ISSN:0003-2654    

DOI:10.1039/AN98005FX017

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BX019

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005FP053

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BP059

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

May 1980 The Analyst Vol. 105 No. 1250 Limiting Non-Nernstian Calibrations of Ion -selective Electrodes Derek M idg ley Central Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey, KT22 7SE The relationship between the e.m.f., E , of an ion-selective electrode and the concentration, C, of deterininand ceases to be logarithmic when C is sinall and eventually a limit is reached in which the relationship is linear; E = q + rC. The values of q and Y are derived tlieoretically for the cases where the cause of the deviation from the logarithmic relationship is (2) the solubility of a salt incorporated in the electrode itself or (ii) the presence of determinand in reagents added to the sample solution. In case ( i ) , y = E o + hlnH and Y = y k / ( x f y ) H , where H = [(x/y)~K]l/(~ + v ) , h = IiT/zF being the Nernst slope factor and K the solubility product of the salt A,B, incorporated in the A-selective electrode. Allowance can be made for the effect of a com- plexing agent on the solubility. In case ( i i ) , when the reagent blank, b, is large compared with both the determinand concentration and the solubility of the electrode components, y = E o + hlnb and Y = h/b.Keywords : I on-selective electrodes ; calibration ; Yeagent-Flank Earlier papers have dealt with how the deviations of ion-selective electrodes from the Nernstian response may be interpreted1 and how they influence the limit of detectione2 This paper is concerned with the region of limiting response, where it will be shown that in many instances the e.m.f. is directly proportional to the concentration when the concentration is low enough.At constant ionic strength, the e.1n.f. of an ion-selective electrode is described by a form of the Nernst equation: E = Eo + Kln[A] , . . . . . . . * . (1) where E is the observed e.m.f., Eo is a constant (the standard potential), [A] is the con- centration of the ion to which the electrode responds and K = RTIxF is the Nernstian slope factor, R being the gas constant, T the absolute temperature, F Faraday’s constant and z the charge (with sign) of the ion A. In analytical methods using ion-selective electrodes, the calibration generally takes the same form as equation (1) : .. .. - ( 2 ) E = Eo + KlnC . . . . where C is the nominal concentration of determinand in the sample solution. Because of the solubility of the materials of the electrode and the presence of a reagent blank or inter- ference, C and [A] cannot be equated when C is small and equation ( Z ) , therefore, no longer describes the relationship between e.m.f.and concentration. Deviation from equation (2) does not preclude the electrode from being used; Tonilinson and Torrance3 and Marshall and Midgley4 have determined chloride with, respectively, silver chloride and mercury( I) chloride electrodes operating in the non-Nernstian response region. In these instances the e.1n.f. was directly proportional to the determinand concentration (at very low values of C): .. . . - . (3) E = q + K .. .. where q and Y are constants. When the limiting response of an electrode is determined by the solubility of a salt with 1 : 1 stoicheiometry, e.g., silver chloride, the e.m.f. can be explicitly expressed in terms of C 417418 MIDGLEY LIMITING NON-NERNSTIAN Analyst, Vol.105 and the solubility product, and hence q and Y can be derived fairly imply.^ This paper presents a more general theory in which q and Y, for electrodes whose responses are limited by the solubility of binary salts, can be calculated for salts of any stoicheiometry. The limiting response determined by solubility is d.istinguishecl from limits set by the conditions in which the electrode is used, such as the presence of a reagent blank or interference. The theory is tested experimentally against measiirements made with a silver - silver sulphate electrode. Theoretical In the following derivations it is assumed that the ionic strength is constant (as is usual for analysis with ion-selective electrodes) and that concentrations can be used where activities would be strictly correct.Con- centrations refer to the solution in which the electrode is immersed, i.e.:, after sample treat- ment in an analytical procedure. Equation (1) is assumed to be valid a t all concentrations. Response Limited by Solubility Product of Electrode Materials Consider an electrode incorporating a salt A,B, and responding to the determinand A. When the electrode is immersed in a solution containing a concentration C moll-1 of A, the concentration, s mol l-l, of A dissolving from the electrode is given by the solubility product .. * - (4) K = [A:/" [B]Y .. . . . . As [A] = C + s and [B] = sy/x, we have K = (C + s)"e)' Differentiating with respect to C, we obtain (;)?!$ = (x + y) s"+v--l(l + ;).;s + s"+Yx ( 1 + - c)""- - c x S) S2 ac S . . - * (6) = 0 (as K is a constant) . . . . Hence, Theref ore, X In the limiting condition, C < s and, therefore, aslac m - ____ x -I- y On integration, we obtain xc X+Y s = H - - . . . . .. where H is an integration constant.May, 1980 CALIBRATIONS OF ION-SELECTIVE ELECTRODES 41 9 When C = 0, H = s, i.e., H is the concentration of ion A arising from equilibrium solubility From equation of the salt A,B, in an ionic medium containing neither ion A nor ion B. (5) with C = 0, therefore, H Substituting for [A] in equation (1) : E=EO+ = Eo + = Eo + L\y/ J Kln(C + s) klnH + kln [ 1 + yc .. .. (x + Y)H Because yC << (x + y ) H , we can expand the final term in equation (11) and, with neglect of higher terms, we obtain the approximate equation E w Eo + klnH + ky x c . . . . . . . (12) (x + Y ) H From equations (3) and (12), therefore, q = Eo + klnH . . . . . . . . . . (13) .. .. . . . . . . (14) y =.-- kY (x + Y)H The constants q and Y in equation (3) can, therefore, be expressed in terms of the constants in equation ( 1 ) and the solubility product and stoicheiometry of the sparingly soluble salt incorporated in the electrode. Curves Q, A and S in Fig. 1 show how the limiting calibration graphs of hypothetical electrodes that consist of salts of equal solubility are influenced by the stoicheiometry of the salt and the charge, z , on the determinand.The value of H is the same for all electrodes and values of Eo were assigned so that for all electrodes E = 0 when C = 0 and there is no reagent blank (see below). The limiting slope of curve S is half that of curve Q, even though the stoicheiometries and solubility products are the same in each instance, because K in equation (4) depends on the charge on the determinand. Curves Q and R show for univalent determinands the difference between electrodes comprising salts of equal solubility but different stoicheiometry (1 : 1 and 2: 1 ) . The electrode, R, with salt containing the divalent counter ion has the lower calibration slope. Conversely, for divalent determinands, an electrode (S) consisting of a salt of 1 : 1 stoicheio- metry has a lower calibration slope than one consisting of a 1 : 2 salt of equal solubility (not shown, but with a limiting calibration slope coincident with that for electrode R).Efect of the reagent blank The presence of determinand in any reagents added to the sample solution will affect the response of the electrode. Let the concentration of determinand originating in the reagents be b,.. If C + b, < s, it can be shown, by repeating the above derivation with the variable M = C + b, substituted for C , that equations (10) and (14) are still valid but that the value of q is now given by q = Eo + klnH + rb, . . .. .. . . (15)420 YIIDGLEI' : LIhlITING NON-NERNSTIAN Analyst, Vol. 105 i.e., the reagent blank does not change the slope of the calibration graph but shifts it along the response axis, as is usual with methods in which the measured quantity is linearly related to the concentration (cf., curves P and Q in Fig.1 ) . If b, m s, the derivation is not valid and equaltion (12) does not apply (the calibration not being linear). If b,, > s, however, another type of linear calibration may be obtained (see below). I P 7 6 5 > E y: Y 4 € 3 w 2 1 0 5 10 Concentration x 1 o 7 h o i I - - ' Fig. 1. Limiting calibrations for electrodes with H = 2 x mol 1-1: ( P ) E" = 337.15 mV, z = 1, x = y = 1, I< = 4 x 10-12 mo12 1-2, b, = 1 x lo-' rnol 1-l; ( Q ) E" = 387.15 niV, z = 1, x = y = 1, 137 = 4 x 10-l2 mo12 1-2, b, = 0 ; (R) Eo = 337.15 mV, z = 1, x r= 2, y =7 1, li = 4 x 10-1* mo13 1-3, b, = 0 ; ( S ) E" = 168.57 niV, z == 2, x = y = 1, I< = 4 'X 1 0 k 1 2 mo12 1-2, b, = 0.Points represent the e.m.f.s corresponding to equation (1) and lines the limiting response:; given by equa- tion ( 3 ) . Eflect of coinplexing agents would otherwise be the case. the stability constant of the complex it forms w!th the determinand. centrations of determinand are related by the equation In the presence of a complexing agent, iiiorc material will dissolve from the electrode than Let 1x1 be the concentration of the complexing agent and p Provided [XI > C 3- s, the free and total con- Conzplex formed with the determinand C + s = [A](] + p[IX]) . . . . . . . . (16) The solubility product is Hence (C + S ) p y = K(1 -4- p[x])z = L,, , . . . . . (18) X Under the conditions stated, LA, is effectively constant an& the derivation can be carried out as a b o i ~ , using LA, instead of K .We would obtain equations (13) and (14) exactly asMay, 1980 CALIBRATIONS O F ION-SELECTIVE ELECTRODES before, but H would be given by the equation 1 42 1 ComjXex formed with the determinand’s counter i0.12. Provided [XI > s, the free and total concentrations of the counter ion, B, are related by the equation F = [ B ] ( l + f i [ X ] ) x . . .. .. .. The solubility product is . . . . . . (21) Hence (C + s)” (;)” = K (1 + p [ X q = L,, . . . . . . (22) L,, being effectively constant, the derivation using I,,, instead of K would again give equations (13) and (la), with H defined by the equation 1 Applicability of the theory (i) Equations (1)-(23) are valid whether ion A is an anion or a cation, the sign of z deter- mining whether the e.m.f.increases with concentration (cations) or decreases (anions). (ii) The theory does not depend on the form of construction of the electrode, e.g., for a chloride-selective electrode based on silver chloride, q and Y would be the same whether the electrode was of the silver - silver chloride electrode type or had either a single-crystal silver chloride membrane or a mixed-crystal silver sulphide - silver chloride membrane. Midgleyl has gathered data that show the equivalence of different forms of chloride electrode. The values of q and Y are determined by the most soluble component of the electrode. For example, in a chloride electrode incorporating silver sulphide (Ag,S) and silver chloride (AgCl), the latter is the more soluble and so q and Y are calculated for 1 : 1 stoicheiometry with K = KA,,,.The theory is not valid if the solubility of the electrode material is governed not by the solubility product, but by a chemical reaction, e.g., the copper(I1) sulphide in copper- selective membrane electrodes dissolves as a result of oxidation rather than through a solubility product mechanism. The theory is not valid unless solubility of the electrode material is the principal cause of non-Nernstian behaviour. If the reagent blank is large compared with the solubility, a linear non-Nernstian response with different characteristics will be obtained (see below). If the reagent blank is comparable in magnitude to the solubility there will be no such linear response region. (iii) (iv) (71) Response Limited by Reagent Blanks The limits to sensitivity discussed above arise primarily from the mechanism of the electrode itself and cannot be avoided under the specified conditions.These limits, however, may not be reached because the necessary conditions have not been attained. If the concentration of determinand introduced with the reagent is large compared with the concentration of determinand dissolved from the membrane, the assumptions involved in deriving equations (8)-(15) are invalid. When b, 3 s, we obtain on substitution in equation (1)422 Analyst, Vol. 105 MIDGLEY : LIMITING NON-NERNSTIAN E = Eo + kln(C + b,) = Eo + klnb, + kln . . .. . . (24) In the region of the limiting response, C < b, and by expansion of ln(1 $- Clb,), with neglect of terms beyond the first, equation (24) can be approximated by .. (25) . . . . k E w Eo + klnb, + - x C . . br Equation (25) represents a linear relationship between e.m.f. and concentration, as shown in Fig. 2 for hypothetical calibration graphs calculated from equation (24). The limiting calibrations calculated from equation (25), represented by solid lines, become valid approxi- mately when b,. > 1OC and show changes in both slope and intercept as b, varies, in contrast to the case where b, < s and changes in b, affect the intercept only. 0 5 10 Concentration/mg I-' Fig. 2 . Univalent elec- trode responses limited by reagent blanks. Lines repre- sent limiting resFionses calcu- lated from equation (25) with E" = 0 mV and points are calculated from equation (24) ., A, b = 100 mg 1-1; B, b = 25 mgl-I; and C, b = 10 mg 1-I. Because the conditional type of limiting sensitivity is independent of the solubility product, equation (25) can be generalised to cover a wider range of circumstances, such as the effect of interferences.Provided that b >> s and b > C, . . . . . . (26) k b E w Eo + klnb + - x C . . where b = b, + Xbi and b, is the apparent determinand concentration arising from the presence of the ith interfering substance introduced with the reagents. Each term bi can be expressed as the product of the concentration of the ith iiiterferent and its selectivity coefficient. In the event of a linear calibration being obtained, it is possible to tell whether it is of the conditional type represented by equation (26) or of the more fundamental type definedMay, 1980 CALIBRATIONS OF ION-SELECTIVE ELECTRODES 423 by the parameters of equations (lo), (13) and (14) by comparing the intercept and slope of the calibration graph with those predicted by equations (13) and (14).If b is large, the intercept will be considerably more positive (for cationic determinands) or negative (for anionic deterniinands) than predicted hy equation (13). In addition, the calibrations can be analysed using the function plots described by Midg1ey.l I t may be noted that equation (26), being independent of the mechanism of the electrode, applies to any type of electrode provided that the conditions are such that b > C and the solubility of the electrode material is negligible.Thus, equation (26) may often describe the limiting response of ion-exchange and glass electrodes, which are more susceptible than solid-st ate electrodes to interferences. Experimental The applicability of the above theory depends on being able to measure small changes in e.m.f. in the region of the limiting response and to keep the temperature sufficiently constant that changes in e.m.f. are caused by changes in the determinand concentration rather than by changes in solubility. As discussed below, results from the literature show that electrodes comprising a sparingly soluble uni - univalent salt (silver chloride) behave according to theory. It was desired to test the theory further by using electrodes incorporating salts that had a stoicheiometry other than 1 : 1 and a moderately high solubility.The silver - silver sulphate electrode fulfils both requirements, as silver sulphate has a solubility of 0.0267 mol kg-l at 25 "C, which is much higher than would normally be considered suitable for an ion-selective electrode. Apparatus The electrodes were a 0.04-in diameter silver wire (Johnson, Matthey and Co.) and a Radiometer K4025 saturated calomel electrode connected to the test solution through a 1 mol 1-1 potassium nitrate salt bridge with a ceramic frit junction. Potentials were measured with a Radiometer PHM 64 digital pH meter. The test solution was contained in a water-jacketed titration vessel and agitated with a mechanically driven stirrer (Radiometer TTA 60 titration assembly).The temperature of the solution was maiiitained at 25.0 & 0.05 "C by passing water from a Techne C-100 circu- lator through the water-jacket of the titration vessel. Reagents (BDH Chemicals). Silver sulphate, sodium sulphate and potassium nitrate were all of analytical-reagent grade Procedure solid silver sulphate was added to produce a saturated solution. immersed in the solution and allowed to attain a steady e.m.f. meter syringe and the e.m.f. was noted once it had reached a steady value. concentration. A 20-ml volume of de-ionised water was pipetted into the titration vessel and enough The electrodes were Portions of 0.956 niol 1-1 of sodium sulphate solution were added from an Agla micro- The values of q and Y were obtained from a least-squares fit of the graph of e.m.f.against Results The results from four titrations are shown in Table I. The theoretical values are derived from equations (lo), (13) and (14) using the standard potentials for silver and saturated calomel electrodes6 and the solubility product of silver s ~ l p h a t e . ~ The standard deviations quoted for the theoretical values of q and Y represent the spread of four results calculated from the solubility products obtained by Marshall and Slusher in two different ionic media together with two methods of calculating activity coefficient^.^ $8 No allowance has been made for liquid junction potentials (<1 mV) or for the variation of up to 2 mV found in the standard potentials of commercially produced saturated calomel electrodes.424 MIDGLEY LIMITING NON-NERNSTIAN Analyst, Vol.105 Discussion Within experimental error, the theory has been verified for an electrode with a moderately soluble non-isovalent salt (silver sulphate), as the results in Table I show. Further con- firmation of the theory comes from the results of Tomlinson and Torrance3 and Florence9 for electrodes incorporating a sparingly soluble j sovalent salt (silver chloride). The limiting sensitivities, Y, calculated for the conditions obtained in their experiments are in good agreement with those observed (Table 11); values of q cannot be calculated from these data. As predicted, the theory works equally well for Tomlinson and Torrance's silver - silver chloride electrode and for Florence's silver chloride - silver sulphide membrane electrode.It should be noted that when x = y in equations (10) and (14) the general theory reduces to the equations that can be derived for electrodes incorporating salts of 1 : 1 stoicheiometry, for which the e.m.f. can be expressed explicitly in terms of the concentration of determinand and the solubility p r o d u ~ t . ~ TABLE I RESULTS FOR THE SULPHATE RESPONSE OF A SILVER - SILVER SULPHATE ELECTRODE AT 25 "C 4/mV r/mV mol-l 1 Titration 1 . . .. . . 468 337 Titration 2 . . . . . . 47 1 334 Titration 3 . . . . . . 477 303 Titration 4 . . . . . . 478 312 Theoretical . . . . . . 476.5 & 1.3 335 f 14 As the response slope can be theoretically predicted, the experimentally observed slope indicates whether factors other than solubility equilibria are influencing the results.This test is additional to those devised by Midgleyl and, while not providing as much information, involves much less calculation. Being able to predict limiting linear responses should encourage the use of ion-selective electrodes outside the Nernstian response region at lower concentrations than are often considered practicable, as has been done for trace chloride analysis3$* and for a number of other determinands that form sparingly soluble silver salts.1° There are few sparingly soluble salts suitable for making electrodes that have Nernstian responses at concentrations below about moll-l; the most soluble salt used at present in commercially produced solid-state electrodes is silver chloride, with a solubility of about moll-l. If measure- ments can be made reliably in the non-Nernstian region, electrodes incorporating salts of moderate solubility ( 10-3-10-4 moll-l) may become practicable.Examples of species for which it may be possible to make useful electrodes working on the above basis are sulphate, chromate and iodate. TABLE I1 RESULTS FOR CHLORIDE-SELECTIVE ELECTRODES Temperature/ Observed Y/ Theoretical Y/ Electrode material "C mV mg-l 1 mV mg-l 1 Reference Ag-AgCl . . . . 10 35.4 35.5 3 Ag- AgC1 . . . . 25 22.6 22.6 3 Ag,S - AgCl . . .. 25 2 2 22.6 9 If conditions can be controlled so that the precision of measurement of e.m.f. is sufficient to allow the region of linear non-Nernstian response to be used, calculation of the limit of detection is simpler than in the general case for potentiometric analysis.2 Taking Roos's definition of limit of detection,ll L , for 95% confidence: L = 4.650, .. .. .. .. . . (27)May, 1980 CALIBRATIONS O F ION-SELECTIVE ELECTRODES 425 where uB is the standard deviation of the blank in concentration units, we obtain in the linear region L = 4.65~v/l~l . . . . .. . . . . (28) where uv is the standard deviation of the blank in e.m.f. units and Y is the calibration slope. For an electrode whose non-Nernstian behaviour is caused primarily by solubility product effects, we obtain from equations (lo), (14) and (28) 1 If the limit of detection occurs above the linear response range, a different expression corresponding to equation (29) is required for each stoicheiometry with and without allowance for reagent blanks.2 If the reagent blank is the only significant cause of non-Nernstian behaviour, we obtain from equation (26) Y = k / b and hence, from equation (28), L = 4.65avb/(kl . . . . . . . . . . (30) Because this type of non-Nernstian behaviour is independent of the mechanism of the electrode, only one form of equation is needed even if the calibration is non-linear.2 Expressing equation (6’) of reference 2 in the notation of this paper: L = b[antilog(4.65av/(kl) - 11 . . . . .. . . (31) Equation (30) can also be derived as a limiting case of equation (31) by using the relationship antilog (x) m 1 $- x for 0 < x << 1, where x = 4.65 a,/lkl. 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. 4. 5. 6. 7. 8. 9. 10. 11. Midgley, D., Anal. Chem., 1977, 49, 1211. Midgley, D., Analyst, 1979, 104, 248. Tomlinson, I<., and Torrance, K., Analyst, 1977, 102, 1. Marshall, G. B., and Midgley, D., Analyst, 1979, 104, 55. Midgley, D., Anal. Chim. ,4cta, 1976, 87, 19. Ives, D. J . G., and Janz, G. J., “Reference Electrodcs, Theory and Practice,” Academic Press, Marshall, W. L., and Slusher, R., J . Inorg. Nucl. Chem., 1976, 38, 279. Davies, C. W., “Ion Association,” Butterworths, London, 1962, p. 41. Florence, T. M., J . Electvoanal. Chem. Interfacial Electrochem., 1977, 31, 77. Bardin, V. V., Shartukov, 0. F., and Tolstousov, V. N., Zh. Anal. Khim., 1971, 27, 25. Roos, J . B., Analyst, 1962, 87, 832. New York and London, 1961, p. 393. Received October 22nd, 1979 Accepted November 23rd, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500417

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

426 Analyst, May, 1980, Vol. 105, p p . 426-431 Effect of Anionic Surfactarrts on Calcium Ion-selective Electrodes A. Craggs, G. J. Moodyand J. D. R. Thlomas and B. J. Birch Chemistry Department, Uniuersitv oj- Wales Institute of S c i e m e and Technology, Cardiff, CF1 3NU Unilever Reseavch Laboratory, Port Sunlight Laboratory, Wirral, Merseyside, L62 4XN The response characteristics of calcium ion-selec tive electrodes to various anionic surfactants have been investigated in test solutions in the presence and absence of calcium ions. Particular attention has been given to the effect of sodium dodecylsulphate (SDS) on the response of PVC membrane calcium ion-selective electrodes based on calcium bis[di( 4-octylpheny1)- phosphate] sensor and dioctyl plienylphosplioiiatc solvent mediator.M , is to lower the e m.f. response of electrodes when in contact with the various test solutions. However, replacement of the dioctyl plienylpliosphonate solvent mediator by decan-1-01 diminishes the surfactant ion interference to a con- siderable extent, but the amount of decan- 1-01 introduced into the electrode membrane needs to be restricted in order to preserve calcium ion selectivity. The niechanistn of anionic surfactant interference sccnis to involve a calcium - surfactant interaction a t the electrode surface and possibly involving dioctyl plicnylpliosplionate rather than complexation of calcium ions in the bulk solution. Alkylsulphates exhibit tlie trends cxpected of a homologous series, and those of alkyl chains shorter than 6 carbon atonis give little interferencc.A “CIG-CIR olefin sulplionate,” although having good surfactant properties, also shows little interference below about 5 x M. The effect of added anionic surfactant, even a t about 2 x Keywords Ion-selective electvodes ; calcium io.tz-selective electvodes ; suvfactants ; anionic surfactants I t is known1t2 that in the absence of calcium ions Orion 92-20 liquid membrane ion-selective electrodes respond to anionic surfactants over limited concentration ranges with an anionic slope. In tlie presence of calcium ions, Llenatlo2 observed that when a small aliquot of Sulframine, a linear-chain alkylbenzene sulphoiiate mixture with 42.5% Cll, 34.6% C,,, 13.80/, C,, and 9.1% <C9 and >C15 in the alkyl chain, was added to a solution containing calcium and in which a newly constructed Orion 92-20 calcium ion-selective electrode was immersed, there was a sudden e.m.f.shift in the negative direction (about 5-15 mV) followed by a deterioration of the signal into a noisy drifting potential. Further study of this effect over a period of 1 h with a newly prepared electrode immersed in a calcium test solution and adding 2 x lo-* M of the sulphonate showed that the initial sharp reaction following the addition of sulphonate is followed by a gradual drift back to the initial e.m.f. upon standing. When the electrode is immersed in a fresh calcium ion solution the same e.m.f. was obtained as if the sulphoriate had had no effect. In independent work3 on the effect of adding suilphonate to calcium ion-containing solutions in which an Orion 92-20 liquid membrane electrode is immersed, the gradual drift back to the initial e.m.f.on standing was not observed. Hence, the present study was undertaken in order to re-examine the interference of anionic surfactants on calcium ion-selective electrodes, including those based on the newer di(4-octylpheny1)phosphate s e n s ~ r s . ~ ~ ~ Experimental Electrodes PVC matrix membrane electrodes were fabricated as previously described6$‘ and the Orion 92-20 liquid membrane calcium ion-selective electrode was assembled according to the manufacturer’s instructions. The compositions of the various PVC in atrix membranes are summarised in Table I.CRAGGS, MOODY, THOMAS AND BIRCH TABLE I 427 COMPOSITIONS OF MASTER PVC MATRIX MEMBRANES USED FOR ELECTRODE CONSTRUCTION 0.17 g of PVC powder (Breon S llO/lO; BP Chemicals International) used in addition t o sensor and solvent mediator in each instance.Electrode designation A . . . . B . . . . c . . . . D . . . . E . . . . F . . . . Sensor and solvent mediator in master membrane Calcium bis [di(4-octylphenyl)phosphate] (0.04 g) + dioctyl phenyl- Di(4-octylpheny1)phosphoric acid (0.04 g) + dioctyl phenylphosphonate Dioctyl pheiiylphosphonate (0.40 g) Calcium bis [di(4-octylphenyl)phosphate] (0.04 g) 4- decan-1-01 (0.36 g) Calcium bis [di(4-octylphcnyl)phospliate] (0.04 g) + dioctyl phenyl- Orion 92-32-02 divalent liquid ion exchanger (0.40 g) phosphonate (0.36 g) (0.36 g) phosphonate (0.12 g) + decan-1-01 (0.24 g) Reagents Calcium bis [di(4-octylphenyl)phosphate] and the parent acid were synthesised as previously described,8 as was dioctyl phenylpho~phonate.~ “Specially pure” sodium dodecylsulphate (SDS) was obtained from BDH Chemicals, other sodium alkylsulphates with C4-C1, chain lengths from Cambrian Chemicals Ltd ., and the remainder from Unilever Research Laboratory.All other materials were of the best laboratory-reagent grade available. Procedure The experiments were carried out in the cell assembly previously de~cribed,~ with the various ion-selective electrodes being used in conjunction with an Orion, Model 90-02-00, double-junction reference electrode with 5 M ammonium nitrate rather than potassium nitrate as the outer filling solution in order to avoid the possible formation of insoluble potassium salts of the surfactant anions. This electrode was checked periodically and had an e.m.f.of 10 mV with respect to the Radiometer K401 saturated calomel reference electrode. All e.m.f. measurements were made with a Radiometer, Type pHM 64, digital pH - milli- voltmeter used in conjunction with a Servoscribe, Model RE 4541, potentiometric chart recorder. Solutions of the sodium salts (0.1 M) of the surfactants in 0.01-cm3 aliquots were added to the test solution (100 cm3) contained in the cell assembly and the e.m.f. was monitored. Results and Discussion Experiments with an Orion 92-20 Liquid Membrane Electrode The st able response evoked by newly constructed electrodes immersed in calcium chloride solutions in the 10-1-10-4 M range suffered negative shifts of about 10 mV with added M DOBS 055 (an alkylbenzene sulphonate) in each instance.This effect was similar to that reported by Llenado2 but the e.m.f. did not return to its original value even on standing for 24 h. The characteristic response112 to anionic surfactant in the absence of calcium was also observed but the desensitisation effect observed by Llenado2 was not evident. Experiments with a PVC Membrane Calcium Ion-selective Electrode Based on Calcium Bis[ di(4-octylphenyl)phosphate] Sensor The first group, the results of which are summarised in Fig. 1, was conducted to determine the effect of anionic surfactants on the PVC matrix membrane electrode4y5 with calcium bis [di(4-octylphenyl)phosphate] sensor and dioctyl phenylphosphate solvent mediator (electrode A).The second group of experiments (Fig. 2) was devoted to the effect of SDS on three different electrodes immersed in 0.1 M sodium chloride solution. The third group (Fig. 3) assessed the effect of replacing the dioctyl phenylphosphonate solvent mediator of the sensing electrode membranes with decan-1-01. Three groups of experiments were carried out.428 AnaZyst, VOZ. 105 Efect of anionic surfactants SDS was used as the main surfactant in this study and it can be seen from curves A and B (Fig. 1) that the effect on the response of the calcium ion-selective electrode based on a PVC matrix membrane of calcium bis jdi(4-octylphenyl)phosphate] sensor and dioctyl phenylphosphonate solvent mediator (electrode A) is to lower the e.m.f. response regardless of whether or not calcium is present in the cell test solution. CRAGGS et aL.: EFFECT OF ANIONIC SURFACTANTS v 5 4 2 psurfactant (EDTA for curves D and F ) Fig. 1. Effect of anionic surfactants on the response of electrode A. Key to curves: A = SDS added to 10-1 M sodium chloride; B == SDS added t o 1 0 - 3 M calcium chloride; C = SDS added t o 2 >( M calcium chloride in pH 5.1 potassium hydrogen phthalate - sodium hydroxide pH bufferlo adjusted to about 0.1 M ionic strength with sodium chloride; D= EDTA (0.1 M) added t o the same solution as for curve C; E= SDS added to a solution containing 10-3 M calcium chloride and 2 x M EDTA, in pH 5.1 potassium hydrogen phthalate - sodium hydroxide pH bufferlo and adjusted to about 0.1 M ionic strength with sodium chloride (“free” [Ca2+] = 2 x M) ; F -L EDTA (0.1 M) added to the same solution as for curve E ; G, H, I, J , I( and I, = anionic surfactant added to an approximately 0.3 M ionic strength solution consisting of 60 cm3 of de-ionised water + 10 cm3 of ammonia - ammonium chloride pH 10 bufferll + 23 cm3 of 0.1 M sodium citrate + 7 cm3 of 0.1 M calcium chloride (“free” [Ca2+] = 10-3 M). Static response times (minutes) are quoted for the data points of curves B-F. The response in curve A is fairly linear up to an SDS concentration of about 4.5 x lo-* M, when the signal became noisy and slowly drifted.This change in signal is presumably associated with micelle formation, although the concentration is considerably lower than the accepted critical micellar concentration limits12 i(8.0-8.4 x The slope of the linear region of the response curve (about 28 mV decade-l) is near to that expected if the response mechanism involved Ca(DS), formation at the membrane - solution interface.The time taken for the e.m.f. to stabilise after addition of an aliquot of surfactant solution was generally between 10 and 20 min compared with the usual 1-2 min for this electrode (see, for example, the response times quoted for curves B, C and E in Fig. 1). As also noted above for the Orion 92-20 electrode, the e.m.f. did not return to its original value. M) for SDS.May, 1980 ON CALCIUM ION-SELECTIVE ELECTRODES 429 In order to establish whether the surfactant interference was a characteristic of the solution conditions or of the electrode, use was made of solutions buffered by EDTA at pH 5.1 with respect to calcium ion levels along the lines previously reported4p5 and kept at 0.1 M ionic strength with sodium chloride.In such solutions where the “free” calcium ion level was 2 x 10-5 M (curves E and F in Fig. 1) the SDS caused a larger drop in e.m.f. (curve E) than was observed for solutions containing the same level of calcium ions (curve C), but the general effect was the same. The addition of EDTA to the “unbuffered” calcium solution causes an expected drop in e.m.f. because of complexation (curve D), but the lowering of e.m.f. caused by SDS in the heavily buffered solution (curve E) suggests that the mechanism of SDS interference does not involve complexation of calcium ions. The use of EDTA demands a fairly low pH in order to obtain a reasonably high “free” calcium ion level arid can lead to possible hydrolysis of the dodecylsulphate.Hence, experi- ments were conducted a t pH 10 and a “free” calcium ion level of M was maintained with sodium citrate. Again there was a lowering of e.m.f. on addition of SDS (curve G in Fig. 1). This system was also used to demonstrate the effect of other anionic surfactants and it can be seen that the sodium alkylsulphates have the trends expected of a homologous series (curves G, H, K and L) with those having alkyl chains shorter than 6 carbon atoms showing no surfactant properties. The “C16-C18 olefin sulphonate” forms a more readily water- soluble calcium salt than SDS and the reduced interference of this surfactant anion, which otherwise has good surfactant properties, is interesting.A I 3 [I N M O T I I 1 5 4 3 pDodecylsulphate (Ca for curve 0 ) Fig. 2. Effect of SDS (curves B, M and N) and Ca2+ (curve 0) on the response of various electrodes in 0.1 M sodium chloride solution. Key to curves: B = electrode A ; M = elec- trode B; N and 0 = electrode C. E f e c t of electrode membrane constituents Calcium was completely eliminated from the system by incorporating di(4-octylpheny1)- phosphoric acid into the electrode membrane instead of the calcium salt, but with 10-1 M hydrochloric acid as the internal solution (electrode B). Curve M (Fig. 2 ) shows only a small drop in e.m.f. on adding SDS to 0.1 M sodium chloride, compared with a large drop when the calcium salt of the sensor is used in the electrode membrane (curve B).I t should be noted that the electrode with the acid form of the sensor gave only a slowly drifting e.m.f. when exposed to calcium ions, but if immersed for several hours in a concentrated (about 0.5 M) calcium chioride solution adjusted to pH 10.5 with sodium hydroxide it can be con- verted into a normally functioning calcium ion-selective electrode when an internal reference solution of calcium chloride is employed. No quantitative conclusion can be made about curve M, but it should be noted that an electrode with a membrane of just dioctyl phenylphosphonate and no sensor (electrode C) responds slightly to SDS [curve N), but the nature of both curves M and N when compared430 CRAGGS et al. : EFFECT OF ANIONIC SURFACTANTS Analyst, Vol.I05 with curve B indicates that calcium has to be present in the electrode membrane for the manifest ation of SDS interference. Replacement of electrode membrane dioctyl phenylj%osphonate i d h decan-1-01 Perchlorate lowers the e.m.f. response of the Orion 92-20 calcium ion-selective electrode in the presence of calcium ions13 and its effect in this work during the series of experiments for plotting curves G-L (Fig. 1) was seen to be between those of octyl- and hexylsulphates (curves I and K in Fig. 1). However, the mechanism of its interference seems to be different from that of SDS since the acid-form electrode (electrode B) exhibits interference from perchlorate similar to the calcium form. It is interesting that the Orion 92-32 divalent cation electrode, which is thought to contain a similar sensor material to its 92-20 calcium partner but with decan-1-01 mediator, suffers no adverse effects from perchlorate.This, and the nature of curve N, suggests that the mediator may also be involved in the interference of anionic surf act ants. 10 rnV 1 5 4.5 4 p D odecy I:iu I p hate Fig. 3. Effect of SDS on electrode response with decan-1-01 instead of dioctyl phenylphosphonate as solvent mediator in electrode membranes. Key t o curves: E = electrode A in a solution of M calcium chloride and 2 x M EDTA in pH 5.1 potassium hydrogen phthalate - sodium hydr- oxide pH bufferlo and adjusted to about 0.1 M ionic strength with sodium chloride (“free” [Ca2+] = 2 x M). P = electrode A in water; Q = electrode E in Ihe same solution as for curve E; R = electrode D in water; S = electrode E in water; T = electrode F in 0.1 M sodium chloride solution; U = electrode F in water.Fig. 3 shows that for an electrode with calcium bis [di(4-octylphenyl)phosphate] sensor and decan-1-01 mediator (electrode D) the change in SDS concentration in water from to 10-4 M causes a lowering of just 10 mV (curve R) compared with 30 niV when using dioctyl phenylphosphonate as the solvent mediator (curve P). The interference effect is even smaller when using Orion 92-32--02 liquid ion exchanger as the sensing component in the electrode membrane (curves T and U). Unfortunately, decan-1-01 is not a PVC plasti ciser and PVC membranes incorporating decan-1-01 readily exude the alcoh01.l~ Hence, in order to lessen this inconvenience, the usual 0.36g of solvent mediator used in the master PVC matrix membranes with calcium bis[di(4-octylphenyl)phosphate sensor was made up of 0.24 g of decan-1-01 and 0.12 g of dioctyl phenylphosphonate (electrode E).Electrodes from such a membrane and used inMay, 1980 ON CALCIUM IOX-SELECTIVE ELECTRODES 43 1 solutions buffered by EDTA at pH 5.1 to give 2 x M “free” calcium ions gave a minimal drop in e.m.f. when SDS was added (curve Q). Similar results were obtained for electrodes with membranes containing other proportions (5: 1, 1 : 1 and 1 : 2) of decan-1-01 to dioctyl phenylphosphonate. However, when water was substituted for the calcium ion- buffer solution large e.m.f. drops were again observed (curve s). Conclusion The effect of SDS on calcium ion-selective electrodes indicates that the formation of a complex or precipitate of calcium surfactant and possibly involving dioctyl phenylphosphonate at the electrode membrane surface is responsible for the observed interference.Certainly the scheme advocated by Llenado,2 involving initial surfactant monolayer formation at the membrane - solution interface followed by extraction of the surfactant into the membrane phase, needs modifying because it predicts that the interference should also be observed when the electrode sensor material is in the acid form, which is contrary to the results of the present work (Fig. 2, curve 11). The apparent decrease in surfactant interference observed when mixed mediator systems of decan-1-01 and dioctyl phenylphosplionate are used may be explained in terms of differing solubilities of the calcium - surfactant species in these systems or even variations in the porosity of the membrane surface.Such systems indicate a possible method for overcoming the anionic surfactant interference, but the loss of selectivity of calcium over other cations associated with decan-1-01, especially sodium with regard to detergent systems, is clearly a limitation. Howevei, the high calcium over sodium selectivity of the calcium bis Cdi(4-octyl- pheny1)phosphatel sensor ensures that part of the dioctyl phenylphosphonate solvent mediator can be replaced with decan-1-01 in order to combat interference by surfactant anions. The authors thank the Science Research Council for a studentship (to A. C.) under the Cooperative Awards in Science and Engineering Scheme in conjunction with Unilever Research, Port Sunlight Laboratory. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Birch, B. J., and Clarke, D. E., Anal. Chim. Acta, 1973, 67, 387. Llenado, R. A., Anal. Chem., 1975, 47, 2243. Birch, B. J., unpublished work. Craggs, A., Moody, G. J., and Thomas, J . D. R., Analyst, 1970, 104, 412. Craggs, A., Moody, G. J., and Thomas, J . D. R., Analyst, 1979, 104, 961. Moody, G. J . , Oke, R. B., and Thomas, J . D. R., AIzalyst, 1970, 95, 910. Craggs, A., Moody, G. J . , and Thomas, J . D. R., J . Chem. Educ., 1974, 51, 541. Craggs, A., Delduca, P. G., Keil, L., Key, €3. J . , Moody, G. J., and Thomas, J . D. R., J . Inovg. Nucl. Craggs, A., Delduca, P. G., Keil, L., Moody, G. J., and Thomas, J . D. R., J . Inorg. Nucl. Chem., Dempsey, B., and Perrin, D. D., “Buffers for pH and Metal Ion Control,” Chapman and Hall, Vogel, A. I., “Quantitative Inorganic Analysis,” Third Edition, Longmans, London, 1962. Mukerjee, P., and Mysels, K. J ., “Critical Micclle Concentrations of Aqueous Surfactant Systems,” “Instruction Manual for Calcium Ion Electrode Model 92-20,” Orion Research Inc., Cambridge, Craggs, A., Keil, L., Moody, G. J., and Thomas, J . D. R., Talanta, 1975, 22, 907. Chem., 1978, 40, 1483. 1978, 40, 1943. London, 1974. NRS Publ. NSRDS/NBS No. 36, 1971. Mass., 1966. Received November 19th, 1979 hccepted Decembev 20th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500426

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

432 Analyst, May, 1980, Vol. 105, pp. 432-440 Acquisition and Processing of Data from a Multiple pH-stat System Monitoring Enzyme Kinetics (Urea - Urease) I. Rousseau” and B. Atkinson? Department of Chemical Engineering, University College of Swansea, Singleton Park, Swansea, SA 2 8PP A complete data acquisition and processing system is described for the study of enzyme kinetics using multiple pH-stat units. The system is based on a digital data logger linked to individual autotitrators and generates a computer compatible record of titrimetric data. The system can be used to follow any reaction in which hydrogen ions are either produced or consumed. An external module has been developed that causes an operating auto- titrator to activate the logger to perform a continuous scan until all auto- titrators return to the passive mode.The operating time, the real time and the channel identification are recorded when any autotitrator executes a pH correction. The raw data generated by the logger are manipulated by com- putational algorithms, which identify the operative titrators and produce computer amenable data on accumulative volumes of titrant added as a function of elapsed time. The data logger - pH-stat system was used in con- junction with the urea - urease reaction and the derived data were processed by numerical methods to yield the characteristic parameters associated with enzymic reactions. Keywords : Data acquisition and processii4g; multi$le PH-stat system; enzyme kinetics ; titrimetric data ; urea - urease In recent years the “pH-stat” has been used advantageously in the analysis of reactions that result in either the production or consumption of hydrogen ions and is rapidly becoming established as a routine laboratory technique for kinetic studies on enzyme-catalysed re- a c t i o n ~ , ~ - ~ for analytical determination~~~’ and for pH-controlled fermentations.In all these applications the instrumental system involves potentiometric detection of a deviation from a controlled pH value and automatic addition of a titrant. Recently developed systems have been equipped with a strip-chart recorder to produce a graph of accumulative volume of added titrant ueysus elapsed time. The few notable departures from the pH-stat system include electrolytic titrant generationg and spectrophotometric monitoring of pH .9 During various kinetic and stability studies10-12 involving both soluble and immobilised enzymes, it was found that the use of conventional pH-stat instrumentation was unsatis- factory owing to the lack of accuracy that resulted from the graphical extraction of reaction rates from titrimetric graphs. This inherent inaccuracy featured most strongly in the early stages of the time course of the reaction, and was notably associated with estimates of the initial rates of reaction.A further less important, but nevertheless unsatisfactory, feature of the graphical interpretation of the raw experimental data is the considerable expenditure of time necessary to extract both the reaction rates and the characteristic parameters of the reaction.To eliminate the slow, inaccurate process of manual data manipulation various possible modifications to the basic pH-stat system were considered with a view to producing a coni- puter compatible output. The most attractive possibility was found to lie in digital data logging equipment, as this has become an effective means of data acquisition from analytical and similar experimental devices.13-15 The data logger enables experimental data to be sampled, collected and subsequently analysed by off -line computer methods. The work described in this paper involved the application of digital logging equipment to the recording of the “on” operations of a multiple pH-stat system and in developing the software * Present address : Department of Chemical Engineering, Technion, Haifa, Israel.f Present address : Department of Chemical Engineering, UMIST, P.O. Box 88, Manchester, M60 1QD; t o whom correspondence should be addressed.ROUSSEAU AND ATKINSON 433 necessary to convert the digital data firstly into titrimetric data and secondly into kinetic parameters. The performance of the data acquisition and processing system was evaluated using the enzyme-catalysed hydrolysis of urea by urease. 7 Chart recorder Adaptation of a Data Logger to a Multiple pH-stat System When operating, the logging equipment receives continuous inputs in analogue electrical form, e.g., voltage, from the experimental device and the scanner controller makes a selection from these signals. This selection is based on the scanning mode chosen, e.g., a single or a group of channels can be scanned either at regular intervals or continuously.The signals selected are measured by a digital voltmeter and converted into the binary coded decimal (BCD) form. In addition, the digital clock provides a record of real time in BCD at pre-determined intervals. Signals generated both by the voltmeter and by the clock are routed sequentially via the gating system of the serialiser into the common output unit and recorded either on punch- tape or by a printer, or fed directly into a computer. Normally available data loggers cannot be used to scan signals generated by a multiple pH-stat system efficiently. In order to log only the “on” times of several autotitrators operating simultaneously the logger “print time” and “skip” modes were modified 3y the addition of an external module developed for this purpose.The external logger control module caused the logger to perform a continuous scan whenever any autotitrator com- menced a pH correction, while termination of all autotitrator “on” signals reverted the logger to the passive mode. The logger control module (LCM)* consisted of two separate gating units and was connected between the logger and the autotitrators as shown in Fig. 1. A block diagram of the main components of a data logger is included in Fig. 1. 1 - controller Digital voltmeter Serialiser Data logger I I I J----GJ ’ experimental I+ -’ j parameters j Fig. 1. Block diagram of the pH-stat - data logger system. Logger Control Module : Titrator Control Gate The titrator control gate (Fig.2) was designed to activate the logger scanner controller and to select only the operative channels for a continuous scan. When all the autotitrators are idle, the unit allows the logger to proceed via the “time scan” mode (TSG, Fig. 2 ) and to provide a record at selected intervals of the real time followed by an identification record of the autotitrators sampled. * Details of the logger control module are available from the authors on request.434 Analyst, vol. 105 When an autotitrator is switched to the “on” position to achieve a pH correction, the signal is routed to the common start input line of the logger control module and, simul- taneously, the skip function on the skip control gate (described in the next section) corres- ponding to this autotitrator is by-passed and the logger “time scan” mode (TSG) is dis- connected.The signal proceeds via the continuous scan gate (CSG) and triggers the logger scanner controller, thus initiating the logger “continuous scan” mode. A delay of 4 ms ensures that the change of mode is completely established. Finally, a trailing edge of 10 ms (pulse B) activates the scanner controller by triggering the manual switch (MSG). The operative autotitrator is repeatedly scanned at a pre-selected rate, e.g., 0.1, 1, 4, 8 or 12 channels per second, and any additional autotitrator that becomes operative is incorporated into the scan via the continuous scan gate without further delay. ROUSSEAU AND ATKINSON : ACQUISITION ,4ND PROCESSING OF DATA I 1 + ve Common start input line - Data logger L - - - - l Fig.2. Block diagram of the logger control module gating system. When an autotitrator signal is terminated, a pulse of 4 ms (pulse A) is generated via the “print time” gate (PTG) to switch to the “print i;ime” mode to record the real time and the autotitrator identification corresponding to this event. The autotitrator channel then reverts to the skip function via the skip control gate. Logger Control Module: Skip Control Gate Logging equipment usually has a manual “skip” facility, which enables any inoperative channels to be omitted during channel scanning. The skip control gate was developed to provide an automatic skip facility when an autotitrator ceased to operate, to prevent “skip” signals interfering with logger operation and to enable the logger to function via the built-in modules when all the autotitrators were idle.Raw Data Prolcessing For the data logger that was available a continuous scan of an operative autotitrator generated a repeated six-digit word. The word format used commenced with a comma, to segregate items of data, followed by two digits to identify the autotitrator and three digits for the voltmeter reading. The autotitrator “o:n” time was obtained by monitoring the total time that elapsed during a continuous scan. The initial scan included the activation of the passive logger by an operative titrator and the time steps involved were as follows: Tstart the time required by the logger control module to activate the logger scanner Tauto the time required to scan and record an operating autotitrator; Tfikip the time required to skip an inoperative channel; and Tre-set the time required to re-set the scanner controller for the next scan.controller ; Thus, the scan elapsed time for the initial scan (SET,) is given byMay, 1980 FROM A MULTIPLE PH-STAT SYSTEM MONITORING ENZYME KINETICS 435 SET, = Tstart + NautoTauto + NskipTskip + Tre-set . . * - (1) where Nauto and Nskip are the number of active and passive autotitrators, respectively. For subsequent scans the scan elapsed time (SET,) is given by SET, = When an autotitrator becomes inactive the position of the scanner is uncertain and the scan elapsed time corresponding to this scan (SET,) may correspond either to SET, or be given by SET3 = SET2 - Tauto + T&ip . . .. . . .. depending on the time at which the active autotitrator reverted to a passive role. of this uncertainty SET, was estimated as Because . . - * (3) SET, = SET, - *(Tauto - Tskip) . . Thus, the “on” time for the initiating autotitrator is given by n - 1 i = 2 . . * * (4) T = SET, + C (SET2)i + SET, .. where n is the number of scans that took place while the given autotitrator was active, i.e., during a pH correction. For autotitrators that became active while a scan was in progress n - 1 . . * (5) T = C (SET,)i + SET, . . . . 2-1 and the numerical values of Nauto and Nskip had to be adjusted accordingly. The various time elements that make up the scan elapsed time, namely Ts+,art, Tskip, Tauto and Tre-set, were obtained by calibration against an electronic timer for scanning rates of 0.5, 1 and 2 channels per second (Table I).TABLE I AVERAGE SCANNING TIMES Scanning rate/ channels ssl Tstart/s TskiplS -I’auto/S T r e - s e t l s 0.5 0.0138 0.042 2 2.161 7 0.233 3 1 .o 0.0143 0.042 1 1.041 6 0.108 4 2.0 0.014 1 0.042 3 0.608 1 0.061 7 Computer Programs* This program, written in PLAN, led to the identification of the operative channels and the scan elapsed times, while further programs, written in FORTRAN, were used to evaluate the titrant volumes equivalent to the scan elapsed times, the volume of titrant added as a function of real time, the rates of reaction and the kinetic parameters. The punched tape for each experiment was identified by a manually punched code number, which was used to re-set the computational procedure following processing of the data.This identification number was followed by a record of the real time of initiation of the experiment. The additional information necessary for data processing was fed into a master program and consisted of the individual scanning speeds (Table I), the number of potentially operative autotitrators (Nauto) and permanently inoperative channels ( N s k i p ) , the normality of the titrant, the pH correction function, which accounted for the buffering capacity of the pro- ducts,lG and the code numbers of the channels to be processed. A computer program was necessary to interpret the data on the punched tape. * .4 list of the computer programs is available from the authors on request.436 ROUSSEAU AND ATKINSON ACQUISITION AND PROCESSING OF DATA Analyst, Vol.105 The raw data processing program was used i:o identify and read the number of digits in each data item on the punched tape. The identification of an itern containing six digits led to the second and third digits, i e . , the channel number, being compared with the number of the channel to be processed. For the required channel the scan elapsed times were calculated using either equation (4) or (5). Identification of a five-digit word indicated the completion of a pH correction and gave a real time record, which was expressed as the experimental elapsed time. Finally, the accumulated scan elapsed times were converted into the equiva- lent titrimetric data using a previously determined calibration graph. A typical calibration graph for one of the pH-stat units is given in Fig..3 and this relates the scan elapsed times to the addition of known titrant volumes. were requirid for each unit. 0.20 5 0.16 2 .e 0.12 a, 0.08 2 0.04 - C CI LC - 5 Sirnil& calibration graphs 0 7.5 15.0 22.5 30.0 37.5 45.0 52.5 60.0 Scan elapsed tirne/s Fig. 3. Calibration of data logger scan elapsed times vevsus acid added (given autotitrator; scan- rate 1 channel per second). The titrimetric data were then expressed as product formcd or substrate consumed and curve-fitted by orthogonal polynomials of the third degree using a least-squares criterion. The corresponding rates of reaction were computed using the Rerezin and Zhidkovl' variant of the Lagrange formulae for the nutnerical differentiation of equi-spaced interpolated data.This procedure allowed the initial rates of reaction to be interpolated with accuracy. The enzyme kinetic parameters Vmax, and Km were obtained from the initial rates of reaction using the statistical method developed by Wilkinson.l* This method involves the Lineweaver - Burk rearrangement of the Michaelis equation as the basis for curve-fitting 0 4 8 12 16 20 24 28 Experimental time/min Fig. 4. Time course of ammonia production by urease for various initial urea concentrations: 0, 0.1; A, 0.5; and x , 1.0 pmol ml-1.MflY, 1980 FROM A hIULTIPLE PH-STAT SYSTEM XONITORING ENZYME KINETICS 437 data for graphs of rates of reaction versus substrate concentration at constant enzyme con- centration. An ICL 1905 computer was used throughout the work although the storage requirements and CPU time were extremely modest.The system could be developed further, particularly for multi-pH-stat operation on a routine basis, whereby data from edited tapes could be stored on disc file, with the data processing carried out later. Experimental Equipment and Enzyme Reaction The catalytic hydrolysis of urea by soluble urease (Worthington Biochemical Corp., Batch No. URC-SHA, prepared from j ack-bean meal) was monitored by automatic additions of acid. The quoted activity of the enzyme was 84 units per milligram of enzyme, where a unit is defined as micromoles of ammonia produced per minute. A single pH-stat system was linked via a logger control model to a data logger (Solatron Electronic Group Ltd., Compact Logger Series 2, LU 1975).The pH-stat consisted of 50 ml of solution completely mixed in a thermostated (38 & 0.1 "C) reaction cell, equipped with a combined pH electrode (Pye Ingold, Type 407 E,7). All potentiometric measurements were made by a pH meter (Yye, Model 290) and an autotitrator - controller (Pye, Cat. No. 11603) was used to activate a titrant flow inducer (Watson-11farlow, Type MHRE 5000) whenever a deviation occurred from a pre-selected pH value. A desk-type potentioinetric chart re- corder (Servoscribe) was connected t o the autotitrator to provide a direct record of the titrator on - off' signals 1250 ,- a, $ 1000 C v 7 I 0 7 - 750 \ 5 -0 2 500 -0 Q m S 0 E 250 E Q 2 .- and' to allow a rapid visual estimate of tl;e progress of the reaction. 0 4 8 12 16 20 24 28 Experimental time/min Fig.5. Time course of ammonia production by urease for various initial urea con- centrations: 0, 3.0; A, 5.0; >!, 8.0; 7, 50; and 0, 100 pmol ml-l. Procedure Urease solutions were prepared by dissolving weighed amounts of the enzyme in 0.1 M tris buffer (pH 7.4) and standing for 1 h at room temperature, to allow for hydration of the enzyme. Enzyme solution was then introduced, by micropipette, into the reaction cell, which contained urea solution of an appropriate concentration at pH 7 0.2. Addition of enzyme solution led to a drop in pH, of about 0.15. The time course of the reaction was followed for at least 20niin, during which time the acid added was recorded by the data logger and measured by burette. A similar series of experiments was performed to establish the influence of pH for a single initial urea concentration of 0.1 M.438 from a single preparation were used for the pH studies.ROUSSEAU AND ATKINSOW : ACQUISITION AND PROCESSISG OF DATA Analyst, VoZ. 105 Freshly prepared urease solutions were used for each urea concentration, while aliquots " 0 50 100 150 200 250 300 350 400 Initial urea concentration/pmoI mI-' Fig. 6. Influence of initial urea concentration on initial rates of react.ion. 0, Experimental data obtained using data logger - pH-stat system; solid line, Michaelis equation [ Vmax, = 84.3 pmol min-l mg-l (enzyme) and K , = 1.8 pmol ml-l]. Results and Discussion The catalytic hydrolysis of urea by urease in free solution has been used as a typical enzymic reaction to examine the use of a data logger in conjunction with a multiple pH-stat system.The time course of ammonia production per unit mass of added enzyme is given in Figs. 4 and 5 for various initial urea concentrations. Each experimental point on Figs. 4 and 5 represents the end of a pH correction, while the continuous lines are curve fits based upon third-degree orthogonal polynomials ; inclusion of higher order terms was found to have little effect on the first polynomial coefficient and therefore on accuracy. A comparison, a t the termination of the exper~~ment, between the cumulative total volume of acid added, computed via the data logger and measured by burette is given in Table I1 for various initial urea concentrations. Initial rates of reaction obtained by numerical differentiation of the interpolated data are given in Fig.6 as a function of the initial urea concentration. Values of the initial rate of reaction, less than 80% of the maximum rate obtained experimentally, were used to estimate the kinetic parameters V,,,. and K , as recommaded by Wilkinson.l* TABLE I1 RESULTS COMPUTED BY DATA LOGGER AND MEASURED BY BURETTE FOR AIZID ADDITIOKS Initial urea concentration/ pmol ml-' 0.1 0.5 1.0 3.0 5.0 8.0 10.0 30.0 50.0 80.0 100.0 150.0 200.0 300.0 400.0 Amount of enzyme used/ mg 0.174 0.148 0.162 0.080 0.072 0.052 0.040 0.052 0.052 0.046 0.040 0.040 0.040 0.042 0.040 Experimental Logger-computed time/min acid added/ml 20 0.45 28 0.61 28 0.91 32 2.02 30 1.92 31 1.56 30 1.29 31 1.78 33 1.86 32 1.65 32 1.47 31 1.42 31 1.25 29 1.43 33 1.28 Burette observed acid added/ml 0.48 0.65 0.93 2.07 1.93 1.61 1.32 1.85 1.85 1.73 1.50 1.46 1.32 1.48 1.36May, 1980 FROM A MULTIPLE PH-STAT SYSTEM MONITORING ENZYME KINETICS 439 The value obtained for Vmax, of 84.3 & 4.05 pmol NH, min-l (mg enzyme)-l is in agree- ment with the nominal quoted activity of the enzyme preparation and the value of K , of 1.80 & 0.26 pmol ml-l can be compared with values quoted in the literature (Table 111). The agreement occurs in spite of the fact that the curve fit is relatively poor in the region where the rate of reaction is insensitive to substrate concentration. This suggests that the pH-stat method is insufficiently accurate to monitor the rate of reaction - concentration characteristic beyond 80% of the maximum rate of reaction.TABLE I11 VALUES OF K m OBTAINED FOR THE UREA - UREASE REACTION WHEN USING pH-STAT METHODS K,/pmol ml-l pH Temperature/"C Reference 1.25 7.05 22 2.25 7.0 30 3.28 7.0 38 19 1 2 The pH characteristic obtained using the data logging system is given in Fig. 7. These data suggest an optimum pH of about 6.5, a value that lies between the optimum values of 719 and 6.36 also obtained by pH-stat techniques. The pH dependence of the rate of reaction at substrate saturation conditions (0.1 M, Fig. 6) is given by . . . . . . K m X . (PHopt.) Ka' + -") Vmax. (PH) = ('+(H+; Kb In equation (6) Ka' and Kb' are the dissociation constants involved in the equilibrium of the substrate - enzyme complex, Vmax. (pHopt.) is the maximum rate of reaction a t the optimum pH and Vmaxe (pH) is the maximum rate of reaction at any given pH." 2 4 6 8 10 PH Fig. 7. pH characteristic of soluble urease (urea concentration 0.1 M). x , Experimental data; solid line, calculated using equation (6) with Ka' = 7 . 1 x 10-9 M, K ~ / = 3.6 x 10-5 M. The value for Ka' of 7.1 x M was obtained by linear regression of the relative rate Similarly, M was obtained from Vma,.(pHopt.)/T/m,x.(pH) vemus [H+] on the acid The values of Ka' and Kb' can be compared with those of Barth and Michellg M when employing a pH-stat data, Vmax.(pHopt.)/Vmax. (pH) V e n u s 1 / [ H + ] , on the alkaline side of pHopt.. Kb' = 3.6 x side of pHopt.. who obtained Ka' = 7.9 x M and Kb' = 7.68 x method.440 ROUSSEAU AN11 ATKINSON Conclusions The data acquisition and processing method based upon data logging and off-line computa- tional analysis in conjunction with pH-stat systems allows a large number and wide range of enzymic experiments to be carried out rapidly arid accurately with output in a form amenable to further analysis.The curve-fitting procedure, based on orthogonal polynomials, has proved to be entirely satisfactory for the analysis of time course data and in particular has proved to be more than an adequate method of estimating initial rates of reaction for subsequent calculation of enzymic reaction rate parameters. The authors thank Dr. J. L. Beckner for his assistance with the computer programming and Mr. T. G. Hughes for his invaluable contribution to the development of the logger control module. 1. 2. 3. 4 . 5. 6 . 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. References Toren, E. C., and Burger, F. J., Mikrochiw. Ada, Suppl. 111, 1968, 1049. Blakely, R. L., Webb, E. C., and Zerner, B., Biochemistry, 1969, 8, 1984. Keijer, J. H., Anal. Biochem., 1970, 37, 439. Vendermeers, h., Lelotte, H., and Christophe, J., AnaZ. Biochem., 1971, 42, 437. Engasser, J . M., and Horvath, C., Biochemistry, 1974, 13, 3849. Malmstadt, H. V., and Piepmeier, E. H., Anal. c‘hem., 1965, 37, 34. Schwabe, K., Adv. Anal. Chem. Instrum., 1974, 10, 495. Matsen, J . M., and Linford, H. B., Anal. Chem., 1962, 34, 142. Karcher, R. E., and Pardue, H. L., CZzn. Chem., 1971, 17, 214. Rousseau, I., PhD Thesis, University of Wales, 1’973. Lester, D. E., PhD Thesis, University of Wales, 1973. Millar, M. H., PhD Thesis, University of Wales, 1974. Allen, P., Kropholler, H. W., and Spikins, D. J., BY. Chern. Eng., 1966, 11, 703. Caisey, J . D., and Riordan, B. D., Analyst, 1973, 98, 126. Larsen, D. G., Anal. Chem., 1973, 45, 217. Atkinson, B., Rott, J., and Rousseau, I., Biotech~ol. Bioeng., 1977, 19, 1037. Berezin, I. S., and Zhidkov, N. P., in Booth, A. D., Editor, “Computing Methods,” Volume I, Wilkinson, G. N., Riochem. J , , 1961, 80, 324. Barth, von A., and Michel, H. J., Biochem. Physiol. Pflaiaz., 1972, 163, 103. Pergamon Press, Oxford, 1965, p. 198. Received Mamh 5th, 1979 Accepted September 24th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500432

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, May, 1980, Vol. 105, $9. 441-447 Gas-chromatographic Study of Methylstyrenes in 441 the Light-oil Fraction of Coal Distillate Using Selective Hydrogenation M. R. Tirgan and N. Sharifi-Sandjani Rubber and Plastic Laboratory, Institute of Standard and Industrial Research of Iran, Karadj, Iran Research Laboratory of Polymer Chemistry, Faculty of Science, University of Tehran, Tehran, Iran Catalytic hydrogenation of aromatic olefins, such as styrene, a- and p-methyl- styrene, 0-, m- and P-methylstyrene and indene, in light oil was carried out by a selective method, the aromatic rings being kept intact. Gas chromato- tography was employed to identify the starting materials and the products. The chromatograms of hydrogenated and non-hydrogenated samples obtained under identical conditions were compared.The qualitative and quantitative determinations of the compounds of interest were based on the shifts of the peaks after the hydrogenation process and calculation of peak areas (e.g., the shift on converting a-methylstyrene into cumene). Keywords : Coal distillate light-oil avzalysis ; uaetliylstyrene identi3cation ; selective hydrogenation; gas chrowaatography The oxygen bridges in the heterocyclic rings in coal are broken down by pyrolysis,l yielding various light compounds that can be measured by gas chromat~graphy.~-~ The light-oil fraction of coal distillate obtained from the pyrolysis of coal at a coking plant is generally a mixture of mono- and bicyclic hydrocarbon~.l-~ The optimum pyrolysis temperature of coal to produce the maximum yield of aromatic hydrocarbons is 700-900 "C,l which is the same as the temperature range required for cleavage of phenolic OH bondslo in the pyrolysis of phenoplast resin to increase the percentage of aromatic hydrocarbons. Generally, platinum,llJ2 p a l l a d i ~ r n l ~ ~ ~ ~ ~ ~ ~ nickel,l5 copperl5?ls and zinc1' have been used as catalysts for hydrogenation, and the copper catalyst prepared according to the method of Sabatier and Sanders15 was used in this work because it gave the best results for selective hydrogenation of ethylenic and dienic bonds, the aromatic and heterocyclic rings remaining unaffected. This work is a modification of the methods of Berozall and Thompson and c o - w ~ r k e r s l ~ ~ ~ ~ with catalytic hydrogenation carried out in the injector block.Experimental Materials Sample coal at the Coking Plant of the Esfahan Steel Mill, Iran. The light-oil (crude benzole) fraction of coal distillate was obtained from the pyrolysis of Standard solution of hydrocarbons Hydrocarbons were purchased from Illerck-Schuchardt, Darmstadt, Germany (laboratory grade). Thiophen (50 pl), toluene (500 pl), ethylbenzene (10 pl), $-xylem (100 pl), m- xylene (100 pl), o-xylene (50 pl), propylbenzene (5 pl), mesitylene (5 pl), 1,2,4-trimethyl- benzene (5 pl), 1,2,3-trimethylbenzene (5 pl), curnene (1 pl), P-ethyltoluene (5 pl), m-ethyl- toluene (5 pl), o-ethyltoluene (5 pl), styrene (5 pl), cc-methylstyrene (10 pl), m- and 9- methylstyrene (10 pl; 60% w-), allylbenzene (5 pl), coumarone (5 pl), indene (50 pl), tetra- hydronaphthalene (50 pl), naphthalene (50 mg) and dicyclopentadiene (15 pl) were each injected separately into a 6-ml vial containing 4 ml of benzene and kept cool. Solid sup$ort Creek, Calif., USA, Chromosorb P (acid-washed, 60-80 mesh) was obtained from Varian Aerograph, Walnut442 TIRGAN AND SHARIFI-SAND JANI : GC ST'IJDY OF MBTHYLSTYRENES Analyst, VOZ.105 17 3 18 25 56 24 23 13 14 20 21 i I 2 N & a 7 x 7 c 0 P ci, 0, a 12 3 56 22 24 8 10 Ill 25 20 7 L I I I 30 60 Ti me/m in 90 Fig. 1. Chromatograms of a standard hydrocarbon blend on a capillary column (a) before hydrogenation ; and (6) after copper catalyst hydrogenation.May, 1980 I N THE LIGHT-OIL FRACTION OF COAL DISTILLATE 443 Copper(II) hydroxide mixed with Chromosorb P adding an excess of sodium hydroxide. mixed with the appropriate amount of Chromosorb P.oven at 50 "C, contained 50% of copper after reduction. Copper(I1) hydroxide was precipitated from a hot solution of copper(I1) sulphate by The precipitate was decanted three times and The product, dried in a vacuum gas chromatograph equipped with a flame-ionisation was hvdrogen at a flow-rate of 1.35 ml min-I, with an Apparatus detector was used. A Varian Aerograph, Model 2800, Operating conditions Capillary column. The carrier gas air flow-rate of 300 ml min-l, a hydrogen Aow-race of 25 ml min-l and an electrometer setting of 10-lo and A mV-l. A stainless-steel capillary column (200 ft x 0.01 in i.d.) was coated with poly(Yuz-phenyl ether) (6-ring) and stabilised at 150 "C with nitrogen at a flow- rate of 1 ml min-l for 48 h.The column temperature was maintained at 50 "C for 15 min, then programmed at 1 "C min-l up to 110 "C and at 4 "C min-l up to 150 "C. The injector temperature was 240 "C and the detector temperature 240 "C. A 0.5-pl volume of sample (or standard) was injected, with a splitting ratio of 1 : 50. The carrier gas was hydrogen at a flow-rate of 30 ml min-l, with an air flow-rate of 300mlmin-l and an electrometer setting of x 16 and 10-l" x 16 A mV-l. A coiled copper column (24 ft x + in i d . ) was packed with 157; FFAP on Chromosorb P and stabilised at 250 "C with nitrogen at a flow-rate of 15 ml min-l for 24 h. The column temperature was progranimed from 60 to 150 "C at 2 "C min-l and then from 150 to 240 "C at 20 "Cmin-l.The injector temperature was 240 "C and the detector temperature 240 "C. &-in FFAP column. The injection volume was 0.5 p1. TABLE 1 COMPOSITION OF THE LIGHT OIL Son-hydrogenatable compounds Hydrogenatable compounds r A A 7 r 1 Concentration. Peak Concentration, Peak Compound Benzene . . . . . . Thiophen . . . . Toluene . . . . . . Ethylbenzene . . . . p-Xylene.. . . . . m-X ylene . . . . o-Xykne . . . . Propylbenzene . . . . p-Ethytoluene . . . . m-Ethyltoluene . . . . Mesitylene . . . . o-Ethyltoluene . . . . 1,2,4-Trirnethylbenzene I ,2,3-Trimethylbenzene Coumarone . . . . Indane . . . . . . Tetrahydronaphthalene Cumene . . . . . . Naphthalene . . . . Hydrogenation unit Yo wlgn 70.0 0.045 13.20 0.09 0.90 2.50 0.45 0.003 0.02 0.045 0.25 0.006 0.28 0.05 0.048 0.03 3.1 - - number 1 2 3 4 5 6 7 8 10 11 12 14 15 16 20 21 22 25 24 Compound yo mlwi number Styrene .. . . . . 1.2 9 Allylbenzene . . . . nz-Methyls tyrene . . 0.20 o-Methylstyrene . . . . 0.025 19 /3-Methylstyrene . . 0.018 - 13 - cc-Methylstryene . . 0.03 p-Methylstyrene . . 0.11 $ Dicyclopentadicne . . 0.35 26 Indene . . . . . . 0.95 23 The injector block of the gas chromatograph was used as a hydrogenation unit. The injector tube (length 136.5 mm, i.d. 3.2 mm) was packed successively with Chromosorb P (0.25 g), copper(I1) hydroxide mixed with Chromosorb P (0.15 g) and Chromosorb P (0.10 g). The catalyst was activated at 400 "C by passing hydrogen through the tube at 50 ml min-l for 1 h. When hydrogenation was not carried out, the injector tube was packed only with Chromosorb P (0.5 g).444 TIRGAN AND SHARIFI-SAND JANI : GC STUDY OF METHYLSTYRENES Analyst, VoZ.105 3 6 5 13 1-21 8 23 19 I .c X 7 0 . . c --I---- c N !5 3w-h. I I , . ____.. 30 60 Timelmi n 90 1 i c . . 12 3 6 E 1 1 ' -___ 6 21 \ 22 I I 39 60 90 Tim e/m i n Fig. 2. Chromatograms of a light oil on a capil.lary column : (a) before hydrogenation; and (b) after copper catalyst hydrogenation.May, 1980 I N THE LIGHT-OIL FRACTION OF COAL DISTILLATE 445 Results and Discussion Throughout this text the numbers in parentheses refer to the peaks in the chromatograms and listed in Table 1. By comparing the chromatograms of hydrogenated [Figs. l ( b ) , 2 ( b ) and 3 ( b ) ] and non- hydrogenated [Figs. l(a), 2(a) and 3(a)] standard solutions and light oil obtained under identical conditions, it was found that after hydrogenation the peaks of styrene (S), allyl- benzene (13), a-methylstyrene (17), m- and P-methylstyrene (18), o-methylstyrene (19) and indene (23) had shifted to the corresponding hydrogenated forms as follows : ethylbenzene (4), propylbenzene (lo), cumene (8), m-ethyltoluene (12), P-ethyltoluene (1 1), o-ethyltoluene (15) and indane (22), respectively.Dicyclopentadiene (26) was decomposed by the hydro- generation process and the main product peak (27) was identified. When the FFAP column was used the peaks of styrene, dicyclopentadiene and an unknown compound (U) overlapped [Fig. 3(a)]. The sytrene peak was shifted to ethylbenzene (4) by the hydrogenation process [Fig. 3 ( b ) ] ; the peak of a-methylstyrene was eliminated and the hydrogenated form of dicyclopentadiene (27) appeared in place of (17), and the unknown compound remained unaffected [Fig.3 ( b ) ] . When using the capillary column no peak was recorded for dicyclopentadiene in the first stage of temperature programming [Figs. 1 (a) and 2 ( a ) ] , but after hydrogenation dicyclo- pentadiene partially decomposed and the main peak (27) and two small peaks (D1 and D,) were observed [Figs. 1 ( b ) and 2 ( b ) ] . Finally, the following results were obtained from the peaks belonging to the compounds involved in the hydrogenation reaction : (i) The shift of the peaks in the chromatograms due to hydrogenation was used to identify the compounds concerned. For example, after hydrogenation a-methylstyrene in the light oil was identified by disappearance of peak (17) and an increase in the area of peak (8) [Figs.l ( b ) , ( b ) and 3 ( b ) ] . (ii) I t was possible to calculate quantitatively even the compounds that gave overlapping peaks by considering the differences in the areas under the peaks before and after hydrogenation. For example, the exact amount of styrene (9) was calculated from the increase in the peak area of ethylbenzene (4) and the reduction in the peak area of styrene [Figs. 3(a) and 3 ( b ) ] . The percentages in Table I are average values from at least three experiments, and were calculated from peak-area measurements by the triangulation method; the three decimal places are related to the 10-l1 sensitivity range. (iii) Because certain peaks of the hydrogenatable and non-hydrogenatable compounds overlapped in the chromatograms, the catalytic hydrogenation process made it possible to eliminate the peaks belonging to the hydrogenatable compounds [e.g., indene, Figs.3(a) and 3 ( b ) ] , and therefore to determine quantitatively the non-hydrogenatable compounds from peaks that remained [e.g., coumarone, Figs. 3(a) and 3(b)]. In addition to the identification and determination of styrene, indene, a-methylstyrene and o-methylstyrene, the actual amounts of fn- and P-methylstyrene [unresolved even with the capillary column, Fig. 2 ( a ) ] were obtained from the increased peak area of m- and 9- ethyltoluene [Fig. 2(b)]. However, the increase in the peak area of propylbenzene is more than would be expected from the hydrogenation of allylbenzene [Fig.2 ( b ) ] and is probably due to the presence of P-methylstyrene in the light oil, the peak of which was not assigned and was not well separated in the chromatograms. After hydrogenation the peaks of coumarone and thiophen remained unaffected owing to the presence of their heterocyclic rings. The percentages of the isomers of methylstyrene, xylene and ethyltoluene in the light oil are shown in Fig. 4, and it is evident that the abundances of the isomers decrease in the order m- > 9 - > 0-. The presence of thiophen or carbon disulphide eliminates the catalytic effect of the injector and metal column wall during the analysis of the samples and standards without hydrogena- tion.446 TIRGAN AND SHARIFI-SANDJANI : GC STUDY OF METHYLSTYRENES Analyst, VoZ.I05 3 2 A 14 6 I 26 3- U 18 !3 & 25 P . I ______--- 30 GO Ti m e/m i in - I J x 0 0 - F 111.12 25 L 30 Time/min 60 Fig. 3. Chromatograms of a light oil on an FFAP column: (a) before hydrogenation ; and ( b ) after copper catalyst hydrogenation.May, 1980 IN THE LIGHT-OIL FRACTION OF COAL DISTILLATE 447 rn-isomer p-isomer o-isomer o-, M - and p-isomer classification Fig. 4. Representative relationships of the composition of xylene, methylstyrene and ethyltoluene in terms of their o-, m- and p-isomers in the light oil. Conclusion By selective hydrogenation of aromatic olefins in light oil using a copper catalyst it is possible to identify methylstyrenes that have not been reported previously in the chromato- graphic analysis.The contents of aromatic components in light oil have been determined, and it has been established that the abundances of the isomers decrease in the order m- > p- > 0-. The authors are grateful to Mt. M. Gaeeni for his assistance. 1. 3. 4. 5. 6. 7 . 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. i) A. References Lowry, H. H., “Chemistry of Coal Utilization,” John Wiley, New York, 1963, p. 379. Krekel, G., and Steinbrecher, F., Rrenmt.-Chenz., 1964, 45 (3), 81. Jaworslti, M., and Bobinski, J . , Chcrn. AvaZ. (Warsaw), 1964, 9, 1003. Eremeeva, I<. T., and Kuznetsov, G. T., K o k s K h i m . , 1965, 12, 59. Bricteux, J., Awn. Mi,nes Bclg., 1966, 12, 1543. Mitsuw, I . , Haruyoshi, T., and Toshiolti, O., Kowtaw, 1966, 18, 514. Romovacek, J . , and Kubat, J., Anal. Chern., 1968, 40, 1119. Roseira, A. N., Rev. Bras. l’ecml., 1971, 2, 1 . Pichler, I I . , Ripperger, W., and Schwarz, G., E’rmulst.-Chetn., 1970, 23, 91. Sharifi, N., and Tirgan, M. I<., J . A p p l . Polym. Sci., 1973, 17, 1113. Beroza, M., Anal. Chem., 1962, 34, 1801. Yasuda, S. I<., J . Chrornatogr., 1967, 27, 72. Iwanow, A., and Eisen, O., J . Chromatop., 1972, 69, 53. Hendifar, A. R., and ‘Tirgar, M. R., J . Ch~ornatogr., 1978, 161, 119. Grignard, V., “Trait@ de Chirnie Organique,” Volume 2, Masson, Paris, 1948, p. 958. Hall, W. K., and Hassell, J . H., J . Phys. Chem., 1963, 67, 636. Stransky, Z., Gruz, J . , and Ruzicka, E., J . Chromatogr., 1971, 59, 158. Thompson, C. J . , Coleman, H. J . , Hopkins, R. L., Ward, C. C., and Rall, H. T., Anal. Chem., 1960, Thompson, C. J . , Coleman, H. J . , Ward, C. C., and Rall, €3. T., Anal. Chem., 1960, 32, 424. 32, 1762. Received August 29th, 1979 Accepted Octobev 25th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500441

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

448 Analyst, May, 1980, Vol. 105, fie. 448-454 Determination of 6-Mercaptopurine and Related Compounds by Phosphorescence Spectroscopy* A. I . Al-Mosawi and J. N. Miller and J. W. Bridges Department of Chemistry, Loughbovough University of T ~ L hnology, Loughbovough, Leicestershire, LE 11 3T U Institute of Industrial and Envivonvtzental Health and Safety, University of Suvrey, Guildfovd, Surrey, GU2 5XH The phosphorescence properties of 6-mercaptopurine and ten related com- pounds have been studied in acidic, neutral and alkaline ethanolic glasses a t 77 K, and adsorbed on thin layers a t 77 K and a t room temperature Nano- gram amounts of most of the compounds can be detected at 77 I(, and sub- ~nicrograin amounts at room temperature A technique combining thin-layer chromatography and phosphorimetric scanning of the chromatograms has been used to determine trace amounts of the compounds in human blood plasma.Keywords : Mercaptopuvine determination ; tJt in-layer chfomatography ; phosphorimetry ; room-temperature phosphorescence The widespread use of 6-niercaptopurine and azathioprine as imniunosuppressive and anti- leukaemic agents has created a demand for the selecti1.e determination of these compounds and their metabolites a t trace levels in biological fluids. The many methods used include paper1 and thin-layer chr~matography,~~~ high-voltage4 and thin-layer5 electrophoresis, gas - liquid6 and high-performance liquid chromatography7 and mass spectrometry.8 A number of luminescence methods have also been developed. Finkelg showed that 6-mercapto- purine could be determined fluorimetrically after oxidising deproteinised plasma samples with potassium permanganate to produce purine-6-sulphonate ; the limit of detection was 1 pg nil-1.The strong phosphorescence exhibited by many purine derivatives has been frequently reported (e.g., reference lo), and Aaron and Winefordnerll showed that 6- methylmercaptopurine and 2-amino-6-methy1me;rcaptopurine showed analytically useful phosphorescence. The same compounds were more recently shown to exhibit a weaker TABLE I MERCAPTOPURINES AND DERIVATIVES STUDIED Compound S O . 1 2 3 4 5 G 7 8 9 10 11 Kame 2-Mercaptopurinc 6-Mercaptopurine 6-Mercap toguanosine 6-Methylmercaptopurine 6-Mercaptopurine riboside 6-Mercaptopurine 2-deoxyriboside 6-Nercaptopurine riboside 5’-phosphate Azathioprine 2-Thioxanthine 6-Thioxanthine 2-hmino-6-mercaptopurine Melting- point/’C > 325 300 226 230 222 214 > 325 254 > 325 > 325 > 325 RF * 0.37 0.44 0.51 0.63 0.63 0.41 0.71 0.66 0.19 0.21 0.32 * On cellulose thin layers.Solvent: 0.1 M hydrochloric acid, except for 2-thioxanthine, for which where the solvent system was propan-2-01 - methanol - water - ammonia (60 + 20 + 20 + 1). * Presented at the Meeting on “Kesearch and Development Topics in Analytical Chemistry,” Heriot- Watt University, Edinburgh, July, 1979.AL-MOSAWI, MILLER AND BRIDGES 449 room-temperature phosphorescence (RTP) when adsorbed on to chromatography paper from alkaline solution.12 Maddocks and Davidson13 reported the detection of picomole amounts of azathioprine, 6-mercaptopurine and seven related compounds after thin-layer chromato- graphy on cellulose, using the luminescence developed when the chromatography plate was cooled to liquid nitrogen temperature.Wong and Maddocks5 later used a similar technique to detect mercaptopurine derivatives after thin-layer electrophoresis on a variety of media. Neither of these papers reported spectroscopic data, and fluorescence and phosphorescence were apparently not distinguished. This paper describes a detailed study of the luminescence properties of 11 mercaptopurine derivatives. The application of thin-layer phosphorimetry,14 both at room temperature and at liquid nitrogen temperature, to their rapid determination at trace levels in deproteinised blood plasma is also described. Experimental The compounds studied are listed in Table I ; all were obtained from Sigma Chemical Co.(Poole, Dorset). Each compound yielded a single spot when analysed by thin-layer chromato- graphy (TLC) on cellulose thin layers (E. Illerck, obtained through BDH Chemicals, Poole, Dorset) using 0.1 M hydrochloric acid as the normal developing solvent,13 and low-temperature luminescence as the detection method. All other compounds were of the highest grades commercially obtainable. Fluorescence measurements at room temperature were performed on a Fluoricord spectro- fluorimeter (Baird-Atomic, Braintree, Essex) using silica cells of 10-mm path length. Phosphorescence studies in rigid glasses at 77 K were performed using a Fluorispec spectro- fluorimeter (Baird Atomic) fitted with a silica Dewar flask and rotating cylinder phosphoro- scope as previously described.15 Thin-layer phosphorimetry was carried out using a specially designed accessory to the Fluoricord fluorimeter; this device and its operation have been described in detail in an earlier paper.16 For room-temperature studies the thin-layer attachment was used in the same way, except that no liquid nitrogen was added to the sample drum.In all luminescence assays the limit of detection of a solute was defined as that concentration yielding a signal two standard deviations above the background signal. Absorption spectra were obtained using an SP 800 spectrophotometer (Pye-Unicam, Cam- bridge) at room temperature. TABLE I1 PHOSPHORESCENCE CHARACTERISTICS OF MERCAPTOPURINES I N ETHANOL GLASSES AT 77 K I Compound Ae,/ No.nm 1 342 2 340 3 341 4 286 5 325 6 332 7 337 8 300 9 295 10 344 11 356 Neutral ethanol - 7 7 -A AP/ Detection limit/ A,,/ nm rig ml-l nm 510 6 330 463 20 316 466 120 340 436 25 295 456 30 323 476 60 320 446 ND* 300 442 104 311 458 104 2 95 480 50 349 484 40 330 * ND = not determined. Alkaline ethanol A,/ Detection limit/ nm ng ml-l 488 1 454 2 46 1 70 446 40 462 50 460 60 45 1 2 600 508 50 479 50 430 105 452 105 TLC of the mercaptopurine derivatives was performed on cellulose thin layers, and on high-performance thin-layer chromatography (HPTLC) plates coated with silica gel, Both types of thin layer were coated on aluminium foil (E. Merck, obtained through BDH Chemicals) to ensure good thermal contact when the plates were subsequently examined in the thin- layer phosphorimeter.Before the samples were applied all of the plates were completely450 AL-MOSAWI et UZ. DETERMINATION O F 6-MERCAPTOPURINE Analyst. VOZ. 105 eluted with 96% ethanol (James Burroughs Ltd., London), and the top 1-2 cm of the plates cut off; the background luminescence of the oven-dried plates was thus substantially reduced. Samples of 0.1-1 p1 were applied to the plates using disposablc micropipettes (ICL Scientific, Fountain Valley, Calif., USA) and chromatography was performed in closed chambers. (The chromatography step was omitted in determinations of detection limits.) When the TLC step was complete the plates were air-driled and sprayed with ethanol. Thin-layer phosphorimetry at 77 K was then performed at once; room-temperature studies were performed after the plates had been dried overnight. Pooled normal human blood plasma samples, obtained from Leicester Royal Infirmary, were spiked with the compounds under study.Deproteinisation was carried out using 10 volumes of cold ethanol, followed by centrifugation. 100 a I0 \ 1 ' 0.1 I------ 0.01 0.1 1 10 100 Concentration/pg mi-' /* Fig. 1. Analytical growth graph for the phosphorimetric determination of B-mercapto- purine riboside (compound 5 ) in ethanol glasses containing 0.1 M sodium hydroxide. Excitation and emission wavelengths, 323 and 462 nm, respectively. 100 a -2 10 1 Amount/ng per spot Fig. 2. Analytical growth graphs for the determination of 6-methylmercaptopurine (compound 4) by thin-layer phosphori- nietry a t (A) 77 K and (B) room temperature.At 77 K the graph was determined using pure solutions of the drug (0) and extracts of spiked blood plasma (m). Excitation and emission wavelengths, 286 and 436 nm, respectively. Results; The ultraviolet absorption spectra of the 6-mei-captopurine derivatives were found to be pH dependent. Spectra obtained in neutral and acidic (0.1 M hydrochloric acid) ethanolic solutions were generally similar, but in ethanol containing 0.1 M sodium hydroxide shifts of absorption maxima and changes in molar absorbance were sometimes observed. 6-Mercapto- purine, for example, exhibited a blue shift of about 18 nm, and a slight decrease in molar absorbance, in alkaline solution. Such changes, similar to those found by other workers,17 suggested the desirability of performing luminescence studies in acidic, alkaline and neutral solutions. Room-temperature studies showed that only one compound, 2-amino-6-mercaptopurine, exhibited significant fluorescence.In alkaline ethanol, with excitation and fluorescence wavelengths of about 316 and 402 nm, respectively, this compound had a limit of detection of 7 pug ml-l. The phosphorescence characteristics of all the compounds in ethanolic glasses at 77 K are given in Table I1 (in these conditions, compounds 3, 11 and, in alkaline ethanol, 4 exhibited relatively feeble fluorescence signals at 350400 nm. The remaining compounds exhibited no fluorescence and could therefore be studied without the rotating cylinder phosphoroscope in the light beam).As expected from the absorption spectra, the phos- phorescence spectra in neutral and acidic ethanolic solutions were very similar; only the data in neutral solution are therefore given in Table 13:. I t is apparent that limits of detection of nanograms per millilitre can be obtained in many instances. For some compounds, alkaline conditions produced the best detection limits ; for others, neutral or acidic solutionsk?oy, 6980 AND RELATED COXPOLT’NDS BI- PHOSPHORESCESCE SPECTROSCOPY 45 1 were preferred. Analytical graphs were found to be linear over at least two orders of magnitude of concentration ; the graph for 6-mercaptopurine riboside in alkaline ethanol is shown in Fig. 1. When the compounds were adsorbed on to cellulose TLC plates at 77 K, the excitation and emission spectra were generally closely similar to those obtained in ethanolic glasses, and again very low levels of many of the compounds could be detected, particularly when the samples were applied as alkaline ethanolic solutions.Table I11 shows that amounts as low as 10 pg could be detected in deproteinised plasma samples. Similar results were obtained when pure solutions of the compounds were studied. The recovery of the solutes using the cold ethanol precipitation method was almost lOOyo ; Fig. 2 shows typical analytical graphs. These detection limits were at least an order of magnitude better than could be obtained by visual observation of the TLC plates. Three of the compounds were also studied using silica gel HPTLC plates, and the limits of detection achieved were similar to those obtained using cellulose thin layers.In attempts to enhance the phosphorescence signals still further, ethanolic solutions of potassium iodide (1% m/V), lead tetraacetate (10% m/V) and thallium acetate (10% m/V) were investigated as spray reagents; none of these compounds produced a “heavy atom” enhancement on cellulose or silica gel thin layers and in most instances a quenching of the phosphorescence was observed. TABLE I11 LIMITS OF DETECTION OF MERCAPTOPURINES IN BLOOD PLASMA USING THIN-LAYER PHOSPHORIMETRY AT 77 K Limit of detectionlng per spot Compound No. 1 2 3 4 5 6 7 8 9 10 11 Cellulose thin layers , Neutral ethanol solvent 45 3 4 1 2 5 30 150 50 1 0.1 Alkaline ethanol solvent 2 0.05 0.1 1 5 3 25 10 50 5 5 Silica gel HPTLC layers, neutral ethanol solvent ND* 0.04 0.2 1 ND ND ND ND ND ND ND * ND = not determined.Several mixtures of mercaptopurines were studied in order to test the precision and selectivity of the combined TLC - phosphorimetry method : all of these measurements were made at 77 K using cellulose thin layers. Fig. 3 shows the separation of four compounds on a cellulose TLC plate: again, the results obtained were the same whether the compounds were dissolved in ethanol or derived from spiked plasma samples. The coefficient of variation when the same TLC plate was scanned repeatedly for 30 min was 8%. Analytically useful signals were obtained only when cellulose was used as the adsorbent, with the highest intensities generally being obtained when alkaline ethanol was used as the solvent for sample application.Even then, the limits of detection (Table IV) were at least an order of magnitude inferior to those determined at 77 K. Room-temperature phosphorescence spectra generally exhibited higher band widths and less vibrational fine structure than spectra obtained at 77 K ; the example of 6-mercaptopurine is shown in Fig. 4. The room-temperature phosphorescence of all of the compounds was studied. Discussion The luminescence of purines has been investigated by a number of workers,ll~l8~lg and it is well established that many purine derivatives exhibit strong phosphorescence signals in the452 AL-MOSAWI et nZ. : DETERMINATION OF ~-MERCAPTOPURINE Analyst, VoZ. 105 wavelength region 400-550 nm.The phosphorescence lifetimes and emission wavelengths indicate that T-T* transitions are responsible. Some purines are also fluorescent, although the nature of the lowest singlet state has been a imatter of controversy.18 It has been estab- lished, however, that the lowest singlet excited state of purine itself is n-n* in nature. In alkaline solution purines form anions by loss of a proton at the 9-position; the phosphorescence of the purine anion is red-shifted compared with the parent molecule, and fluorescence is also observed. Aaron and Winefordnerll showed that 6-methylmercaptopurine and its 2-amino derivative could be detected phosphorinietrically a t low levels (less than 1 ng ml-I), and were more phosphorescent than purines lacking the mercapto group.This may be due to the “heavy atom” effect of the sulphur substituent. Indirect confirma.tion of this came from the results of Vo-Dinh et aZ.,20 who found that an external heavy-atom perturbant (sodium iodide) did not enhance the phosphorescence of 6-methylmercaptopurine, when it was observed at room temperature adsorbed on filter-paper. The limit of detection under these conditions was inferior to that at 77 K, but was still of analytical value. I Fig. 3. Thin-layer phosphorimetry at 77 K of a mixture of compounds 2, 3, 4 and 11 on a cellulose thin layer. The developing solvent during the thin-layer chromatographic step was 0.1 M hydrochloric acid. Each of the 14 sequential scans took 1 min. S marks the point of sample application and F the solvent front; for RF values see Table I.In (a) the excitation and phosphorescence wavelengths were 342 and 485 nm, respectively, and in (b) 320 and 448 nm, respectively. In this work all of the compounds studied showed analytically useful phosphorescence, but few were found to be measurably fluorescent, again presumably because of the heavy atom effect on the inter-system crossing rate constant. Although the phosphorescence properties of the compounds are generally simil.ar, there are considerable differences in phosphorescence intensity. The position and substitution of the thiol group in the purine system are clearly of importance. Compounds with this group at the 6-position are generally more strongly phosphorescent than 2-substituted purines : 6-mercaptopurine and 6-thio- xanthine can be detected at lower levels than 2-mercaptopurine and 2-thioxanthine, respec- tively.Both 6-methylmercaptopurine and azathioprine have excitation and phosphorescenceMay, 1980 AND RELATED COMPOUNDS BY PHOSPHORESCENCE SPECTROSCOPY 453 maxima blue-shifted compared with the parent compound. weakly phosphorescent, possibly because of its nitro group : Azathioprine is only very Azathioprine The presence of a ribose group has, as expected, little effect (compounds 3, 5 and 6), but the phosphate group in 6-mercaptopurine riboside 5-phosphate also has a strong quenching effect. In agreement with earlier studies,20 external heavy-atom effects could not be induced. r - P 250 350 450 500 W avel ength/n rn Fig. 4. Excitation (E) and phosphorescence (P) spectra of 6-mercaptopurine (compound 2) in a neutral ethanol glass a t 77 K (solid line) and adsorbed on a cellulose thin layer a t room temperature (broken line).The spectra obtained on a cellulose thin layer a t 77 K were indistinguishable from those obtained in an ethanol glass. It is apparent that thin-layer phosphorimetry at 77 K is a powerful method for deter- mining 6-mercaptopurine and its metabolites. In many instances the limits of detection are very low, the drugs may be readily and completely recovered from deproteinised plasma and both the chromatographic and luminescence steps contribute to the selectivity of the method (Fig. 3). Further, the TLC step is rapid (especially using HPTLC plates) and utilises a simple, non-luminescent eluting solvent.Room-temperature phosphorescence was, as expected, weaker than the phosphorescence observed at 77 K. Nonetheless, it may be of some use where higher levels of the mercaptopurines are under study; the thin-layer phosphorimetry method is, of course, easier when no liquid nitrogen is required. The finding that cellulosic thin layers give optimum room-temperature phosphorescence effects is in agreement with the results of several previous workers (reviewed in reference 21). This study has shown that the combination of thin-layer chromatography and phosphori- metry is a valuable analytical method, especially when several structurally similar phosphorescent compounds (e.g., a drug and its metabolites) are to be analysed. It complements the numerous existing techniques that combine TLC and fluorimetry, and further work in the authors' laboratories is extending the method to the analysis of other groups of compounds.454 AL-MOSAWI, MILLER AND BRIDGES TABLE IV LIMITS OF DETECTION OF MERCAPTOPURINES IN PURE SOLUTION USING THIN-LAYER PHOSPHORIMETRI' AT ROOM TEMPERATURE Compound No.1 2 3 4 5 6 7 8 9 10 11 Limit of detection/ng per spot r Neutral ethanol solvent NP* 28 20 6 NP N P NP NP NP 40 7 3 Alkaline ethanol solvent 25 4 20 5 40 20 300 50 700 200 100 * NP = No phosphorescence detectable on cellulose thin layers. We are grateful to the Medical Research Council for a Project Grant in support of this work. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. References Salser, J . S., and Balis, M. E., Cancer Res., 1965, 25, 539.Loo, T. L., Luce, J. K., Sullivan, M. P., and Frei, E., Clin. Pharmacol. They., 1968, 9, 180. Zimmerman, T. P., Chu, L.-C., Bug&, C. J . L., Nelson, D. J., Lyon, G. M., and Elion, G. B., Cancev Bennett, L. L., and Allen, P. W., Cancer Res., 1971, 31, 152. Wong, P. C.-P., and Maddocks, J. L., J . Chromatogr., 1978, 150, 491. Bailey, D. G., Wilson, T. W., and Johnson, G. E., J . Chromatogr., 1975, 111, 305. Day, J . L., Tterlikkis, L., Niemann, R., Mobley, A., and Spikes, C., J . Pharm. Sci., 1978, 67, 1027. Rosenfeld, J . M., Taguchi, V. Y., Hillcoat, B. L., and Kawai, M., Anal. Chem., 1977, 49, 725. Finkel, J . M., Anal. Biochem., 1967, 21, 362. Longworth, J . W., Kahn, R. O., and Shulman, R. G., J . Chem. Phys., 1966, 45, 2930. Aaron, J . J . , and Winefordner, J. D., Anal. Chem., 1972, 44, 2127. Wellons, S. L., Paynter, R. A., and Winefordner, J . D., Spectrochim. Acta, Part A , 1972, 30, 2133. Maddocks, J . L., and Davidson, G. S., Brit. J . Clin. Pharmacol., 1975, 2, 359. Gifford. L. A., Miller, J. N., Burns, D. T., and Bridges, J . W., J . Chromatogr., 1975, 103, 15. Gifford, L. A., Hayes, W. P., King, L. A., Miller, J. N., Burns, D. T., and Bridges, J . W., Anal. Miller, J . N., Phillipps, D. L., Burns, I>. T., and €<ridges, J. W., Anal. Chem., 1978, 50, 613. Breter, H.-J., and Zahn, R. K., J . Chromatogr., 1977, 137, 61. Cohen, B. J., and Goodman, L., J . Am. Chem. Sot., 1965, 87, 5487. Drobruk, J., and Augenstein, L., Photochem. Photobiol., 1966, 5, 13. Vo-Dinh, T., Yen, E. L., and Winefordner, J . D., Anal. Chem., 1976, 48, 1186. Vo-Dinh, T., and Winefordner, J . D., Appl. Spectrosc. Rev., 1977, 13, 261. Res., 1974, 34, 221. Chim. Acta, 1972, 62, 355. Received September loth, 1979 Accepted January 22nd, 1980

ISSN:0003-2654    

DOI:10.1039/AN9800500448

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, May, 1980, Vol. 105, pp. 455-467 455 Quality Control of Prednisolone Sodium Phosphate N. Stroud, N. E. Richardson, D. J. G. Davies and D. A. Norton Centre for Drug Formulation Studies, School of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath, BA2 7 A Y Prednisolone sodium phosphate is believed to undergo hydrolysis of the phosphate ester group as its primary degradation pathway. Most published assay methods do not determine the phosphate ester directly, and therefore a high-performance liquid chromatographic method has been developed for prednisolone sodium phosphate in the presence of its breakdown products, which has been validated in the presence of excipients used in ophthalmic solutions. Stability data are presented that are comparable to those obtained for related steroid phosphate esters.The stability data indicate that a simpler ultraviolet spectrophotometric assay method can be used for routine stability testing. Keywords : Pvednisolone sodium plzospliate determination ; high-performance liquid chvomatograplay ; prednisolone sodium plaosphate stability The efficiency of corticosteroids such as prednisolone for the treatment of ocular inflammatory conditions is now well established. Prednisolone has a low solubility in water and for aqueous formulations the more water-soluble phosphate ester is used, which can be formulated for both parenteral and topical administration. Several workers have reported that cortico- steroids such as prednisolone undergo thermal degradation in aqueous solution, involving the 17-dihydroxyacetone ~ide-chain.l-~ Transformations and eliminations have been shown to occur in both the presence and absence of air.In the presence of air under alkaline conditions, the predominant reaction appears to involve cleavage of the C17 side-chain to yield the corresponding etianic acid. In the absence of air, two reactions predominate, yielding the 17-keto steroid and the hydroxy acid. Degradation of the A ring has also been shown to occur in a related steroid, hydrocortisone, formulated in a polyethylene glycol base.5 However, the A ring is an inherently stable structure and the rate of degradation was much slower than that for the C,, side-chain. The degradation of steroid phosphate esters has not been studied as extensively, although Marcus6 has reported that the degrada- tion of hydrocortisone phosphate in aqueous solution involved hydrolysis as the only signifi- cant degradative pathway and was dependent on the hydrogen-ion concentration.It would appear, therefore, that the thermal degradation of prednisolone sodium phosphate in aqueous solution would involve the pathways illustrated in Fig. 1 and that hydrolysis of the phosphate group on the C17 side-chain would be predominant. Both prednisolone sodium phosphate and the parent prednisolone possess the 3-keto group and related conjugated system and have similar absorption spectra in the ultraviolet region. Kaplan and Levine7 have developed a column chromatographic method for separating the two compounds using ion-pair formation between the ester and trihexylammonium chloride.However, the method is tedious and lengthy for routine analysis. The C,, side-chain of prednisolone has been determined by complexation with tetrazolium blue followed by spectrophotometric determination of the coloured complex.* This method, however, is specific for the C,, side-chain and prednisolone sodium phosphate would require preliminary hydrolysis to the parent alcohol, which is difficult to achieve quantitatively. Other methods reported are the determination of the inorganic phosphate produced as the ester hydrolysese and gas - liquid chromatography. Upton et aL9 have reported a high-performance liquid chromatographic (HPLC) method for steroid phosphates using a reversed-phase column. However, preliminary work in our laboratories indicated that prednisolone sodium phosphate was eluted immediately after a non-retained compound (potassium dichromate) on a Spherisorb S5 ODS reversed-phase column. It is essential that any assay method distin- guishes between the parent compound and its degradation products and we have therefore developed an HPLC assay for prednisolone sodium phosphate using an anion-exchange column, as the phosphate ester is present in an anionic form in aqueous solution.Most published assay methods do not determine the phosphate ester directly.456 STROUD et d. : QUALITY CONTROL Amzlyst, Vol. 105 Experimental Apparatus Chromatograms were determined routinely using a Pye LC20 system, which has a fixed- wavelength detector set at 254nm. Injections were made on-column with a Pye Unicam fixed-volume 10-pl loop valve.All measurements were made at ambient temperature in replicate. Spectrophotometric determinations were made using a Pye Unicam SP1800 spectrophoto- meter. pH determinations were performed using either a Pye Unicam 291 pH meter or a Radio- meter Type 27 pH meter fitted with a PHA 630P scale expander. Both pH meters were used in conjunction with Pye-Ingold combined glass - silver electrodes. All pH measure- ments were carried out on solutions equilibrated to 25 & 0.1 "C; meters were standardised with two appropriate standard buffers. CH,OPO, Na2 I Prednisolone sodium &=:H phosphate / I 0 CH20H I Pred n isolone I5 , &=:H Prednisolone COOH I Fig. 1. Thermal degradation pathways of prednisolone sodium phosphate.Mat eri a1 s Prednisolone sodium phosphate was a gift from Smith and Nephew Ltd. and was used as received. All buffer salts were of analytical-reagent grade and other reagents were of at least laboratory-reagent grade. Potassium hydrogen phthalate was of an NPL certificated grade supplied by BDH Chemicals Ltd. Solvent:; were of analytical-reagent grade. Water was freshly distilled from an all-glass still. Poly(viny1 alcohol) (Gohsenol N300, Nippon Goshei) was supplied by British Traders and Shippers Ltd. Trihexylammonium chloride was prepared from trihexylamine (Eastman Kodak Co.) according to the method of Kaplan and L e ~ i n e . ~May, 1980 OF PREDNISOLONE SODIUM PHOSPHATE 457 The stationary phase was Partisil 10 SAX (Whatman), an anion-exchange material, packed into either 250 x 4.6 mm or 100 x 4.6 mm stainless-steel columns.The mobile phase consisted of a 1 + 9 V/V mixture of methanol and one-fifth strength McIlvaines citrate - phosphate buffer (pH 5.2), and was de-gassed before use. The actual pH of the mobile phase was 5.6. Chromatography of Prednisolone Sodium Phosphate in Aqueous Solution Aqueous solutions are prepared to contain 0.004-0.024~0 m/ V prednisolone sodium phosphate and 0.12% m/V potassium hydrogen phthalate as internal standard and 10 pi are injected on to the column with a mobile phase flow-rate of 1.5 ml min-l. The chromato- gram is recorded at a detector wavelength of 254 nm. The concentration of prednisolone sodium phosphate is then determined by comparing the peak-height ratio of drug to internal standard with that obtained with a standard solution containing 0.02% m/V of prednisolone sodium phosphate and 0.12% m/V of potassium hydrogen phthalate.Chromatography of Prednisolone Sodium Phosphate in the Presence of a Viscoliser Place 2 ml of prednisolone sodium phosphate solution (concentration range 0. 1-0.5y0 m/V) in a 10-ml glass centrifuge tube containing 3 ml of double-strength Sprrensens phosphate buffer (pH 5.0), mix and add 5 ml of a 5% V/V solution of trihexylammonium chloride in dichloromethane. Stopper the tube, shake it vigorously for 30s, then centrifuge it at 4000 rev min-l for 15 min. Remove the aqueous phase, transfer 3 ml of the organic phase into a fresh centrifuge tube containing 4 ml of 0.1 M sodium hydroxide solution, shake for 30 s and then centrifuge a t 4000 rev min-l for 5 min.Pipette 3 ml of the aqueous phase into a 25-ml calibrated flask containing 3 ml of 0.1 M hydrochloric acid and 3 ml of 1% m/V potassium hydrogen phthalate solution and dilute to volume with water. The solution in the flask is then chromatographed as described above. Thermal Degradation of Prednisolone Sodium Phosphate priate buffer were sealed into glass ampoules and heated in an oil-bath. removed after known periods and assayed for residual drug. Aliquots of 10 ml of a 0.5% m/V solution of prednisolone sodium phosphate in an appro- The ampoules were Results and Discussion Prednisolone sodium phosphate forms an anion in aqueous solution and should therefore be retained by an anion-exchange column to an extent dependent on the pH of the mobile phase, which would be reflected in longer retention times with increase in pH.Fig. 2 shows the capacity factors for 0.02yo m/V of drug using benzyl alcohol as the non-retained com- 1 1 1 I 1 1 3 4 5 6 Influence of pH of mobile phase on capacity factors of: A, pred- nisolone sodium phosphate; and B, potassium hydrogen phthalate. PH Fig. 2.458 STROUD et al. : QUALITY CONTROL Analyst, Vol. 105 pound, over the mobile phase pH range 2.95-6.10. It is apparent that the capacity factor decreases with increase in pH, and this is probably due to competition by the buffer com- ponents dominating the extent of the interactions between column and solute. Knox and VasvarilO have shown that the capacity factor of phthalic acid on an anion-exchange column can be selected by manipulation of the mobile phase pH.Fig. 2 also shows the capacity factors for the more water-soluble potassium hydrogen phthalate over the mobile phase pH range 3.50-6.10, and again it is observed that an increase in pH decreases the retention time. However, it is apparent that over this pH range good resolution is obtained between predniso- lone sodium phosphate and potassium hydrogen phthalate, and the latter was therefore selected as the internal standard using a mobile phase pH of 5.2. The addition of 10% V/V of methanol as organic modifier was found to reduce the analysis time and improve peak symmetry, although the final pH of the mobile phase increased slightly to 5.6. Initially, chromatograms were obtained using; the 250-mm column with a mobile phase flow-rate of 1.5 ml min-l.Subsequently, a 100-mm column with a mobile phase flow-rate of 1.2 ml min-l was shown to give improved pleak symmetry and a reduction in analysis time. A typical chromatogram using this system is shown in Fig. 3(a). 10 5 0 -L 10 5 0 Time/mi n C) i 1 C t 10 5 0 Fig. 3. Chromatograms of prednisolone sodium phosphate and its degradation pro- ducts. (a) Prednisolone :sodium phosphate, 0.02'70 m / V ; ( b ) prednisolone sodium phos- phate, 0.02% m/V + prednisolone, 0.018 6% m/ V ; and (c) prednisolone sodium phosphate, 0.02% m/V in pH 8 buffer heated a t 110 "C for 24 h. Conditions: ambient temperature; flow-rate 1.2 ml min-I; stationary phase 100 x 4.6 mm of Partisil 10 SAX (10 pm); mobile phase 10% V / V methanol in McIlvaines citrate - phosphate pH 5.2 buffer and ionic strength 0.1 M (final pH of mobile phase was 5.6) ; detector, ultraviolet a t 254 nm; sensi- tivity 0.16 a.u.f.s.1 , Prednisolone sodium phosphate ; 2, potassium hydrogen phthalate (internal standard) ; 3, prednisolone and 4, degradation products. The principal degradation product of prednisolone sodium phosphate in aqueous solution is probably the parent steroid, prednisolone, which may break down further to give the corresponding etianic acid, hydroxy acid or 17-keto steroid. Prednisolone is un-ionised inMay, 1980 OF PREDNISOLONE SODIUM PHOSPHATE 459 aqueous solution and should be non-retained by an anion-exchange column. Fig. 3(b) shows a chromatogram of 0.02y0 m/V prednisolone sodium phosphate in the presence of 0.0186~0 m/V of prednisolone, and it is apparent that the peak due to the latter is non- retained.A 0.018670 m/V solution of prednisolone buffered at pH 7.4 was heated at 100 "C for 48 h and the chromatogram again showed only one non-retained peak, indicating that any further degradation products would not interfere in the assay for prednisolone sodium phosphate. Finally, a 0.5% m/V solution of prednisolone sodium phosphate buffered at pH 8.0 was heated at 100 "C for 20 h and, after appropriate dilution, the chromatogram [Fig. 3(c)3 shows that adequate resolution is obtained between drug, internal standard and degradation products. The linearity of the response was checked by injecting solutions of prednisolone sodium phosphate over the concentration range 0.004-0.024~0 m/V in the presence of 0.12% m/V of potassium hydrogen phthalate and calculating the peak-height ratios.Replicate cali- bration graphs constructed on two consecutive days were linear, with slopes of 63.08 [standard deviation (s.d.) 0.181 and 63.8 (s.d. 0.22) and intercepts of -0.014 (s.d. 0.01) and -0.039 (s.d. 0.18), respectively. Comparison by a Student's t distribution showed them to be not significantly different (tt;:: = 0.42; tintercept ca,c. - 1.20; ttabulated = 2.45; n = 10, p = 0.05). Simple aqueous formulations of prednisolone sodium phosphate containing only the drug, buffer and a preservative such as benzalkonium chloride can be chromatographed directly after appropriate dilution and the addition of an internal standard.The peak-height ratios of drug to internal standard can then be compared between the sample and a standard solution of prednisolone sodium phosphate. In the presence of formulatory excipients such as polymeric viscolisers, pre-extraction of the drug is necessary. Preliminary extraction experiments were monitored by determining the absorbance of prednisolone sodium phosphate in the aqueous phase at 248 nm. An aqueous phase consisting of double-strength McIlvaines citrate - phosphate buffer (pH 5.0) and an organic phase consisting of a 5% V/V solution of trihexylammonium chloride in methylene chloride was found to transfer 98.0% (s.d. O.lyo, n = 3) of the drug to the organic phase. The extent of subsequent re-extraction into 0.1 M sodium hydroxide solution was 1 0 l .l ~ o (s.d. 0.86y0, n = 3). Formulations of 0.1% m/V and 0.5% m/V prednisolone sodium phosphate containing 4.25% m/V of Gohsenol N300 as viscoliser, O.Olyo m/V of benzalkonium chloride and 0.01 yo m/V of EDTA, disodium salt, buffered at pH 8 were extracted and assayed by HPLC as described. Recoveries were 100.8% (s.d. 1.7%, n = 3) and 98.6% (s.d. 0.62y0, 12 = 3) compared to injection of the standard aqueous prednisolone sodium phosphate solutions, which was considered to be satisfactory. Applicability of the Assay to Stability Studies Stability studies were carried out as described above and residual prednisolone sodium phosphate was determined using the simple non-extraction HPLC assay procedure. Degradation was generally followed to below 50%.Preliminary experiments showed that in Smensens phosphate buffer (pH 6.1 and 8.2) at 90 "C the data for the degradation of prednisnlone sodium phosphate could be fitted to first-order rate plots, leading to values for the rate constants of 4.7 x and 6.52 x 10-3h-1, respectively, which are close to the values of about 4 x h-I determined by Marcus6 for the hydrolysis of hydrocortisone phosphate at pH 6 and 7.5 and 91 "C. However, it was observed that prednisolone sodium phosphate underwent an initial rapid degradation of about 5y0, and this was overcome by the addition of O.Olyo m/V of EDTA disodium salt. The influence of temperature on the degradation of prednisolone sodium phosphate was determined over the range 80-110 "C in Smensens phosphate buffer (pH 8) containing 0.01% m/V of EDTA, disodium salt. When the data (Table I) were plotted according t o the Arrhenius relationship, a value for the activation energy of 126.2 kJ mol-l was obtained, which compares with a value at pH 7.5 of 113 kJ mol-l for methylprednisolone phosphate reported by Flynn and Lambll and 71 k J mol-I determined by Marcus6 at the same pH for hydrocortisone phosphate.The low value of 71 kJ mo1-I for hydrocortisone phosphate reported by Marcus has been attributed to the reaction system being of a significant micellar character a t the drug con- centration studied and that the micellar fraction changes with temperature.ll and 8 x460 STROUD et al. : QUALITY CONTROL Analyst, Vol. 105 Applicability of the Assay to other Dosage Forms and Related Drugs Most liquid formulations of prednisolone sodium phosphate are simple aqueous solutions, which may contain a buffer (pH 5-8), a preservative such as benzalkonium chloride, EDTA, disodium salt, or a viscoliser, and the drug can be determined by the procedures described.The method should also be applicable to solid dosage forms utilising the extraction procedure described. Related steroid phosphate esters can also be determined directly by the HPLC procedure. Fig. 4 shows the chromatogram of 0.02% m/V dexamethasone sodium phosphate, which is similar to that obtained for prednisolone scldium phosphate. The retention time for prednisolone sodium phosphate on this column it; the same as that for dexamethasone sodium phosphate. Burgess,12 however, has reported good resolution of related steroidal esters using an elevated temperature and reversed-phase ZIPLC.Whether the method reported here can also be used for this purpose must await further work, which is proceeding in our laboratories. FIRST-ORDER RATE CONSTANTS FOR THE DEGRADATION OF PREDNISOLONE SODIUM PHOSPHATE AT pH 8.0 AT DIFFERENT TEMPERATURES IFirst-order rate Temperaturel'C constaat/h- 80 1.13 x 90 4.145 x 100 1.271 x 110 3.323 x Routine Quality Control of Aqueous Prednisolone Sodium Phosphate Formulations During the stability studies it was observed thad if the degraded drug solution was extracted prior to HPLC assay, the chromatographic peak due to the degradation products was negligible down to about 60% of residual drug. It was thought possible, therefore, that the residual drug could be determined by UV spectroscopy of the extraction solution.A solution con- taining 0.5:/, m/V of drug, O.Olyo m/V of benzalkonium chloride and O.Olyo m/V of EDTA, disodium salt, in Sorensens phosphate buffer (pl3 7.4) was prepared and the degradation of the prednisolone sodium phosphate a t 100 "C was determined using the following assay procedures : Solutions were appropriately diluted with water, internal standard was added and the mixture was subsequently assayed by HPLC. Solutions were extracted and assayed by IIPLC as described above. Solutions were extracted and 2 ml of the aqueous 0.1 M sodium hydroxide phase were added to a 100-ml calibrated flask containing 2 ml of 0.1 M hydrochloric acid, diluted to volume with water and the UV absorbance of a 1 cm layer of this solution was determined at 248 nm against an appropriate blank.The percentage residual drug concentrations were calculated relative to the values of the drug to internal standard peak-height ratios or absorbance at 248nm determined at zero time. Fig. 5 shows graphs of the percentage residual concentration on a logarithmic scale against time for the three assay procedures. Both the direct HPLC and extraction - HPLC methods gave straight lines, leading to rate constant values of 1.87 x and 1.88 x 10-2 h-l, respectively. With the extraction - 1JV method the percentage residual con- centration of drug is in agreement with that determined by the HPLC techniques down to about 60% of residual drug. I t is apparent, t'herefore, that until 20-30% of the drug is degraded, insufficient degradation products are transferred in the extraction procedure to interfere significantly in the UV determination, and this method can be used for routine quality control purposes. 1.2. 3.May, 1,980 1 C 0 +J al K .- .- - J OF PREDNISOLONE SODIUM PHOSPHATE 461 aJ t 0 6+ tII 10 5 0 Time/min Fig. 4. Chromatogram of 0.02% m/V dexameth- asone sodium phosphate. Conditions as for Fig. 3; detector, ultraviolet a t 254 nm, sensitivity 0.16 a.u.f.s. 1, Dexameth- asone sodium phosphate ; and 2, potassium hydro- gen phthalate (internal standard). 20 40 Ti me/m in Fig. 5. Comparison of assay techniques to determine the degradation of 0.5% prednisolone sodium phosphate a t pH 7.4 and 100 "C. 0, Direct HPLC; 0, extraction - HPLC; and 4, extraction - ultraviolet. Conclusions It has been shown that an anion-exchange column can be used for the HPLC assay of prednisolone sodium phosphate in the presence of its degradation products and that the assay is suitable for stability studies. Using a simple extraction procedure, prednisolone sodium phosphate can be separated from interfering formulatory excipients such as viscolisers. The extraction - UV assay procedure described is particularly useful for the routine analysis of prednisolone sodium phosphate in quality control laboratories. References 1. 2. 3. 4. 5. 6. 7. I). 9. 10. 11. 12. Mason, H. L., J . Biol. Chem., 1938, 124, 475. Hertzig, P. T., and Ehrenstein, M., J . Org. Chem., 1951, 16, 1050. Wendler, N. L., and Graber, R. P., Chem. Ind. (London), 1956, 549. Guttmann, D. E., and Meister, P. D., J . A m . Pharm. Assoc., Sci. Ed., 1958, 47, 773 Allen, A. E., and Das Gupta, V., J . Pharm. Sci., 1974, 63, 107. Marcus, A. D., J . Am. Pharm. Assoc., Sci. E d . , 1960, 49, 383. Kaplan, G. B., and Levine, J . , J . Assoc. Ofl. Anal. Chem., 1973, 57, 735. Mader, W. J., and Buck, R. R., Anal. Chem., 1952, 24, 666. Upton, L. M., Townley, E. R., and Sancilio, F. P., J . Pharm. Sci., 1978, 67, 913. Knox, J . H., and Vasvari, G., J . Chromatogr. ScZ., 1974, 12, 449. Flynn, G. L., and Lamb, D. J., J . Phavm. Sci., 1970, 59, 1433. Burgess, C., J . Chromatogr., 1978, 149, 233. Received September loth, 1979 Accepted December 12th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500455

出版商:RSC    

年代:1980

数据来源: RSC