(1) the analyte must be transferred completely from the matrix into the bulk SF;

(2) the analyte must be transferred from the extraction vessel to the accumulation device in a quanti­tative fashion;

(3) the analyte must be collected efficiently by the device.

The evaluation of the extraction (step 1) cannot be addressed until the third step is op­timized. Unfortunately, many workers have reported recovery data without clear­ly demonstrating the effectiveness of the accumulation device (e.g., trap) being utilized. Without a doubt, many low recoveries reported in the past can be traced to inadequate trapping and not to insufficient extraction. The efficient purging of extracted analyte from the extraction vessel (step 2) also cannot be evaluated ac­curately unless trapping is reproducible and quantitative. Therefore, it appears that step 3 is the most critical one for achieving the goal of 100% recovery. Natu­rally, if one is developing a process (e.g., drug from a natural product), is not in­terested in quantitative results, is more concerned with the decontaminated matrix (e.g., raffeinate) than the extracted component (e.g., polymer purification), the reliability of step 3 is not as important. Since our focus is on trace analysis of the extracted analyte, a closer look at trapping after SFE is justified.

Unfortunately, there is not one universal trap which will effectively accumu­late all chemical classes under varied temperature, pressure, and fluid composi­tion parameters. Volatile materials will probably be more difficult to quantita­tively trap than nonvolatile ones. High percentages of certain modifiers will also create trapping problems. Some understanding of the chemical and physical na­ture of the analyte(s) to be trapped is a valuable guide when selecting an accumu­lation device. Extraction hardware (e.g., the outlet restrictor) and extraction para­meters (e.g., density, temperature, and time) may also influence the perceived efficiency of the trap. For example, high flow rates can be detrimental, especially in cases where the trap is not very active.

It is important to realize that the accumulation device and restrictor are intri­cately intertwined in the SFE experiment. Low recoveries and nonreproducible results are often due to erratic fluid flow arising from plugged restrictors during the run. Restrictors are usually fixed-diameter tubes or electronically controlled variable orifices. As the name implies, the fixed restrictor does not change inner diameter during the SFE run; the variable orifice responds to partial restrictions in the opening caused by precipitation of the extracted analyte prior to the trap. This is not to say that a variable restrictor will not plug, but experience has shown that plugging is a smaller problem with this type of restrictor. Restrictor plugging is especially prevalent with samples containing large amounts of inor- • ganic sulfur, lipids, water, or other highly extractable material. Fixed restrictors have advantages, however, in that they are cheaper, can be manufactured in house, may be used with liquid traps, and are amenable to all vendor instrumen­tation. Variable restrictors are more expensive and currently cannot accommo­date direct liquid solvent trapping. Our discussion of trapping after SFE must first review the physical relationship of the restrictor and the trap.