Direct Determination of Oxygen by HPLC. 2. Chamber and Sample

All oxygen measurement systems so far available are characterized by a lack of suitable precision and/or required limit of detection, which would be e...
0 downloads 0 Views 164KB Size
Anal. Chem. 1997, 69, 4476-4481

Direct Determination of Oxygen by HPLC. 2. Chamber and Sample Application System for Determination of O2 at Trace Levels T. Seppi,† G. Stubauer,‡ D. Obendorf,*,‡ and P. Lukas†

Institut fu¨ r Analytische Chemie und Radiochemie, Universita¨ t Innsbruck, Innrain 52a, A-6020 Innsbruck, Austria, and Universita¨ tsklinik fu¨ r Strahlentherapie und Radioonkologie, Universita¨ t Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria

All oxygen measurement systems so far available are characterized by a lack of suitable precision and/or required limit of detection, which would be essential for a great variety of applications. In this paper, a novel oxygen chamber together with a completely new concept of sample application (“two-chamber-siphon technique”) is presented which can be used in combination with the previously reported chromatographic oxygen sensor (part 1). This new oxygen-sensing assay exhibits several advantages in comparison to conventional oxygen measurement systems: e.g., the uncontrollable influence of the surrounding atmosphere as well as oxygen consumption and storage processes are excluded. For the first time, measurements of molecular oxygen below 1 × 10-7 mol L-1 can be performed. Reliable quantification of oxygen in liquids and also in gaseous and solid samples can be achieved with utmost sensitivity (LOD 4.9 × 10-9 mol L-1 O2 ) 98 fmol of oxygen on column) and precision (RSD ) 0.7%, n ) 8). Despite the long history of oxygen determination by means of various methods and apparatus, it is still an unsolved problem to determine O2 in trace level concentrations. However, in many fields, e.g., medicine, microbiology, cellular physiology, biochemistry, respirometry and in analytical chemistry, reliable measurements of oxygen concentrations below 1 × 10-6 mol L-1 are important. Usually, polarographic sensors (membrane covered electrodes, e.g., Clark electrodes) are used for the detection of oxygen, because they are basic laboratory equipment. They can be easily handled and are cheap compared to more sophisticated methods (ESR, NMR, PET, fluorescence and luminescence techniques). Although these oxygen sensors are sufficiently sensitive and accurate for most applications, they usually suffer still from at least one of two major problems: the distortion of the analytical response by the measuring system itself and the predominant influence of the surrounding atmosphere.1,2 One problem of polarographic O2 sensors arises from oxygen consumption of the electrode due to the electrochemical measuring principle. This leads to an undesired decrease in the oxygen concentration in the sample chamber. Another critical point in measuring oxygen at trace level is the construction and quality * Corresponding author: (phone) (+43)512-507-5178; (fax) (+43)512-580519; (e-mail) [email protected]. † Universita ¨tsklinik fu ¨ r Strahlentherapie und Radioonkologie. ‡ Institut fur Analytische Chemie und Radiochemie. (1) Stone, H. B.; Brown, J. M.; Phillips, T. L.; Sutherland, R. M. Radiat Res. 1993, 136, 422-434. (2) Haller, T.; Ortner, M.; Gnaiger, E. Anal. Biochem. 1994, 218, 338-342.

4476 Analytical Chemistry, Vol. 69, No. 21, November 1, 1997

of a suitable measurement chamber. A variety of chambers with different designs have been developed and published which try to solve the tricky problems of easy application of samples and reagents to the chamber, sensor calibration and gas-tight connection of the oxygen electrode to the sample solution.2-7 But none of the chamber designs published so far was able to cope with all of these special requirements and all of them are limitated in gas tightness. As a consequence of this flaw, uncontrolled diffusion of atmospheric oxygen into the chamber occurs, e.g., during sample transfer through a sealed needle port or due to inadequate attachment of the electrode to the chamber. The diffusion rate cannot be calculated, because it depends on the O2 concentration gradient between the sample solution and the environment.8 The materials constituting a typical conventional measurement chamber represent further sources of error. Porous materials (e.g., Teflon, PEEK, and other polymers), including sealings, polymer coatings, and the embedding of the magnetic stirrer bar, form a pool for reversible storage and release of oxygen. Materials that are not inert toward oxygen corrosion (e.g., ignoble metals) are oxidized, and this process leads again to uncontrolled oxygen depletion in the sample. Corrosion and reversible storage or release of oxygen were previously suspected to be the reason for distortional effects on oxygen readings obtained in conventional test chambers.2 But these effects could not be detected experimentally because of the overwhelming oxygen influx into the measurement chamber compensating any occurring oxygen consumption processes. Considerable errors in oxygen readings are the consequences of these effects, especially at very low oxygen concentrations (