423
Anal. Chem. 1991, 63,423-427
Determination of Gad0Iinium in BioIogicaI MateriaIs Using Graphite Furnace Atomic Absorption Spectrometry with a Tantalum Boat after Solvent Extraction Lian Liang, Patrick C. D’Haese, Ludwig V. Lamberts, Frank L. Van de Vyver, and Marc E. De Broe* Department of Nephrology-Hypertension, University of Antwerp, Antwerp, Belgium
A method was developed for the determination of gadolinium (Gd) in biological material using graphite furnace atomic absorption spectrometry (GFAAS). The element Is first extracted into methyl isobutyl ketone and then reextracted into hydrochloric acid. Factors influencing the recovery of extraction such as pH, choice of chelating agents, and hydrochloric acid concentration have been investigated. The eiement is determined under STPF (stabilized temperature platform furnace) conditions with atomization from a tantalum boat. Under optimized furnace conditions, the use of the tantalum boat improved sensitivity substantially compared to the use of pyrolytically coated graphite tubes. Around 150 measurements could be performed with 1 boat. Memory effects, being a common problem in the GFAAS determination of ianthanoids, were no longer observed after insertion of the boat. The characteristic mass and detection limit (2SD; SD = standard deviation) of the Gd determination are 1000 and 2060 pg, respectively. The precision evaluated as the relative standard deviation (RSD) of six analyses was below 10% for tissue Gd concentrations ranging from 0.92 to 72.0 pg g-‘. The recovery of added anaiyte ranged between 92.0 % and 99.3%. The method was found to be suitable for studying the pharmacokinetics and biodistribution of Gd in rats.
INTRODUCTION The gadolinium tetraazacyclododecanetetraacetate (GdDOTA) complex has recently been introduced as a contrast agent for magnetic resonance imaging ( I ) . In order to investigate its toxicity, a study of the biodistribution and kinetics of Gd in the human body is necessary. This requires the availability of reliable analytical methods to determine the element in biological fluids and tissues. Some workers have determined the element by atomic emission spectrometry (2). Others have used radiolabeled gadolinium complexes (3) to study the biodistribution of the element. T o the best of our knowledge, no report exists on the application of graphite furnace atomic absorption spectrometry (GFAAS) for the determination of Gd in blood, urine, or tissues. However, because of its ability to use small sample sizes and its relatively low cost and simplicity, the applicability of GFAAS for the measurement of Gd should be considered. The GFAAS determination of lanthanoids in aqueous solutions was first shown by Grobenski ( 4 ) . Although a substantial increase in sensitivity was noted when the graphite tube was coated with pyrolytic graphite, the lowest detectable amounts of Gd were not small enough for pharmacokinetic studies. Moreover, severe memory effects were observed. The low sensitivity of the GFAAS determination of lanthanoids with conventional graphite tubes has been attributed to the formation of refractory carbides ( 5 ) and pyrocarbon shells on microparticles of the samples (6) and to the large dissociation constants of the gaseous monoxides of some of these elements, Le., Lu, La, and Gd ( 4 ) . In order to circumvent carbide formation, some workers have used tantalum-lined graphite tubes to prevent 0003-270019 1/0363-0423$02.50/0
interaction of the analyte with the graphite tube (5-9). Using this device, Haines ( 5 ) observed a substantial improvement in the efficiency of atomization of some of the rare earth elements. However, Haines stated that even with tantalum lining the determination of Gd could not be performed with sufficient accuracy because of severe interferences. Others, however, noted increased Gd sensitivity (8)and elimination of memory effects when a tantalum foil liner was used inside a pyrolytically coated graphite tube (7). L’vov and co-workers (6) reported that the use of tantalum-lined graphite tubes and tantalum platforms gave a 50-70-fold increase in sensitivity compared to that obtained with conventional pyrolytically coated graphite tubes (4,IO)for 17 elements, the majority of them being rare earth elements. Although significant increases in sensitivity were observed, many workers found the tantalum-lining procedure to be very critical (9). Also, the usable lifetime of the tantalum foil was only 5-40 firings (5,8). In addition, the GFAAS determination of lanthanoids is prone to severe matrix interferences (4, 5). This is perhaps the reason for the lack of reports dealing with the GFAAS determination of Gd in biological materials. In view of the foregoing considerations, it has been our aim to develop a simple procedure involving the insertion of a tantalum boat into a graphite tube, to find the optimal analytical conditions for the GFAAS determination of Gd with a tantalum boat, and to develop a separation and preconcentration technique for the determination of Gd in biological materials.
EXPERIMENTAL SECTION Instrumentation. A Perkin-Elmer Zeeman 3030 AAS equipped with an HGA-600 graphite furnace, an AS-60 autosampler, and an Anadex silent scribe printer was used. Tantalum boats were made by hand out of a 0.25-mm-thick tantalum foil. Boats were inserted into nonpyrocoated graphite tubes (Figure 1). The ends of the boats were compressed to prevent sample solutions from flowing out of the graphite tube. The hollow cathode lamp (Perkin-Elmer) current was 25 mA, and measurements were performed at the 407.9-nm resonance line using a spectral slit width of 0.2 nm. No background correction was used, and signals were processed in the peak area mode. Argon (99.9999%) was used as the purge gas. A 5/95 hydrogen/argon mixture was used to study the influence of hydrogen on the atomization of Gd. Materials and Reagents. Five-milliliter or 10-mL polypropylene tubes (Biolab, Belgium), an automatic 1.0-5.0-mL Finnpipette (Labsystems, Finland), and a 200-1000-pL transferpette (Brand, FRG) with disposable tips (Labsystems) were used for solvent extraction. Stoppered 12-mL Teflon tubes (VEL, Belgium) were used for sample digestion. The caps of these Teflon tubes were not screwed but coned. As such, they loosely fit in the Teflon test tubes, allowing vapors to escape from the tube at elevated pressure. Gd-DOTA (Guerbet, France) was used for studying the biodistribution of Gd in the rat. Methyl isobutyl ketone (MIBK; Merck, FRG) was used for solvent extraction of Gd. The synergic 0 1991 American Chemical Society
424
ANALYTICAL CHEMISTRY, VOL. 63,NO. 5, MARCH 1, 1991 CROSS SECTIONS 1 0 0 0 n g o f Gd atomized a t 26OO'C
tantalum boat st 2 5 0 0 ' C
from the wall 01 a
,/
/ A t u b e 30
T I M E is)
T a n t a l u m boat
Figure 2. Comparison of absorbance profiles and sensitivity of the Gd determination using either a graphite furnace with tantalum boat (80 ng of Gd in 1 % HCI) and a pyrolytically coated graphite tube without tantalum boat (1000 ng of Gd in 1 % HCI).
T a n t a l u m foil
Figure 1. Graphite furnace with tantalum boat.
t
k-++
thenoyltrifluoroacetone (?TA)/tributyl phosphate (TBP) (Janssen Chimica, Belgium) system and 4-benzoyl-3-methyl-l-phenyl-2pyrazolin-5-one (PMBP) (Aldrich Chemical Co., Inc, Milwaukee, WI) were compared as possible chelators for complexing the Gd prior to solvent extraction.
Methods. Sample Preparation. One gram of wet tissue or 1.0 mL of biological fluid was digested with 1 mL of concentrated HNO, in the stoppered Teflon tubes for 6 h in an oven a t 120 "C. The clear digestate was quantitatively transferred into a 25-mL plastic container. The pH was adjusted to 6.7 with ammonia and acetic acid by using a pH meter (Beckman Model 41, Fullerton, CAI. The solution was quantitatively transferred into a polypropylene test tube and diluted to 5 mL with high-purity H20. One milliliter of a 6 M ammonium acetate buffer (pH 6.7) and 2.0 mL of MIBK containing 0.01 M PMBP were added. The test tube was shaken for 2 min and then centrifuged for 5 min a t 3000 rpm. After centrifugation, 1.5 mL of the MIBK phase was pipetted into a 5-mL polypropylene tube containing 0.5 mL of 0.15 M HCl. Tubes were again shaken for 2 min and centrifuged for 5 min at 3000 rpm. About 0.4mL of the aqueous phase was then transferred into 5-mL polypropylene tubes and centrifuged under vacuum by using an ultracentrifuge (Beckman Model 52-21] at 5000 rpm for 10 min in order to remove any remaining MIBK in the aqueous phase prior to GFAAS analysis. After centrifugation, this solution was transferred to polypropylene autosampler cups for GFAAS analysis. Standards. Five-milliliter standard solutions containing absolute amounts of 0.0, 2.0,4.0, and 8.0 pg of Gd, respectively, were prepared out of the 1 g L-' stock standard solution (Sigma, St. Louis, MO). Standard solutions were processed as if they were samples.
RESULTS AND DISCUSSION Preparation and Use of the Tantalum Boat. Atomization of Gd from the wall of a pyrolytically coated graphite tube is prone to severe chemical interferences even when the element is determined in H 2 0 or in 1% HCI. As illustrated in Figure 2 , absorbance signals of Gd in these atomizers are strongly tailed and prone t o severe memory effects, leading t o erroneous results. However, after the insertion of a tantalum boat (Figure l),there was neither peak broadening nor memory effects (Figure 2 ) , and the sensitivity was substantially increased. These data support the hypothesis of Haines ( 5 ) , suggesting that atomization of Gd from a graphite and from a tantalum surface must take place according to two different mechanisms. In the presence of graphite, carbides may be formed that do not atomize readily (5). L'vov e t al. (6) found it difficult to vaporize alkaline-earth and rare earth elements
09~
'
"
'
w 50
"
"
'
100
"
"
160
"
"
'
200
Number of Firings Figure 3. Influence of the number of firings on t h e sensitivity and reproducibility of the GFAAS determination of Gd (80 ng) with tantalum boat using either an uncoated ( X ) or a pyrolytically coated tube (0). Each point represents either the mean (upper) or the RSD (lower)of 10 determinations. from a graphite surface due t o the formation of pyrocarbon shells on microparticles of the samples. Our modification of the graphite tube involved the use of a tantalum boat. In our experience, boats could easily be made by hand using a 0.25-mm-thick tantalum foil (Figure 1). Compared to tube lining with tantalum (7,9),tantalum boats can be made with foils of greater thickness, which prolongs the lifetime. Indeed, when lining was done with 0.025-mmthick tantalum foils, damage already occurred after one measurement. This was also reported by Haines ( 5 ) using tantalum foils with a thickness varying between 0.018 and 0.062 mm. Moreover, it has been claimed that the tantalum-lining procedure is very critical ( 5 , 7 ) . Indeed, in order to extend the lifetime of the tantalum lining and thus obtain results within an RSD
