the extra- polated Nernstian graph N.In all instances, 0, was taken as 1.0 mV, resulting in the lines E, = 2.3 mV and E , = 4.6 mV. -6.5 -6.0 -5.5 Logarithm of concentration Fig. 1. Limit and criterion of detection for hypothetical univalent electrodes with non-Nernstian calibrations. A, b = 2 x 10-8 mol 1-1 (interference or reagent blank determinand, no solubility effect) ; B, K = 4 x 10-l4 mol* 1-8 (solubility effect only) ; C, b, = 10-6 moll-1, K = 10-1* mola l-a (solubility effect and interference); D, b, = 1.5 x moll-1, K = lo-'* mola 1-8 (solubility effect and reagent blank) ; E, K = 4 x 10-l8 molS l-s [non-isovalent (2 : 1) solubility effect] and N, ideal Nernstian response.Lines EQ and E t show the criterion and limit of detection in e.m.f. terms.March, 1979 ANALYSIS WITH ION-SELECTIVE ELECTRODES 255 In Fig. 1 the abscissa represents the blank e.m.f. reading and the lines E, and E, the criterion and limit of detection in e.m.f. terms. The intersections of E, and E, with the calibration graphs give the criteria and limits of detection in concentration terms, which vary widely between electrodes even though all have identical responses at C = 0 and C > C , and identical standard deviations in e.m.f., i.e., the lines E,, E,, N and the abscissa are common to all the electrodes.Table I11 shows the limits and criteria of detection obtained for these electrodes both from Fig. 1 and by use of equations (6)-(14). The differences between the two sets of figures arise from the approximations involved in deriving the equations. Better agreement would be obtained at lower values of a,, because the errors from the approximations become smaller at concentrations more removed from the Nernstian region. TABLE I11 CRITERIA AND LIMITS OF DETECTION FOR UNIVALENT ELECTRODES WITH a, = 1.0 mV Curve in Fig. 1 r A * * 106br/mol 1-1 . . .. 106Xb,/mol 1-1 . . .. Equation used . . * - (6) 107cs/mo1 1-1- Graphical . . . . 1.90 Calculated . . . . 1.90 107c,/m01 1-1- Graphical . . .. 3.98 Calculated . . . . 3.97 - 10x2 K , ... .. * 106(br + Cbs) = 2. B 0 0 4 (8) 3.63 3.47 7.41 7.28 C 0 1 1 (10) 3.43 4.28 6.76 8.23 D 1.5 0 1 (12) 2.29 1.99 4.79 4.03 E 0 0 4 x 10-6 (14) 5.25 5.21 9.66 10.06 Table I11 also shows that the ratio L/Q (equal to 2.0 in the conditions adopted) need not equal C,/C,; this arises from the nature of the calibration of e.m.f. against the logarithm of the concentration. When the calibration reaches the limiting stage where the e.m.f. is linearly related to concentrati~n,~~~ the two ratios will be equal. In practice, a, is obtained from a series of measurements with standard solutions, including a zero standard as a blank, but the water used to prepare these standard solutions will itself contain some concentration of determinand, bw, which can be termed the water blank.The value of a, obtained will not, therefore, be the standard deviation of the true blank, but only the best approximation we can make. Because, for convenience, we have already assumed that the standard deviations are constant over the range C = 0 to C = C,, any error in a, from this source should be negligible. The values of Q and L, which are multiples of o,, are, therefore, unaffected by the existence of the water blank. If the value of b in equations (6) and (12) is derived from the potentiometric data, e.g., by any of the methods described by Midgley,20 it will be overestimated by the amount of the water blank unless b, is determined in an independent experiment and then subtracted. In practice both b, and b, may be negligible, e.g, some of ihe chloride results quoted by Midgley.20 Discussion By the IUPAC definition,l the limits of detection for all the hypothetical electrode responses shown in Fig.1 would be 2 x 10-6 moll-1 regardless both of the factors governing the response and of the precision of the measurements. In contrast, the statistical definition yields a different limit of detection for each electrode, reflecting the different sources of non-Nernstian behaviour for the various electrodes. With real electrodes the limits of detection might have appeared to be higher in both instances, because of the experimental difficulty of identifying the equilibrium e.m.f. at low concentrations. With the IUPAC limit there is also the problem of identifying precisely the linear non-Nernstian segment of the calibration graph.256 MIDGLEY: LIMIT O F DETECTION I N Analyst, Vol.104 For instance, chloride-selective electrodes have lower IUPAC limits of detection at lower temperatures, because the calibration graph changes as a result of the temperature dependence of the solubility product. When the two sets of calibration data are analysed statistically, it is possible for the limit of detection to be no better or even worse at the lower temperature because the increased sensitivity of the calibration may be offset by a loss of precision. An example9 of this is the operation of a mercury(I1) sulphide - mercury(1) chloride membrane electrode at 25 and 4 “C, where the limits of detection were almost the same at the two temperatures, despite the %fold increase in sensitivity at 4 “C.The use of the equations for calculating C, enables one to predict the effect that changes in the solubility product or in the purity of the reagents might have on the limit of detection. The same equations can be used to predict the precision required (in e.m.f. units) if a given limit of detection is to be attained under specified conditions. IUPAC and statistical limits of detection are compared in Table IV for different types of electrodes. When, as with the glass electrod~es,lOJl the precision of measurement is poor (standard deviations of 1.5-3 mV) and there is a high reagent blank, the two definitions differ little, but when measurements are made more precisely (standard deviations of 0.04- 0.4mV) and the reagent blank is small, as with the two types of chloride electr0de,~19 the IUPAC definition gives an unrealistic indication of the performance of an electrode in analysis.Even as a qualitative guide, a non-statistical definition may be misleading. LIMITS OF DETECTION OF ION-SELECTIVE ELECTRODES IUPAC: limit/ Statistical limit/ Electrode Determinand l l g 1-1 l l g 1-1 Reference Hg2C12 - HgS c1- 60 1 9 Ag - AgCl c1- 530 15 5 Glass NH,+ 11 3 6 11 Glass Na+ 3 1.4 10 Both the IUPAC and statistical limits arise from the performance of the electrode itself, but in considering an analytical method as a whole, other factors may make these limits irrelevant. For instance, if the concentration is obtained from the measured e.m.f., not by use of a calibration graph but from the Nernst equation, either explicitly as equation (4) or implicitly through the use of a pIon or direct-activity scale on a pH meter, the limit of detection of the method is given by the limit of the Nernstian response.A statistical pro- cedure for finding this limit has been given by ILiteanu et aZ.,22 although it might be improved by the use of weighted data in calculating the linear regression line representing the Nernstian calibration and in the subsequent tests for goodness of fit. In certain circumstances,1s~23~24 e.g. , when high ( 10-4-10-2 moll-1) concentrations of metal ion are in equilibrium with an excess of a strong complexing agent, electrodes can respond to low determinand activities ( 10-15-10-9 mol 1-l), but the existence of this response does not contradict the above definition of limit of detection, which is valid for analysis at trace concentrations of determinand.The operation of ion-selective electrodes in their non-Nernstian ranges is sufficiently well established in the l a b ~ r a t o r y ~ ~ ~ p ~ ~ ~ ~ and inside industrial plant5y6 to warrant a more rigorous definition of limit of detection than the IUPAC approach can provide, although the latter can be used as a convenient rule of thumb when it is desired to compare the performance limits of electrodes without collecting large amounts of data. This work was carried out at the Central Electricity Research Laboratories and is published by permission of the Central Electricity Generating Board. References 1. 2. 3. International Union of Pure and Applied Chemistry, Pure AppZ.Chew., 1976, 48, 127. International Union of Pure and Applied Chemistry, ‘(Recommendations for Nomenclature of Ion-selective Electrodes (Recommendations, 1975),” Pergamon Press, Oxford, 1976. Wilson, A. L., Talanta, 1973, 20, 725.March, 1979 ANALYSIS WITH ION-SELECTIVE ELECTRODES 257 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Torrance, K., Analyst, 1974, 99, 203. Tomlinson, K., and Torrance, K., Analyst, 1977, 102, 1. Goodfellow, G. I., “Proceedings of the International Conference on Water Chemistry of Nuclear Reactor Systems, Bournemouth, 24-27 October,” British Nuclear Energy Society,: London, 1977, p. 127. Bardin, V. V., Shartukov, 0. F., and Tolstousov, V. N., Zh. Analit. Khim., 1971, 27, 25. Roos, J . B., Analyst, 1962, 87, 832. Marshall, G. B., and Midgley, D., Analyst, 1979, 104, 55. Hawthorn, D., and Ray, N. J., Analyst, 1968, 93, 158. Goodfellow, G. I., and Webber, H. M., Analyst, 1972, 97, 95. Louw, C. W., and Richards, J. F., Analyst, 1972, 97, 334. Baumann, E. W., Analyt. Chem.. 1974, 46, 1345. Midgley, D., and Torrance, K., “Potentiometric Water Analysis,” Wiley, Chichester and New York, Lilliefors, H. W., J . Am. Statist. Ass., 1967, 62, 399. Ames, A. E., and Szonyi, G., in Kowalski, B. R., Editor, “Chemometrics: Theory and Practice,” Conover, W. J ., “Practical Nonparametric Statistics,” Wiley, New York, London, Sydney and Marshall, G. B., and Midgley, D., Analyst, 1978, 103, 438. Midgley, D., Analytica Chim. Acta. 1976, 87, 7. Midgley, D., Analyt. Chem., 1977, 49, 1211. Midgley, D., Analytica Chim. Acta, 1976, 87, 19. Liteanu, C., Popescu, I. C., and Hopirtean. E., Analyt. Chem., 1976, 48, 2010. Vesel$, J., Jensen, 0. J., and Nicolaisen, B., Analytica Chim. Acta, 1972, 62, 1. Baumann, E. W., Analytica Chim. Acta, 1971, 54, 189. 1978, p. 71. ACS Symposium Series 52, American Chemical Society, Washington, D.C., 1977, p. 219. Toronto, 1971, p. 302. Received September 20th, 1978 Accepted October 24th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400248

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

258 SHORT PAPERS Analyst, Mcarch, 1979 Spectrophotometric Method for the Determination of Paraquat M. Ganesan, S. Natesan and V. Ranganathan Department of Chemistry, United Planters’ Association of Southern India, Tea Research Station, Cinchona 642 106, India Keywovds : Pavaquat determination ; spectvofihotounetry The divalent l,l’-dimethyl-4,4’-bipyridyliuni ion (paraquat) gives an orange - yellow pre- cipitate of composition paraquat.Hg1, with neutral potassium tetraiodomercurate(I1) (K,HgI,) solution. At low concentrations of paraquat, a yellow colloidal solution is formed. A spectrophotometric method, after stabilisation of the colloidal solution with acetone, ethanol or starch solution, has been developed for the determination of paraquat. The dichloride and dimethylsulphate salts of the divalent 1,l ’-dimethyl-4,4’-bipyridylium ion are used extensively as herbicides1 Extensive work has been carried out on the deter- mination of paraquat and diquat by Calderbank and co-worker~,~-~ and a spectrophoto- metric method described by Yuen et aL5 has been adopted by both CIPAC6 and AOAC.’ Their method is based principally on reduction with alkaline sodium dithionite solution to a blue free radical, which is not very stable and requires immediate determination of the absorbance.A liquid-chromatographic method with an ultraviolet detector has also been reported for determining paraquat dichloride.8 The present investigation was concerned with the development of a spectrophotometric method for the routine determination of paraquat in formulations and in mixtures with other herbicides.The colour developed is stable for up to 60min and hence lends itself for adoption in routine work, requiring less attention than the dithionite method. Experimental Reagents A neutral solution of potassium tetraiodomercurate(I1) is prepared by mixing 22.8 g of mercury(I1) iodide with 17.5 g of potassium iodide, dissolving them in distilled water and diluting to 1 1 with water. The solution is allowed to stand for 1-2 d before the supernatant liquid is withdrawn for use. Pure ethanol, acetone or 1% m/V starch solution is required as a stabiliser of the colloidal solution. Procedure To a 25-ml calibrated flask is added a suitable aliquot of solution of the formulation, containing up to 70pg of paraquat, in the form of its dichloride or dimethylsulphate salt, and then 5 ml of ethanol, 1.5 ml of acetone or 1 ml of 1% m/V starch solution are added.The solution is mixed well, diluted to about 20 ml, 1 ml of neutral potassium tetraiodo- mercurate(I1) solution is added and the solution is mixed well and made up to the mark with water. The absorbance of the yellow solution is measured at 400420 nm after 15 min, but within 60min after the colour development. Pure paraquat dichloride is used as a standard to analyse samples within an unknown paraquat content. The reaction is rapid and the optimum time for taking readings, after colour development, is between 15 and 60 min (see Table I). The determination of mercury and iodine (in the reaction product) was carried out using the standard procedures described by Vogel.!’SHORT PAPERS TABLE I VARIATION OF ABSORBANCE WITH TIME Average of a number of determinations by two different workers on both paraquat salts.Values given are for absorbance in Klett units, with the pH of the final solution between 7 and 8. Time/h 259 r S tabiliser 0 0.25 0.50 0.75 1.00 2.00 4.06 None .. . . . . 118 127 130 128 134 90 80 Starch . . .. .. 110 112 114 113 115 85 60 Acetone . . .. .. 146 113 113 115 114 87 66 Ethanol . . .. .. 109 110 112 111 110 68 60 Results and Discussion The absorbance is linearly related to paraquat concentration up to 3pgml-1. Pure paraquat dichloride is used only for standards. All interference tests and reliability tests are carried out with solutions of paraquat dichloride and paraquat dimethylsulphate. When the concentration of starch in the final solution is between 0.016 and 0.060% m/V, ethanol between 15 and 23% V/V or acetone between 2 and 10% V/V, there is little variation in the colour intensity with the variation in concentration of tpe stabiliser (see Table 11).The TABLE I1 OPTIMUM STABILISER CONCENTRATION Average of 5-10 determinations by two different workers on both paraquat salts. Stabiliser Starch Ethanol Acetone f A 7 . . Concentration, Absorbance, Concentration, Absorbance, Concentration, Absorbance, Klett units % mIV Klett units % VlV Klett units % v/v 0 68 0.008 50 0.0 1 6-0.060 53 0.080 57 0.100 60 0 96 0 12 111 1 15-23 143 2-10 24 128 12 40 24 18 106 110 113 87 14 colour diminishes rapidly at higher concentrations, beyond the upper optimum limit of ethanol or acetone, and almost disappears beyond 40% V/V of ethanol or 18% V/V of acetone.No interferences are observed with Na+, K+, NH,+, Hf, C1-, NO3-, SO,” or CH3S04-, with common herbicides such as sodium 2,2-dichloropropionate, sodium 2,4- dichlorophenoxyacetate and monosodium methanearsonate when added to the solutions of the two paraquat salts (see Table 111). Hydroxyl ions interfere when the pH is above TABLE I11 INTERFERENCE TESTS All concentrations given are in the final solution. Results are the average of 4-6 determinations, carried out by three different workers, on both paraquat salts. The results given are paraquat concentration in milligrams per litre. Concentration of species/mmol l-l r - Species 0 2 5 Na+ .... K+ .. .. . . c1- .. .. NH,+ .. .. NO,- . . .. s0,a- . . .. Ansar . . .. 2,4-D . . .. Dalapon . . .. 1.596 1.596 1.596 1.596 1.596 1.596 1.596 1.596 1.596 1.583 1.596 1.586 1.596 1.609 - 1.596 1.596 1.583 1.683 1.583 1.610 1.596 - 1.596 - 1.596 -260 SHORT PAPERS Analyst, Vol. 104 8. This is overcome by maintaining the pH between 6 and 8 by neutralising with dilute hydrochloric acid and by using neutral potassium tetraiodomercurate(I1). The effects of pH and time on the absorbance are shown in Table IV. TAEILE IV EFFECT OF pH OF THE FINAL SOLUTION AND TIME ON ABSORBANCE All results are the average of 4 determinattions, and the values given are absorbances in K1e:tt units. Time/h A r \ pH range 0 0.25 0.50 0.75 1 2 4 3-5 78 88 8' 5 76 65 62 44 6-8 69 80 8.0 80 80 77 53 9-1 1 97 115 117 120 120 121 117 Diquat (1 ,l'-ethylene-2,2'-bipyridylium ion) interferes with the paraquat determination ; it also gives a yellow colloidal solution with absorption in the range 420-440nm.The presence of diquat can be detected by the formation of a red colour with sodium hydroxide solution. Diquat is hydrolysed completely in dilute alkaline solution in the pH range 12.5-13.0 (2 ml of 0.2 N sodium hydroxide solution is sufficient for a l-ml aliquot containing up to 50 pg ml-l of diquat) within 5-6 h. Diquat interference can be overcome completely when determining paraquat (containing diquat) by keeping it in alkaline solution at pH 12.5-13.0 for 5-6 h and developing the colour in the same manner as described earlier, after neutralising with acid using phenolphthalein as an indicator.Table V shows the results of paraquat determinations in the presence of diquat. TABLE V ELIMINATION OF DIQUAT INTERFERENCE Each result is the average of 3 or 4 d.eterminations by three workers on both paraquat salts. Paraquat, after Paraquat* takenil Diquat* added/ diquat elimination, Paraquat, % mlm mg 1-1 mg 1-I % mlmt 21.14 1.713 2 21.41 20.95 1.676 3 21.00 21.00 1.680 2 21.00 21.00 1.680 3 21.13 * Concentration in the final solution. t By the procedure described in the text. At high concentrations, paraquat gives an orange - yellow precipitate with neutral potassium tetraiodomercurate(I1) solution. Theoretical considerations suggest paraquat.- HgI, as the probable product of any reaction between the two ionic substances, paraquat.Cl, and K,HgI,.On analysis, the precipitated substance was found to contain 20.5 & 0.3% of paraquat, 22.3 5 0.3% of mercury and 56.6 & 0.2% of iodine, compared with the theoretical values of 20.8% of paraquat, 22.5% of mercury and 56.7% of iodine, confirming the formula paraquat.Hg1,. This is unlike the instance of ammonia, when a compound NH,Hg,I, is formed.10 The precipitation is quantitative and there is a distinct possibility of developing this method further for the gravimetric determination of paraquat for quality control purposes. The spectrophotometric method described is accurate, with coefficients of variation of 0.33, 0.41 and 1.41% with ethanol, acetone and starch, respectively, as stabilisers. The method is comparable to the CIPAC method6 within &lyo of the paraquat content, with the added advantage that the colour is stable for up to 60min, unlike the standard methods that require colour development just before measurement of the absorbance (see Table VI).March, 1979 SHORT PAPERS 261 TABLE VI COMPARISON OF METHODS Method Coefficient of variation, yo CIPAC method .. .. .. .. . . 0.45 Present method without stabiliser . . . . 2.47 Present method with starch as stabiliser . . 1.41 Present method with acetone as stabiliser . . 0.41 Present method with ethanol as stabiliser . . 0.33 The authors thank Dr. C. S. Venkata Ram, Director of Research, for his keen interest in the work and Alkali and Chemical Corporation of India Ltd. for providing technical-grade paraquat dichloride for the investigation. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Martin, H., Editor, “Pesticide Manual,” Third Edition, British Crop Protection Council, London, Calderbank, A., Morgan, C. B., and Yuen, S. H.. Avzalyst, 1961, 86, 569. Calderbank, A., and Yuen, S. H., Analyst, 1965 90, 99. Calderbank, A., and Yuen, S. H., Analyst, 1966, 91, 625. Yuen, S. H., Bagness J. E., and Myles, D., Analyst, 1967, 92, 375. Raw, G. R., Editor, “CIPAC Handbook. Volume I. Analysis of Technical and Formulated Pesti- cides,” W. Heff er and Sons, Cambridge, for Collaborative International Pesticides Analytical Council, 1970, p. 547. 1972. Carlstrom, A. A., J . Ass. Off. Analyt. Chem., 1968, 51, 1306. Kawano, Y., Audino, J., and Edlund, M., J . Chromat., 1975, 115, 289. Vogel, A. I., “A Text Book of Quantitative Inorganic Analysis,” Third Edition, Longmans Green, Partington, J. R., “Inorganic Chemistry,” Macmillan, London, 1965. Harlow, 1962. Received May 25th, 1978 Accepted September 27th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400258

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

March, 1979 SHORT PAPERS 261 Titrimetric Determination of Reducing Sugars with Copper(l1) Sulphate T. H. Khan Department of Industvies (Chemical Directorate) , 58, Dilkusha Commercial Area, Dacca-2, Bangladesh Keywords : Reducing sugar detevwzination ; copper(II) sulphate reduction ; titrimetry Of the copper reduction methods for the determination of reducing sugars, that of Lane and Eynonl is still considered to be a simple and rapid method and is often the most a ~ c u r a t e . ~ - ~ Much work5-12 has also been carried out on the automatic determination of reducing sugars but little use has been made of these methods. The use of a constant volume modification of the Lane and Eynon method has been reported,*,l3 but it does not seem to be based on any systematic investigation such as we have carried out in order to obtain information lacking in the literature on the determination of reducing sugars.During the work on the determination of reducing sugars by using the Lane and Eynon method, the following observations were made and necessary modifications were made to the method. Modifications to Lane and Eynon Method T i m e to Teach boiling-point. The total boiling time has been fixed at 3 min (2 + 1 min) in the Lane and Eynon method but the time required to reach boiling has not been fixed. As the reaction of the reducing sugar with the Fehling’s solution is non-stoicheiometric and the method is empirical, it is necessary to maintain this time constant in order to minimise errors.262 SHORT PAPERS AnaLjlst, Vol. 104 After boiling the mixture of sugar and Fehling’s solution for 2 min before addition of the indicator, it was found to be red or orange - red owing to the presence of coagulated copper(1) oxide.When the reaction mixture of sugar and Fehling’s solution was boiled for 5 min before addition of the indicator, the copper(1) oxide settled at the bottom of the flask, leaving a clear solution. The end-point in the published method was given by the appearance of a bright red or orange colour, but when a clear solution was produced after boiling for 5 min, it was found possible to see the last trace of the blue colour of the indicator and a colourless end-point was obtained. Efect of indicator concentration. When a clear reaction mixture was produced only 1 drop of 0.1% methylene blue solution was required in order to give a distinct blue colour, thus giving a sharp colour change at the end-point.This also eliminates any errors resulting from the use of a high concentration (3-5 drops of a 1% solution of methylene blue) as methylene blue itself oxidises some sugar or those errors caused by the back-oxidation of methylene white to methylene blue. During the addition of the sugar solution to the boiling reaction mixture it has been found useful to allow some time for completion of the reaction and observation of the colour after the addition of each drop, and increasing this time from 1 to 3 min avoids errors caused by the addition of excess of the sugar solution. The use of a 200-ml Erlenmeyer flask in place of a 300- or 400-ml flask improves the detection of the end-point as the height of the liquid layer is greater in the smaller flask.Fehling’s solution. Fehling’s solution was allowed to stand overnight and was then used without filtration through treated asbestos.14 The tedious process of filtering the solutions through treated asbestos has been found to be unnecessary, as shown in Table I. A series of experiments was performed with standard solutions of invert sugar, glucose, fructose and lactose of different concentrations using 5 or 10 ml each of Fehling’s solutions A and B as in the procedure described belovv. Total boiling time. Colour of end-point. Reaction time. Advantages of the use of a 200-ml $ask. Experiment a1 Co$Per(II) sulphate (CuSO4.5H,O). AnalaR grade. Potassium sodium tartrate (KNaC4H40,.4H,0).Sodium hydroxide. AnalaR grade. Methylene blue. Sucrose, glucose and lactose ( C1,H2,0,,.H,0). Maltose (Cl2H,,O1,.H2O). Laboratory-reagent grade. Fructose. Fehling’s solutions. Invert sugar. Distilled water was used in all the experimental work. Reagents AnalaR grade. AnalaR grade. Laboratory-reagent grade, dried under vacuum. Prepared by Soxhlet’s modified method.14 A standard solution of invert sugar was prepared by hydrolysing sucrose with hydrochloric acid at room temperature, Method First determination Fehling’s solutions A and B (5 or 10 ml of each) were pipetted into a 200-ml Erlenmeyer flask, mixed well by shaking and approximately 25 ml of water were added. The flask, with its contents, was placed on an asbestos wire gauze on a hot-plate and allowed to boil.A sugar solution was added to this boiling mixture while observing the colour of the mixture. When the colour became a very light blue, 1 drop of 0.1% methylene blue solution was added and drops of the sugar solution were added slowly until the colour of the methylene blue disappeared, leaving a colourless solution. Second determination about 98% of the total volume of sugar solution required in the first determination. A second determination was then made by adding to the cold mixture of Fehling’s solution AMarch, 1979 SHORT PAPERS 263 measured amount of water was then added to make the total volume (sugar solution plus water) up to 50 ml and the flask put on the hot-plate, which had previously been adjusted so that boiling started in exactly 3 min, and was continued for 5 min.One drop of 0.1% methylene blue solution was then added and dropwise addition of the sugar solution was continued until the methylene blue colour disappeared. The final addition of sugar solution after the addition of the indicator was completed within 3 min. Final determination A final determination was then made in the same way as in the second determination except that 98% of the total volume of sugar solution required in the second determination was added initially and, after adding 1 drop of the indicator, addition of the sugar solution was continued until the end-point was nearly reached. A calculated amount of water was then added so that the volume of sugar solution or sugar solution plus water added at the final stage remained approximately constant at 1 ml.The determination was then com- pleted by the addition of further amounts of the sugar solution. Results and Discussion By carrying out the determination using the above procedure, it was found that the factor, i.e., the amount of any particular sugar required for the reduction of a fixed volume of Fehling’s solution (either 10 or 20 ml), was always virtually constant, irrespective of the concentration of the sugar solution (Table I). In the Lane and Eynon method, this factor was found to vary with the variation in the concentration of any particular sugar solution being examined. By following this procedure the reducing sugar in an unknown sample can be determined without the use of Lane and Eynon sugar tables.A volume of sugar solution less than 15 ml can also be used, and in these studies volumes as low as 5 ml have been used. By the addition of 98% of the volume of the sugar solution in the cold during the final determination, it was possible to maintain the concentration of any particular reducing sugar, including the concentration of the reaction mixture in the flask, virtually constant, irrespective of the concentration of the sugar solution used. TABLE I FACTORS FOR THE DETERMINATION OF DIFFERENT SUGARS Equal volumes, either 5 or 10 ml, of Fehling’s solution A and B were used. Sugar Invert sugar Glucose . . Fructose . . Maltose . . Lactose . . Concentration range (mg per 100 ml) covered by 10 ml of solution (a) 104-700 . . 210-1 800 .. 200-2 000 104-1 000 .. 210-1 800 108-736 .. 320-4 000 160-2 000 . . 270-3 000 140-1 400 Volume of each Fehling’s solution/ 10 5 10 5 0 5 0 5 0 5 ml ( b ) Factor (a x b/100) A 7 7 Mean 102.30 102.24 102.27 Maximum Minimum 52.74 99.10 51.13 105.48 54.41 160.03 80.01 134.49 67.38 52.68 99.00 51.06 105.40 54.36 159.94 79.92 134.40 67.30 52.71 99.06 51.09 05.42 54.38 60.00 79.96 34.43 67.34 In the determination of invert sugar in the presence of sucrose, it has been found that the factor, i.e., the amount of invert sugar required to reduce a fixed volume of Fehling’s solution, decreases with the increase in the amount of sucrose in the flask (Tables I1 and 111). Experi- ments were performed with different concentrations of invert sugar and with different amounts of sucrose added to the flask, and it was found that the variation of the amount of invert sugar required was due to the variation in the amount of sucrose present (Table 111) but independent of the concentration of the invert sugar itself (Table 11).264 SHORT PAPERS Analyst, Vol.104 For the determination of the invert sugar content of any sample containing sucrose, correction factors for 1 4 g of sucrose present in the flask are given in Table 111. These correction factors were calculated from the observed factor of invert sugar in the presence of sucrose (Table 111) and the true factor for invert sugar determined with pure invert sugar (Table 1). In the determination of invert sugar, the virtual result obtained is to be multi- plied by the corresponding correction factor to obtain the correct results.TABLE I1 EFFECT OF THE PRESENCE OF SUCROSE O:N THE DETERMINATION OF INVERT SUGAR Concentration range of invert sugar Volume of each Amount of Factor used/mg per Number of Fehling's solution sucrose - 7 100 ml test solutions used/ml added/g Maximum Minimum Mean 200-900 5 10 2 96.12 96.05 96.09 120-400 5 5 1 49.20 49.08 49.14 TABLE I11 CORRECTION FACTOR FOR THE DETERMINATION OF INVERT SUGAR I N THE PRESENCE OF SUCROSE Amount of sucrose present/ g 1 2 3 4 1 2 3 4 Volume of Fehling ' s soluticln, A + B/ml Observed factor 10 + 10 98.52 96.09 93.60 91.20 5 + 5 49.14 47.40 45.77 44.10 Correction factor 0.963 0.939 0.915 0.892 0.932 0.899 0.868 0.837 A comparative study of this method with the Lane and Eynon method was made with standard invert sugar solutions of different concentrations and the results are given in Table IV. It can be seen that the precision of the present method is always above 99.90% and is greater than that of the Lane and Eynon method.TABLE IV COMPARATIVE STUDY OF THE LANE AND EYNON METHOD WITH THIS METHOD Concentration of the invert sugar solution used/ mg per 100 ml 120 140 160 200 240 320 Invert sugar content r By Lane and Eynon method/mg per 100 ml 119.30 137.66 159.30 200.98 235.48 314.34 By present method/ mg per 100 ml Error, yo -0.58 119.91 - 1.67 140.01 - 0.44 159.94 +0.49 200.05 - 1.88 240.20 - 1.76 320.31 Error, yo + 0.007 +0.025 + 0.083 + 0.097 - 0.075 -0.038 This work was completed at the Food Science and Technology Unit, National Council for Scientific Research, Lusaka, Zambia. The author thanks Dr.S. Nkunika, Secretary General of the National Council for Scientific Research, Lusaka, for the facilities made available for these studies and Mr. R. T. Shantina for skilful technical assistance. The author also thanks Mr. Sirajul Karim, Assistant Professor of Analytical Chemistry, Bangladesh Agricultural University, for his kind co-operation and technical help.SHORT PAPERS References 265 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Lane, J. H., and Eynon, L., J. SOC. Chem. Ind., Lond., 1923, 42, 32T. Jeslyn, M. A., “Methods in Food Analysis,” Academic Press, New York, 1970, p. 486. Jacobs, M. B., “The Chemical Analysis of Food and Food Products,” Van Nostrand Reinhold, New Pearson, D., “The Chemical Analysis of Food,” Churchill, London, 1970, pp. 129, 130. Porter, D. G., and Sawyer, R., Analyst, 1972, 97, 569. Folin, O., J. Biol. Chem., 1928, 77, 421. Hoffman, W. S., J. Biol. Chem., 1937, 120, 51. Baum, E. H., Ann. N.Y. Acad. Sci., 1960, 87, 894. Fingerhut, B., Ferzola, R., and Marsh, W. H., Clinica Chinz. Acta, 1963, 8, 953. Fuller, K. W., i n “Automation in Analytical Chemistry,” Technicon Symposia, 1965, Mediad Inc., White Plains, N.Y., 1966, p. 78. Fuller, K. W., in “Automation in Analytical Chemistry,” Technicon Symposia, 1966, Volume 11, Mediad Inc., White Plains, N.Y., 1967, p. 57. Sawyer, R., i n “Automation in Analytical Chemistry,” Technicon Symposia, 1967, Volume I, Mediad Inc., White Plains, N.Y., 1968, p. 227. J. R. Nickolls, “Aids to Analysis of Food and Drugs,” Seventh Edition, Ballikre, Tindall and Cox, London, 1952, p. 45. Honvitz, W., Editor, “Official Methods of Analysis of the Association of Official Agricultural Chemists,” Ninth Edition, Association of Official Agricultural Chemists, Washington, D.C., 1960, pp. 426 and 427. Received June 2nd, 1975 Amended December 9th. 1977 Further amended July 24th, 1978 Accepted October 5th, 1978 York, 1958, p. 437.

ISSN:0003-2654    

DOI:10.1039/AN9790400261

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

SHORT PAPERS 265 Determination of Dimetridazole in Pig and Poultry Feeds by High-performance Liquid Chromatography A. D. Jones, I. W. Burns and S. G. Sellings Ureilever Research Laboratory, Colworth House, Sharnbrook, Bedfordshire MK44 ILQ Keywovds : Dimetridazole determination ; animal feeds ; high-performance liquid clzrounatogvaph~y Dimetridazole (1,2-dimethy1-5-nitroimidazole) is an ingredient used in pig and poultry feeds for the control of swine dysentery, and blackhead in turkeys, chickens and game birds. It is also a growth promoter for both pigs and poultry. The usual content in feedstuffs is 50-300 mg k g l . Available analytical methods include polarography,l spectroph~tometry~~~ and thin-layer chr~matography.~ One high-performance liquid chromatographic (HPLC) method has been developed by Buizer and Se~erijnen,~ which, although specific, requires a lengthy analysis time.The aim of this work was to develop an HPLC method that is both specific and rapid as an alternative to the presently available procedures. Experimental Reagents All reagents were of analytical-reagent grade. Dimetridazole. A cetonitrile. Dichloromethane. Pyridine. Hexane. Dried over sodium. Xylene. Dried over sodium. Chloroform. Hexamethyldisilazane. Dococy lrnethy ldichlorosilane .266 SHORT PAPERS Analyst, Vol. 104 Apparatus The liquid chromatograph consisted of a Model 6000 reciprocating pump (Waters Associates Inc., Milford, Mass., USA) and. a Cecil Instruments, Model 212, ultraviolet - visible monitor set at 309 nm. Sample injection was achieved using a Rheodyne injection valve (Magnus Scientific, Stoke-on-Trent, Staffordshire) fitted with a 10-p1 loop.The column was made from seamless stainless-steel tubing (HSCP, Bourne End, Bucks.) of dimensions 15 x 0.49 cm i.d. and was palcked with dococylmethyldichlorosilane (Magnus Scientific) bonded to Partisil 10 (Whatman). Other equipment used included a rotary evaporator, a grinder and a mechanical shaker. Procedure Column preparation Reflux l o g of dry Partisil 10 silica (dried in an oven overnight at l l O O C ) , 12.7g of dococylmethyldichlorosilane and 1 ml of pyridine in 50 ml of dry xylene. After 30 min add 2ml of hexamethyldisilazane (HMDS) and reflux for a further 10min. Filter the phase through a sintered-glass funnel under vacuum and wash with dry xylene (2 x 50 ml) and dry hexane (2 x 50 ml).Dry the phase in an oven at 60 “C for 30 min and then pass it through a 50-pm mesh sieve. Pack the column using the previously described slurry packing procedure.6 Slurry 2.7 g of phase in 30 ml of chloroform and compress the slurry to a compact bed using acetonitrile at a pressure of 3000 lb inA2 for 20 min. Method Grind the sample in a grinder such that all material will pass through a 1-mm mesh. Weigh 5 g of sample containing between 50 and 300 mg kg-1 of dimetridazole into a 100-ml conical flask and extract with 50 ml of acetonitrile - water (2 + 3) by shaking on a mechani- cal shaker for 15 min. If the dimetridazole content is outside these limits the amount of sample should be adjusted accordingly. Filter and inject an aliquot of the filtrate on to the column using a mobile phase of acetonitrile - water (3 + 7) at a flow-rate of 2 ml min-l.Determine the dimetridazole concentration of the extract by reference to a calibration graph. Make up solutions containing 5, 10, 20 and 40 pg ml-1 of dimetridazole in acetonitrile - water (2 + 3). Inject these solutions on to the column and plot a cali- bration graph (which should be linear) with the absorbance values on the ordinate and the corresponding amounts of dimetridazole in micrograms on the abscissa. The dimetridazole content of the feed is given by F Y / M mg k g l , where F pg ml-1 is the concentration of dimetridazole in the extract, Y ml is the volume of extraction solvent and M g is the mass of sample taken.Calibration graph. Calculation. Results and Discussion Interferences A variety of drugs and other ingredients are present in pig and poultry feeds, depending on the type of animal for which they are intended, the age and the possible diseases it may carry. It is possible that one or more of these constituents may elute with the same elution volume as dimetridazole during chromatography, thus preventing quantitation. Ingredients that commonly occur in these types of feed were therefore dissolved in acetonitrile - water (2 + 3) and their retention volumes determined on the column in order to see if they constituted a source of interference. Both furazolidone and 3,5-dinitro-o-toluamide were found to interfere chromatographically with dimetridazole. A feed containing 100 mg kg-1 of furazolidone was found to give an “apparent” dimetridazole level of 4 mg kg-l while 100 mg kg-1 of 3,5-dinitro-o-toluamide gave 9 mg k g l .Fortunately, these two drugs are very rarely found in combination with dimetridazole in a pig or poultry feed. No interference was found from the following ingredients: amprolium, ethopabate, sulphaquinoxaline, decoquinate, robenidine, monensin, halquinol, nitrovin, zinc bacitracin, vitamin A, vitamin D,, vitamin E and copper sulphate. None of the feeds studied contained bentonite (a hardening ingredient), which has been cited in earlier work1 as a probable cause OE low recovery figures.March, 1979 SHORT PAPERS 267 Recovery of Dimetridazole from Feeds and Pre-mixes Dimetridazole was incorporated into pig feed 1 at levels between 45 and 450 mg kg-l by adding the drug to the feed base in dichloromethane, shaking for 30min and then removing the solvent under reduced pressure, The range of feeds were then extracted and analysed by HPLC.In a similar manner the recovery of dimetridazole from a range of different pig and poultry feeds and pre-mixes was determined. Dimetridazole was incorporated into the feeds at the level of 45 mg kg-1 and at 50 g k g l for the pre-mixes. (A pre-mix is generally added to a feed base at the level of approximately 4 kg tonne-l.) Pre-mixes were extracted in a manner identical with that of a feed but the extracts were diluted by a factor of 1000 with acetonitrile - water (2 + 3) just prior to chromatography. Recoveries of between 96 and 109% were obtained for dimetridazole in pig feed 1 (Table I).The blank value for the feed was found to be less than 1 mg k g l and was subtracted before quoting the recovery figures. TABLE I RECOVERY OF DIMETRIDAZOLE FROM PIG FEED Level of dimetridazole Average recovery, additionlmg kg-l Recovery, yo % 45 109, 109 109 110 105, 109 107 220 99, 96 98 340 103, 108 106 450 98, 94 96 A similar range of recoveries was obtained for the different pig and poultry feeds and pre-mixes. The results of this study are presented in Table 11. TABLE I1 RECOVERY OF DIMETRIDAZOLE FROM PIG AND POULTRY FEEDS AND PRE-MIXES Pre-mix or feed type Poultry feed 1 . . Poultry feed 2 . - Poultry feed 3 . . Poultry feed 4 . . Pig feed 1 . . .. Pig feed 2 . . .. Pig feed 3 . . .. Poultry pre-mix 1 .. Poultry pre-mix 2 . . Poultry pre-mix 3 . . Poultry pre-mix 4 . . .. .. .. .. Pig pre-mix 1 . . .. Average recovery, Recovery, yo % .. 101, 106 104 . . 99, 99 99 . . 94, 99 97 . . 96, 96 96 .. 102, 102 102 . . 97, 97 97 . . 92, 92 92 .. 110,102 106 .. 103, 101 102 .. 100, 103 102 . . 97, 99 98 .. 97, 97 97 Comparison of Methods The results obtained by the HPLC method were compared with those obtained by a spectrophotometric procedure developed by Stone and H ~ b s o n . ~ For this study a range of commercially available feeds were used that had theoretical dimetridazole levels of 200 mg k g l . The HPLC results on finished feeds were found to be consistently lower than those obtained by spectrophotometry although the agreement in general was good (Table 111).As occasional changes occur in the composition of feeding stuffs, it is possible that the higher results obtained by using the spectrophotometric procedure are a result of interference from a co-extractive.268 SHORT PAPERS TABLE 111 DETERMINATION OF DIMETRIDAZOLE IN COMMERCIAL FEEDS BY HPLC AND BY A SPECTROPHOTOMETRIC METHOD Dimetridazole content/mg kg-l A r \ HPLC method Spectrophotometric method Feed Individual results Average Individual results Average 1 158, 158 158 173, 173 173 2 283, 291 ,287 300, 300 300 3 190,190 190 203, 215 209 4 125, 121 123 139, 135 137 6 174, 174 174 201, 194 197 --- A \ A typical chromatogram of a feed extract is shown in Fig. 1. The elution time for dimetridazole is approximately 2 min, which permits the analysis of approximately 30 samples (in duplicate) in one day by one experienced technical assistant. ect Time/min Fig. 1. Determination of dimetridazole in a commercial pig feed. Sol- vent, acetonitrile - water (3 + 7); flow-rate, 2 ml min-l; detector, ultra- violet at 309 nm; sensi- tivity, 0.05 a.u.f.s.; injec- tion volume, 10 p1 (loop). References 1. 2. 3. 4. 5. 6. Analytical Methods Committee, Analyst, 197 1, 96, 746. Analytical Methods Committee, Analyst, 1969, 94, 925. Stone, L. R., and Hobson, D. L., J . Ass. 08. Analyt. Chew., 1974, 57, 343. Hammond, P. W., and Weston, R. E., Analyst, 1969, 94, 921. Buizer, F. G., and Severijnen, M., Analyst, 1975, 100, 854. Jones, A. D., Burns, I. W., Sellings, S. G., and Cox, J. A,, J . Chromat., 1977, 144, 169. Received September llth, 1978 Accepted October 17th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400265

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

Analyst, March, 1979 Book Reviews 269 TOPICS IN ENZYME AND FERMENTATION BIOTECHNOLOGY. Volume 2. Edited by ALAN WISEMAN. Pp. 308. Chichester : Ellis Horwood. Distributed by John Wiley, Chichester. 1978. Price Ll5. This volume contains reviews of unequal length contributed, with one exception, by authors who have already appeared in the first volume and in its precursor, the “Handbook of Enzyme Biotechnolog y . ” The exception is a very comprehensive review by Kent, Rosevear and Thompson of “Enzymes Immobilised on Inorganic Supports.” The important features of the physical form of the matrix and the preparation and properties of the immobilised enzymes are clearly described and illustrated by numerous examples. The engineering aspects of batch and continuous enzyme reactors using such catalysts are discussed including a section on reactor modelling and the review concludes with a useful summary of recent work on the industrial, medical and analytical applications of such catalysts.This introduction should be of great help to anyone wishing to utilise immobilised enzyme technology efficiently. A short article on “Stabilisation of Enzymes” by Wiseman should be read in conjunction with this review and collects a great deal of scattered information on the use of additives, the effects of immobilisation and of crosslinking. Enzyme electrodes have now come of age and are effectively reviewed by Barker and Somers and many examples are given based on electrochemical sensors. However, possibly the most significant non-electroactive sensor has resulted from the development of the enzyme thermistor in which the heat of the reaction can be registered and this is only briefly mentioned.Antibiotic resistance is an ever present problem and the study of the mechanism of resistance has led to the discovery and design of more effective antibacterial agents. Apart from permeability changes the production of antibiotic inactivating enzymes is an important mechanism and three therapeutically important types of enzymes, those that inactivate the /I-lactams, the amino- glycosides and chloramphenicol, are reviewed by Melling. The second longest review is that by Winkler and Thomas on the “Biological Treatment of Aqueous Wastes.” The problems of oxygen transfer and desludging are discussed and the theoreti- cal and practical design and operation of biological film systems for waste treatment and the activated sludge process are illustrated and described.The anaerobic treatment of waste and the removal of nitrogen, a major pollutant of water, are more briefly but admirably covered. This one review justifies the purchase of the book. If one accepts the proposition that editors of books bringing together reviews have a responsi- bility to select articles that summarise existing knowledge but also look to the future to provide the reader with the vital pointers and stimulus for future research then I found the present volume lacking the latter. The summaries of existing knowledge have been done conscientiously and well and will be valuable to those looking for a general introduction to a topic, or the general reader.I. D. FLEMING FUNDAMENTAL RESEARCH IN HOMOGENEOUS CATALYSIS. Edited by MINOURU TSUTSUI and RENATO UGO. Proceedings of the First International Workshop 0% Fundamental Research in Homogeneous Catalysis held at Santa Fluvia, Italy, December, 1976. Pp. x $. 242. New York and London: Plenum Press. 1977. Price $33. The nature of the workshop or conference described in the title was unusual and the consequent presentation in this volume is also unusual. However, it provides an illuminating insight into the use of inorganic compounds as catalysts in homogeneous organic reactions connected with developments in the use of energy resources. It must be stressed that the book has no contact with analytical chemistry or analysis.Part I contains the texts of eight invited lectures, all of which review important areas of homogeneous catalysis such as the activation of oxygen and nitrogen, the use of metal clusters and the application of catalysts supported on the surface of polymers or silica based materials, In the second part of the conference, delegates divided into five working groups to discuss current developments on different topics and to define areas in which further research would produce beneficial results. In each instance, a report was produced270 BOOK REVIEWS Analyst, Vol. 104 by the chairman, discussed further by the complete conference and subsequently published with relevant recommendations. As someone with only a peripheral interest in the subject, I found these reports exciting in both style and content.The working groups covered hydrocarbon conversion, homogeneous selective oxidation, carbon monoxide reactions, frontier areas between homogeneous and heterogeneous catalysis and catalytic processes. To anyone with an interest in the theory or industrial applications of homogeneous catalysts this book can be highly recom- mended. J. M. OTTAWAY HANDBUCH DER ANALYTISCHEN CHEMIE. DRITTER TEIL. QUANTITATIVE BESTIMMUNGS- UND TRENNUNGSMETHODEN. Band 6by. ELEMENTS DER SECHSTEN NEBENGRUPPE . WOL- FRAM. By G. WUNSCH. Pp. xiv + 286. Berlin, Heidelberg and New York: Springer- Verlag. 1978. Price DM146; $73. Tungsten is an element that most analysts do not encounter often, so that detailed methods of determination, when they aye required, are often not to hand.The present text is particularly valuable in this respect, in that it provides a comprehensive selection of complete analytical procedures, applicable to numerous matrices, and using all of the commonly encountered classical and instrumental techniques. The book also describes methods of dissolution of tungsten and its compounds, a wide variety of separation procedures and some special topics such as hetero- polytungstic acids, tungsten bronzes, interactions with periodate ions, cluster compounds and cyanotungstate ions. The book continues the high standard of previous volumes in this series and should be available in all non-routine analytical laboratories. A. TOWNSHEND HANDBUCH DER ANALYTISCHEN CHEMIE. DRITTER TEIL. QUANTITATIVE ANALYSE.Band Pay. ELEMENTE DER VIERTEN HAUPTGRUPPE. 21”. (In English.) By J, W. PRICE and R. SMITH. Pp. xv + 262. Berlin, Heidelberg and New York: Springer-Verlag. 1978. Price DM146; $73. I t is reasonable to assume that at some time or other most analysts have been involved in the determination of tin, especially at “low” levels, and most would agree that each type of sample has its own, often unpredictable, analytical problems. Even when the metal is present as a major constituent, its determination is by no means straightforward, and a consensus of opinion is likely to reveal that of all the metals that are regularly determined at “high” levels, up to loo%, the determination of tin is the most difficult; that has certainly been my experience. Titrimetric determination of the metal is applicable over a very wide range, from the high levels just indicated, down to about O.l%, and the method is usually, though not always, based on the formation of Sn( 11) followed by iodimetric re-oxidation.Other titrimetric procedures that do not involve this preliminary reduction tend to be less reliable, for a variety of reasons. Essentially, therefore, only two apparently simple basic stages are involved in this titrimetric method, yet I cannot cite any other analytical procedure that is more liable to give such erroneous (often reproducible) results than this one. So much for the more important aspects of the unusual analytical proclivities of tin that are so admirably dealt with in appropriate chapters of this excellent, up-to-date and timely publica- tion.Throughout the book, sampling, a very important precursor to most analytical procedures, receives due attention, especially in, for example, such chapters as “Analysis of Tin Ores and Concentrates” and “Analysis of Secondary Materials and Intermediates.” Other chapter headings include “Gravimetric Methods,” “Photometric Methods,” “Electro- chemical Methods,” “Solvent Extraction,” “Atomic-Absorption Spectroscopy,” “Emission Spectroscopy, ” “X-ray Fluorescence,” “Radiochemical and Mossbauer Methods,” “Analysis of Ingot Tin,” “Tinplate,” “Organotin Compounds” and “Tin and Tin-Alloy Electroplating Solutions”; the final chapter (21) deals with “Tin Chemicals.” Apart from its fly-cover, the book is presented entirely in English-a new departure in the series ? The authors make no claim to have catalogued all of the work published on the related subject matter; their objective has been to present a reliable account of currently accepted practice in various associated fields, influenced, and rightly so, by their personal preferences in the light of their combined long and wide experience.March, 1979 BOOK REVIEWS 271 The book has all the attributes of an authoritative publication and can almost be guaranteed a prominent place at the right hand of anyone who has an occasional or regular need to determine tin.W. T. ELWELL PRACTICE OF THIN LAYER CHROMATOGRAPHY. By JOSEPH C. TOUCHSTONE and MURRELL F. New York, Chichester, Brisbane and Toronto: John Wiley. DOBBINS. Pp. xxiv + 383. 1978. Price j514.05; $27.30.The main difference between this book and the many other published volumes on the practice of thin-layer chromatography is that here the authors aim to provide details of all the basic techniques needed to apply this form of chromatography as it is used currently in biochemical and pharmaceutical research and quality control. There is no doubt that the language used is very easy to understand and to follow and, on the assumption, presumably, that the reader will wish to read the book from beginning to end before commencing practical work (as recommended by the authors), there is a glossary of terms at the beginning rather than at the end of the book. It deals with the basic description and uses of the technique, the preparation of the plates (including a chapter on commercial pre-coated plates), preparation and application of the sample, the mobile phase, development techniques, detection procedures, documentation, quantitation and reproducibility.Each of these subjects is dealt with thoroughly. In addition, there are chapters covering radioactive procedures, preparative thin-layer chromatography and the combination of this with the other analytical techniques of column and gas chromatography, mass spectrometry and infrared spectroscopy. The special techniques of vapour-programmed development, radial chromatography, hot-plate and programmed multiple development are mentioned, as also are pyrolysis TLC, bioautography and enzyme inhibition techniques. Brief mention is also made of autotransfer chromatography and the combination of differential thermal analysis and of photoacoustic spectrometry.While there is not a separate section on applications, these are referred to throughout the book, including amino acids, alkaloids, steroids, pesticides and antibiotics. Possibly more applications would have been advantageous, but for ease of following the instructions this method scores highly. If the reviewer had to recommend three books to a student about to work in this field, this would be one of them. The volume has 14 chapters and over 400 references. D. SIMPSON DIOXIN : TOXICOLOGICAL AND CHEMICAL ASPECTS. Edited by FLAMINIO CATTABENI, ALDO Monografihs of the Giovanni Lorenzini Foundation, New York and London: SP Medical and Scientific Books. CAVALLARO and GIOVANNI GALLI. Volume 1.Distributed by Halsted Press, New York. 1978. Price j514.50; $26.50. Pp. xiv + 222. This book represents a collection of 21 papers presented at a 2-day workshop on dioxin held in Milan, Italy, in October, 1976. It has been organised into three broad sections, viz., Chemistry and Analysis ; Toxicology ; and Decontamination. There are original contributions from many eminent scientists in each of these areas. The book starts with an introductory review of the events that occurred a t Seveso in July, 1976. Facts are presented on the geographic extent of the pollution and on numbers of persons living in the most contaminated zones. This section also deals with some of the chemistry involved in the formation of dioxin and tabulates previous industrial accidents.There are several excellent chapters on the analysis of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) using gas chromatography - mass spectrometry (GC - MS). The approaches range from packed-column gas chromatography with low-resolution mass spectrometry through to high- efficiency capillary column gas chromatography combined with high-resolution mass spectro- metry. In this section I found the chapter by Hans-Rudolf Buser, illustrating the use of capillary GC - MS, particularly impressive. The need for maximum sensitivity and specificity in TCDD analysis is a clear message from this section of the book. For those unfamiliar with GC-MS there are a number of simple introductory treatments with definitions of important terms such as mass fragmentography. Isolation and clean-up procedures are covered in this section and also the analysis of closely related and highly toxic chlorinated dibenzofurans, which are shown to be present in many commercial polychlorinated biphenyl formulations.272 BOOK REVIEWS Analyst, Vol.104 The toxicology section brings home the disturbing fact that although many penetrating studies have been carried out on TCDD, the reason for its incredible toxicity remains somewhat of a mystery. Liver enzyme induction measurements are used by several authors as an index of potency and are used to illustrate the vast differences in toxicity between the various halogenated dibenzodioxin isomers. A useful review chapter lists and summarises the findings of many of the animal studies on TCDD that have appeared in the literature.The final section deals with the problem of decontamination, with particular reference to the Seveso incident. The encouraging results of the field trials described offer a ray of hope for the future. In general, I found this an interesting and worthwhile book, which contains many examples of elegant analytical work. This is undoubtedly the consequence of independently prepared papers, but I believe that the Editors should have been able to remove some of this overlap. The book also might have benefitted from the inclusion of some of the discussion that presumably took place at the workshop. However, i t suffers from a significant amount of repetition. G. T. STEEL EMISSION SPECTROCHEMICAL ANALYSIS. By TIBOR TOROK, JOZSEF MIKA and ERNO GEGUS. Pp.692. Bristol: Adam Hilger. Budapest: Akadhmiai Kiad6. 1975. Price k31.50. The authors of this text have made very considerable efforts to include between its covers informa- tion relating to all aspects of the practice of emission spectrochemical analysis. They must be congratulated on their systematic treatment of the basic principles of sampling and sample prepara- tion, sources, conditions affecting line intensities, semi-quantitative and quantitative spectrographic analysis, equipment installation, safety and economics and evaluation of spectrographic data. The book is augmented with tables containing physico-chemical data for spectrochemical analysis, instructions and descriptions of procedures for qualitative, semi-quantitative and quantitative analysis, conversion tables and numerical examples. The authors indicate in the introduction that they were aware of the difficulties involved in attempting to produce a comprehensive treatise devoted to a subject that is undergoing rapid development ; nevertheless, the book contains a wealth of useful information and data that will remain relevant to practising spectroscopists well into the future. This volume can be recommended as a companion text to the earlier book by the same authors (“Analytical Emission Spectroscopy-Fundamentals,” Akademiai Kiad6, Budapest, and Butter- worths, London, 1973). I t should take its place with other volumes concerned with practical emis- sion spectrochemical analysis on the shelves of libraries of reference. The book is well written and translated ; unfortunately, however, the binding and production quality of the copy in the hands of the present reviewer left much to be desired in such a high-priced text. G. F. KIRKBRIGHT

ISSN:0003-2654    

DOI:10.1039/AN9790400269

出版商:RSC    

年代:1979

数据来源: RSC