| Author | Message | ||
     David W. Hopkins (Dhopkins) |
I'm interested in putting a new system online for measuring moisture at 1 to 5% levels and acetone at 10 to 60% in acetone-hexane-water mixtures. Sometimes pure acetone or hexane will be used, and the important variable to control is the low level of water. The samples are clear, and the temperature range is generally 50-100F (10-38C), at about 30 psi pressure in the line. We want to determine all 3 solvents when they present. Currently measurements are taken at 4 hour intervals and measured by GC and KF methods, so NIR would be a real benefit. Is anyone currently doing NIR measurements like this? Can anyone recommend instrumentation for this application? Will it be necessary to use a thermostated measurement, or can we expect to generate a temperature-compensated calibration? Are there any other problems we are likely to encounter? Thanks for your suggestions. Regards, Dave | ||
| hlmark |
Dave - water will mix with the acetone very nicely, but at higher levels of hexane I think you'd be likely to find that the water will start precipitating out and make the sample cloudy - scattering, you know! \o/ /_\ | ||
| David Russell (Russell) |
My colleagues would tell you to get to forget NIR and get a GC for reasons such as the issue Howard suggested. But if you were to try NIR you would need temperature control, not compensation. | ||
| Stephen Medlin (Medlin) |
Dave, We're working on a similar application as what you mentioned. Water in acetic acid, for example, works quite well. We also have done water in hexane. Several problems as Howard mentions. First, we were able to see very low levels of water (ppm) in hexane (which was our objective). The problem was getting very dry hexane to start with (I believe that the best that we could find was 20ppm). We were able to spike the hexane semi-quantitatively to see water, but it was easy to get a saturated solution that was cloudy/immiscible. We also used a thermostatted sample compartment of our FT-NIR instrument (Bruker MPA) to minimize the water peak "walking" around on us as a function of temperature. | ||
| Gabi Levin |
Sorry, it is so late to join - We shared similar issues with very low levels in dichlorobenzene. Max solubility is ~400ppm. The issue was complicated further because of the tendency to form HCl in the solvent through the reactions that led to the formation of the water in the matrix. To demonstrate the capability to measure moisture to low ppm (~30) we needed 1. Very dry DCB. 2. Absolutley sealed bottle 3. Means to inject microliter quantities of water into the bottle without opening 4. Means to extract 1ml sample without exposing the inside solvent to air. 5. Fast (below 4 seconds) measurement to prevent moisture from air to get into the cuvette. We used sureseal with rubber plug dry DCB, with certified max of 20ppm water. We used microliter syringes to inject water into the bottle, and 1 ml syringe to extract sampls and the Luminar 3010 Brimrose AOTF for fast ( 3 seconds) spectrum colletcion. The cuvette was covered with a stopper. The calibration was very good down to 20ppm. The limit of solubility was reached, and we definitely saw that beyond that the spectrum showed no additional increase in water absorption, due to the water separating and settling to the bottom. Next we did was to obtain a solution of very dry HCl in an organic solvent, and we added water as before. The results showed that even in the presence of increasing HCl in the solution, the moisture can be determined with similar results. Thus, to use the NIR for low moisture in organic solvents is a relatively easy thing, with care paid to the solubility limit, and possible cloudiness if the excess water separates from the solution in such fashion. We were able to do very similar measurement, in the 40 - 500 ppm level in Methanol, and here of course the solubility is not an issue. Hope this contributes soemthing. Gabi Levin Brimrose | ||
| David W. Hopkins (Dhopkins) |
Gabi, Thanks for your input. I am still interested in what instrumentation is available to do the job. Best regards, Dave |