Anal. Chem. 1984, 56,24-27
24
centrator column to ensure the necessary sensitivity.
ACKNOWLEDGMENT We thank John Dixon for performing t h e spectrophotometric chloride analyses. We are also grateful to Southern California Edison for providing boric acid samples and verification test plans. Registry No. Boric acid, 11113-50-1.
LITERATURE CITED (1) Rocklin, R. D.; Johnson, E. L. Anal. Chem. 1983, 55, 4-7. (2) Green, L. W.; Woods, J. R. Anal. Chem. 1981, 53, 2187-2189. (3) Pohlandt, C. S A f r . J. Chem. 1982, 35, 96-100.
(4) Small, H.; Stevens, T. S.;Bauman, W. C. Anal. Chem. 1975, 47, 1801-1809. (5) Gjerde, D. T.; Frltz, J. S.;Schmuckler, G. J. Chromarogr. 1979, 186, 509-519. (6) Gjerde, D. T.; Schmuckler, G.; Fritz, J. S. J. Chromafogr. 1080, 187, 35-45. (7) Verification Test Procedure for Safety Related Borlc Acid; Southern Callfornia Edison Go., San Onofre Nuclear Generating Station, San Onofre, CA. (8) Iwasaki, I.; Utsumi, S.; Ozawa, T. Bull. Chem. SOC.Jpn. 1952, 2 5 , 226-229. ~~. (9) Wetzel, R. A.; Anderson, C. L.; Scheicher, H.; Crook, G. D. Anal. Chem. W79, 51, 1532-1535. ~~
RECEIVED for review June 30,1983. Accepted September 19, 1983.
On-Line Extraction, Evaporation, and Injection for Liquid Chromatographic Determination of Serum Corticosteroids Kitaro Oka, Kazuo Minagawa, and Shoji Hara*
Tokyo College of Pharmacy, Horinouchi, Hachioji, Tokyo 192-03, Japan Makoto Noguchi, Yasuo Matsuoka, Michinori Kono, and Shoichiro Irimajiri
Kawasaki City Hospital, Kawasaki 210, Japan
To construct an on-line system for cleanup of bloiogicai fluids, closed-bed columns packed with fine diatomaceous earth granules were developed. Three cascaded columns were coated with neutral water, aqueous sodium hydroxide, and sulfuric acld since the extraction of neutral components with a hlgh degree of selectivity is possible by such stationary phase iiqulds. The recovery of steroid solutes from sera was almost quantltatlve, using aqueous diethyl ether as the carrier. The effluent from the aqueous liquid-iiquld partition columns was introduced Into a line evaporator made of glass tubing. The injection of the sample into a liquid chromatography column was performed by inserting a steel stick for eliminating the dead space in the evaporator to resolve the residuie A model of an on-line equipped extraction, evaporation, a n d injection system was constructed by the incorporation of a valve switching procedure. This technique was applied to a clinical assay of corticosteroids In sera and was found suitable as a highly sensitive method for monitoring steroidal drugs.
For determination of constituents in biological fluids such as serum and urine samples, preliminary fractionation and/or enrichment of target compounds is usually recommended. Elimination of contaminants from the fluids consisting of a complex matrix enhances the sensitivity of analysis of the compounds under consideration. In many instances, the cleanup of biological fluids has been carried out by manual batch operation prior to high-performance liquid chromatography (HPLC) analysis. Recently, two new procedures have been introduced: (1) liquid-liquid distribution using columns packed with diatomaceous earth support and (2) prechromatography involving either a normal or reversed phase system employing a silica gel or octadecylsilyl silica columns. Commercially available disposable short columns such as Extrelut by Merck, Darmstadt, and Sep-PAK cartridges by Waters, Milford, MA, are being widely accepted by analytical chemists. However, it is difficult to transfer the
effluent from the cleanup column directly into the main analytical system. Sample preparation by collecting the fraction and evaporation of the solvent is often necessary. Hence these prechromatographic techniques require off-line processes. At our laboratory, column switching extraction and chromatography for programmed flow preparation (PFP) have been developed (1,2). PFP has improved the preparation system and avoids the defects encountered in former procedures using batch processing in chemical laboratory experiments. Through the PFP concept, an attempt has been made to develop an on-line technique for fractionation, evaporation, a n d injection in microchemical analysis. First, a suitable design for the liquid-liquid distribution column system for on-line procedure was investigated. Instead of the open-bed cartrige columns commonly employed in the laboratory, the closed-bed columns slurry packed by fine diatomaceous earth granules were introduced ( 3 , 4 ) . These cleanup columns were found to have high efficiency and could be used repeatedly if the columns are flushed by water soon after the injection of certain amounts of biological fluids. Secondly, a new type of evaporator made by glass tube was developed to afford quantitative recovery of small amounts of solutes. There was hardly loss of any sample when the residue in the evaporator was injected into the analytical column (3). As a result, construction of the on-line system for switching extraction, evaporation, and injection into the HPLC column (LEEI system) was achieved. Several applications of this system revealed high sensitivity and efficiency for HPLC analysis of biological samples. In this paper, the application of this technique to the clinical analysis of serum corticosteroids is reported.
EXPERIMENTAL SECTION LEEI System. A model for a switching fractionation and chromatography system is illustrated in Figure 1. This sytem is integrated by four functional operations: (a) presaturation of water for the mobile phase solvent; (b) sample application and extraction of the constituents; (c) evaporation of the effluent; (d)
0003-2700/84/0356-0024$01.50/00 1983 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 56, NO. 1, JANUARY 1984
L_1
back IIsSh
I
I
a
b
C
d
Flgure 1. Schematic diagram of the on-line extraction-evaporationinjection (LEEI) system equipped with HPLC. See text for details.
HPLC analysis. The construction materials of the system consist of glass tubes, PTFE plugs and tubes, valves, pumps, and detectors. Thick-wall pressuretight glass tubes were connected by PTFE plugs so designed as to be easily inserted into the inside of the glass tubes and removable by hand (5). The upper pressure limit of the glass tube and PTFE plug connection with a metal clip is 50 kg/cm2 and the PTFE plugs are connected by PTFE tubing. System Operation (a). The solvent was pumped into glass tubes 1, 2, and 3 (2.2 cm i.d. X 30 cm) filled with water and an aqueous solution for equilibration of the solvent and water. The solvent was passed through the three cascaded tubes by a droplet current. Carrier solvent used was a commercial product of diethyl ether containing phenol derivatives as a stabilizer. Tube 1 was filled with a 15% sodium hydroxide solution to remove the phenolic additives. Tube 2 was filled with 25% sodium chloride solution for washing and tube 3 with distilled water to obtain a water-saturated mobile phase. The outlet of the solvent flow was connected to valve 7 of operation b using PTFE tubes (1mm i.d.). System Operation (b). Crude sample fluid was applied by use of a glass syringe and valve 7. A 0.5 mL, or a lesser amount, of serum sample was introduced into the cascaded columns (excolumns 4,5, and 6) for fractional extraction of the constituents. These columns, provided with thin PTFE filters, were packed with fine diatomaceous earth granules (particle size, 10 wm, AquEx-10, DSD, Tokyo) by a slurry packing procedure using a gastight syringe. The column efficiency was determined by using caffeine as the standard sample and aqueous diethyl ether as the carrier. Excolumns 5 (4 mm X 20 cm) and 6 (4 mm X 20 cm) afforded approximately 300 theoretical plates. The packing support in the excolumn 4 (1 cm i.d. X 7 cm) was coated with distilled water. Water was injected into the column with a gastight syringe. The aqueous phase of excolumn 4 was washed out after each analytical run to remove any hydrophilic substances distributed in the aqueous phase, such as serum proteins, by injecting about 5 mL of distilled water from the top of the column. A subsequent fast flow of diethyl ether (10 mL/min) in a direction from top to bottom to remove excess water was required for conditioning prior to the next run. This backflash procedure is indicated by the arrow lines in Figure 1. In this figure, several valves for the conditioning process are abbreviated for clarity. Diatomaceous earth in excolumn 5 was coated with 1% sodium hydroxide solution as the stationary phase. Support material in excolumn 6 was coated with 1% sulfuric acid. The aqueous stationary phases in these columns were conditioned by adding the corresponding aqueous solution from the top of the column and a fast flow of diethyl ether to remove the excess aqueous solution. Reconditioning of excolumn 5 was required for each 1 mL of serum injected. After the serum was fed into excolumn 4, aqueous diethyl ether was injected to extract the ether-soluble components in the
25
column. The mobile phase solvent was then moved through the cascaded columns step by step to eliminate acidic and basic contamination and to fractionate the neutral components at a flow rate of 0.8 mL/min. System Operation (c). Evaporator 8 was constructed of a single glass tube (4 mm i.d. X 7 cm). Evaporation of the effluent was achieved in the inside wall of the tube at a reduced pressure (20 mmHg and 40 OC). The pressure was regulated by a water aspirator via valve 9. The temperature was maintained by an electric heater attached to the outer surface of the glass tube. Evaporator 8 was connected to excolumn 6 via valve 10 and a thin PTFE tube (0.1 mm i.d. X 1m) which allowed switching to the higher pressure site of system operation b and the lower pressure site of operation c. At specified switching times of valve 10, the effluent containing neutral components was introduced into the evaporator and concentrated to dryness. Monitoring of the effluent by a UV detector was necessary for the pilot run, using a larger amount of authentic samples added to the serum. Following stipulation of the extraction time, it was conventional and sufficient to evaporate the proper amount of effluent containing all the desired materials without UV monitoring. The effluent emerging from the column within 5 to 17 min was collected for corticosteroid fractionation. In order to minimize the diffusion of the fractionated solute, the inner dead volume of evaporator 8 was reduced. It was found that insertion of a stick within the glass tube served this purpose quite well. Consequently, a mechanically driven steel stick (3.9 mm i.d. X 6.9 cm) was attached to evaporator 8. On completion of the evaporation, the inner dead volume of the glass tube was reduced by insertion of this steel stick. The outlet of the tube was connected to HPLC column 11 (Lichrosorb S-100,4 mm i.d. X 25 cm) via valve 12. The mobile phase solvent for HPLC was then introduced into the evaporator 8 via valve 10 at a flow rate of 0.1 mL/min and the inner air was pushed out through valve 12. When the solvent reached valve 12, it was directly switched to analytical column 11. System Operation (a). The HPLC equipment consisted of a reciprocal pump (13) (Twincle, Jasco, Tokyo), a UV detector (14) (Uvidec 100-111, Jasco, Tokyo) set at 240 nm, and a pen recorder (15) (RC-150, Jasco). Generation of gasified solvent at the plunger site of the pump caused base line waving unfavorable for high sensitive nano-level analysis. Insertion of a glass tube (1cm i.d. X 7 cm) as a damper between the pump and evaporator eliminated this line waving completely. Furthermore, the reciprocal pump generated other types of base line wavings owing to pressure pulse in the plunger. This noise could be reduced remarkably by a condensed air damper functioning as a neutralizer of pressure change. Base line stabilization was necessary for sensitive analysis within a detection range of 0.005 and recorder range of 2 mV. Glass and PTFE components were supplied by Kusano Scientific, Tokyo. Reagents and Solvents. Reagent grade chemicals were used. Diatomaceous earth (AquEx-10) for the column extraction was supplied from DSD, Tokyo. Authentic steroid hormones were supplied from Teikoku Zoki Co., Kawasaki, Japan). HPLC Procedure. The eluent was a mixture of 0.1% water, 4% methanol, and 30% dichloromethane in n-hexane (v/v). The solvent system was designed according to ref 6. The silica gel column was washed with 1% sulfuric acid and then methanol. This treatment was suitable for analysis of acidic or polar neutral compounds such as corticosteroids and related substances. Results similar to ours have been reported recently by Engerhart and co-worker (7). Calibration for Quantitative Analysis. Calibration curves of cortisol and prednisolone were made on the basis of peak height measurements using sample quantities between 1 ng and 100 ng/mL serum. The peak heights were reproducible with the coefficient of variation (CV) values being less than 4%, using the authentic samples injected from evaporator 8. Cortisol was recovered from the sera in a 102 f 6.8% yield and prednisolone, in a 93.3 f 1.4% yield, when known amounts of the authentic specimens (1-100 ng) were added to the sera. Analysis of Sera. A blood sample ( 5 mL) was centrifuged for 5 min at 3000 rpm. A 0.5-mL sample of the supernatant was injected into excolumn 4 through valve 7. The mobile phase for
28
ANALYTICAL CHEMISTRY, VOL. 56, NO. 1, JANUARY 1984
the extraction was introduced into the column system at a flow rate of 0.8 mL/min. The eluate, collected between 5 and 17 min, was then introduced into evaporator 8. After insertion of the stick into the evaporation tube, HPLC solvent was introduced into the evaporator at a flow rate of 0.1 mL/min followed by switching valve 12 to HPLC column 11. The HPLC flow rate was 1 mL/min.
9i
RESULTS AND DISCUSSION In order to formulate an on-line cleanup procedure for the analysis of biological fluids, a system of a closed-bed column packed with fine diatomaceous earth powder was developed ( 3 , 4 ) . This column-packing material was found to support an aqueous stationary phase and was capable of constructing a two layer aqueous liquid-liquid distribution system for a water-saturated organic solvent. It was also found that protein and other high molecular polar components in serum were trapped in the aqueous stationary phase on the diatomaceous earth support and that contaminants in the fluid were almost completely eliminated from the column by a flushing out procedure using a fast flow of water following injection of a certain amount of serum. Hence, this cleanup column is reusable. To achieve efficient extraction and fractionation of serum corticosteroids, an extraction column system comprised of three diatomaceous earth columns was introduced. The first column contained a water-coated support, the second and third, aqueous stationary phases of sodium hydroxide solution and dilute sulfuric acid. In the neutral column (excolumn 4 in Figure l),lipophilic substances in human serum were distributed within the carrier solvent. Diethyl ether was the most conventional extracting solvent because of its low boiling point and high solubility toward various solutes except for high molecular and high polar compounds such as proteins, peptides, amino acids, and carbohydrates. Although diethyl ether is inflammable, its use poses a slight safety problem since our apparatus is a totally closed system. The manner of distribution was consistent with the partition coefficients of substapces mentioned earlier, using an acid-treated silica gel column and aqueous diethyl ether system (8). Even a nearly insoluble compound in diethyl ether could be quantitatively distributed within the mobile phase on using an extraction column with a wide surface area. The effluent from neutral excolumn 4 was introduced into the next alkaline excolumn 5 in which the acidic solutes were distributed in the basic stationary phase and retained there. In the third excolumn 6, organic bases still present in the mobile phase were removed to the acidic stationary phase. Thus, the effluent from the third column consisted of only neutral diethyl ether soluble materials. For a highly sensitive analysis of the neutral components in sera, a cleanup process incorporating three columns was found to be useful. The durability of each excolumn was maintained completely except for changes in the pH of the stationary phase. The pH was easily measured by the fast flow of the mobile phase from the top to the bottom of the column, resulting in the extrusion of droplets of the stationary phase. A new stationary phase liquid was supplied from the top of the column and the subsequent fast flow of the carrier solvent, diethyl ether, created a condition of column control for the next run. This condition was necessary occasionally even when the pH remained unchanged, since the stationary phase liquid moved slowly, eventually flowing over the top of the column during analysis. An overflowed aqueous phase may result in an uncontrolled column condition and technical difficulties in the following evaporation stage. The excolumn 4 loaded with human sera should be conditioned for each analytical run. This is also necessary for the extrusion of serum protein from the stationary phase, which degenerates slowly on the support surface and gradually clogs the column. Thus, a neutral column can be used re-
I
r
0
I
IO
2'0
I
3'0 min
t
5 rnV
tomv
recorder range
Figure 2. Chromatogram of serum corticosteroids obtained by the LEE1 technique using 0.5 mL of serum from a systemic lupus eryth-
ematosus patient. The flow rate was 1 mL/min. peatedly for hundreds of runs without any technical problems. The effluent from the excolumns was concentrated in the line evaporation system connected to excolumn 6 through a fine PTFE tube functioning as a pressure converter between the high and low pressure sides. In the glass evaporator, 8, the effluent was concentrated to dryness at 40 "C, at which all corticosteroids tested were intact and stable. The residue was redissolved in a chromatographic solvent and introduced via an attached valve after reducing the dead space in the glass evaporator by the insertion of a steel stick, causing the resulting dead volume to be no more than about 50 pL. The quantitative amount of neutral components derived from the serum was injected into the HPLC column. Solvent system was optimized for HPLC equipped with a silica gel column, A multicomponent system consisting of n-hexane and dichloromethane as diluents and methanol and water as stronger solvents was employed for the simultaneous analysis of several corticosteroids. The results suggest that unfavorable diffusion of the solutes does not occur in the process of resolving a residue in the evaporator and transferring the solution to the analytical column. The all-unit operations such as extraction, evaporation, and injection were completely linearized with good reproducibility and efficiency, resulting in no need for an internal standard, usually indispensable for HPLC clinical analysis.
27
Anal. Chem. 1884, 56, 27-29
The calibration curves for the determination of prednisolone and hydrocortisone were made on the basis of the peak heights of the corresponding steroids. Recovery of these steroids from the sera was approximately quantitative and reproducible as shown by a comparison of the injection of the standard sample solutions and sera containing known amounts of steroids (1-100 ng/mL serum). We applied the technique presented in this paper to the nano level analysis of serum corticosteroids such as hydrocortisone suppressed by prednisolone therapy for colagen disease and especially SLE patients. The results in Figure 2 show that our method is quite adequate for clinical assay before the use of prednisolone therapy or medication design determined by pharmacokinetics.
ACKNOWLEDGMENT The authors express their appreciation of M. Shibuya,
Nippon Medical College, for his suggestions and comments. Registry No. Cortisol, 50-23-7;prednisolone, 50-24-8; prednisone, 53-03-2; corticosterone, 50-22-6; cortisone, 53-06-5.
LITERATURE CITED (1) Hara, S.; Oka, K.; Dobashi, Y.; Ohkuma, T. Yakugaku Zasshi 1982, 102, 107-109. 12) . . Oka, K.; Dobashi, Y.; Ohkuma, T.; Hara, S . J. Chromatogr. 1981, 217, 387-398. (3) Hara, S.; Ijitsu, T.; Oka, K. Yakugaku Zasshi f982, 102, 895-897. (4) Hara, S.;Ijltsu, T.; Dobashi, Y.; Oka, K. Yakugaku Zasshi 1982, 102, i .oss-iodci - - - .- , -. (5) Hara, S. US. Patent 4 289 620, 1981. (6) Rose, J. Q.; Jusko, W. J. J . Chromatogr. 1979, 162, 273-280. (7) Engelhardt, H.;Mueller, H. J. Chromatogr. 1981, 218, 395-407. (8) Hara, S.; Dobashl, Y.; Oka, K. J. Chromatogr. 1982, 239, 677-685.
RECEIVED for review October 13,1982.
Resubmitted June 28,
1983. Accepted September 9,1983.
Preconcentration with Dithiocarbamate Extraction for Determination of Molybdenum in Seawater by Neutron Activation Analysis W. M. Mok and C. M. Wai* Department of Chemistry, University of Idaho, Moscow, Idaho 83843
Molybdenum in seawater can be quantltatively extracted with pyrrolldlnedlthlocarbamate and dlethyldlthiocarbamate at pH 1.4 into chloroform, for neutron activatlon analysis. Uranium in seawater cannot be extracted at this pH, and hence ellmMo reaction. Inhates the interference from the 23sU(n,f)gg terferences from matrlx specles in seawater, such as sodlum and bromlne, are also removed durlng the extraction. The proposed method, with good accuracy and sensitivity, Is sultable for the determlnatlon of molybdenum in natural waters.
The determination of certain trace elements has become important in studies of marine environments. Among these, molybdenum, as a biologically active element for growth and as a micronutrient in the aquatic environment, is often measured. Because of the low levels of molybdenum involved, neutron activation analysis (NAA) appears to be a suitable technique. Molybdenum can be identified in NAA through %Mo (tlIz= 66 h) or ita daughter 99”Tc (tl12= 6.05 h). Direct application of NAA for trace element determination in complex systems is generally difficult because of serious matrix interferences. Chemical separation methods have been used either before or after neutron irradiation to eliminate such interferences. Besides matrix interferences, another complication in the molybdenum determination by NAA is the contribution of 99Mofrom the 23SU(n,f)94Mo reaction, which is significant in systems with high uranium contents. Separation before neutron irradiation has the advantage of avoiding this complication if the method is selective for molybdenum. Kulathilake and Chatt reported a preconcentration method for the NAA determination of molybdenum in seawater by cocrystallization with P-naphthoin oxime (1). This method
is selective for molybdenum but requires a couple of days to complete the separation. Molybdenum has also been determined in seawater after preconcentration on activated charcoal in the presence of ammonium pyrrolidinedithiocarbamate (APDC), followed by neutron activation of the charcoal sample (2). Since activated charcoal is an efficient adsorber, a significant fraction of the uranium in seawater was also found with the molybdenum in the charcoal. Consequently, a correction for the uranium contribution to 99Momust be considered in this charcoal adsorption method. This paper describes a simple solvent extraction method using dithiocarbamates for preconcentration of molybdenum in seawater. The solvent extraction method not only is fast and efficient but also eliminates the interferences caused by uranium and other matrix species in the determination of molybdenum by NAA.
EXPERIMENTAL SECTION Ammonium pyrrolidinedithiocarbamate (APDC) and sodium diethyldithiocarbamate (NaDDC) were obtained from the Fisher Scientific Co. Chloroform used in the extraction was Baker Analyzed Reagent. Nitric acid used in the experiments was of Ultrex grade from Baker Chemical Co. Deionized water was obtained by treatment of distilled water through an ion exchange column (Barnstead Ultrapure Water Purification Cartridge) and a 0.2-pm filter assembly (Pall Corporation Ultipor DFA). A stock solution containing 1000 pg/mL molybdenum was prepared by dissolving 1.8401 g of dry ammonium molybdate (analytical reagent grade) in deionized water and diluted to a liter in accordance with the EPA’s instructions (3). Standard working solutions were prepared by appropriate dilution of the stock solution. Surface seawaters were collected from the southwest corner of Discovery Park in western Seattle. The salinity of the seawater was 27%0.The seawater was filtered through a 0.45-pm membrane filter, acidified to pH 2, and stored in precleaned polyethylene bottles. All containers used in this study were
0003-2700/84/0356-0027$01.50/00 1983 American Chemical Society
