Evaluation of eimac lamp as excitation source for molecular

Evaluation of eimac lamp as excitation source for molecular fluorescence with application to quantitation of ergotaminine in plasma by high-pressure l...
0 downloads 0 Views 887KB Size
the basic information needed for calibration has been obtained for a particular chromatographic system, this information can be used repeatedly as long as the proper chromatographic conditions are adhered to. Through the future development of proper separation modes, it is anticipated that the general approach employed for the calibration procedure described herein can be extended to include nearly all types of classwise separations as long as the necessary prior compositional knowledge, often attainable by the application of other methods and techniques, can be realized. The inability to obtain such compositional information for certain types of materials does, however, present one of the most severe limitations to this technique.

ACKNOWLEDGMENT The authors gratefully acknowledge R. W. Simpson who prepared the diagrams in this paper and C. E. Lehman who completed the regression analyses on analytical data. Sincere gratitude is also expressed to H. Sinsel for services rendered in preparing this paper.

LITERATURE CITED (1) P. Vogel and T. Wieske, Anal. Cbaract. Oils. Fats, Fat Prod., 2, 99 (1968). . (2) G.Jurriens. Anal. Cbaract. Oils, Fats, Fat Prod., 2, 217 (1968). (3) K. Haukawa, Kagaku ToKogyo(Osaka), 45, 206(1971). (4) A. A. Y. Sheata and J. M. DeMan, Con. lnst. Food Techno/., J., 4, 38 (1971). (5) A. Yamanoka, Yukagaku, 22, 533 (1973). (6) A. Yamamoto, S. Adachi, and T. Ishibe, Rinsbo Byari, Rinjizokan, 19, 141 (1972). (7) Y. Isoda, Yukagaku, 22, 4756 (1973). (8)J. P. Wolff, Aliment. Vie, 59, 15 (1971). (9) A. Kuksis, Fette, Seifen, Anstrichrn., 73,332 (1971). (10) K. J. Bombaugh, Mod. Pract. Liquid Chromatogr., Mater. Course 1970, Pub. (1971), 237-85. (11) L. D. Metcalfe, A. S. Schmitz. and J. R. Pelka, Anal. Chern., 38, 514 (1966). (12) 8. Borgstrom, Acta Pbysica Scand., 25, 111 (1952). (13) P. Savary and P. Desncrelle, Bull. SOC.Cbim. Fr., 1954, 936. (14) J. Hirsch and E. H. Ahrens, Jr., J. Biol. Cbem., 233, 31 1 (1958). (15) R. E. Kirk and D. F. Othmer, “Encyclopedia of Chemical Technology”, Vol. 6, Interscience Publishers, New York, 1951, pp 142-144.

RECEIVEDfor review April 30, 1975. Accepted July 1,1975.

Evaluation of Eimac Lamp as Excitation Source for Molecular Fluorescence with Application to Quantitation of Ergotaminine in Plasma by High-pressure Liquid Chromatography Robert J. Perchalski,’~~ James D. W i n e f ~ r d n e r and , ~ B. J. Wilder1p2 Veterans Administration Hospital, Gainesville, Fla., and Departments of Chemistry and Neurology, University of Florida, Gainesville, Fla.

A 150-W Eimac illuminator (a miniature, high-pressure xenon arc lamp with an integral collimating mirror) Is shown to give two to three times the photon flux of a standard xenon lamp in an ellipsoidal condensing system, when used for excitation of condensed phase molecular fluorescence. In the case studied, detection limits are almost an order of magnitude lower with the Eimac lamp. Incorporation of this lamp Into a high-pressure liquid chromatographic detection system Is discussed, and a basis for quantitatlon of plcogram amounts of the ergot alkaloids in blological samples is presented.

A brief account of the history and chemistry of the ergot alkaloids is given by Brazeau ( 1 ) . The primary medical uses of ergot are in the treatment of migraine headaches and, more recently, in the treatment of cerebrovascular insufficiency (reduced cerebral blood flow) in geriatric patients. Ergotamine tartrate is given a t the onset of a migraine headache to alleviate pain. This action is due to the vasoconstrictor effect of the natural amino acid alkaloids, which brings about a reduction in the amplitude of pulsations of cranial arteries; this amplitude is proportional to the intensity of the headache pain (2). The 9,lO-dihydro derivatives of the ergotoxine group (ergocornine, ergocristine, and ergocryptine) are given in very low doses to treat cerebrovasResearch Service (MRIS-9403-01), Veterans Administration Hospital, Gainesville, Fla. 32602. Department of Neurology, University of Florida, and Neurology Section, Veterans Administration Hospital, Gainesville, Fla. Department of Chemistry, University of Florida, Gainesville, Fla. Author to whom correspondence should be directed.

cular insufficiency. In contrast to the natural alkaloids, the overall effect of these hydrogenated alkaloids is very slight vasodilation. However, the action of these drugs, in this instance, stems from a modification (normalization) of ganglion cell metabolism, as explained by Emmenegger and Meier-Ruge ( 3 ) , and not from a direct cardiovascular effect. The ergot alkaloids have been routinely analyzed by paper and column chromatography ( 4 ) , and thin-layer chromatography (5-7). The gas-liquid chromatographic behavior of the amine alkaloids, including lysergic acid diethylamide (LSD), was investigated by Agurell and Ohlsson (8).LSD has been quantitated after solvent extraction from blood plasma with B lower limit of detection of 1 ng/ml by fluorescence spectrophotometry (9, 10). A recent modification ( 1 1 ) increased the precision and sensitivity of this technique by comparing the fluorescence signal of the LSD to the response obtained after photochemical conversion of the LSD to the nonfluorescent “lumi” derivative (12). This hallucinogen was detected in human plasma after oral doses of 2 pg/kg of body weight. Bowd et al. ( 1 3 ) measured fluorescence quantum yields and lifetimes of LSD and other ergot alkaloids. Hooper et al. (14), attempting to establish a technique for the determination of ergotamine in biological samples, described a solvent extraction method with measurement of fluorescence in absolute ethanol. A lower limit of 2 ng/ml was claimed for extraction of 5 ml of an aqueous standard. Several groups have published high-pressure liquid chromatographic (HPLC) assays of the ergot alkaloids. Jane and Wheals (15) and Heacock et al. (16) used fluorescence detection with excitation at 350 nm, almost at the edge of the excitation band for most of these alkaloids. Absolute

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12, OCTOBER 1975

1993

detection limits were in the microgram range, adequate only for the analysis of pharmaceutical preparations and street samples. Wittwer and Kluckhohn (17),using ultraviolet absorption detection, obtained a detection limit of 50 ng of LSD. The ergot alkaloids are all high-molecular-weight, thermally unstable compounds, which are administered in doses varying from 0.5 to 10 mg/day, depending on the route of administration and the alkaloid. Such doses generally give plasma concentrations lower than 1 ng/ml, as determined by radioactive tracer studies (18). Recent reports of a thin-layer amperometric liquid chromatographic detector (19-22) have shown that detection limits in the low picogram range are possible with ion exchange HPLC. However, for molecules that are not electrochemically active or are insoluble in nonconducting mobile phases, the fluorometric detector holds the most promise for high sensitivity analysis. A considerable amount of quantitative work has already been done with fluorometric HPLC detectors. Polynuclear aromatic hydrocarbons (23) and physiologicany important indoles in urine (24) have been separated by HPLC and quantitated directly by fluorometry; pesticides (25, 26), phenols in urine (27), and barbiturates in plasma (28) have been separated and quantitated after fluorigenic labeling. Detection limits are routinely in the low nanogram range. Frei and Lawrence (29) recently reviewed the subject of fluorigenic labeling for HPLC quantitation of carbamates, ureas, organophosphorous compounds, aliphatic amines, aldehydes, ketones, biphenyls, and pharmaceuticals, giving a broad survey of applications. Cassidy and Frei (30) described a detector incorporated into a low-cost fluorimeter with a detection limit of 2 ng/ml for quinine sulfate. Various commercially available systems have also been evaluated (31-33). Although detection limits for commercial spectrophotofluorometers are quoted a t parts-per-trillion of quinine sulfate in 0.05M H2S04, these limits have not yet been reported for fluorescent compounds extracted from a biological matrix, even when the detection step has been preceded by a selective chromatographic step. The purpose of the present work is to evaluate a high-pressure xenon arc lamp with an integral parabolic-collimating mirror (Varian Eimac Division, San Carlos, Calif.) as a source for high sensitivity molecular fluorescence studies in the condensed phase, and to propose a high-pressure liquid chromatographic detection system, based on this source, which will be comparable in sensitivity to the electron capture detector in GLC. The Eimac illuminator has been used previously as a continuum source for atomic fluorescence spectrophotometry (34-37). Luthjens (38) briefly mentioned using a 500-W Eimac illuminator in a pulsing circuit as a substitute for a 450-W Osram lamp to measure short-lived transients in absorption spectrometry. He calculated a 20-fold increase in usable photon flux for the Eimac lamp, based simply on the geometry of the light-collecting optics. In this study, a 150-W Eimac lamp is compared to a standard xenon lamp in an ellipsoidal condensing system. The primary advantage of the new system would be increased sensitivity, due to more efficient light gathering at the source. This would allow the use of narrower monochromator slit widths to increase resolution and selectivity, and minimize scatter from the highly curved surfaces of low dead volume flow cells, which are necessary for efficient detection and resolution in a chromatographic system. Finally, a procedure for the quantitation of ergotaminine (a pharmacologically inactive isomer of ergotamine) in plasma by high-pressure liquid chromatography with fluo1994

rescence detection is presented. Based on the detection limits achieved in this study with on-line fluorescence monitoring, a new detection system is proposed, in which an Eimac lamp is used for excitation, to give detection limits below 1 ng/ml for this representative of the ergot family of alkaloids, extracted from plasma.

EXPERIMENTAL Fluorescence Instrumentation. The light source evaluated in this study was a Model VIX-15O-UV, high-pressure, short-arc, 150-W xenon illuminator (Varian Eimac Division). The integral parabolic mirror was coated with aluminum-magnesium fluoride for enhanced UV output. The only dimension of this reflector known exactly was the 25.4-mm diameter; however, the position of the arc was estimated to be between 2 and 4 mm from the vertex (3-mm focal length), giving an approximate depth to the mirror of 13.4 mm. With these values, the solid angle of radiation collected by the mirror into a collimated beam was calculated to be 7.9 sr (2.5 T sr). The lamp was routinely operated in the cw mode a t 12.0 A, 12 V, dc. The power supply and lamp holder with forced-air cooling were Models P15OS-7 and R150-7, respectively (Varian Eimac Division). The collimated beam was directed through a simple quartz lens (108-mm focal length, 32-mm illuminated diameter), located 127 mm from the arc plasma and 108 mm from the entrance slit of the excitation monochromator. The lamp used for comparison was a standard 150-W, high-pressure xenon arc (Hanovia Lamp Division, Newark, N.J.), operated in an ellipsoidal condensing system (American Instrument Co., Silver Spring, Md.). The aluminized off-axis ellipsoidal mirror has an effective area of 2667 mm2 a t a primary focus of 32 mm and collects approximately 2.6 sr (0.83 T sr) of radiation from the lamp. This lamp was powered by a Model 422-829 dc power supply (American Instrument Co.) a t a fixed current of 7.5 A. Both optical systems focused the exciting light on the entrance slit of the Aminco-Bowman Spectrophotofluorometer (American Instrument Co.). The dispersive system with slit positions is shown in Figure 1. Fluorescence was detected with a potted 1P21 photomultiplier tube (American Instrument Co.), powered by a highvoltage dc power supply (Keithley Instrument Co., Cleveland, Ohio). The signal was fed into a nanoammeter, designed by O’Haver and Winefordner (39),and the output was recorded on a 10-mV recorder (Houston Instruments Co., Austin, Texas). The Aminco solid-sample accessory (American Instrument Co.), the optical path of which is shown in Figure 2, was used to obtain the relative spectral dispersion curves of each lamp. A temperature-controlled cell compartment, adapted for 3-mm pathlength square cells, was used for all other work. Because of restrictions on the movement of instrumentation, it was necessary to use an Aminco SPF-125 (American Instrument Co.) with an 85-W, medium-pressure, mercury vapor lamp as the HPLC fluorescence detector. For comparison with the Eimac and Hanovia systems, this instrument was adapted for use with the same 3-mm square cuvet as was used in the Aminco-Bowman spectrofluorometer. This cell compartment was then replaced with a flow-cell adapter (American Instrument Co. No. B19-63019), designed to hold a 25-mm X 2-mm i.d. (78.5-11) quartz flow-cell (American Instrument Co. No. B18-63019). Dead volume in the cell was reduced to about 16 pl by reducing the length of the cell to 5 mm. Teflon spacers, made of %-inch o.d., high-pressure Teflon tubing, were used to position the cell in the adapter so that the entire length of the cell was illuminated by the excitation light beam and monitored for fluorescence. An effective amplification of 5 times was obtained by monitoring the 50-mV output of the instrument with a 10-mV recorder (Houston Instrument Co.). High-pressure Liquid Chromatography. The high-pressure liquid chromatograph-consisting of a gradient pumping system, digital pressure monitor, and multiwavelength UV absorbance monitor-was constructed from commercially available components (Instrumentation Specialties Co., Lincoln, Neb.). The column (Varian Aerograph, Walnut Creek, Calif.) was a 25-cm X 2.2mm i.d. stainless steel tube, packed with a 10-pm (average particle diameter), totally porous silica gel adsorbent. The UV absorbance monitor was replaced by the fluorescence detector described above for high sensitivity studies of ergotaminine. Sample injection was made through a modified septum injection port (Precision Sampling Corp., Baton Rouge, La.), using a Model B-110 high-pressure microsyringe (Precision Sampling Corp.) with a 51-mm needle, which introduced the sample about 1 mm above the head of the column in a 0.8-mm i.d. channel.

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12, OCTOBER 1975

Flgure 1. Dispersive system of the Aminco-Bowman spectrophotofluorometer showing mirrors (M), gratings (G), cell compartment (C), positions of adjustable slits (a-f), and points of entrance of the exciting light (E) and detection (D)

Solvents were degassed for 15-20 min immediately before use in a simplified version of the ultrasonic degasser described by Dell'Ova et al. (40),consisting of a 1000-ml glass reagent bottle, fitted with a small reflux condenser, which is attached to a laboratory vacuum source (200 mm Hg). This unit containing the mobile phase is then placed in a standard ultrasonic cleaner with enough water to almost fill the bath. Observation of the solvent indicated that the most vigorous evolution of gas occurred in the first 30 sec of degassing. Reagents. Solvents and reagents for fluorescence and highpressure liquid chromatography were analytical reagent grade. Quinine sulfate and reserpine (ICN Nutritional Biochemicals Corp., Cleveland, Ohio), Rhodamine B (Eastman Organic Chemicals, Rochester, N.Y.) and ergotaminine (E. M. Merck & Co., Elmsford, N.Y.) were used as purchased. Procedure. Signal-to-Noise S t u d y for the Eimac L a m p System. The three parts of this study were designed to show the effects on signal-to-noise ratio (S/N) of variation of lamp current, photomultiplier voltage, and excitation monochromator entrance 350 nm; A, 450 nm) slit width. A solution of quinine sulfate (L,,, in 0.05M H2S04 (200 ng/ml) was used for all measurements. The sample was contained in a 3-mm-pathlength square cell that was maintained a t 24.3 f 0.5 "C in the temperature-controlled cell compartment. The amplifier time constant was 0.5 sec, and all measurements were recorded for a t least 120 sec. Variation of S I N with L a m p Current. Slit System 1 (Table I) was used, and the Eimac lamp current was varied from 12.5-7.0 A. Sample and blank measurements were taken a t each setting, after allowing a t least a 2-min equilibration time, after a change in lamp current. The lamp was allowed to stabilize a t 13.0 A for a t least 60 min before any measurements were taken. Variation of S / N with Photomultiplier Voltage. After an initial warmup period of 60 min a t -700 V, measurements were taken over the range -700 to -400 V in 60-V increments. Slit System 2 was used, and lamp current was set a t 12.0 A. Variation of S I N with Excitation Monochromator Entrance Slit W i d t h . With the lamp current set a t 12.0 A and the photomultiplier powered a t -580 V, measurements of S/N were made a t slit widths from 5-0.5 mm. Slit System 3 was used. Comparison of the Eimac and Hanovia L a m p Systems. In the three comparative studies of the Eimac and Hanovia lamps, the Eimac lamp was operated at 12.0 A. The photomultiplier was powered at -700 V, and the amplifier time constant was set a t 1sec. Relative Spectral Distribution. Spectral distribution curves for the Eimac system and the Hanovia lamp in the ellipsoidal condensing system were obtained by the procedure of Melhuish (41). The path of the exciting and emitted light within the Aminco solid-sample accessory, used in this study, is shown in Figure 2. Front-surface fluorescence from a solution of Rhodamine B (3 mg/ml in ethylene glycol) in a 1-cm cell was monitored a t an emission wavelength of 615 nm, while the excitation monochromator was scanned from 200-600 nm a t 2.29 nm/sec. The nanoammeter was set to give a full-scale deflection a t A. Slit system 4 was used. Both curves were run within a 60-min period on the same day. The only instrumental change made between runs was the substitution of the Hanovia lamp and optical system for the Eimac system. Photodecomposition of Quinine Sulfate. The time required to reduce the fluorescence signal of a 10-ng/ml solution of quinine sulfate in 0.05M HzS04 (100 fi1 in a 3-mm cuvette) to 90% of the original intensity, was determined for both systems. The slit settings used were Systems 5 and 6 for the Eimac and Hanovia lamps, respectively. The cuvette was tightly closed with a Teflon stopper and maintained a t 24.3 f 0.5 O C in the temperature-controlled cell compartment for 30 min, while the degradation of the fluorescence

Flgure 2. Optic?l path of Aminco solid sample accessory showing the positions of the sample cell (C), plane mirror (M), and adjustable slits (cl,c2, dl, d2)

Table I. Slit Settings for the Aminco-Bowman Spectrophotofluorometer in Millimeters System l a

System 2'

System 3a

System d n

( a ) 5.0

( a ) 4.0

( a ) Varied

( a ) 2.5

( b ) 1.0 ( c ) 0.1 ( d ) 0.1

( b ) 1.0 ( c ) 0.1 ( d ) 0.1 ( e ) 1.0

( b ) 1.0 ( c ) 0.1 ( d ) 0.1 ( e ) 4.0

(f)0.5

(f)0.5

System 6'

System 7'

System

( a ) 5.0

( a ) 5.0

( a ) 2.5

( e ) 5.0 (f)1.0

( b ) 1.0 ( c l ) 0.5 (c2) 0.1 ( d l ) 0.5 ((12) 0.5 ( e ) 1.0

(f)0.5 System

sa

( a ) 5.0

( b ) 3.0

( b ) 3.0 ( b ) 3.0 ( c ) 2.0 ( c ) 3.0 (c) 3.0 ( d ) 2.0 ( d ) 0.5 ( d ) 0.5 ( e ) 3.0 ( e ) 3.0 ( e ) 5.0 (f)5.0 (f)0.5 (f)4.0 See Figures 1 and 2 for designation of slits.

sa

( b ) 1.0

( c ) 0.1 ( d ) 3.0 ( e ) 5.0 ( f ) 5.0

signal was continuously monitored. The signal-to-noise ratios a t the beginning and end of this time period were calculated and were used to determine the rate of decrease of the signal. Analytical Curues. Using the temperature-controlled cell compartment (maintained a t 24.3 & 0.5 "C) and the 3-mm pathlength sample cell, analytical curves were determined for quinine sulfate in 0.05M H2S04 and for ergotaminine in absolute ethanol. Slit settings were optimized for signal stability, rather than sensitivity, and were kept narrow to avoid dilution problems a t low concentrations. Slit Systems 7 and 8 were used for the determinations of quinine sulfate and ergotaminine, respectively. Quinine sulfate fluorescence was measured a t 450 nm after successive excitation a t 250 nm and 350 nm. Ergotaminine fluorescence was observed a t 402 nm with excitation a t 308 nm. To compare detection limits of the on-line HPLC fluorescence detector (SPF-125) with those of the Eimac and Hanovia systems, an analytical curve was determined for ergotaminine in ethanol, using the 3-mm cell in the SPF-125. Slit settings (positions approximate to those of the Aminco-Bowman in Figure 2) were: ( a ) 4 mm, ( b ) 4 mm, (c) 2 mm, ( d ) 3 mm, ( e ) 4 mm, ( f ) 4 mm; and fluorescence was monitored a t 402 nm after excitation at 312 nm (Hg line). Chromatographic Analysis of Ergotaminine i n Plasma. EXTRACTION. Standard solutions of ergotaminine and reserpine (internal standard) were made up in chloroform. Aliquots of the standards were added to the extraction tubes (20- X 125-mm culture tubes with Teflon-lined screw caps) and evaporated to dryness. Before extraction, 10-ml batches of outdated plasma from the blood bank were deproteinized by the method of Mondino et al. (421, with 3.75% sulfosalicyclic acid in distilled water (4 parts to 1 part plasma by volume). The deproteinized supernatant was added to the tubes containing the drug residues (4 ml equivalent to 1 ml plasma) to give standard samples containing the desired amounts of drugs. To the deproteinized plasma were added 2 ml of 3M KzC03 and 5 ml of the extracting solvent (benzene:ethyl acetate, 4:l by volume). The mixture was shaken mechanically for 10 min, centrifuged a t 700 X g for 5 min, and the organic layer was transferred to

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12, OCTOBER 1975

1995

Table 11. Signal-to-Noise Ratios for Various Eimac Lamp Current Settings Lamp

15

current,A

20

12.5 12.0 11.5 11.0 10.5 10.0 9.0

s

IS

8.0 7.0

b

100

600 500 PHOTOYULIIPLI~RVOLTAGt, -Vdc

400

Figure 3. Signal-to-noise ratio vs. photomultiplier voltage, Eimac system.

AlSnsc

2 04 186 180 174 180 181 165 150 162

,bast

sc

NC

SIN

48 42 46 40 54 50 42 28 29

1484 1430 13 53 1270 1193 1147 967 873 82 6

41.9 38.1 37.2 35.7 37.6 37.6 34.1 30.6 33.0

35.3 37.5 36.4 35.6 31.7 30.5 28.4 28.5 25.0

(E

a

*

36.2 3.3) Sample signal noise. Blank signal noise. Relative units.

rescence-HPLC quantitation procedures was calculated, based on four replicate extractions of 870 pg ergotaminine/ml from the deproteinized equivalent of 2 ml of plasma.

RESULTS AND DISCUSSION

L

To determine the limiting noise component of the Eimac-fluorescence system, the S/N, calculated by the method of St. John et al. ( 4 3 ) , was measured during separate variation of the lamp current, photomultiplier voltage, and excitation monochromator entrance slit width. The results of these determinations are shown in Table I1 and Figures 3 and 4, respectively. The data in Table I1 indicate that the optimum lamp current is about 12.0 A, not significantly different from the manufacturer's recommended operating current of 12.5 A. The S/Ndecreases with decreasing lamp current, as would be expected; however, the data show that this is primarily because of a reduction in signal. Over the range of current settings used, the noise component remained essentially constant. This is indicative of a photomultiplier shot noise limited system and is confirmed by the data shown in Figure 3, which indicate that a decrease in photomultiplier voltage has a greater effect on the noise component than on the signal, resulting in an increase in S/Ndown to the shut-down voltage of the photomultiplier. Figure 4 shows the relationship of S/Nto the flux of the exciting light (governed in this case by the excitation monochromator entrance slit width). The noise component decreases a t a lower rate than the signal component, resulting in a direct dependence of S/N on slit width. At high light levels, this rate difference is slight because of the higher noise level associated with the more intense sample fluorescence signal. However, as the flux of radiation reaching the detector becomes weaker at smaller slit widths, the limiting detector noise becomes more obvious, causing a rapid decrease in S/N. Relative performance of the Eimac and Hanovia systems was evaluated with respect to spectral distribution of the lamp intensity, tendency to cause photodecomposition of a light-sensitive molecule, and linear dynamic range and detection limits, to determine the feasibility and advantages, if any, of converting from the standard ellipsoidal condensing system to the Eimac system. Relative spectral distributions of the light intensity for the Eimac and Hanovia systems from 200-600 nm are shown in Figure 5. Use of the Eimac lamp results in a twofold to threefold increase in photon flux over most of this spectral region. The greatest differences occur in the visible region, above 500 nm, where the Eimac photon flux approaches 4 times that of the standard lamp. In the photodecomposition study, the signal from a 10ng/ml solution of quinine sulfate in 0.05M HzS04, illumi-

l ' # # o # L 5

4

3 2 1 SLIT WIDTH, nm

0

Figure 4. Signal-to-noise ratio vs. excitation monochromator entrance slit width, Eimac system

I14 $41 fX(ITAII0I W A V ~ l I N 6 T H . nm

All

Figure 5. Relative spectral distribution of light intensity for the Eimac (A) and Hanovia (4 lamp systems

a 5-ml amberized Mini-Vial (Alltech Associates, Inc., Arlington Heights, Ill.). While the first extract was being evaporated a t 55 OC under a stream of nitrogen, the aqueous phase was reextracted with another 5 ml of the extracting solvent. After centrifugation, the second organic layer was transferred to the Mini-Vial and taken to dryness. The residue was dissolved in the mobile phase (15 ll/ml of plasma extracted), and 10 J (two-thirds of the extract) were injected into the liquid chromatograph. Chromatographic conditions for the analysis of ergotaminine with reserpine as internal standard were as follows: drugs were eluted with a mixed mobile phase of isopropyl ether:acetonitrile: methanol (69.5:30:0.5, by volume), delivered a t a constant flow rate of 20 m l h r . ANALYTICALCURVE.An analytical curve was determined with fluorescence detection for ergotaminine extracted from plasma with reserpine as an internal standard. This curve covered the range of 3.63-0.224 ng of ergotaminine/ml of plasma, with 208 ng/ml of reserpine added. The deproteinized equivalent of 2 ml of plasma was extracted. REPRODUCIBILITY.Reproducibility of the extraction and fluo1996

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12. OCTOBER 1975

10'~ 10-1 (ON(EW1ILTIOW Of PUlNlWE SULFATE IN 0.05

IO-' IO 10-1 IO ' (OllCENTKAlIOW OF ERGOTAHININf IN ETHANOL, W/ml

10' H2S0,, M/ml

Figure 6. Analytical curves of quinine sulfate in 0.05M H2S04, excited at 350 nm by the Eimac (0)and Hanovia (A)lamps. Attenuation settings of the nanoammeter (amperes full scale) are shown at each

Doint

Figure 8. Analytical curves of ergotaminine in absolute ethanol, excited by the Eimac (0)and Hanovia (A)lamps in the Aminco-Bowman spectrophotofluorometer, and by a medium pressure mercury

vapor lamp ( 0 )in the Aminco SPF-125

I, 10-3

10-1

(Ol(tW1IAlIOW OF PUllllt SULlAlE IN 0.0s

1 ik

IO'

H2SOk, N/nl

Flgure 7. Analytical curves of quinine sulfate in 0.05M H2SO4, excited at 250 nm by the Eimac (0)and Hanovia (A)lamps. Attenuation

settings of the nanoammeter (amperes full scale) are shown at each point

A

1

nated by either lamp, decreased linearly over the period of observation. The irradiated and observed volumes of solution were identical for both excitation systems. As expected, to.9 (the time a t which the fluorescence signal reaches 90% of its original value) was much less for the Eimac system (14.5 min) than for the Hanovia system (42.5 min). This indicates that, for light-sensitive molecules in a static system, the additional flux of the Eimac lamp might not be so useful for increasing detectability as for increasing resolution and selectivity. The analytical curves for quinine sulfate in 0.05M H2S04, irradiated a t 350 nm and 250 nm, are shown in Figures 6 and 7 , respectively. These figures show full logarithmic plots (analytical curves) of relative fluorescence intensity vs. concentration, rather than SIN vs. concentration. Electrical blank subtraction was not used; and the noise levels of all signals, including that of the blank, were not measurable when the Eimac lamp was used for excitation, even at the highest amplification used for the determination. For excitation a t 350 nm (Figure 6), the fluorescence signal (sample signal minus blank signal) generated by the Hanovia lamp is greater a t each concentration than that generated by the Eimac lamp. The upper limit of the linear dynamic range occurs at a greater concentration for the Hanovia lamp. For excitation a t 250 nm (Figure 7 ) , where fluorescence of quinine sulfate is weaker, the curves are identical in upper and lower limits of linearity and fluorescence intensity. However, the amplification settings, used to monitor the fluorescence produced by the Hanovia lamp, were 1-2 orders of magnitude higher than those used with the Eimac lamp. Therefore, although the blank reading is higher with the Eimac lamp, if electrical blank subtraction is used, as it normally would be in any practical analysis or

I

3

5

7

9 B

KElElllION IlHt, min

Figure 9. Chromatograms (UV detection at 254 nm) of extracts of 1

ml of drug free deproteinized plasma (B) and deproteinized plasma with reserpine (I) and ergotaminine (11) added at concentrations of 258

and

171 ng/ml, respectively (A)

L

- 12 4 6 1 1 4 6 8 IfTtWlIOI TIME, min

Flgure 10. Chromatograms (fluorescence detection at 402 nm after

excitation at 312 nm) of extracts of 2 ml of deproteinized plasma with reserpine (I) added at 208 ng/ml (A) and reserpine and ergotaminine (11) added at 208 ng/ml and 870 pg/ml, respectively (B) (Aminco SPF-125 with Hg vapor lamp and flow cell)

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12, OCTOBER 1975

1997

/ L 2110-' 51lO-' 10' 1110~ 5i10° CONCtNTRAlION OF fRGOlAMININt IN PLASMA, a g l m l

Figure 11. Analytical curve for HPLC analysis (fluorescence detection) of ergotaminine with reserpine as internal standard

in a chromatographic detection system, the detectability for the Hanovia system would be limited by noise or allowable (practical) amplification before that of the Eimac system. The curves in Figure 8 were calculated on the basis of signal-to-noise ratio. The plots are full logarithmic, showing S/N vs. concentration of ergotaminine in ethanol (pg/ ml), for the Eimac lamp, the Hanovia lamp, and the Hg vapor lamp in the Aminco SPF-125 spectrophotofluorometer, which was later used with a flow cell as the highpressure liquid chromatographic fluorescence detector. At the lower limits of the linear dynamic ranges of the first two systems, which occur at the same concentration of sample, the S/N for the Eimac system is greater than 2 times that for the Hanovia system. Assuming that by more careful sample handling, the linear dynamic ranges of both systems could be extended to a value of S/N equal to 10 (limiting S/N for 2 determinations a t a 99% confidence level), the limits of detection would be 2.8 X loe4 pg/ml and 3.0 X lom5kg/ml for the Hanovia and Eimac systems, respectively. In this static-sample situation, even at maximum slit settings, the detection limit with the SPF-125 is much higher, being 1.2 X lo-' pg/ml (S/N = 10). Representative chromatograms of extracts of drug-free deproteinized plasma and of deproteinized plasma with added reserpine and ergotaminine are shown in Figures 9 and 10, for UV absorption and fluorescence detection, respectively. To obtain trace A of Figure 9, the deproteinized equivalent of 1 ml of plasma, with reserpine and ergotaminine added at 258 ng/ml and 171 ng/ml, respectively, was extracted and chromatographed according to the outlined procedure. Extraction of nondeproteinized plasma resulted in a solvent front, which completely obscured the peaks of interest. Under the chromatographic conditions employed, reserpine and ergotaminine had reproducible retention times of 5.9 and 7.4 min, respectively. Both extracts of the deproteinized equivalent of 2 ml of plasma, shown in Figure 10, contained added reserpine (208 ng/ml). The sample that gave trace A contained no added ergotaminine, but that which gave trace B contained the equivalent of 870 pg/ml. Since excitation and emission wavelengths (312 and 402 nm, respectively) were well off the excitation and emission maxima for reserpine (300 and 380 nm, respectively), a concentration about 200 times that of the ergotaminine was used to obtain adequate response. The extremely clean base line and minimal solvent front, obtained with fluorescence detection, in comparison with those obtained with UV detection, are indications of the tremendous selectivity of the HPLC-fluorescence combination. The impurity in Figure 10, trace A , which elutes at the exact retention time of ergotaminine, may be a plasma im1998

Table 111. Limits of Detection of Ergotaminine for the Instrumental Systems Studied Light s a u c e

Detection limit

Drug extracted from plasma-

Drug ~n

HPLC -fluor,

pure ethanol

W l 1 0 ul

fluor, u g l m l

injected

of ethanol

Hg vapor' 2 . 7 x 10-3d 1.2 x 10-1: Hanoviab 6.3 x 2.8 x 10-4 EimacC 6.8 x 3.0 x 10-sd In Aminco SPF-125. * In ellipsoidal condensing system, Aminco-Bowman SPF. In Aminco-Bowman SPF. Experimentally determined. e Calculated from results in column 2 and the Hg-vapor result of 2.7 X fig/lOfilinjected. purity; however, preliminary results indicate that it is probably due to residual ergotaminine in some part of the instrumental system. Successive injections of 10 pl of the mobile phase, after one injection of 10 p1 of a solution of ergotaminine in the mobile phase, resulted in a decrease in this peak to a very low, but still detectable, level after about 8 injections. After each of these injections, the syringe was rinsed with 5 X 50 pl of the mobile phase. Because of this impurity, when fluorescence was used, a more extensive rinsing procedure, including rinses with 5 x 50 pl of methanol, 5 X 50 pl of methylene chloride, and 5 X 50 pl of the mobile phase (in order), was performed after each injection of a sample containing this drug. The fluorescence analytical curve is shown in Figure 11. Five deproteinized 2-ml plasma samples, each containing 208 ng/ml of reserpine and from 3.63 to 0.224 ng/ml of ergotaminine, plus a blank, were extracted as outlined in the procedure and chromatographed, beginning with the blank and proceeding to the highest concentration. For this curve, the peak height ratio of ergotaminine to reserpine was calculated. Peak-to-peak noise was measured on the base line 2 min after the elution of the ergotaminine peak. Base line (blank) and drug peak (sample) noise were assumed equal (valid for low sample concentrations). The linear correlation coefficient of the best line (least squares fit) was 0.998, and the limit of detection for ergotaminine at S/N = 10 was2 ng/ml. Reproducibility of the HPLC fluorescence procedure at the 870-pg/ml level was 11%for four replicate determinations, although the relative standard deviation for three of these samples was only 4%. In Table 111, the fluorescence detection limits are given for ergotaminine for the various instrumental systems used in this study. These detection limits for the purely spectrophotometric systems are for pure ethanol solutions of the drug and were calculated from the experimental data shown in Figure 8 for the Eimac and Hanovia lamps with the Aminco-Bowman dispersing system, and for the medium pressure mercury vapor lamp in the Aminco SPF-125 dispersing system. The combined chromatographic-fluorescence detection limits are given for ergotaminine, extracted from 2 ml of deproteinized plasma. The value for the SPF125 was experimentally determined and calculated from the data in Figure 11. Theoretical values for the Eimac and Hanovia chromatographic systems were calculated from the experimental values. These values indicate that detection limits in the low picogram range are possible with either of these systems, with the limit for the Eimac lamp being almost an order of magnitude lower than that for the Hanovia system. It is possible that these detection limits are not practically attainable, simply because of sample handling problems a t high dilution. However, these low

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12, OCTOBER 1975

theoretical limits hold promise for an extremely versatile, sensitive, and selective analytical system. The Eimac illuminator, when used as a light source for excitation of fluorescence from molecules in the condensed phase, has been shown to give a signal-to-noise advantage over a standard Hanovia xenon arc lamp in an ellipsoidal condensing system, used under identical conditions. The cost of the Eimac system, including lamp, lamp housing, and power supply, i s approximately equal to that of the ellipsoidal condensing system (lamp and power supply are purchased separately at an additional cost). Also, the collimating mirror, being an integral part of the Eimac lamp, is replaced each time the lamp is changed. Replacement Eimac lamps are 1.5 times more expensive than standard xenon arc lamps, and lifetimes for both lamps are approximately equal (75% of peak radiant flux at 1000 hr). As with any more intense light source, the Eimac lamp will destroy photosensitive molecules more quickly than the Hanovia source, if both are used a t the same excitation monochromator slit settings. However, the photon flux of the Hanovia system can be equalled by the Eimac system at narrower slit widths, giving greater resolution and selectivity. If incorporated into a high-pressure liquid chromatographic detection system, in which the dwell-time of each molecule in the light beam is relatively short, photodecomposition becomes a very minor, or nonexistent, problem, and the full light flux can be used to decrease detection limits. The analysis of ergotaminine, presented in this study, shows that high-pressure liquid chromatography can be as sensitive as gas-liquid chromatography. Although there are more sophisticated means of increasing the sensitivity of fluorescence analysis, such as ratio recording, photon counting, and synchronizing amplification and excitation frequencies, the least expensive method is to increase the usable photon flux through the excitation monochromator. The Eimac illuminator does this by simply capturing more of the radiation which is available in the xenon arc. If addit,ional S/N enhancement is desired, the more sophisticated methods can again be applied.

LITERATURE CITED (1)P. Brazeau, in "The Pharmacological Basis of Therapeutics," 4th ed., L. S. Goodman and A Gilman. Ed.. The Macrnillan Co., New York, N.Y., 1970,pp 898-900. (2)J. R. Graham and H. G. Wolff, Arch Neurol. Psychiatry, 39, 737 (1938). (3)H. Ernmenegger and W. Meier-Ruge, Pharmacology, 1, 65 (1968). (4)T. G. Alexander, J. Assoc. Off. Agr. Chem., 43, 224 (1960). (5) J. L. McLaughlin, J. E. Goyan, and A. G. Paul, J. Pharm. Sci., 53, 306

(1964).

(6) S.Keipert and R . Voigt, J. Chromatogr., 84, 327 (1972). (7)R. Fowler, P. J. Gomm, and D. A. Paterson, J. Chromatogr.. 72, 351 (1972). (8) S.Agurell and A. Ohlsson, J. Chromstogr.,81, 339 (1971). (9)J. Axelrod, R. 0. Brady, B. Witkop, and E. V. Evarts, Ann. N.Y. Acad. Sci., 86, 435 (1957). (IO) G. K . Aghajanian and 0. H. L. Bing, Clin. Pharmacol. Ther., 5 , 611 (1964). (11) D. G. Upshall and D. G. Waiiling. Clln. Chim. Acta., 38, 67 (1972). (12) A. Stoll and W. Schlientz, Helv. Chim. Acta, 38, 585 (1955). (13)A. Bowd, J. B. Hudson, and J. H. Turnbull, J. Chem. SOC.,Perkin Trans. 2,I O , 1312 (1973). (14)W. D. Hooper, J. M. Sutherland, M. J. Eadie, and J. H. Tryer, Anal. Chim. Acta, 89, 11 (1974). (15)I. Jane and B. B. Wheals, J. Chromatogr., 84, 181 (1973). (16)R. A. Heacock. K. R. Langille, J. D. MacNeil, and R . W. Frei, J. Chromatogr., 77,425 (1973). (17)J. D. Wittwer and J. H: Kluckhohn, J. Chromatogr. Sci., 11, 1 (1973). (18) R. D. Venn, Hanover, N.J., personal comrnunicatlon, August 1973. (19)P. T. Kissinger, C. Refshauge, R . Dreiling, and R. N. Adams, Anal. Lett.. 8,465 (1973). (20)P. T. Kissinger, C. Refshauge, R . Dreiling, and R. N. Adams, ACS Abstracts, 166th National Meeting, Chicago, Ill., August 26-31, 1973. (21)C. Refshauge, P. T. Kissinger, R. Dreiling, L. Blank, R. Freeman, and R. N. Adams. Life Sci., 14, 31 1 (1974). (22) P. T. Kissinger, L. J. Felice, R. M. Riggin, L. A. Pachia. and D. C. Wenke, Clin. Chem., 20,992 (1974). (23)E. D. Peilizzari and C. M. Sparachino, Anal. Chem., 45, 378 (1973). (24) D. D. Chilcote and J. E. Mrochek, Clin. Chem., 18, 778 (1972). (25)R. W. Frei, J. F. Lawrence, J. Hope, and R. M. Cassidy, J. Chromatogr. Sci., 12,40 (1974). (26) J. F. Lawrence and R. W. Frei, J. Chromatogr., 98, 253 (1974). (27)R. M. Cassidy, D. S. LeGay, and R. W. Frei. J. Chromatogr. Sci., 12, 85 (1974). (28)W. Dunges, G. Naundorf. and N. Seiler, J. Chromatogr. Sci., 12, 655 (1974). (29)R. W. Frei and J. F. Lawrence, J. Chromatogr., 83, 321 (1973). (30)R. M. Cassidy and R . W. Frei, J. Chromatogr., 72,293 (1972). (31)D. R. Baker, R. C. Williams, and J. C. Steichen. J. Chromatogr. Sci., 12, 499 (1974). (32)H. Hatano, Y. Yarnarnoto, M. Saito, E. Mochida, and S. Watanabe, J. Chromatogr., 83, 373 (1973). (33)J. C. Steichen, J. Chromatogr., 104, 39 (1975). (34)M. P. Bratzel, Jr.. R. M. Dagnall, and J. D. Winefordner, Anal. Chim. Acta, 52, 157 (1970). (35)R. L. Miller, L. M. Fraser, and J. D. Winefordner, Appl. Spectrosc., 25, 477 (1971). (36)F. W. Piankey, T. H. Glenn, L. P. Hart, and J. D. Winefordner, Anal. Chem., 46, 1000 (1974). (37) D. J. Johnson, F. W. Plankey, and J. D. Winefordner, Anal. Chem., 48, 1898 (1974). (38)L. H. Luthjens, Rev. Sci. hsfrum., 44, 1661 (1973). (39)T. C. O'Haver and J. D. Winefordner, J. Chem. Educ., 48, 241 (1969). (40)V. E. Dell'Ova, M. B. Denton, and M. F. Burke, Anal. Chem., 48, 1365 (1974). (41)W. H. Melhuish, J. Opt. SOC.Am., 52, 1256 (1962) (42)A. Mondino, G. Bongiovanni. S. Fumero, and L. Rossi, J. Chromatogr., 74,255 (1972). (43)P. A. St. John, W. J. McCarthy, and J. D. Winefordner, Anal. Chem., 39, 1495 (1967).

RECEIVEDfor review May 5, 1975. Accepted July 7 , 1975. This work was supported in part by the U.S. Public Health Service Grant No. GM-11373-11, and the Epilepsy Research Foundation of Florida, Inc.

Quantitative Analysis of Cation-Exchange Resins by Infrared Spectrophotometry and Pyrolysis-Gas Chromatography J. R. Parrish Department of Chemistry, Rhodes University, Grahamsto wn, South Africa

Two series of sulfonic acid resins, one with varying capacity and one with varying cross-linking, have been synthesized. These were used to develop methods for the determination of capacity as well as cross-linking by pyrolysis-gas chromatography and infrared spectrophotometry. The relative amount of styrene produced by pyrolysis of a resin is inverseiy related to its capacity, and capacity can be deter-

mined over a small range with a standard deviation of 0.1 mequiv g-'. Capacity can be determlned over a wider range with a standard deviation of 0.2 mequiv g-' by measurement of the IR absorbance at 1008 cm-'. Cross-linking is more difficult to estimate, and variations of published methods are compared. The standard deviations lie in the range 0.4-296 dlvlnylbenzene.

ANALYTICAL CHEMISTRY, VOL. 47, NO. 12, OCTOBER 1975

1999