|
 |
Monday, June 1, 2009
|
 |
|
|
Should You (Heat) or Cool Your Spray Chamber?
There is a large body of research which shows that a modern argon plasma can handle a mass load of about 40 - 60 mg/min. If you do the math, and assume the nebulizer-spray chamber is 5% efficient, at 1 mL/min you are putting about 50 mg/min into the plasma. Solvent loading is the term generally used to describe changes in the plasma when too much solvent reaches it. Essentially, more plasma energy must be expended for vaporization of the solvent, leaving less energy for decomposition, excitation and ionization of the analyte. In ICP-AES sensitivity and detection limits will be degraded; additionally, in ICP-MS oxide levels will increase if the solvent is water. This solvent loading is a major reason why “more efficient” nebulizers really don’t deliver better instrument detection limits. Short answer: cooling is a good idea, heating isn’t, and here is why.
The spray chamber has a number of roles. The foremost role is to reduce the aerosol size to something less than 10 um on average (<5 um is most desirable). It achieves this through gravitational settling and through impaction on surfaces. In a cyclonic spray chamber, inertial impaction occurs as the aerosol undergoes a cyclonic swirl before it leaves the spray chamber. Larger droplets hit the walls and, for the most part, collect and travel to the drain. This ordinarily accounts for about 95% of the sample that goes into the nebulizer. But some droplets will impact the walls and splatter, putting material that was on the walls back into the aerosol stream. In a cyclonic spray chamber, or a conical with impact bead, this is a cause of memory or carryover which is manifested as a longer washout time. The second role is to very rapidly and smoothly remove the collected liquid from the walls to minimize this effect. The current cyclonic and conical spray chamber designs utilize improved geometry and aerodynamics as a result of research. But if the spray chamber is “dirty” and does not drain well, you can expect an increase in washout time. Third, it should be understood that desolvation of the aerosol begins in the spray chamber. If a droplet moves through an atmosphere of dry argon, it will rapidly lose solvent as vapor to the argon.
How rapidly it desolvates depends on the nature of the solvent, solvent-solute interactions, the size (surface area) of the droplet, the surface area of the spray chamber, and the temperature of the spray chamber (among other things). Cooling the spray chamber will reduce the solvent load by condensing solvent vapor on the walls, and by removing the condensed liquid along with the “excess” aerosol. Clearly, heating the spray chamber will enhance desolvation and will put more vapor into the plasma unless there is a downstream mechanism for solvent removal such as a cold condensor or membrane desolvator as, for example, in the Apex from Elemental Scientific.
This spray chamber desolvation process is the source of a couple of interferences that lead to bias in the analytical result. In an elegant set of experiments, Stewart and Olesik (J. Anal. At. Spectrom. 1998, 13, 843) showed that the difference in analytical response between a 2% HNO3 solution and a 10% HNO3 solution of Mn results from the time it takes the newly-formed aerosol to achieve vapor-phase equilibrium with the liquid that remains on the walls. If a spray chamber is washed with 10% HNO3 then the sample is introduced in 2% HNO3, there will be an initial surge in signal of perhaps 15%, which might take several minutes to stabilize at some level that is lower than the peak, while HNO3 on the walls and the vapor in the spray chamber slowly decreases from 10% to 2%.
Practically, what this means is that the rinse matrix as well as the calibration standard matrix should match the sample matrix as closely as possible.
Desolvation and equilibration in the spray chamber depend on temperature. If the temperature changes, signal will drift. In fact changes in spray chamber temperature account for a large fraction of the long-term drift in modern ICP spectrometers because the amount of analyte and the amount of solvent change with time. Some instruments have been designed with a heater in the sample intro area in order to stabilize the spray chamber temperature and improve long-term stability. Generally, this will only be a few degrees above the normal high ambient temperature that the instrument is expected to encounter. This approach is less desirable in ICP-MS since, for aqueous solutions, the oxide level will increase. But in ICP-AES, the improved stability comes at a very small price.
For aqueous solutions, the ideal temperature is just above 0 C to remove the most solvent vapor. Usually chillers are set to +2 C; you do not want solvent to freeze in the spray chamber; you do not want to induce crystallization or precipitation so that solids collect and block the drain. Organic solvents are generally far more volatile than water, and lead to a more significant solvent loading problem. Moreover, in ICP-AES there is hydrocarbon band structure below 300 nm that can overwhelm analyte emission, again because so much solvent is reaching the plasma relative to analyte. Band structure that is not stable is highly problematic. Toluene is more of a problem than kerosene, and gasoline is much more of a problem than toluene. Cooling the spray chamber to less than 0 C dramatically reduces vapor transport and plasma loading. If there is no water present (kerosene can carry several percent water), one might surmise – the colder the better. And one might be wrong.
There are practical limits to how well the spray chamber can be protected. If condensation forms on the outside and freezes between the spray chamber and its mount, it could crack a glass spray chamber or move it enough to diminish the integrity of the spray chamber – injector connection. With organic solvents, viscosity can be nearly as sensitive to temperature as is vapor pressure. For the analysis of lubricating oils, even at 1:10 dilution, a very cold spray chamber will not drain very well, and will leave a lot of material on the walls of the spray chamber. Usually, very little is gained by cooling to less than -5 C unless the solvents are highly volatile, like methanol, acetone, or methylene chloride. Finally, in the analysis of organic solvent solutions, it is imperative to let nebulization control the rate of analyte transport to the plasma. At room temperature, slightly volatile organometallic compounds will be transported both as vapor and as aerosol. If the material used to calibrate is not chemically identical to the analyte in the sample, there can be significant bias in the analytical data. Among the more obvious examples would be the determination of sulfur – anyone who has ever handled an organo-sulfur compound has smelled it. That vapor would be sample transported to the plasma with essentially no control, and no hope of analytically useful data.
Article courtesy of Geoff Coleman, Meinhard
Return to Spectron, Inc. Mini-Site
|
|
|
|
|