HS-GC for residual solvent with high B.P. and very low level

[wzfumaa]

I am developing an HS-GC method for the following four residual solvents as IPC for washing/dry of an API (wash with MeCN, then IPA, then water). The spec limit is 2% (20,000 ppm) for IPA and MeCN, only 20 ppm for reaction reagents cyclopropyl methyl chloride (CPM-Cl) and cyclopropyl methly iodide (CPM-I).

Name / B.P./ Spec Limit
IPA /82/ 2%
MeCN /82/ 2%
CPM-Cl /89/ 20 ppm
CPM-I /142/ 20 ppm

The problems we have:
1. We tried several columns and diluents (DMF, DMSO, Tolune, etc). CPM-I is always co-elute with diluent.
2. The FID sensitivity for CPM-Cl/CPM-I seems very low. We also found that the CPM-I being such a high boiler did not have consistent area counts when injected at low levels.
3. In order to be able to detect low level of CPM-Cl/CPM-I, sample amount needs to be very high (e.g. 500 mg/mL). However, the high amount of MeCN in the sample caused severe carry over in the headspace requiring service calls to change out the 6 port valve and sample loop in order to get it out.
4. We tried to use two HS-GC methods, one for high level of IPA/MeCN (pretty straightforward), another one for low level CPM-Cl/CPM-I, but we still have the same co-elute issue for CPM-I with diluent DMF, DMSO as said above and same carryover issue of MeCN when using MeCN as diluent.
5. We also tried direct injection GC for CPM-Cl/CPM-I, the LOD is ~8 ug/mL with peak area and height only 5 PA. That is 800 ppm if the sample conc is 10 mg/mL (By the way, typically how much of a sample (ex. 1 mg/mL) can be injected directly onto the GC? do not care the peak distortion of API).

Any suggestions would be greatly appreciated!

[chromatographer1]

Try making standards of CPM-Cl and chlorobutane in DMAc? and inject both stds for comparison.

If the peak areas are drastically different (and I think they will be) then the cyclopropane based reagents are probably too reactive to use a headspace method.

To confirm the HS conclusion, add both to a headspace vial of water (max partition) and compare the peaks of each.

best wishes,

Rod

[wzfumaa]

Rod,

Thanks for your suggestion. Here are the results I just got:

Peak area of 0.1 mg/mL of 1-Chlorobutane, CPM-Cl, and CPM-I in DMA: 310, 209, and 29. CPM-Cl seems OK with HS, but the response of CPM-I is way too low. Not sure it is because high boiling point of CPM-I (142C), not stable with HS, or my HS-GC conditions were not right.

Also prepared 1-Chlorobutane and CPM-Cl in 50/50 DMA/H2O, the peak area of 0.1 mg/mL of 1-Chlorobutane and CPM-Cl : 884 and 271, respectively. Addition of water can improve the response of 1-Chlorobutane a lot but not much for CPM-Cl. There is also an unknown peak with area of 99, which may came from CPM-Cl when prepared it in water.

The HS Conditions I used:
Vial size: 20 mL
Sample volume: 1 mL
Oven Temp: 90 °C
Loop: 120 °C
Tr. Line: 130 °C
Shaking: High
Oven Stab.: 1.00 min
GC Cycle: 25.0 min1
Sample Equilib.: 10.0 min
Vial Press: 0.20 min
Loop Fill: 0.20 min
Loop Eq. time: 0.10 min
Sample Inject: 2.0 min

GC conditions:
Oven Temp (initial): 60 °C
Run Time: 20.0 min
Equilib. Time: 1.0 min

Temperature program:
Rate / Next oC / Hold (min)
Initial / 60 / 7.00
30 oC/min / 220 / 7.67

Inlet
Mode: Split
Inlet Temp: 180
Split Ratio: 10:1

Column
J&W, DB-624, 75 m x 0.53 mm, 3.0 µm film
Mode: Constant Pressure: 15.0 psi

These conditions have not been optimized yet, the sensitivity seems very low, especially for CPM-I. In order to meet the spec (20 ppm of CPM-Cl and CPM-I), the sensitivity need to be imporved at least 10 times after optimizing the HS-GC conditions.

Any suggestions would be greatly appreciated!

[Peter Apps]

You need a 10 times improvement in sensitivity and you have a 10:1 split. With a 530 um dimater column you can run at 10 ml/min carrier flow (although the pressure would be quite high with such a long column - do you really need all that length ?) which will transfer your 1 ml loop volume to the column in 6 s, so you will see no significant band spreading. You will need a connections with less dead volume that an ordinary split/splitless inlet liner. The easiest for a quick tryout would be a 1 mm i.d. SPME liner but for a permanent installation I suggest a zero dead volume connector between transfer line and column.

Shortening the column and increasing the gas flow rate to make the peaks narrower will also make them higher, and easier to detect.

peter

[chromatographer1]

Some of the low sensitivity is due to the high weight potion of the I in the molecule, but much of it is due to the reactivity of the chemical.

Heat at a low temp as possible for a short a time as possible and cross your fingers.

Also a HID would be useful instead of a FID for increased sensitivity.

Good luck, your problem is not trivial.

Rod