Method Development with Modern HPLC Phases

Method Development with Modern HPLC Phases Frank Michel, Sigma-Aldrich Germany R. A. Henry, W. K. Way, C. T. Santasania, D. S. Bell Supelco/Sigma-Aldrich, USA

Three Factors Control HPLC and UHPLC Resolution 9 R s = N k k+1 4 α-1 α 3.0 2.5 α N = 16 (t R /w) 2 or N = 5.5 (t R /δ) 2 k = (t R t 0 )/t 0 α = k 2 /k 1 Resolution (R) 2.0 1.5 1.0 0.5 N k All factors are important, but selectivity is considered the most powerful term. 0.0 1.00 1.05 1.10 1.15 1.20 1.25 0 5K 10K 15K 20K 25K 0 5 10 15 20 25 9. Yun Mao, PhD Dissertation, University of Minnesota, 2001. 2 α N k

Selectivity Variables in Reversed-Phase * Continuous variables (solvent): type (organic, water) solvent strength ph (especially ionizable solutes) additives (type and concentration) temperature Discontinuous variable (column): type (phase and substrate) More predictable (modeling software available) Some analysts may spend too much time here force feeding C18 columns. Less predictable (screening required) * Excerpted with permission from John Dolan, 2009 Minnesota Chromatography Forum Spring Symposium; adapted by R. Henry. 3

Alkyl Bonded Phases (C18 and C8) C18 reagents are large and can leave some silanols unreacted (ca. 50%). C8 reagents are smaller and provide better silanol coverage At ph >4, silanols can ionize and add cationexchange character. Primary phase reagent Hydrophobic selectivity Endcapping reagent R Free silanol Si Si OH O OH O 6

Alkyl Amide Bonded Phases (abbreviated: RP-Amide) Possible Solute Shielding (basic solutes)- alkyl phases with embedded polar groups (EPG) have been reported to correlate much better with logp data than C18 columns due their higher base deactivation 1. Possible H-bonding with solutes that are good H- bond donors (acids, etc.) OH Endcapping- Reagent Free Silanol groups Si O C 9 H 19 NC=O OH Hydrophobic OH-Backbinding Amide-Carbonyl: strong H-Bonding acceptor O Si C 9 H 19 HN O Si H-Bonding C=O 7

Phenyl Bonded Phases Phenyl is a Lewis base or electron donor; π-π interaction can occur with solutes that are deficient in electrons (Lewis acids). Due to the rigid nature of the aromatic ring, solute shape can dictate selectivity (how closely solutes can approach the ring). Free silanol Si π base Endcapping reagent Primary phase reagent Si OH O OH O 8

Pentafluorophenyl (F5) Bonded Phases PFP is a Lewis acid or electron acceptor; π-π interaction can occur with solutes that are rich in electrons (Lewis bases). Due to the rigid aromatic ring, solute shape can dictate selectivity. F F π acid and dipole (H-bond) Endcapping reagent F F F Primary phase reagent Dipolar (possible H-bond) interactions can occur Si Si Free silanol OH O OH O 9

Classification by Possible Chemical Interactions a Bonded Phase Hydrophobic H-Bonding Dipolar π-π Steric b Ionic b C18 Very Strong Weak No No No Moderate C8 Strong Weak No No No Weak RP- Amide Strong Strong Acceptor Phenyl Strong Weak Acceptor F5 or PFP Cyano Moderate Light to Moderate Moderate Acceptor Weak Acceptor Moderate No Weak Very weak Weak Strong Strong Donor Strong Acceptor Strong (Rigid) Strong (Rigid) Weak Moderate Strong Weak No Moderate a. Using Euerby 2 variation of Snyder-Dolan-Carr Hydrophobic Subtraction Model 3. b. Steric and Ionic probe data are not very helpful in predicting or interpreting steroid selectivity results; however, they are always underlying factors with silica bonded phases. 10

Visualization of Selectivity Determination of retention factors of different compounds at different chromatographic conditions Diagram of logarithmized retention factors of different conditions (kappakappa plot) 11

Columns with Same or Similar Phases Should Show High Correlation: C18 vs C8 R 2 = 0.97 13mM ammonium acetate, 30:70 H20:ACN, ph 6.91, 50 solutes (acid, base, neutral) 12

Selectivity Stationary Phases: RP-Amide C18 1,1 0,9 0,7 log k RP-Amide 0,5 0,3 0,1-0,8-0,6-0,4-0,2-0,1 0,0 0,2 0,4 0,6 0,8 1,0-0,3-0,5 log k (C18) 13

Columns with Different Phases Should Show Orthogonality: C18 vs Phenyl R 2 = 0.41 13mM ammonium acetate, 30:70 H 2 0:ACN, ph 6.91, 50 solutes (acid, base, neutral) 14

Columns with Different Phases Should Show Orthogonality: C18 vs F5 (PFP) R 2 = 0.63 13mM ammonium acetate, 30:70 H 2 0:ACN, ph 6.91, 50 solutes (acid, base, neutral) 15

Comparison of Selectivity Hydrophobic Retention 1 1 2 4 4 3 Retention 3 2 1 Pentafluorophenyl, F5 1 3 2 1 2 4 Phenyl 4 2 3 C8 3 4 Dispersive Interactions mainly drive the separation of hydrophobic molecules. Retention of butyl benzene and amyl benzene (4) directly depends on dispersive interaction with the stationary phase. Additional selectivity by steric impact (2, 3). RP-Amide C18 1. Butylbenzene 2. o-terphenyl 3. Triphenylene 4. Amylbenzene 15x4.6 5um; 65% ACN 35% water 1.5mL/min, 220nm, 5uL 16

Comparison of Selectivity Nitro compounds π π interactions NB NB TNB 2,4-DNT 3NT 2 & 4NT TNB C8 RP-Amide Pentafluorophenyl 2,4-DNT Phenyl Nitro groups highly polar and electron drawing; aromatic compounds are a strong π acid Pentafluorophenyl, F5 Phase is a strong π acid. Selectivity is similar to C18, but with increased retention due to polar interactions Phenyl π-π interactions change both selectivity and retention. Di- and trinitro compounds are retained most strongly. C8, RP-Amide & C18 Low hydrophilic retention NB TNB 2,4-DNT 17 C18 0 10 20 15x4.6; 5um; 50% Methanol 50% water; 1.0mL; 210nm; 40C; 10uL

Amide Hydrogen-Bonding Selectivity Example Ascentis Express C18 Phenols and other Natural Antioxidants RP-Amide retention enhanced vs C18 likely due to amide hydrogen bonding 1 3 2 4 6 5 7 Phenols (H-bond donors): 1. p-methoxyphenol 2. phenol 3. p-cyanophenol 4. p-fluorophenol 5. p-methylphenol 6. p-nitrophenol 7. p-chlorophenol Ascentis Express 10cm x 4.6mm, 35:65 ACN:H 2 O, 0.2% formic acid, 1.8 ml/min, 230nm, 35 C, 5uL. 1 2 3 4 5 6 7 Ascentis Express RP-Amide 1.0 2.0 3.0 Time (min) 18

Fast Method Development: Using Selectivity Solv. A Solv. B HPLC- Pump Injector Solv. C Solv. D PC Det. C18 C8 RP-Amide F5 Phenyl Ascentis C18 Ascentis C8 Ascentis RP -Amide Discovery HS F5 Ascentis Phenyl Column Selector 19

Fast Method Development: Using Selectivity Solv. A Solv. B HPLC- Pump Injector Solv. C Solv. D PC Det. C18 C8 RP-Amide F5 Phenyl Ascentis C18 Ascentis C8 Ascentis RP -Amide Discovery HS F5 Ascentis Phenyl Column Selector 20

Application Development at Customer: Drug for Flu/cold (SmithKline Beecham) Compound mg/tablet Acetaminophen 500 Dextromethorphan HBr 15 Phenylpropanolamine HCl 12.5 Chlorpheniramine Maleate 2 A: 0.1% HCO2H/Water (ph ~2.5) B: ACN Gradient: 0 to 35% B in 2.5 min Flow: 4mL/min Temp: 35º C UV: 254nm O O Cl NH H NH 2 N N OH OH Acetaminophen Dextromethorphan Phenylpropanolamine Chlorpheniramine

Fast generic Gradients 22

Experimental Design Five different stationary phases on Fused Core particles Ascentis Express C18 Ascentis Express C8 Ascentis Express RP-Amide Ascentis Express F5 (Pentafluorophenyl) Ascentis Express Phenyl-Hexyl have been tested with two mobile phases consisting of either Methanol/water Acetonitrile/water as organic modifier on different herbal drugs such as Withania, Panax Ginseng and Milk Thistle 23

Whitania potential marker compounds, C18 methanol/water peak qty = 7 min resl = 1.28 Withanone Withanolide A Withanolide B Withaferin A 12-Deoxywithastramonolide Withanoside IV Withanoside V acetonitrile/water peak qty = 7 min resl = 0.76 10.0 11.0 Time (min) Ascentis Express C18 7.0 8.0 9.0 10.0 Time (min) 24

Whitania potential marker compounds, PFP methanol/water peak qty = 6 min resl = 1.40 Withanone Withanolide A Withanolide B Withaferin A 12-Deoxywithastra- monolide Withanoside IV Withanoside V 10.0 11.0 12.0 Time (min) Ascentis Express F5 acetonitrile/water peak qty = 7 min resl = 2.76 6.0 7.0 8.0 9.0 Time (min) 25

Withania results Sorted by peak quantity then min. resolution phase organic peaks min resl F5 acetonitrile 7 2,76 Phenyl-Hexyl acetonitrile 7 2,01 Phenyl-Hexyl methanol 7 1,36 C18 methanol 7 1,28 C8 acetonitrile 7 0,98 C18 acetonitrile 7 0,76 RP-Amide acetonitrile 7 0,46 C8 methanol 6 2,55 F5 methanol 6 1,40 RP-Amide methanol 5 0,90 26

Panax ginseng results Sorted by peak quantity then min. resolution phase organic peak qty min resl C18 methanol 7 0,99 C8 methanol 7 0,94 Phenyl-Hexyl methanol 7 0,77 F5 methanol 7 0,49 RP-Amide methanol 6 1,63 RP-Amide acetonitrile 6 1,39 C8 acetonitrile 6 1,24 C18 acetonitrile 6 0,71 F5 acetonitrile 6 0,58 Phenyl-Hexyl acetonitrile 5 1,94 27

Extract of Chinese Panax Ginseng Column: Ascentis Express C18, 150 x 4.6 mm Mob. Phase A: water Mob. Phase B: acetonitrile Gradient: 0-1.5 min at 25% B, 1.5-13.5 min to 85% B, 13.5-14.5 min at 85% B Flow rate: 1.5 ml/min Temp.: 60 oc Det.: 205 nm Injection: 10 µl Sample: extract of Chinese Ginseng with water/ethanol (50:50, v/v) Ginsenoside Re Ginsenoside Rg1 Ginsenoside Rf Ginsenoside Rb1 Ginsenoside Rc Ginsenoside Rb2 Ginsenoside Rd 0 2 4 6 8 Time (min) 28

Milk Thistle (Silybum marianum) results Sorted by peak quantity then min. resolution phase organic peak qty min resl C18 methanol 9 - F5 methanol 8 - C18 acetonitrile 8 - Phenyl-Hexyl methanol 7 - Phenyl-Hexyl acetonitrile 7 - F5 acetonitrile 7 - C8 methanol 7 - C8 acetonitrile 7 - RP-Amide methanol 6 - RP-Amide acetonitrile 6-29

Milk Thistle (Silybum marianum) Markers (red) and Herbal Medicinal Product (black) taxifolin silychristin apigen 7- glucoside silydianin Milk Thistle Xtra 250 mg, extracted w/ water:ethanol 50:50 quercitrin silibin A silibin B isosilibin A isosilibin B Column: Ascentis Express C18, 100 x 3 mm A: MeOH B: Water + 0,1 % FA Gradient: 0-3 min at 35% B, 3-13 min to 45% B, 13-15 min at 45% B Flow rate: 0.6 ml/min. Temp.: 35 o C Det.: 254 nm 0 10 20 Time (min) 30

Conclusions Column phases having optimum retention and selectivity should be identified by fast screening experiments before serious mobile phase optimization begins. Selectivity has highest impact on resolution Stationary phase impacts selectivity most strongly Early screening of stationary phases with highest diversity Subsequent method optimization by mobile phase Use of dedicated software Ascentis Express columns are excellent choices for developing selective, high speed separations. Broad range of very different selectivities savailable (α). Higher efficiency observed at lower pressure vs porous packings; comparable efficiency to sub-2 µm particles (N). 31

Selectivity References 1. D. Benhaim and E. Grushka, Amide Phase for LogP Values, JCA, 1217 (2010) 65-74. 2. M. R. Euerby, et. al., Classification of Phenyl Columns, JCA, 1154 (2007), 138-151. 3. L.R. Snyder, J.W. Dolan and P.W. Carr, Hydrophobic Subtraction Model for Classification of Reversed-Phase Columns, JCA, 1060 (2004), 77. 4. J. W. Dolan and L. R. Snyder, Selecting an Orthogonal Column, JCA,1216 (2009), 3467-3472. 5. M. R. Schure, et. al., Molecular Level Comparison of Alkyl and Polar- Embedded Systems, Anal. Chem. 2008, 80, 6214-6221. 6. M. Yang, et. al., Impact of Methanol and Acetonitrile on Phenyl Selectivity, JCA, 1097 (2005), 124-129. 7. D. H. Marchand, et. al., Phenyl Column Selectivity, JCA, 1062 (2005) 65. 8. Y. Kazakevitch, et. al., Surface Studies of Phenyl Modified Adsorbents, JCA, 1082 (2005), 158-165. 9. Mao, Y., Selectivity Optimization in Liquid Chromatography Using the Thermally Tuned Tandem Column (T3C) Concept, Ph.D. Thesis (P. Carr), The University of Minnesota, 2001. 32

Dziękuję za uwagę! 33