Fast Analysis of USP 7 Residual Solvents using the Agilent 789A GC and Low Thermal Mass (LTM) System Application Note Pharmaceutical Author Roger L Firor Agilent Technologies, Inc. 8 Centerville Road Wilmington, DE 988 USA Abstract A dual column residual solvent analysis according to USP 7 (8-9 revision) was performed using the Low Thermal Mass (LTM) system installed on an Agilent 789A GC system. The G888 Automated Headspace sampler connected to the volatiles interface was used for sample introduction. A Capillary Flow Technology (CFT) two way splitter was used to split the sample equally to a inch 7 M x. mm x. µm Agilent J&W DB- column and a inch 7 M x. mm x. µm Agilent J&W HP-INNOWax column. Each column module was connected to its own FID by retention gaps. Aqueous solutions of Class, Class A, and Class B solvents were analyzed. Sensitivity, linearity, and resolution met the requirements of USP 7. Overall cycle times for the analysis of all specified Class and Class A and B solvents were reduced to min.
Introduction Residual solvents in pharmaceuticals may remain from the manufacturing process of the active pharmaceutical ingredients (API) or final product. The level of residual solvents are monitored and controlled for a number of reasons that include safety, effect on crystalline form, solubility, bioavailability, and stability. All drug substances, excipients, and products are included under USP 7. The LTM (Low Thermal Mass) chromatographic system is combined with static headspace sampling for the analysis of residual solvents in pharmaceuticals according to USP 7 revised general chapter 8. [] This chapter follows guidelines set by the International Conference for Harmonization (ICH) QC. [] Residual solvents are divided into three classes based on possible toxicity. Class solvents are considered the most toxic and should be avoided in manufacture. These solvents may also pose an environmental risk. Class solvents (A, B, and C) are less toxic with limited use. Class C solvents have higher boiling points and some of them require analysis by non-headspace methods. Class are least toxic and should be used as solvents where practical. Headspace GC is used for determination of Class and Class solvents, while most Class solvents are analyzed by a nonspecific method such as loss on drying. Each Class solvent has a "permitted daily exposure" (PDE) limit. If a given solvent tests below the PDE limit then further testing is not required (daily dose below grams). Option of the general chapter, which looks at the total solvent added for all components of the drug product, is used for daily amounts above g. This work follows the guidelines of the method with the exception of column dimensions and GC oven programs. Column dimensions and program rates were optimized to gain a significant reduction in analysis time and overall cycle time. Alternate methodologies such as those described here can be used, however, validation and comparison to the original USP monograph may be required. The FDA also requires that any new ANDA provide the data necessary to prove control of residual solvents prior to a generic drug approval. USP 7 specifies three procedures as follows for Class and Class residual solvents:. Procedure A: Identification and limit test. Procedure B: Confirmatory test. Procedure C: Quantitative test Procedure A uses a G phase (Agilent J&W DB- column in this work) and Procedure B uses a G phase (Agilent J&W HP-INNOWax column in this work). In general a particular co-elution that occurs on one of these phases will not occur on the other. Experimental The water soluble procedures were used for standard sample preparation to demonstrate system performance. Insoluble articles require use of DMSO, DMF, or other suitable nonaqueous solvent. The methodology used is very similar for both solvent systems. A diagram of the dual column system used is shown in Figure. The setup splits the effluent from the headspace equally to the J&W DB- and J&W HP-INNOWax columns for a simultaneous dual channel analysis. Previous work has been described using conventional oven heating and two-hole ferrules with the split/splitless inlet for dual column residual solvents. [] Configuration and parameter settings for the LTM system are given in Table. Agilent G888 Headspace Sampler Volatiles Interface LTM : 7 M. mm. µm Agilent J&W DB- FID CFT Splitter Agilent 789A GC System LTM : 7 M. mm. µm Agilent J&W HP-INNOWax FID Figure. System diagram showing CFT splitter use in dual LTM column configuration.
Table. Residual Solvent System Parameters Standards Class : p/n 9-, equivalent to USP Mixture RS Class A: p/n 9-9, equivalent to USP Mixture A RS Class B: p/n 9-9, equivalent to USP Mixture B RS Software ChemStation: B.. Headspace: G9AA, A.. LTM: G8AA 789A Configuration and Method Parameters Inlet: Volatiles Interface, C Pressure program: psig ( min) to psig ( min) at. psi/min Split ratio: :, He carrier Detectors: Dual FID CFT: Two-way splitter, G8B 789A oven: Isothermal at C LTM Module : 7 M. mm. µm J&W DB- LTM Module : 7 M. mm. µm J&W HP-INNOWax Module connections to CFT splitter and FID's:. M. mm deactivated retention gap LTM module programs: See Table. G888A Headspace Parameters Oven: 8 C Loop: 9 C Transfer line: C Cycle time: LTM program dependent Vial Equilibration time: min Pressurize time:. min Loop fill:. min Loop equilibration: min Inject:. min Vials: ml, high shake Vial pressure:. psig for 789A Aux channel Standard solutions of the Class, Class A, and Class B solvents were prepared in pure water according to the USP 7 procedures shown in Table. These stock solutions can be stored for to months at room temperature in a well sealed volumetric. Two grams of sodium sulfate was added to each headspace vial to assist with headspace extraction. Table. Standard Preparation Class solvents.. ml stock solution vial plus 9 ml DMSO diluted to ml.. ml from step diluted to ml with water. ml from step diluted to oo ml with water.. ml step and ml water in ml HS vial Class A solvents.. ml stock solution vial, diluted to ml.. ml from step in ml water in ml HS vial Class B solvents.. ml stock solution vial, diluted to ml.. ml step in ml water in ml HS vial Headspace samples were also prepared for the Class A solvents at other concentrations ranging from about times below USP limit values to times above to demonstrate linearity. Results are shown in Figure. For example, according to USP Procedure A, the limit concentration (in prepared headspace vials) for, dioxane is.7. The concentrations used () for linearity were.9,.7,.9,.7, and 9. in water. Methanol R = Acetonitrile R =.9999 Dichloromethane R =.9998 7 trans-,-dichloroethene R =.9999 Figure. Calibration curves for Class A solvents from approximately times below limit values to times above. (Continued)
8 7 cis-,-dichloroethene R =.9999 THF R =.9997 8 8 Cyclohexane R =.999 Methylcyclohexane R =.999 7 8 7 R =.9999 R =.9999,-Dioxane Chlorobenzene 7 R =.9999 R =.9998 Toluene Ethylbenzene m-xylene, p-xylene o-xylene R =.9999 R =.9998 8 Figure. Calibration curves for Class A solvents from approximately X below limit values to X above.
Discussion In temperature-programmed gas chromatography, which is required for residual solvent analysis, oven cool down time has a major impact on overall cycle time. LTM column modules cool down at considerably faster rates compared to air bath ovens due to their very low thermal mass and cooling fan configuration directly below the column assembly. LTM columns are also capable of much higher temperature programmed ramp rates, which shorten cycle time further. Maximum practical program rates will depend on a number of factors including column dimensions, phase ratio, carrier gas, and the separation required. When translating a conventional air bath method to the LTM format, Agilent Method Translation software can be used to calculate starting conditions. An example is shown in Figure for translating from a standard M column to a 7 M LTM module. LTM program rates ranging from C/min to C/min gave acceptable results in terms of meeting required resolution of specific solvent pairs. Previous work describing the use of LTM technology for a generic set of residual solvents employed a m.8 mm,. µm J&W DB- column. [] A comparison of various column dimensions and program rates are shown in Table. Air bath and LTM methods are included. The table includes entries for the same column (7 M) dimension in air bath and LTM configurations, which allows a valid comparison of overall cycle time. Note that the maximum air oven program rates possible for the 789A V and V GC systems, over the range needed for this application ( C to C), are C/min and C/min, respectively. When comparing against a V 789A GC, the LTM still achieves a % reduction in cycle time. Throughout this work, both LTM columns were controlled from the LTM ChemStation Software add-on module and operated with identical oven programs. However, the LTM columns can each have unique programs that assist with optimization. The only restraint is that both column programs start at the same time. Ending times may be different Figure. Method translation from standard M column to an Agilent J&W DB- 7 M LTM column. See www.agilent.com/chem/methodtranslator to download this tool.
Table. Cycle Times for Various Column and Oven Type Configurations Heating Column Program Cool down Cycle time 789A (V) M. mm. µm C ( min) to C min sec with min 7 min Agilent J&W DB- ( min) at C/min oven equil. 789A () 7 M. mm. µm C ( min) to C 8 min sec with min min Agilent J&W DB- ( min) at C/min* oven equil. sec 789A () 7 M. mm. µm C ( min) to C 8 min sec with min min Agilent J&W DB- ( min) at C/min** oven equil..sec LTM (Fast) 7 M. mm. µm C ( min) to C min sec (one min Agilent J&W DB- ( min) at C/min module system) sec LTM (Faster) 7 M. mm. µm C ( min) to C min sec (one min Agilent J&W DB- ( min) at C/min module system) sec LTM (Fastest) 7 M. mm. µm C ( min) to C min sec (one min Agilent J&W DB- ( min) at C/min module system) sec LTM chromatograms of Class solvents are shown in Figures A and B on Agilent J&W DB- and J&W HP-INNOWax modules, respectively. Resolution between two Class A solvents (acetonitrile and methylene chloride on J&W DB- columns) meets method requirements as shown in Figure. Signal-to-noise ratio's for all Class solvents are greater than at specified limit concentrations. Chromatograms for Class A and B solvents on both J&W DB- and J&W HP-INNOWax phases are shown in Figures and 7. All Class, A, and B solvents combined at limit concentrations are shown in Figure 8. Peak identifications and limit concentrations in prepared headspace vials are shown in Table. Note that operating at C/min yields a cycle time of. minutes. Headspace vial equilibration times were kept at min in this work, following USP 7. However, it should be noted that equivalent results can be obtained with min heating times []. Additional benefits in sensitivity and repeatability are possible using electronic back pressure control of the headspace vial venting (loop fill) process. This is discussed at length in Application Note 989-79EN []. Norm....,-dichloroethene.,,-trichloroethane. Carbon tetrechloride. Benzene.,-dichloroethane. 7 Figure A. Class residual solvents at limit concentration on an Agilent J&W DB- column at C/min program rate.
Norm. 7,..7.7..,-dichloroethene.,,-trichloroethane. CCl. Benzene.,-dichloroethene.9...7...7. min Figure B. Class residual solvents at limit concentration on an Agilent J&W HP-INNOWax column at C/min. Resolution, R =.9 USP requirements, R = or greater. Acetonitrile. Methylene chloride Figure. Acetonitrile/methlyene chloride resolution. 7
Norm. 8 7 7 8. Methanol. Acetonitrile. Methyl cyclohexane. trans,-dichloroethene/thf. cis,-dichloroethene. Tetrahydrofuran 7. Cyclohexane 8. Methylene chloride 9.,-dioxane. Toluene. Chlorobenzene. Ethylbenzene. m-xylene, p-xylene. o-xylene 9 DMSO Agilent J&W DB-, C/min, Class A Figure A. Class A solvents at limit concentration on an Agilent J&W DB- column, C/min. Norm. 7 8 9. Cyclohexane. Methyl cyclohexane. trans,-dichloroethene/thf. Methanol. Methylene chloride. cis,-dichloroethene 7. Acetonitrile 8. Toluene 9.,-dioxane. Ethylbenzene. p-xylene. m-xylene. o-xylene. Chlorobenzene DMSO Resolution, R =. for acetonitrile and cis,-dichloroethene Agilent J&W HP-INNOWAX, C/min, Class A Figure B. Class A solvents at limit concentration on an Agilent J&W HP-INNOWax column, C/min. 8
8 7. Hexane. Nitromethane. Chloroform.,-dimethoxyethane. Trichloroethene. Pyridine 7. -hexanone 8. Tetralin 7 DMSO 8 7 8 Agilent J&W DB-, C/min, Class B Figure 7A. Class B solvents at limit concentration on Agilent J&W DB- column, C/min 7. Hexane.,-dimethoxyethane. Trichloroethene. Chloroform. -hexanone. Nitromethane 7. Pyridine 8. Tetralin 8 7 Agilent J&W HP-INNOWAX, C/min, Class B Figure 7B. Class B solvents at limit concentration on Agilent J&W HP-INNOWax column, C/min. 9
Norm. 7 8 9 8 9 DMSO 7 Class, A, B on an Agilent J&W DB- column at C/min Figure 8. Class, A, and B solvents at limit concentration on Agilent J&W DB- column, C/min. Peak IDs in Table. Table. Peak Numbering for Figure 8 and Actual Headspace Vial Concentrations Class Conc () Class A Conc () Class B Conc ().,-dichloroethene.7. Methanol.. Hexane.8 9.,,-trichloroethane 8.. Acetonitrile. 7. Nitromethane.8 9. Carbon tetrachloride.. Methylene chloride. 8. Chloroform..,-dichloroethane.7. trans-,-dichloroethene 7.8.,-dimethoxyethane.7. Benzene.7 7. cis-,-dichloroethene 7.8. Trichloroethene. 8. Tetrahydrofuran.. Pyridine. Coelutions on DB- 9. Cyclohexane. 7. -hexanone.8 cis-,-dichloroethene, nitromethane. Methylcyclohexane 9.8. Tetralin.7 THF, chloroform.,-dioxane.7 Cyclohexane, CCl,,,-trichloroethane. Toluene 7. Benzene,,-dichloroethane 8. Chlorobenzene. 9. Ethylbenzene.7. m, p-xylene.8. o-xylene.
Conclusions A X overall reduction in cycle time is possible when converting from the standard methodology to a LTM based system for residual solvent analysis. Capillary flow technology can be employed to conveniently analyze on two column phases (Agilent J&W DB- and Agilent J&W HP-INNOWax columns) simultaneously from a single headspace injection. LTM column dimensions of 7M x. mm provide a good compromise among speed, ease-of-use, and capacity while meeting the resolution requirements of USP 7. This general methodology using LTM technology should be particularly attractive to new drug development where variations to the USP procedures are appropriate and fast method optimization is desired. For More Information For more information on our products and services, visit our Web site at www.agilent.com/chem. The Chemstation method integrates Agilent 789A GC/Agilent G888A Headspace, and LTM control through add-on software modules for ease of setup, operation, method integration, and compliance. References. USP -NF 7, General Chapter USP <7> Organic Volatile Impurities, United States Pharmacopeia. Pharmacopoeia Convention Inc., Rockville, MD, 8/9.. International Conference on Harmonization (ICH) of Technical Requirements for the Registration of Pharmaceuticals for Human Use, QC (R): Impurities guideline for residual solvents, Step, July 997.. Joseph M. Levy and Michael Kraft,"Simultaneous Dual Capillary Column Headspace GC With Flame Ionization Confirmation and Quantification According to <USP7>, Agilent Technologies publication 989-88EN, 8.. Frank David, Roman Szucs, Jay Makwans, and Pat Sandra, "Fast Capillary GC using a Low Thermal Mass Column Oven for the Determination of Residual Solvents in Pharmaceuticals," Pfizer Analytical Research Centre, Ghent University, Krijgslann, Ghent, Belgium, J. Sep. Sce., 9, 9-98,.. Roger L. Firor,"The Determination of Residual Solvents in Pharmaceuticals Using the Agilent G888 Network Headspace Sampler," Agilent Technologies publication 989-EN,.. Albert E. Gudat and Roger L. Firor, "Improved Retention Time, Repeatability, and Sensitivity for Analysis of Residual Solvents," Agilent Technologies publication 989-79EN, 7.
www.agilent.com/chem Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc., Printed in the USA January, 99-9EN