ANALYST, JUNE 1993, VOL. 118 649 Phase-selective Alternating Current Adsorptive Stripping Voltammetry of Aminopterin on a Mercury Thin Film Carbon Fibre Ultra m icroelect rode* Michael A. Malone,+ Agustin Costa Garcia and Paulino Tunon Blanco* Department of Physical and Analytical Chemistry, University of Oviedo, Asturias, Spain Malcolm R. Smyth School of Chemical Sciences, Dublin City University, Dublin 9, Ireland The electrodeposition of mercury thin films onto carbon fibres for the determination of aminopterin and its analogues has been optimized following an investigation of the electrochemical reduction processes of aminopterin obtained at a static mercury drop electrode. The advantageous characteristics of ultramicroelec- trodes combined with adsorptive preconcentration and phase-selective a.c.stripping voltammetry were found to yield a very sensitive and reproducible method. By using this electrode, accumulation was performed at five different concentrations of aminopterin ranging from 5 x IO-lO to 5 x rnol dm-3. The electrode yielded a calibration graph from 2 x IO-IO to 8 x rnol dm-3 ( r = 0.994) with a limit of detection [signal-to-noise ratio (S/N) = 31 of 1 x IO-lO rnol dm-3 aminopterin in aqueous solutions. The reproducibility of the signal was evaluated at three different concentrations of aminopterin producing relative standard deviations ranging from 3.57% at the 5 x 10-10 rnol dm-3 level to 2.49% at the 1 x lo-* rnol dm-3 level ( n = 10). The electrode was applied to the determination of aminopterin in urine resulting in a limit of detection (S/ N = 3) of 2.5 x Keywords: Mercury thin film ultramicroelectrode; adsorptive stripping voltammetry; aminopterin; urine rnol dm-3 without the employment of any pre-treatment of the urine.Much of the work reported to date relating to mercury thin film electrodes has been carried out using conventional sized graphite or glassy carbon electrodesl-3 due to their inertness and simplicity of use. It was found that the deposition of mercury occurs at sites of varying activity and that the deposition of larger amounts leads to formation of mercury droplets, the size and distribution of which depends on the deposition potential. More recently, the popularity of microelectrodes has grown rapidly due to the recognition that many of the undesirable aspects of electrochemical and electroanalytical techniques can be reduced or eliminated by virtue of their use.Microelectrodes possess several advantages over conventional sized electrodes ,4-6 including reduced capacitative charging currents and increased mass transport rates due to the radial component of mass transport; these consequently lead to excellent signal-to-noise (S/N) characteristics. Thus, much work has been directed towards the formation and characteri- zation of mercury films of microelectrodes using various substrates. Various workers studied the possibility of using iridium as a substrate7-11 and found that it was suitable for application to the adsorptive stripping analysis of several metals without problems of intermetallic compound formation.Platinum has also been widely studied as a substrate12-14 and applied to the flow injection anodic stripping voltammetry of various heavy metal ions.15 Silver has also been used to support mercury depositsl"17 and was reported to yield a coherent surface of the entire mercury deposit and to cover the base disc completely . Numerous studies and applications of mercury thin film carbon fibre electrodes to the anodic stripping voltammetricl8- 22 and potentiometric stripping analyses23 of heavy metals have appeared in the literature. Carbon fibre has been * Presented at SAC '92, an International Conference on Analytical + Permanent address: School of Chemical Sciences, Dublin City * To whom correspondence should be addressed. Chemistry, Reading, UK, September 20-26, 1992.University, Dublin 9, Ireland. reported to be a suitable inert substrate and yields stable and reproducible films in the form of mercury micro-droplets. Despite the numerous studies of mercury thin film carbon fibre microelectrodes using inorganic substances no reports have been located relating to their use in cathodic stripping voltammetry of organic compounds. Aminopterin was the first antifolate compound to show proven success in the treatment of cancer24 and in the past has been used widely for the successful treatment of acute leukemia. Today, it has limited use in cell fusion experiments to select for hybrid cells by killing the unfused cells, which are deficient in enzymes for nucleotide salvage pathways.25 However, more importantly the analogues of aminopterin are under constant study26-31 in an effort to develop more active and less toxic drugs for the treatment of various cancers, including leukemia, lung cancer and head and neck cancer.Thus aminopterin analogues continue to hold investigational interest and many are currently undergoing phase 1 and phase 2 clinical trials. The majority of the analogues maintain the pteridine ring of aminopterin intact as substitution at the N-10 position is one of the main routes to increasing the antileukemic effectiveness.32 Thus the first electrochemical reduction process, which is the process of analytical importance (described later) will be common to all. Few methods have been reported in the literature for their determination in biological fluids. A promising approach has been reported however by Tellingen et aZ.33 using high- performance liquid chromatography with fluorimetric detec- tion.In this paper the parent compound, aminopterin, was studied as a model for the aminopterin analogues bearing in mind that the parent compound and the analogues of major investigational interest have the analytically important revers- ible reduction process in common. The cyclic voltammetric behaviour of aminopterin at a static mercury drop electrode (SMDE) is presented. The conditions for optimum mercury deposition on carbon fibre electrodes, optimum phase-selec- tive a.c. voltammetric conditions and optimum methodology for the determination of aminopterin in aqueous solutions and urine samples are subsequently presented and discussed.650 ANALYST, JUNE 1993, VOL. 118 Experimental Reagents and Materials Aminopterin was purchased from Sigma and used without further purification.Stock solutions (1 x rnol dm-3) were prepared daily in 1 x 10-* rnol dm-3 sodium carbonate and stored at 4 "C in the dark. A 0.1 rnol dm-3 ammonium acetate buffer (pH 5 ) was prepared by adjusting 0.1 rnol dm-3 acetic acid to pH 5 using ammonia solution and used as the background electrolyte throughout. All other reagents were of analytical-reagent grade, including: Hg(N03), (Merck); hydrochloric acid (Panreac); and acetic acid (Panreac). All solutions were prepared using de-ionized water obtained by passing distilled water through a Milli-Q water purification system (Millipore). All de-aerations were carried out using purified (N-48) nitrogen (<1 ppm 0,) (Sociedad Espaiiola del Oxigeno).The urine analysed consisted of pooled human urine samples obtained from healthy individuals spiked with various amounts of aminopterin stock solutions to achieve the desired concentration of the drug in urine. Instrumentation All cyclic voltammograms were obtained using a Metrohm VA scanner (E-612) linked to a Metrohm VA detector (E- 61 l ) and a Graphtex x-y recorder (wx-4421). A Metrohm EA- 290 (Kemula) SMDE was used as the working electrode. All potentials are referred to an Ag-AgC1-KCI (3 rnol dm-3) reference electrode. The microelectrode studies were carried out using a Metrohm (Herisau) Model E-506 Polarecord. A 20 cm3 electrochemical cell was used, which allowed the working electrode, reference electrode, counter electrode and nitrogen delivery tube to be fixed in position through a Plexiglas cover. The carbon fibres used were supplied by Donnay and had a nominal diameter of 7.5 pm.All pH measurements were made using a Crison micropH Model 2001 pH meter. Microelectrode Preparation Initially all fibres were washed sequentially in 10% v/v nitric acid, water and acetone, respectively, and then dried at 70 "C. The electrodes were then prepared by microscopically insert- ing a single fibre in the eye of a 100 mm3 plastic micropipette. The eye was sealed either by heat sealing using a soldering iron or by using low viscosity resin, kit TK4 (A. R. Spurr). This resin was polymerized by placing in an oven at 70°C overnight. Following polymerization the electrode was back- filled with mercury and the electrical contact was made using a copper rod that had been filed to remove surface oxides.The electrode was then sealed using the low viscosity resin. Various lengths of fibre were studied with respect to stability of the mercury film and peak height and shape of the aminopterin cathodic stripping peak, and a length of 3 mm was found to be optimum. By using shorter fibres very low currents were produced. Longer fibres yielded higher currents but lower stability of the mercury film, particularly during medium exchange as a result of physical vibration. Determination of Aminopterin Aminopterin was found to precipitate in the presence of mercury(I1) because of the formation of insoluble mercury salts; therefore, two separate cells were employed.After prior activation of the carbon fibre, by initially dipping in chromic acid for 2 min followed by 2 min in concentrated nitric acid, the mercury film was generated under the optimum conditions (described in Results and Discussion). Once formed, the mercury thin film microelectrode was rapidly transferred to the analytical cell containing the ammonium acetate (pH 5 ) electrolyte. This transfer, inevitably, caused a decrease in the amount of mercury present in the film due to both physical detachment of the mercury droplets and air oxidation of the mercury. The transfer procedure was studied in order to minimize these losses. The optimized procedure involved the rapid but careful removal of the first cell in a downward direction, followed by careful cleaning of the reference and counter electrodes and placement of the analytical cell in an upward direction making sure that vibrations of the fibre, which is the main potential cause of loss of mercury film, were minimized.This transfer lasted approximately 10 s. Through- out the procedure a closed circuit was maintained with the potential set at the film deposition potential of -800 mV. Once the electrodes had been placed in the analytical cell the potential was moved to -1400 mV for 30 s to ensure that a clean film was present for the first analysis. Both solutions were de-aerated using nitrogen for 15 min prior to film formation and analysis, respectively, and a nitrogen blanket was employed over the electrolyte throughout the analysis.Results and Discussion Cyclic Voltammetry In this work voltammograms were recorded in a 0.1 rnol dm-3 ammonium acetate (pH 5 ) medium containing 1 x mol dm-3 aminopterin. At this low concentration, the responses seen were adsorption controlled. The cyclic vol tam- metric behaviour of aminopterin in this medium is shown in Fig. l(a), which shows three reduction processes, the first process being reversible, followed by two irreversible pro- E N versus Ag-AgCI- -0.4 -0.8 EN Fig. 1 (a) Cyclic voltammogram of 1 x 10Ph rnol dm-3 aminopterin in 0.1 mol dm-3 ammonium acetatc (pH 5 ) electrolyte using an SMDE of drop area = 2.2 mm2; and scan rate = 100 mV S Y ' . ( b ) Direct current adsorptive stripping voltammctry of a uiesccnt solution of 1 x rnol dm-3 aminopterin in 0.1 mol dmq ammonium acetate electrolyte on an SMDE of drop area 2.2 mm2; scan rate = 100 mV s-l; E,,, = 0 V; accumulation times: A.0; B , 60; and C, 240 sANALYST, JUNE 1993, VOL. 118 65 1 cesses. It has been proposed34 that the first process (1c at -620 mV) is a 2e-/2H+ reduction of the pteridine ring to yield the 5,8-dihydro derivative. The smaller current of the anodic response ( l a at -530 mV) is due to the subsequent tautomeri- zation of the 5,8-dihydro derivative to yield the 7,8-dihydro derivative, which cannot be re-oxidized back to aminopterin. The second (2c at - 1030 mV) and third (3c at - 1210 mV) reduction processes are under kinetic control because they are dependent on the chemical rearrangement outlined above. The second process is due to the 2e-/2H+ reductive cleavage of the dihydro derivative, produced during the first reduction process, between the C-9 and N-10 positions to yield R-NHI and the 7,8-dihydro derivative of the pteridine moiety.The third reduction process is then due to the subsequent 2e-/2H+ reduction of this 7,8-dihydro derivative to the 5,6,7,8- tetrahydro derivative. If the direction of the potential scan is switched immediately after the first process, the anodic response increases because the majority of the 5 &dihydro derivative has not yet been tautomerized to the 7,8-dihydro derivative. The effect of accumulation at 0 V using both open circuit and applied electrolysis was then studied with respect to the first (reversible) process. Similar accumulation rates were observed with and without applied electrolysis.Fig. l(h) demonstrates accumulation under the influence of applied electrolysis. Accumulation in open circuit would obviously present practical advantages in terms of selectivity. However, with a view to working with a mercury thin film microelec- trode, accumulation in open circuit is not possible because electrolysis must be applied at all times to ensure the stability of the mercury film. Mercury Thin Film Optimization The formation of the optimum mercury thin film is a critical factor in the development of a sensitive and reproducible electrode. Mercury deposition on carbon fibres can be in the form of a thin film or can be increased to an almost spherical shape. The deposition of larger amounts of mercury for adsorptive stripping voltammetric applications might seem to be advantageous, facilitating the adsorption of more analyte; however, in this work it was seen that larger amounts of mercury produced unstable films and it seemed that the physical structure of the film was more important than the actual amount of the mercury present. Following a study of the literature and previous studies (unpublished) carried out in these laboratories, mercury nitrate in hydrochloric acid was selected as the mercury salt solution to be studied for the film formation.Therefore, a series of experiments were carried out to find the optimum concentrations of both the mercury salt and the hydrochloric acid. The mercury salt concentration was varied between 1 x 10-I and 1 x rnol dm-3 and a film was formed at each concentration by applying a deposition potential of -1200 mV €or 30 s.After formation the film was anodically stripped and the stripping peak was studied as an indication of the morphology of the film. At high conccntra- tions of the mercury salt (1 X 10-I rnol dm-3) the anodic stripping peak was broad and short. As the concentration was decreased to 1 x mol dm-3 the stripping peak became sharper and the peak height increased dramatically. At this concentration the mercury is thought to exist as numerous micro-droplets of high surface area. Below 1 x rnol dm-3 the peak height decreases as there is a decrease in the number of droplets on the carbon fibre surface. By using the same experimental criteria the molarity of the HC1 was studied in the range 1-6 rnol dm-3.The peak height of the mercury stripping peak increased with molarity up to 5 rnol dm-3 HCI and then began to decrease. Thus, it is thought that at a concentration of 1 x mol dm-3 Hg(N03)I in 5 rnol dm-3 HCI the mercury exists in the form of numerous micro- droplets on the fibre surface. At different concentrations the quality of the film deteriorates in terms of a decrease in the number of droplets or a growth in the size of the droplets with an overall effect of a decrease in the surface area of mercury. Following this, two important parameters, namely, the electrodeposition potential (Efilm) and deposition time (tfilm) were studied. By employing a solution of 1 x rnol dm-3 Hg(N03)2 in 5 rnol dm-3 HCI a series of films were formed sequentially at various deposition potentials using a depo- sition time of 60 s.Each film was subsequently stripped and studied as before. The electrodeposition potential was varied between -100 and -1200 mV in 100 mV intervals. As the applied potential became increasingly more negative the peak height of the mercury stripping increased down to a potential of -900 mV at which point it began to decrease again. It would seem that as the potential becomes more negative the number of surface active sites on the carbon fibre becomes increasingly larger and hence the number of mercury droplets increase. At potentials more negative than -800 mV hydrogen gas production, which potentially causes droplet detachment, became evident. Fig. 2(a) represents the results obtained.In order to ensure a stable film a deposition potential of -800 mV was chosen for further studies. The deposition time (tfilm) was then studied using the same criteria as in the previous study and was followed by a study of the aminopterin phase-selective a.c. cathodic stripping res- ponse with respect to tfilm. As tfilm was increased the height of the mercury stripping peak increased up to a ttilm of 90 s after which time it began to decrease again. This is represented in Fig. 2(b). At each deposition time the electrode was 2t 0 1 ’ ’ ’ I ’ 1 4 - d 1200 1000 800 600 400 200 0 0 50 100 150 200 250 300 50 100 150 200 250 & i d s t .. (+) EImV - Fig. 2 ( a ) Optimization of the mercury thin film formation on carbon fibre in terms of the applied deposition potential (Eel,) using direct current adsorptive stripping voltammetry.Mercury salt solution = 1 X rnol dmP3 Hg(N03)2 in 5 rnol dm-3 HCI; film deposition time (ftilm) = 60 s; film was stripped anodically at 100 mV s-l. (b) Optimization of the mercury thin film formation on carbon fibre in terms of deposition time (tfilm) using direct current adsorptive stripping voltammetry (series 1). Mercury salt solution = 1 X lo-’ rnol dm-3 Hg(N03)2 in 5 rnol dm-3 HCI; Etilm = -800 mV; film was stripped anodically at 100 mV s-l. (c) Optimization of the mercury thin film formation on carbon fibre in terms of tfilm versus aminopterin cathodic stripping response using phase-selective alternating current adsorptive stripping voltammetry. Mercury salt solution = 1 X lo-’ rnol dm-3 Hg(N03)2 in 5 rnol dm-3 HCI; Etilm = -800 mV.5 X mol dm-3 aminopterin solution in 0.1 rnol dm-3 ammonium acetate elcctrolyte (pH 5 ) ; aminopterin accumulation time = 60 s at an applied potential of 0 V. (For procedures and a.c. voltammetric conditions see text). (d) Two typical anodic stripping peaks of the mercury film from the carbon fibre electrode using direct current adsorptive stripping voltammetry. Mercury salt solution = 1 X loP3 rnol dm-’ Hg(N03)2 in 5 rnol dm-3 HCI; ttilm = 90 s; Etit, = -800 mV; scan speed = 100 mV S-I652 ANALYST, JUNE 1993, VOL. 118 transferred to a 5 x mol dm-3 aminopterin solution in pH 5 ammonium acetate medium and aminopterin was accumulated for 60 s at an applied potential of 0 V and then cathodically stripped. Similarly a mercury deposition time of 90 s produced the best aminopterin stripping peak.As outlined in Fig. 2(c) the aminopterin response began to decrease when higher mercury deposition times were employed. It seems that at deposition times of greater than 90 s a growth in droplet size rather than droplet number occurs, thus causing an overall decrease in the surface area of mercury available for analyte adsorption, and also producing a less stable film. Therefore, an optimum deposition time of 90 s was chosen. Between each film formation the fibre was regenerated producing a clean surface free of mercury; this contributed to the reproducibility of the film. The optimum conditions for this regeneration were found to be the application of a potential of +790 mV for 40 s, which oxidized all the Hg(0) on the fibre surface.Fig. 2(d) shows two typical anodic stripping peaks of the mercury film after formation under optimum conditions. The reproducibility of film formation under these optimum conditions was studied by forming and subsequently anodically stripping six consecutive films. Measurement of peak current (ip) of the stripping peaks yielded a relative standard deviation (RSD) of 1.11% (n = 6). To study the reproducibility of medium exchange of the film, the film- coated electrode was rapidly transferred to a blank 5 mol dm-3 HCI solution and stripped. The transfer reproducibility yielded an RSD of 13.3% (n = 6). Therefore, this dictated the necessity to use the same film throughout the whole analysis run using regeneration of the film surface between each measurement.Optimization of Phase-selective a.c. Conditions Once the mercury thin film microelectrode had been trans- ferred to the analytical cell, phase-selective a.c. voltammetry (AC1-Tast) was employed to exploit the reversibility of the reduction process of interest. Thus the conditions were optimized to yield the best aminopterin response possible. A study of the effect of the phase angle (0) was carried out by measuring the peak characteristics of a 5 x mol dm-3 aminopterin solution over the complete range of angles. It was found that at angles close to 90" the current was essentially the charging current component, whereas angles closer to 0" produced the best analytical signals. An optimum angle of 18" was chosen as under the experimental conditions employed it was seen to produce the best discrimination of the faradaic current against the capacative current, allowing easier measurement of the analytical signal.As the phase angle was increased to angles approaching 90" a dramatic increase in the background current was observed resulting in an increasingly poor analytical signal. The applied a.c. voltage amplitude (AE) was studied under the above conditions and a linear relationship between the signal (aminopterin cathodic strip) and AE was observed up to 20 mV, according to the following equation: ilnA = 1.27 x 10-1 AElmV -0.015, (Y = 0.9998). The instrument used worked at a fixed frequency of 75 Hz, and did not permit any frequency variation. A scan speed of 10 mV s-l was chosen as in a.c.voltammetry the scan speed does not significantly affect the signal. Aminopterin Accumulation Studies Accumulation studies were carried out in quiescent solutions for five different concentrations of aminopterin, i.e., 5 X l O - l O , 1 x 5 x 1 x 10Ws and 5 x rnol dm-3. An accumulation potential of 0.0 V was applied in all cases as it produced the best response. More negative accumulation potentials produced slower rates of accumulation and more positive potentials led to oxidation of the film, which is stripped at +260 mV in the acetate medium. Activation of the mercury film between each measurement was carried out by holding the potential at -1.4 V for 30 s. This removed the adsorbed reduction products of aminopterin producing a fresh film for the next measurement.At all concentrations studied the electrode responded in the same way. Initially the response increased linearly with time, then at a certain point the slope decreased and the response began to increase again in a linear fashion at this lower slope. For example, this slope change occurred at an accumulation time (tact) of 15 s for 5 x 10Wg rnol dm-3 aminopterin and at a tacc of 360 s for 5 x lo-'* rnol dm-3 aminopterin. After this second linear portion the slope again decreased to a point where the response became virtually independent of accumulation time (lace). Fig. 3 represents the first two linear regions of the curves. The third region of the curves can be interpreted as saturation of the electrode surface. However, the response still increases with increasing concentration. The changes in slope of the accumulation curves can be explained in terms of the surface of the mercury film becoming modified at higher concentra- tions of the drug, because at higher concentrations the potential of the aminopterin cathodic stripping peak shifts to a few millivolts more negative.A comparison of the curves indicates that at higher concentrations of aminopterin, the slope growth was faster. This phenomenon was previously explained35 as being due to molecular interactions between adsorbed molecules on the electrode surface. Careful exam- ination of the accumulation curves facilitated the selection of suitable accumulation times (tact) for further studies. Calibration Graphs By carefully selecting accumulation times, different concen- tration ranges of aminopterin could be studied.Obviously, at higher accumulation times lower concentrations can be detected, but the linear dynamic range was smaller than with shorter accumulation times. Thus, a compromise between limit of detection and linear range must be made to suit the analysis of interest. By using an accumulation time of 180 s and an applied potential of 0.0 V, a linear calibration plot was obtained between 2 x 1O-IO and 8 x 10-9 rnol dm-3 aminopterin in aqueous solutions according to the following equation: i h A = 2.88 x lo-' x caminoplmol dm-3 + 0.360 (Y = 0.9994). By employing a shorter accumulation time higher concen- trations of the compound can be studied. For example, by employing an accumulation time of 10 s, the electrode could be used in the 1 x 10-8-1 x rnol dm-3 region yielding the 0 50 100 150 200 250 300 350 400 450 t X C l S Fig.3 Aminopterin accumulation curves on the mercury thin film carbon fibre electrode using phase selective a x . adsorptive stripping voltammetry; E,,, = 0 V; electrolyte = 0.1 mol dm-3 ammonium acetate (pH 5 ) . (For procedure and a.c. voltammetric conditions see text). A, Series 1, 5 x rnol dm-3 aminopterin; B, series 2, 1 X mol dm-3 aminopterin; C , series 3, 5 x mol dm-3 aminopterin; D, series 4, 1 x rnol dm-3 aminopterin; and E, series 5 , 5 x lo-'" rnol dm-3 aminopterinANALYST, JUNE 1993, VOL. 118 653 following equation: ilnA = 5.74 x X caminodmol dm-3 + 0.525 ( I - = 0.9991). By using an accumulation time of 180 s the detection limit based on the concentration that yielded a stripping peak three times greater than the background level was 1 x 10-lO mol dm-3 of aminopterin.Fig. 4(a) represents the increase in peak current with the increase in aminopterin concentration at low aminopterin concentrations. This figure also outlines how the baselines were constructed at these low concentrations. It can be seen from Fig. 4(b) that baseline construction becomes increasingly easier due to the improved peak shape at higher aminopterin concentrations. Reproducibility and Stability The reproducibility of the aminopterin cathodic stripping response was studied at low (5 X rnol dm-3), medium (1 x rnol dm-3) and high (1 X rnol dm-3) concentrations. This study involved recording ten consecutive voltammograms at each concentration and calculating the RSD of the response.At a concentration of 5 x 10-lO rnol dm-3 an accumulation time of 120 s was employed and an I I t .- -(-I E -(-I E Fig. 4 (a) Phase-selective a.c. adsorptive stripping voltammograms of aminopterin in aqueous solution using a mercury thin film carbon fibre electrode; E,,, = 0 V, t,,, = 180 s. A, Blank pH 5 acetate medium; B, 2 x 10-lO rnol dm-3; C, 4 X mol dm-3; D, 6 X 1O-IO rnol dm-3; E, 8 x 10-'0 mol dm-3; and F, 1 X rnol dm-3 aminopterin. (For a.c. voltammetric conditions see text). (b) Phase- selective a.c. adsorptive stripping voltammogram of 5 X rnol dm-3 aminopterin in aqueous solution using a mercury thin film carbon fibre electrode. Accumulation potential = 0 V; accumulation time = 10 s.(For a.c. voltammetric conditions see text) Fig. 5 Standard additions analysis of urine spiked with 5 x rnol dm-3 aminopterin using phase-selective a.c. adsorptive stripping voltammetry. 0.5 cm3 of a 1 + 5 dilution of the spiked urine was injected into the 20 cm3 analytical cell followed by standard additions of a 1 x mol dm-3 stock aminopterin solution. E,,, = 0 V; tact = 40 s. A, Blank urine; B, sample; C, 2 mm3 of standard; D, 4 mm3 of standard; E, 6 mm3 of standard; F, 10 mm3 of standard; and G, 14 mm3 of standard. (For a.c. voltammetric conditions see text)654 ANALYST, JUNE 1993, VOL. 118 RSD of 3.57% (n = 10) was obtained. At an aminopterin concentration of 1 X mol dm-3 an RSD of 2.79% ( n = 10) was obtained using an accumulation time of 60 s.At a concentration of 1 X mol dm-3 the RSD was 2.49% (n = 10) using an accumulation time of 20 s. This reproducibility is also an indication of the stability of the mercury film using the optimized deposition and regeneration conditions. Thus, the electrode proved to be very sensitive and reproducible in aqueous solutions of aminopterin. The same electrode could be used for a period of at least 4-6 weeks. The shape of the anodic stripping peak of the mercury film was a good indicator of the state of the fibre surface and when this stripping peak was seen to broaden and decrease in peak height the surface of the fibre was regenerated by dipping in chromic acid for 30 s. In this way the chromic acid re-oxidized any reduction products of aminopterin that might have adsorbed on the surface of the carbon fibre, thus providing a cleaned surface. Aminopterin Determination in Urine Pooled urine samples were spiked with appropriate amounts of aminopterin stock solution to achieve final concentrations of 5 x 1 x and 5 x rnol dm-3 aminopterin in urine.Various dilutions of these urine samples were made using the ammonium acetate electrolyte so as to reduce the effect of interfering compounds naturally present in urine. Dilutions of 1 + 4 , l + 9 and 1 + 19 were made and 0.5 cm3 of the resulting solutions was injected into the 20 cm3 analytical cell. Blank urine samples were injected to ensure that no interference was seen at the same potential as that of aminopterin. In each case after injection of the spiked urine to the cell, appropriate standard additions were made and these standard additions plots were used to calculate the concentra- tion of aminopterin in the original sample, by extrapolation. Two separate analyses were carried out at each spiked concentration involving two separate dilutions.At a spiked concentration of 5 x 10-6 mol dm-3 aminopterin an RSD of 1.75% ( n = 2) was achieved by employing a 1 + 10 dilution using a 90 s accumulation time and a 1 + 19 dilution using 120 s accumulation. When a spiked concentration of 1 x mol dm-3 was studied, dilutions of 1 + 4 (tact = 60 s) and 1 + 9 (tact = 120 s) were executed and an RSD of 2.45% (n = 2) was obtained. At the lower concentration of 5 x mol dm-3 aminopterin, a 1 + 4 dilution was always used with an accumulation time of 120 s as at lower dilutions the signal was difficult to measure.The RSD of the signal at this concentra- tion was 16.5% (n = 2). A seemingly obvious way to improve this signal is to increase the accumulation time. However, at higher accumulation times adsorption of naturally occurring compounds in urine becomes more evident causing increased electrode passivation. Throughout the analysis of urine the gradual build up of these compounds on the electrode was evident. This phenomenon can clearly be seen in Fig. 5 where initially two extra peaks can be seen in addition to the aminopterin peak. These peaks from urine gradually increase and eventually merge. However, these responses occur at a different potential to that of the analyte. Fig. 5 is a standard additions analysis of a 1 + 4 dilution of a 5 x rnol dm-3 spiked urine sample.Baseline construction was carried out in the same way as in Fig. 4. Adsorption of these compounds from urine was less evident at the lower dilutions of 1 + 9 and 1 + 19. As with the analysis of aqueous solutions fibre regeneration was attempted using chromic acid. Although this regeneration resulted in some improvement, the fibre never returned to its original activity as the other compounds in urine caused gradual passivation of the fibre surface. Therefore, the electrode usually had to be replaced after approximately 20 measurements. The use of a pre-analysis extraction of the analyte from the urine using CI8 solid-phase extractisn cartridges would undoubtedly lower the limit of detection and minimize the passivation of the electrode surface.However, the objective of this work was to develop a system that could be applied to the determination of aminopterin analogues without the employment of pre-extraction procedures. Conclusion In this paper the cyclic voltammetric behaviour of aminopterin was presented and discussed. A mercury thin film carbon fibre ultramicroelectrode was developed so that the advantageous features of ultramicroelectrodes, adsorptive preconcentration and phase-selective a.c. voltammetry were combined. This system was applied to the determination of aminopterin as a model for its analogues that retain the pteridine ring within their structure. It yielded a sensitive and reproducible method, facilitating the possibility of both pre-administration and post-administration analysis of these compounds.This ultramicroelectrode system could potentially be applied to other organic compounds that exhibit accumulative behaviour on mercury. The incorporation of this type of system into a flow-injection system format offers a potentially useful area of investigation. Related work is currently underway in these laboratories. The authors thank the ERASMUS programme and the Spanish Ministry of Education and Science (DGICYT) project PB87-1041 for their 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 22 23 24 25 26 27 References Stulikova, M.. J. Electroanal. Chem. Interfacial Electrochem., 1973, 48, 33. Lieberman, S. H., and Zirino, A., Anal. Chem., 1974, 46, 20.Loung, L., and Vydra, F., J. Electroanal. Chem. Interfacial Electrochem., 1974, 50, 379. Pons, S . , and Fleischmann, M., Anal. Chem., 1987,59. 1391 A. Wightman, R. M., Anal. Chem., 1981, 53, 1125A. Johnson, D. C., Ryan, D. M., and Wilson, S. G., Anal. Chem.. 1988, 60, 147R. Golas, J., Galus, Z., and Osteryoung, J., Anal. Chem., 1987, 59, 389. Golas, J., and Kowalski, Z., Anal. Chim. Acta, 1989,221,305. Kounaves, S. P., and Deng, W., J. Electroanal. Chem., 1991, 301, 77. Kounaves, S. P., and Buffle, J . , J. Electroanal. Chem., 1987, 216, 53. Kounaves, S. P., and Buffle, J., J. Electroanal. Chem., 1988, 339, 113. Gunawardcna, G., Hills, G., and Scharifker, B . , J. Electroanal. Chem., 1981, 130, 99. Wehmeyer, R. K., and Wightman, R. M., Anal. Chem., 1985, 57, 1989. Stojekand, Z., and Osteryoung, J., Anal. Chem., 1988,60,131. Tay, E. B.-T., Khoo, S.-B., and Loh, S.-W., Analyst, 1989,114, 1039. Stojek, Z., and Osteryoung, J., Anal. Chem., 1989, 61, 1305. Ciszkowska, M., Penczek, M., and Stojek, Z., Electroanalysis, 1990, 2, 203. Schulze, G., and Frenzel, W., Anal. Chim. Acta, 1984,159,95. Golas, J., and Osteryoung, J . , Anal. Chim. Acta, 1986, 181, 211. Jennings, V. J., and Morgan, J. E., Analyst, 1985, 110, 121. Sottery, J. P., and Anderson, C. W., Anal. Chem., 1987, 59. 140. Baranski, A. S., Anal. Chem., 1987, 59, 662. Hua, C., Jagner, D., and Renman, L., Talanta, 1988, 35, 597. Farber, S., Diamond, L. K., and Mercer, R. D., New Engf. J . Med., 1948, 238. 787. Fehlner, P. F., Bencsath, A . , Lam, T., and King, T. P., Immunol. Methods, 1987, 101, 141. Souhami, R. L., Rudd, R. M., Spiro, S. G., Lamond, P., and Harper, P. G., Cancer Chemother. Pharmacol., 1992, 30,465. Green, M. D., Sherman, P . , and Zalcberg, J., Invest. New Drugs, 1992, 10, 31.ANALYST, JUNE 1993, VOL. 118 655 28 29 30 31 32 Sirotnak. F. M., Degraw, J. I., Chello, P. L., Moccio, D. M., and Dorick, D. M., Cancer Treat. Rep., 1982, 66, 351. Rosowsky, A., Forch, A. R., Bader, H., and Freisheim, J . H., J. Med. Chem., 1991, 34, 1447. Rosowsky, A., Forsch, A. R., Moran, G. R., and Freisheim, J. H . , J . Med. Chem., 1991, 34, 227. Rosowsky, A.. Bader, H., and Freisheim, H., J. Med. Chem., 1991, 34, 203. Holland, J . F., and Frei, E., in Cancer Medicine, LEA Febiger, Philadelphia, 1982, pp. 77.5-789. 33 Tellingen, 0. V., Beijnen, J. H., van de Woude, H. R., Bruning, P. F., and Nooyen, W. J., J. Chromatogr. Biomed. Appl., 1990, 529, 135. 34 Dryhurst, G., in Electrochemistry of Biological Molecules, Academic Press, New York, 1977, pp. 324-357. 3.5 Laviron, E. J., J . Electroanal. Chem., 1979, 100, 263. Paper 2105157E Received September 25, 1992 Accepted November 23, 1992