ADV06-15_1-412

347 / 413

Analyzing Residual Solvents in Water-Soluble Articles
Dynamic Headspace Sampling Enhances Sensitivity by GC
By Rick Lake,Pharmaceutical Innovations Chemist
• Sensitivity increased 13X-30X for residual solvent s (Ovls) in water.
• Excellent resolution and stable retention times , using an Rtx"'-G43 column.
• Greater sensitivity makes smaller samples possible.
Residual solvents, or organic volatile impurities (Ov ls), in pharmaceuticals are trace-level
leftover solvents that were used in the manufacture of drug products or excipients. The
International Conference on Harmonization (ICH) provides guidelines that summarize the
allowable concentrations of common solvents. However, some of the detection limits in the ICH
guidelines are not easily achieved thro ugh the normal sampling technique, static headspace
analysis, and pharmaceutical manufacturers are becoming concerned with attaining greater
mo re toxicity data become available, maximum allowable concentration limits are being
lowered. And, as active ingredient and excipient markets are becoming more global, tighter control of
impurities is needed.
In our investigation s, we have found that coupling a dynamic headspace samp ling techn ique with analysis
on an Rtx®-G43 column greatly increases sensitivity for residual solvents, and maintains stable retention.
Analyses for residual solvents typically are performed using headspace sampling coupled with GC/FID. In the
commonly used static headspace techniq ue, a pressurized or ballast loop system is used to extract a portion of the headspace in the
samp le vial for introd uction into the Gc. Another, more novel, technique for headspace sampling is the dynamic headspace technique .
In this technique, the entire content of the vial headspace is swept onto an activated trap, which collects and concentrates the target
analytes, then desorbs the analytes into the GC carrier flow. Dynam ic headspace increases the sensitivity of the analysis,but high con­
centrations of organi c solvents will cause contamination and lifetime problems with the trap and, therefore, this technique is not com­
patible with the use of organic solvents as diluents for water-insolubl e articles. On the other hand, the techniq ue is well suited to, and
easily performed in, analysesof residual solvents in water-soluble articles.
We evaluated the sensitivity of th e static and Table 1 Dynamic headspace sampling greatly increases sensitivity
dynamic headspace techniques, using solvents in an fo r Ovls.
aqueous matrix, to compare responses as they
Sample Concentration Increase in might relate to pharmaceutical analysis of residual
Cone. at Regulatory Mean Peak Area Response Sensitivity with
solvents in water-soluble articles. We prepared ref­
Analyte (ppm) Limit (ppm) Static Headspace Dynamic Headspace Dynamic Headspace
erence standards containing the USP467 solvents at
!lli:_hlorqrJletha_I!!LI?~0 600 --,,619 I~.6 9 --",3OX __,-- ,,-,7-<.
their regulatory limits in water, by adding 100ilL of chloroform 1.2 60 39 783 20X
ou r USP 467 Calibration Mix #5 (cat.# 36007) to benzene 0.04 2 15 313 21X
5mL of deionized water in a 22mL headspace samtrichloroethene 1.6 80 141 3479 25X
14-doxane 7.6 3~0 20 272 13X pling vial. We also added approximately I gram of
an inor ganic salt, sodium sulfate, to each sample to
decrease the solubility of polar compounds. This is
Table 2 Solvent retention times and resolution are equivalent for static
critical for highly water-soluble volatiles, like 1,4­
or dynamic headspace sampling and analysis on an Rtx"'-G43column.
dioxane, as it promotes analyte tran sfer into the
gaseous phase in the sample vial. Static Headspace Dynamic Headspace
Retention Retention
First, we used a traditional static headspace (loop ) Solvent Time (min.) Resolution Time(min.) Resolution
technique to assay a system suitability set com dichloromethane Mean 5.092 5.139
prised of 6 replicates (Figure IA). The sample vial Std. Dev. 0.01 > 0.00
%RSD 0.25 0.04
was heated , mixed, and pressurized. A six-port
chloroform Mean 9.250 23.02 9.263 22.18
valve was used to fill a specified loop volume with Std. Dev. 0.02 0.26 > 0.00 0.07
an aliquo t of the headspace, then the valve was %RSD 0.23 1.11 0.04 0.31
switched to redirect the gas flow, flushing the sambenzene Mean 11.134 7.67 11.145 7.72
ple into the transfer line and ultimately mixing Std. Dev. 0.03 0.08 > 0.00 0.01
%RSD 0.23 1.04 0.03 0.11
with the GC carr ier gas flow. Next, we used a
trichloroethene Mean 14.592 11.87 14.599 11.86
dynamic headspace (trap) technique to analyze an Std. Dev. 0.03 0.06 > 0.00 0.01
equivalent 6-replicate system suitability set (Figure %RSD 0.23 0.46 0.04 0.10
IB). The sample vial was heated and mixed und er l A-dioxane Mean 17.388 7.91 17.411
Std. Dev. 0.04 0.10 0.09
the same conditions as used in the loop method,
%RSD 0.20 1.23 0.50
then a gas flow was int roduced into the headspace
2006 vol. 1 • 14 •
Website : E-mail : TelNo : 03 9762 2034 . . . in AUSTRALIA