DEVELOPING HPLC METHODS WHEN C18 COLUMNS DON T WORK

Transcription

1 DEVELPING HPLC METHDS WHEN C8 CLUMNS DN T WRK Richard A. Henry, Technical Consultant, 98 Greenbriar Drive, State College, PA 680 (hyperlc@comcast.net) Stephanie A. Schuster, William L. Johnson, and Conner W. McHale Advanced Materials Technology, Inc. Silverside Road, Suite -K, Quillen Bldg. Wilmington, DE 980 Thomas J. Waeghe, MAC-MD Analytical, Inc. 0 Commons Court Chadds Ford, PA 97 Contact Dr. Stephanie Schuster for information (SSchuster@advanced-materials-tech.com) Pittcon 07 RAH When C8 Columns Don't Work

2 CHSING PHASES THER THAN C8 Why choose a C8 phase? Very versatile and retentive for RPLC with simple binary gradients. Proven stability with wide range of mobile phase and additives. Low drift and bleed with LC-MS and other detectors. Past success with C8 columns. When to try another phase chemistry? Isomers or very similar analytes can t be resolved. Too much retention or selectivity with C8 for desired analysis time. Polar analytes not well retained with low or no organic modifier. Polar analytes not well resolved even if retained. A C8 method already in use is not rugged enough (revalidate). Multiple column and solvent screening is recommended for resolving difficult mixtures and achieving rugged system performance. Pittcon 07 RAH When C8 Columns Don't Work

3 FACTRS CNTRLLING HPLC/UHPLC RESLUTIN* R s = N k k+ 4 α- α.0. N = 6 (t R /w b ) or N =. (t R /w / ) k = (t R t 0 )/t 0 α = k /k Resolution (R) N k All factors are important, but selectivity is considered the most effective term to change. Pittcon 07 RAH K 0K K 0K K * Yun Mao, PhD Dissertation, Prof. Peter Carr, University of Minnesota, 00. When C8 Columns Don't Work N k

4 HAL PHASES FR SMALL MLECULE SEPARATIN Phase USP Particle Sizes ( m) Structure C8 L.7 C8 L7.7 Industry Big Five Phenyl-Hexyl RP-Amide L L PFP L4.7 ES-CN L0.7 Phases For Special Situations HILIC (silica) L.7 Penta-HILIC L9.7

5 FEATURES AND APPLICATINS F HAL PHASES HAL Column Features and Benefits Best Applications C8, C8 Excellent performance for broad range of analyte polarities Phenyl-Hexyl Complementary selectivity to alkyl phases Enhanced selectivity for aromatic compounds Analytes differing by an aliphatic or aromatic group Aromatic molecules with electron-withdrawing groups (N, CH, CR, halogens), heterocycles, benzodiazepines, highly aqueous conditions RP-Amide PFP ES-CN HILIC Penta-HILIC Complementary selectivity to alkyl phases Enhanced stability for minimum bleed and long life Complementary selectivity to alkyl phases Enhanced selectivity for stereoisomers Can be used in RPLC and HILIC modes Complementary selectivity to alkyl phases More retention for polar analytes and much less retention for non-polar analytes Can be used in HILIC and normal-phase modes Ideal for separation of highly polar compounds that are not retained in RPLC Acidic and basic analytes, heterocycles, proton donors and acceptors, highly aqueous conditions Basic analytes at low ph, stereoisomers, steroids, taxanes, substituted aromatics, highly aqueous conditions Aromatic molecules with electron-withdrawing groups (N, CH, CR, halogens), heterocycles, benzodiazepines, highly aqueous conditions Enhanced sensitivity and peak shape for LC-MS, analyses of basic analytes in HILIC mode, normalphase analysis with non-aqueous mobile phases Polar acids, bases and zwitterions that are not retained or are poorly retained with RPLC Pittcon 07 RAH When C8 Columns Don't Work

6 VARIABLES THAT AFFECT SELECTIVITY FR RPLC METHD DEVELPMENT Column & Instrument System Column (dictated by method) Stationary phase Particle morphology Particle composition Particle pore size Instrument (not dictated) Delay/Dwell volume Methods will be easier to reproduce between labs if selectivity comes from the stationary phase with simple mobile phases. perational variables contribute more to error because they are cumulative. Keep mobile phases as simple as possible. perational (Mobile Phase) Mobile phase ph rganic modifier choice CH CN, CH H, CH CN/CH H blend, IPA, etc. % rganic (isocratic) or gradient slope (gradient) Column temperature affects ph and ionization Additive type TFA, H P 4, phosphate, formate, acetate, NH 4 H, etc. Additive concentration ionic strength, ionization suppression Pittcon 07 RAH When C8 Columns Don't Work 6

7 RPLC SELECTIVITY RANKING VS HAL C8 AT DIFFERENT PH VALUES - Phases Most Different from HAL C8 Most similar Most different ph.8 ph 7 C8 C8 Phenyl-Hexyl PFP ES-CN Phenyl-Hexyl RP-Amide ES-CN PFP RP-Amide ph impacts charge on solutes that have acid-base properties and can also affect silica substrates. Silanol groups that are not covalently bonded will have neutral charge at low ph and negative charge at ph 7. Column rankings will also differ greatly between CH CN and CH H and be very analyte dependent. Pittcon 07 RAH When C8 Columns Don't Work 7

8 PHASE CMPARISN FR CARBXYLIC ACID SELECTIVITY mau mau mau H NH 7,4,6 8 HAL C H H HAL RP-Amide ,4,7 N CH H F 6 H 8 HAL Phenyl-Hexyl HAL columns, 4.6 x 0 mm,.7 µm :7 ACN/water with 0.% formic acid, C,.0 ml/min H H 9 9 F H C H 8 9. uracil. p-aminobenzoic acid. acetylsalicylic acid 4. -fluorobenzoic acid. benzoic acid 6. phenoxyacetic acid 7. -cyanobenzoic acid 8. -fluorobenzoic acid 9. m-toluic acid H-bond acceptor (carbonyl) of HAL RP- Amide group attracts H-bond donor (acidic H) and imparts more retention and better selectivity than either C8 or phenyl. Pittcon 07 RAH When C8 Columns Don't Work 8

9 Critical pair FAST SEPARATIN F ANTICAGULANTS N HAL FUSED-CRE PACKINGS 4 Extra retention of HAL RP- Amide does not always guarantee the best selectivity 6 7 HAL C8 i Test Conditions: Column size: 4.6 x 0 mm Mobile phase: A/B--40/60 A=0.% formic acid, ph=.6 B=Methanol/Acetonitrile-0/0 Flow=.0 ml/min, 0 Bar, 4 C UV=4 nm, µl injection Absorbance Critical pair 7 4 HAL RP-Amide i HAL Phenyl-Hexyl i Peak Identities:. Uracil. 4-Hydroxycoumarin. Coumarin 4. 6-Chloro-4-hydroxycoumarin. Warfarin 6. Coumatetralyl 7. Coumachlor i=impurity Minutes Pittcon 07 RAH When C8 Columns Don't Work 9

10 PHASE CMPARISN FR BENZDIAZEPINES Test Conditions: Column size: 4.6 x 0 mm A = mm Ammonium Acetate B = Acetonitrile Flow rate =. ml/min. Gradient = 4 6 %B in. min. Pressure = 00 bar Temperature = C UV=4 nm, µl Injection Peak Identities:. xazepam. Lorazepam. Nitrazepam 4. Alprazolam 4. Clonazepam 6. Temazepam 7. Flunitrazepam 8. Diazepam HAL Phenyl phases are -bases (electron-rich) and attract heterocyclic rings and other rings that are acids (electron-poor) by shape-selective - interaction. HAL Phenyl-Hexyl Absorbance HAL C8 4,,6 7 8 HAL RP-Amide 4,6 7.. Time (min.). 4 Pittcon 07 RAH When C8 Columns Don't Work 0 8 HAL PFP

11 CYAN PHASES FFER ALTERNATIVE SELECTIVITY AND INCREASED RETENTIN FR PLAR ANALYTES Absorbance i 4 i 4 i 6 8 HAL columns, 4.6 x 00 mm,.7 µm. ml/min, 0 C, Detection: UV at 4 nm HAL C8 CH H/water 60: HAL ES-CN CH H/water 0:0 PEAK IDENTITIES:. Resorcinol. Benzyl alcohol. Phenylacetonitrile 4. -Indanol.,4-Dinitrotoluene 6.,-Dinitrotoluene 7.,4-Dinitrotoluene 8. Anisole 9. -Chloro-4-nitrobenzene 0. Toluene i = impurity Minutes HAL Cyano phases interact by mild hydrophobic attraction plus dipoledipole and dipole-aromatic forces and should be more popular in methods. Cyano requires only 0% less organic for equivalent retention, is very orthogonal to C8 for polar compounds, and is quite stable. Pittcon 07 RAH When C8 Columns Don't Work

12 HAL PFP FASTER SEPARATIN AND BETTER RESLUTIN USING HAL PFP CLUMN VS. HAL C8 0 4, Time, (min) HAL C Time, (min) 6 7 6,7 8 8 PEAK IDENTITIES:. δ-tocotrienol. β-tocotrienol. γ-tocotrienol 4. α-tocotrienol. δ-tocopherol 6. β-tocopherol 7. γ-tocopherol 8. α-tocopherol Tocotrienol Tocopherol HAL PFP phases are -acids and can attract solutes by - and several other mechanisms. HAL C8 has excess retention and selectivity overall, but cannot distinguish between isomer peaks, and 6,7. HAL columns, 4.6 x 0 mm,.7 µm 90:0 CH H/water, C,. ml/min, Detection: Fluorescence: Excitation, 96 nm; Emission, nm Pittcon 07 RAH When C8 Columns Don't Work

13 Nucleosides and nucleobases are often used as void markers for C8 columns, but they can be easily retained by Halo Penta-HILIC. Nucleosides, glycosides and other sugar-related compounds are ideal for Halo HILIC and Penta-HILIC. If RP and HILIC both show resolution, compare peak shape, equilibration, loading, and other factors. PHASE CMPARISN USING HILIC MDE, Test Conditions: Column = 4.6 x 0 mm Mobile Phase = 90/0 Acetonitrile/ 0. M ammonium formate buffer, ph =.0 Flow rate =.8 ml/min Temperature = 0 C UV=4 nm, µl Injection Peak Identities:. acenaphthene HAL C8. adenosine. cytosine adenosine Time, min. HAL Penta-HILIC Time, min. Pittcon 07 RAH When C8 Columns Don't Work

14 DEVELPING AN RPLC SEPARATIN METHD bjective: to develop a good separation with the desired R s and the necessary robustness for the desired purpose Method needs vary from relatively simple to very demanding Simple: one lab for only several days or weeks Moderate: one lab for several months Demanding: method applied in multiple labs country-wide or worldwide and used for -0 years or more Sample complexity (known analytes) impacts method development time < -6 < 0 > 0 Robustness and ruggedness required for demanding methods Robust- insensitive to small changes in variables Rugged- reproducible during use by many labs, instruments, analysts, etc. Pittcon 07 RAH When C8 Columns Don't Work 4

15 METHD DEVELPMENT CAN VARY FRM RELATIVELY SIMPLE T QUITE CMPLEX Separations that are simple often need little or no column screening so you can develop a separation quickly by varying solvent strength at low ph. Separations that are more difficult may not require a different phase, but you should compare different organic modifiers at multiple ph values. Demanding methods and more complex samples (e.g., stability-indicating method), usually require screening multiple phases, multiple organic modifiers, and several ph values to generate the most rugged separation. C8 CH CN ph.8 Broad Range Gradients 9% Fine Tune Gradient or Isocratic C8 RPA ES-CN CH CN CH H ph.8 ph.8 ph 4.8 Broad Range Gradients 9% Select best combo of phase and modifier ptimize ph ptimize gradient slope and temperature Pittcon 07 RAH When C8 Columns Don't Work

16 APPLICATIN F MULTIPLE PHASES FR STABILITY INDICATING METHD DEVELPMENT Atorvastatin Calcium 0 mg active/0 mg tablet Generate HCl-degraded and NaHdegraded samples Pool acid- and base-treated samples together Compare five different HAL phases using both CH CN and CH H at one ph (.8, ammonium formate) Compare results and identify best option(s) for further development and optimization Use x 0 mm,.7 µm HAL geometry Initially screened HAL C8 column using broad gradient with CH CN (locate primary drug) Fine tuned to narrower ranges Compared all phases using narrower range using both CH CN and CH H Pittcon 07 RAH When C8 Columns Don't Work 6

17 BRAD RANGE GRADIENT WITH C8 IS USEFUL T ESTABLISH HYDRPHBICITY F CMPLEX SAMPLES Mixture of acid- and base-treated atorvastatin calcium x 0 mm,.7 µm HAL C8 90% CH CN/pH.8 0 mm NH 4 CH min initial hold, 0 min gradient Initial screen with HAL C % CH CN/pH.8 0 mm NH 4 CH min initial hold, 0 min gradient Raise starting % organic modifier % CH CN/pH.8 0 mm NH 4 CH min initial hold, 0 min gradient Truncate both ends of gradient Pittcon 07 RAH When C8 Columns Don't Work 7

18 ATRV 0-7% CH CN ATRV 60-90% CH H # Peaks HAL C8 HAL C HAL Phenyl-Hexyl HAL Phenyl-Hexyl HAL RP-Amide HAL RP-Amide 6 4 HAL ES-CN 4 6 HAL ES-CN HAL PFP HAL PFP Pittcon 07 RAH When C8 Columns Don't Work 8

19 HW D YU CHSE WHICH CMBINATIN T DEVELP AND PTIMIZE FURTHER? Compare chromatograms for number of peaks observed. Compare shapes for detected peaks. Select most attractive phase/modifier combination(s) on points of merit: # peaks separated minimum R s for critical peak pair shortest analysis time most peaks with acceptable USP T f If no clear winner emerges, run several gradients having differing slopes. For example, 0 minutes and minutes for Phenyl-Hexyl (0 7% ACN) and C8 (60-90% CH H). Assess whether either combination stands out vs. criteria Compare separation and pressure on longer columns with higher efficiency. Check location of impurities for easier integration. HAL Phenyl-Hexyl HAL C8 0 7% CH CN 60-90% CH H CH CN CH H Pittcon 07 RAH When C8 Columns Don't Work 9

20 SUMMARY AND RECMMENDATINS C8 columns are widely preferred in HPLC methods and may be satisfactory for most simple methods, but a C8 phase may not be the best solution for complex samples in demanding methods. Polar analytes can be separated better on phases that have greater polarity than C8, but selection of preferred phases and operating conditions to resolve complex mixtures may require column screening. A development strategy for demanding situations should be holistic and document all system parameters that control selectivity. This is the reason behind testing method variables for robust behavior before the method is validated. A robust method is likely to be rugged in use. The best approach for HPLC method development is to systematically screen multiple stationary phases and operating parameters. Special software and instrumentation (column, solvent selection) is available to assist with automated screening, selection, and optimization. Shortcuts can lead to fragile methods. Pittcon 07 RAH When C8 Columns Don't Work 0

21 REFERENCES AND ACKNWLEDGEMENTS. Adapted from book, Introduction to Modern Liquid Chromatography, rd Edition, L. R. Snyder, J. J. Kirkland and J. W. Dolan, 00, p.9, Wiley & Sons.. rthogonal separations for reversed-phase liquid chromatography; J. Pellett, P. Lukulay, Y. Mao, W. Bowen, R. Reed, M. Ma, R.C. Munger, J.W. Dolan, L. Wrisley, K. Medwid, N.P. Toltl, C.C. Chan, M. Skibic, K. Biswas, K. A. Wells, and L. R. Snyder; Journal of Chromatography A, 0 (006).. Column selectivity in reversed-phase liquid chromatography I. A general quantitative relationship; N.S. Wilson, M.D. Nelson, J.W. Dolan, L.R. Snyder, R.G. Wolcott and P.W. Carr; Journal of Chromatography A, 96 (00) Column selectivity in reversed-phase liquid chromatography IV. Type-B alkyl-silica columns; J. J. Gilroy, J. W.. Dolan and L. R. Snyder; Journal of Chromatography A, 000 (00) ( ColumnMatch ). 8. D.V. McCalley and U.D. Neue, J. Chromatogr. A 9, 9 (008). 9. A.J. Alpert. Anal. Chem. 80, 6 76 (008). 0. D.V. McCalley, J. Chromatogr. A 7, 46 6 (007).. A.J. Alpert et al., Anal. Chem. 8, 9 (00). Acknowledgements: Thanks to Dr. Joseph DeStefano of AMT, Inc. and to Dr. Mark Schure of Kroungold Analytical, Inc. for constructive comments and edits. HAL and Fused-Core are registered trademarks of Advanced Materials Technology, Inc. Pittcon 07 RAH When C8 Columns Don't Work