Polar analyte analysis

[Hollow]

(How)/Can a logP (octanol/water) be used for a rough estimation on what HPLC mode could be used? (or log D for ionisable compounds)

Values calculated with ACDLabs ChemSketch 12 with logP Add-On (Freeware)

• Acetophenon 1.66 +/- 0.22
  Propiophenon 2.2 +/- 0.22
  Butyrophenon 2.73 +/- 0.22
  
  Acetone -0.16 +/- 0.19
  Uracil -0.71 +/- 0.29
  Thiourea -1.05 +/- 0.19
  1-Nitroguanidine -1.31 +/- 0.66
  2-Nitroguanidine -0.89 +/- 0.19

Can this statement be used as a rule of thumb?
The bigger the logP, the more will it be retained on a RP-column, and the smaller the logP, the more will HILIC or NP mode be the choice.

[bisnettrj2]

A.) I contacted Waters re: Atlantis T3 column. We'll see if the boss says yea or nay, depending on what Waters' terms are. Based on their flowchart (http://www.waters.com/waters/nav.htm?cid=513211), I wouldn't think it would work. I would think, based on the chart, they would recommend XBridge HILIC or Atlantis HILIC Silica, or as Uwe has suggested, normal phase.


B.) Normal phase seals are cheaper than doing a lot of worthless work, I agree.

C.) Vlad - $3000/week? I think I might be underpaid....

[HW Mueller]

Well, the ref I gave above says that putting this molecule in water gives about a neutral solution, thus it doesn´t dissociate. Still, I really don´t understand why HILIC wouldn´t work. Is this thing not polar enough? But too polar for RP? Salting out on RP doesn´t go either?

(On acid or base: Had it a pKa of 4 it would be a weaker base than acetic acid, or, relative to water an acid)

[Vlad Orlovsky]

If your first bare silica column did not work xBridge is not going to work either (in HILIC mode)..and I was under impression that I told you to use normal phase (even sent you a method)...now I am upset and not going to post on this thread

$3000/week is direct and indirect compensation. If you consider your salary, SS and medicare company pays for you, health insurance, workers compensation insurance and 10 other things you will arrive at $3000 if not more. I did not realize this either until I start running my own company

[Bryan Evans]

Just to be clear about this 'HILIC' term:

'HILIC' is normal phase, but with water. You saw no retention when water
was used as the strong solvent. Vlad suggested normal phase, meaning you
use a 'weaker solvent' (less polar) so that your solute is retained.

Imtakt used hexane/THF/aqueous for NP of nucleosides on Unison UK-Amino:
http://www.imtaktusa.com/site_media/fil ... TI406E.pdf

They call this 'sandwich' technology - meaning water is miscible with hexane
when THF is added to the eluent. Similar conditions (as Vlad suggested) may
work for your solute on your (current) bare silica column.

[H.Thomas]

It is a quite old method (1989) but apparently worked:

RP-HPLC analysis

Our analysis was conducted as described by Walsh (1989) on a modular
system composed of a Spectra Physics SP8810 Precision Isocratic pump with the flow rate set at 1.5 mL/minute, and the rise time and range set at 0.10 and 0.001, respectively. We used a 4.6- × 250-mm Mixed Mode RP18 Cation 100A-7-U column (Alltech) and a Spectra Physics 100 variable wavelength detector set at 263 nm. The autosampler (Dynatech Precision Sampling Corporation) was programmed to cycle each sample through in five minutes, beginning with a 45-second flush, and then overfilling a 100-μL precision sample loop injector.

The water used as eluent was prepared by vacuum-filtering through a microfiber filter nylon membrane (0.45 μm). Air was removed from the eluent by continuous helium-sparging throughout the RP-HPLC process.
Prior to sample analysis, we ran calibration standards ranging in concentration from 0.1 to 5.0 mg/L. Between each time series of samples, we analyzed a blank water sample, and a 0.8-mg/L standard. At the end of the run, a final set of calibration standards was analyzed, ranging from 0.1 to 5.0 mg/L. The Day0 through Day8 samples all were analyzed together eight days after spiking, thereby eliminating any uncertainty due to slight day-to-day calibration differences. The NQ-peak elution times ranged from 3.05 to 3.12 minutes during the seven-hourlong analysis period.

quoted from
"Stability of Nitroguanidine in Moist, Unsaturated Soils", ERDC/CRREL TR-05-2 http://libweb.erdc.usace.army.mil/Archimages/58848.PDF

[HW Mueller]

Curiosity overcame me so I googled some more, this time with "HPLC of Nitroguanidine". The first one was an armed forces lab method using a ODS column, apparently not getting retention, but getting enough selectivity via voltammetry etc. detection (they think). here the link:


http://www.ornl.gov/info/reports/1987/3445601535257.pdf


There follow an array of other armed forces methods which are similar, all with typical curiosities. Later comes a review from 1992, apparently including NP:

http://books.google.de/books?id=lcVwgGR ... ne&f=false

[bisnettrj2]

Thank you HW Mueller! Your search gets a gold star! Normal phase it is (confirming Uwe's idea), using a silica column and isooctane:ethanol (second Google books link got me there - and I found a used version of that book on Amazon for $5, so I bought it). That only questions I have now are:

Since my current extraction is with water, would I still inject the water extract on a silica column running a 90:10 iso-octane:ethanol mobile phase? Or would I need to exchange the extract into ethanol first? And do you treat normal phase like HILIC, where you basically run a gradient to the higher polarity mobile phase constituent?

[Uwe Neue]

I am not that terribly concerned about the water in your sample solvent. It is not any different than trying to inject a DMSO sample in RP: as long as your injection volume is small, it is not a big concern. You also have a large amount of ethanol in the mobile phase, so a rapid deactivation of the silica is also not of concern.
I would first try to find isocratic conditions. Start with the proposed mobile phase from the literature, and then add or subtract ethanol.

[bisnettrj2]

Thank you very much. I really appreciate the thoughts and suggestions everyone has had on this topic, and as soon as I get a chance to set up an instrument in normal phase for this application, I'll try to post some results.

[Uwe Neue]

Well, good luck! You had some of the best brains on this board work on your problem. BTW: if a normal-phase method works, the honor for this suggestion should go to Vlad, who proposed it first.

[bisnettrj2]

I have no doubt the best brains worked on this, and I am appreciative of Vlad's suggestion. I'm just glad there are great brains out there to help those of us who need the helping.

I have the new normal phase seals on order, as well as the solvents necessary. With the holidays here in the US, I don't expect to receive them until the first full week of January. I assume I should wash my Kinetex silica HILIC column with water (to clear out ACN and any remaining buffer), then IPA, then iso-octane and ethanol? I also assume I should start isocratically at perhaps 95:5 iso-octane:ethanol, and go in 5% greater increments of ethanol to find a reasonable k' value?

Thanks again to everyone who helped. Getting this figured out is my Christmas present for this year!

[bisnettrj2]

FYI, NP is a pain in the rear. Using 95:5 isooctane:ethanol at 1 mL/min, I was able to inject 1 uL of water with 2 ppm Nitroguanidine, and I got a decent peak at around 5 minutes (started a gradient to 50:50 at 2.5 min) on a Kinetex 50x4.6mm HILIC (bare silica) column. I only get to mess with NP for a while, but I couldn't inject enough water extract to get a decent peak at my low standard without have a massive pressure spike. I had to set the analysis aside for other work, but it's not too promising at this point.

I think the solution may be to concentrate my water extracts in order to correlate to a higher standard in my calibration while still maintaining my current reporting limits, but that adds a whole other bag of cats to the problem.

[bisnettrj2]

Thanks to Vlad for suggesting using a different solvent for injection. I found a solubility table for nitroguanidine that showed methanol as being relatively equal to water in terms of nitroguanidine solubility.

Today I added 100 uL of a 100 ppm nitroguanidine standard to 400 uL of ethanol and 500 uL of isooctane and set up the instrument as follows:

Kinetex 50x4.6 mm silica HILIC column
1.0 mL/min
40 C
20 uL injection

Gradient program - A = Ethanol, B = Isooctane
t - A : B
0 - 5 : 95
2.5 - 5 : 95
10 - 50 : 50
12.5 - 50 : 50
15 - 5 : 95
23 minute total runtime

My peak elutes at about 5.7-5.9 minutes, and is a little distorted due to the solvent mismatch.

I've run out of ethanol, so I need to get more to run tests on extracting nitroguanidine from soil using ethanol, so I can make a pre-injection dilution to match the initial mobile phase composition I will eventually use.

Things are looking up in the world of nitroguanidine!

Because of what I've done today, I have a new question - I'm running a gradient, but I'm still getting a lot of tailing on the peak (even when I evaporated a methanol-based nitroguanidine standard and reconstituted in 95:5 isooctane:ethanol) - what can I due to reduce this tailing? Steepen the gradient? I'll post a chromatogram tomorrow when I get a chance to show the tailing.

[bisnettrj2]