ACD/AutoChrom Assisted Method Development for Challenging Separations Vera Leshchinskaya February 7, 2018
Introduction Resolution of isomeric species is one of the most challenging areas of separation science due to the inherent similarity of the molecules and their respective energies. A number of pharmaceutical active ingredients and intermediates contain moieties that may exist as isomers that could potentially contaminate the material. The accepted approach for finding optimal separation conditions initiates with the screening of multiple columns/conditions followed by gradient optimization. Traditional trial-and-error studies to optimize gradient slope for a given separation may require several days. In comparison, computer simulation can deliver the optimized conditions within minutes.
Typical Reversed Phase HPLC Method Development Workflow for Isomeric Separations Search literature/available databases for existing methods Method available? Yes Prepare mixture of markers Test literature method No Prepare mixture of markers Screen generic columns/generic gradient at multiple phs/choose the best column Screen multiple column at best ph/choose the best column Yes Finalize method Yes Method criteria met? No Yes Method criteria met? Further screening of organic modifiers/fine tuning ph No Specialized method development 1. Specialty columns; 2. Chiral columns 3. Chiral GC 4. CE 5. Normal Phase, etc Peak co-elution check (UV and MS) No Method criteria met?` Gradient and/or Temperature optimization through modeling
Typical Instrumentation/Software Use with Generic Approach to Method Development ph screen Manual evaluation of results LC-MS ChemStation Initial condition Generic method Empower Drylab or manual twicking Manual transfer of data from Empower or file conversion Column Screening LC/MS/UV Peak Purity Optimization Meduza/Shimadzu Class VP
Automated Method Development What is ACD/AutoChrom? A multiple software tools in one convenient and fully automated package. Offers calculations of Phys-Chem properties for analytes with known structures prior to MD (pka, LogP). Controls HPLC, UHPLC, and LC/MS instruments - fully automated, needs minimal programming. Spectrus: LC-GC-UV-LC/MS-GC/MS-SPECTROSOPY Processor Intellextract automated peak mapping capabilities based on MS/UV spectra, identifies and matches relevant peaks based on spectral information (MS; UV) thus allowing to see peak distribution and choose the best separation based on user defined criteria and the number of peaks found. ACD/LC & GC Simulator More than.000 chromatograms in Chromatography Application Database offer search for starting conditions for known structures. Guides workflows that facilitate the entire method development process: runs screening/maps peaks based on UV and MS data/provides optimization/modeling/suggests best results
ACD/AutoChrom Method Development Cycle 1. Acquire data 5. Suggest Next Experiment Start Column, Buffer & Solvent Screening Select Best 2. Track peaks between injections Optimize Gradient & Temperature End 4. Acquire data to model separation 3. Generate peak table Gradient Temperature
Generic Approach vs AutoChrom HPLC (Agilent or Waters) Column screening system (Shimadzu) LC-MS Drylab ACD/AutoChrom
Method Development for Cis/Trans isomers and Closely Eluting Impurities Separation: Case Study 1
Background Information N R 1 O [H*] N R 1 OH + N R 1 OH Et 3 SiH N R 1 X N racemate Ar X N Ar ~99%e.r. ~99%d.r cis isomer X N Ar trans isomer X N S-Isomer, >98%e.r. Ar Control of trans-isomer is critical since this step is a quality gate-keeper for API
Best Results by Traditional Column Screening
ACD/AutoChrom: Strategy for Method Development 2 phs, 11 columns Weak Phase Comment (0.05 mm TFA) ph 2 Water/MeCN (95:5) (+ mm AcONH4) ph 6.99 7 Strategy: Origin Column & Buffer Screening Take Next Optimize Gradient Result Experiment Rs Score Total Zorbax ECLIPSE PLUS & ph 2-0/26 ZORBAX SB AQ C18 & ph 2 0.44 13/26 ABZ PLUS SUPELCO, ser# 123990-05 & ph 2-0/26 AQUA SEP ES & ph 2-0/26 ACE C18-AR ser#a79091 & ph 2-0/26 BONUS RP & ph 2-0/26 ACSENTIS EXPRESS C18 & ph 2-0/26 FORTIS DIPHENYL, ser# A05111301 & ph 2-0/26 Phenomenex Aqua C18 & ph 2-0/26 ATLANTIS T3 & ph 2-0/26 Poroshell 120 SB-C18 & ph 2-0/26 Zorbax ECLIPSE PLUS & ph 6.99-0/26 ZORBAX SB AQ C18 & ph 6.99-0/26 ABZ PLUS SUPELCO, ser# 123990-05 & ph 6.99-0/26 AQUA SEP ES & ph 6.99-0/26 ACE C18-AR ser#a79091 & ph 6.99-0/26 BONUS RP & ph 6.99 0.519 20/26 ACSENTIS EXPRESS C18 & ph 6.99-0/26 FORTIS DIPHENYL, ser# A05111301 & ph 6.99-0/26 Phenomenex Aqua C18 & ph 6.99-0/26 ATLANTIS T3 & ph 6.99 0.56 15/26 Poroshell 120 SB-C18 & ph 6.99-0/26
Three Best Columns: Choice for Optimization Response 48 40 32 24 16 8 Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -90B (30 min) Column Name: BONUS RP 290 C_09 260 C_14 294_2 308 294 * * 294_1 296 336 336_1 345 338 ML Sample Name: BMT-082676_ML 48 40 32 0 2 4 6 8 12 14 16 18 Retention Time (min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -90B (30 min) Column Name: ATLANTIS T3 294 294_2 20 336 336_1 22 24 26 28 30 32 ML Sample Name: BMT-082676_ML Response 24 16 * * 308 8 290 260 403 391 C_13 294_1 296 345 338 48 40 32 0 2 4 6 Mobile Phase A: Water; 0.05mM TFA Mobile Phase B: ACN/Water (95:5 v/v); 0.05mM TFA -90B (30 min) Column Name: ZORBAX SB AQ C18 8 12 294 294_2 14 16 18 Retention Time (min) 336_1 336 20 22 24 26 28 30 32 ML Sample Name: BMT-082676_ML Response 24 16 8 290 260 C_09 * * 294_1 308 329 345 296 C_11 0 2 4 6 8 12 14 16 18 Retention Time (min) 20 22 24 26 28 30 32 * Critical pair
Atlantis T3: Separation Predicted by the Software 0.00015 Simulated Chromatogram 294 0.000 336_2 0.00005 336_1 290 260.02 403 294_5 308 447 294_1 296.04 345 338 2 4 6 8 12 14 16 min 18 20 22 24 26 28 30 Based on the modeling data, Atlantis T3 was abandoned since Bonus RP showed better separation for the isomeric pair
Bonus RP Gradinet Optimization Response 24 16 8 Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -90B (30 min) Column Name: BONUS RP 290 260.02 C_19 447 C_20 294 308 294_1 294_5 296.04 336_2 336_1 C_23 C_24 Initial run Sample Name: BMT-082676 R=1.42 RXN MIX Response 15 5 0 2 4 6 8 12 14 16 18 Retention Time (min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate 19-57.8B (30 min) Column Name: BONUS RP 290 260.02 C_23 447 294 308 294_1 294_5 20 296.04 22 24 336_2 C_20 336_1 C_24 26 28 ACD requested 30 32 Sample Name: BMT-082676 R=1.62 RXN MIX Response 15 5 0 2 4 6 8 12 14 16 18 Retention Time (min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -57.9B (36 min) Column Name: BONUS RP 290 C_19 260.02 C_23 447 C_20 294 294_1 352.02 294_5 20 22 296.04 24 26 336_2 336_1 C_24 28 ACD requested 30 32 Sample Name: BMT-082676 R=0 RXN MIX Response 15 5 0 0 5 15 20 Retention Time (min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate 3-30B (6.8 min); 30-56B (26.8 min) Column Name: BONUS RP 25 30 35 Sample Name: BMT-082676 ACD suggested R=1.75 2 4 6 8 12 14 16 18 20 Retention Time (min) Baseline separation of trans-isomer and other impurities form the main peak were achieved. Accurate correlation between modeled and experimental data. 22 24 26 28 30 32 34 36
Method Development for Multiple Diastereomers Mixture Case Study 2
Background Information Develop a method suitable for IPC rxn completion and release of an intermediate (diastereomers separation) preferably in one run; Challenge: multiple chiral centers, no pure markers available, time constrain. Strategy: Negotiate critical pairs ph Evaluation: Run generic method at different phs Further screening No separation Baseline separation Transfer method
Initial screening : ph Evaluation Using Generic Method Day 1 Origin Buffer Screening Response 80 60 40 20 Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -90B (25 min) Column Name: Zorbax ECLIPSE PLUS Sample Name: 98395-042-F4 BMT-123934 363_2 363_1 362.82 362.75 Compound 1 (Product) mixture of all isomers Experiment Take Next Optimize Gradient Result Status Rs Score ph 2 Complete 0.164 Response Response 240 160 80 0 2 4 6 8 12 14 Retention Time (min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -90B (25 min) Column Name: Zorbax ECLIPSE PLUS Sample Name: 98395-040-F2 BMT-129167 0 2 4 6 8 12 14 Retention Time (min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate -90B (25 min) Column Name: Zorbax ECLIPSE PLUS 200 Sample Name: 98395-040-F3 BMT-129167 150 0 50 360.95 360.95 360.75 360.75 16 16 18 342.75 18 20 SM 20 22 22 SM diastereomer marker 24 24 Continue screening since starting material and Compound 1 co-elute ph 6.99 Complete 0.167 0 2 4 6 8 12 14 Retention Time (min) 16 18 20 22 24
Strategy of Method Development Day 1 Strategy for overnight screening Since generic column screening scored better at neutral ph, further column screening was conducted only with NH4Ac. Strategy for Automated Processing Process only promising separations, choose best scored chromatogram for further optimization Column Screening Results Origin Column Screening Take Next Optimize Gradient Result Experiment Status Rs Score Min Rs Zorbax ECLIPSE PLUS Complete 0 - ZORBAX SB AQ C18 Complete 0 - ABZ PLUS SUPELCO Complete 0.680 1.23 AQUA SEP ES Complete 0 - ACE C18-AR Complete 0 - BONUS RP Complete 0 - ACSENTIS EXPRESS C18 Complete 0 - FORTIS DIPHENYL Complete 0 - Phenomenex Aqua C18 Complete 0 - ATLANTIS T3 Complete 0 - Poroshell 120 SB-C18 Complete 0.613 0.165
Building an Accurate model MD Day 2 Origin Column Screening Take Next Optimize Gradient Result RETRO-RXN Compound 1_diast 2 Compound 1_diast 3 SM-diast Compound 1- H20_1 SM Compound 1_diast 1 Compound 1 Compound 1- H20_2 RETRO-RXN Compound 1_diast 2 Compound 1_diast 3 SM-diast Compound 1- H20_1 SM Compound 1_diast 1 Compound 1 Compound 1- H20_2 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5.0.5 11.0 11.5 12.0 12.5 min 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5.0.5 11.0 11.5 12.0 12.5 min
Software Optimized Final Method MD Day 2 150 0 50 ABZ PLUS SUPELCO, ser# 123990-05 RETRO RXN Compound 1_diast 2 Compound 1_diast 3 * ) Gradient: 5-32B (5.6 min); 32-32B (.4 min); 32-90B (9 min) Mobile Phase A: Water/ACN (95:5 v/v); mm Ammonium Acetate Mobile Phase B: Water/ACN (5:95 v/v); mm Ammonium Acetate SM-diast Compound 1-H20_1 SM BMT-123934-diast 1 Compound 1 Compound 1-H20_2 Compound 1_diast 1 Acquired knowledge based on MD: 1. All the peaks were mapped by the software by mass; 2. 3 pairs of isomers were identified; 3. Degradants identified 0 0 2 4 6 8 12 14 16 18 20 Retention Time (min) Baseline separation of all impurities of interest were achieved within 2 days. * ) Separation of diastereomers 2 and 3 of Compound 1 is not critical for the process
The Importance of Software-Assisted Peak Coelution Check in Addressing Root Cause of Reaction Stalling Case Study 3.
Background Information During the course of development for a BMS compound, an improvement to the yield of one of the early intermediates was desired due to an expensive starting material (SM 1). This improvement would not only reduce the cost of the entire synthesis but also increase the throughput. Systematic analysis of in-process reaction samples showed that the reaction stalled at ~80% yield, regardless of process conditions, as both starting materials were not completely consumed and an acid of SM 2 was observed as a major side product. H R 1 N R 2 O OMe R 2 O OH SM 1 SM 2 Side product
Initial Screening results In-process monitoring of the reaction completion was challenging due to the lack of chromophore and highly polar nature of SM1. A generic HPLC method was assessed initially with a combination of Evaporative Light Scattering Detector (ELSD) and UV detection since SM1 could not be detected by UV while SM 2 was not visible by ELSD. Chromatogram of Markers Overlay Hypothesis: the apparent reaction stalling effect is due to an unknown impurity co-eluting with the SM1.
Peak Purity Check with ACD/AutoChrom Software ACD/AutoChrom General Method Development cycle 1. Acquire data Initial Screening Results Gradient used 5. Suggest Next Experiments Start Column, Buffer & Solvent Screening Select Best Optimize Gradient & Temperature End 2. Track peaks between injections based on MS data or /and UV spectra similarity Results of LC/MS column screening Gradient 4. Acquire data to model separation 3. Generate peak table, identify peak co-elution, suggests best separation based on the resolution score Zorbax Eclipse Plus column was assigned the best overall resolution score by the software. Temperature Peak purity was checked based on the results of LC/MS column screening.
Findings UV (Top)+MS ACD/AutoChrom Processed Data (Bottom) Simulated composite chromatogram 480 400 320 240 160 80 0 Zorbax ECLIPSE PLUS 0 2 4 6 SM1 296 355.25 328 422.25 8 Relative Intensity 15 5 3.25 4.5 C_04 C_19 490.25 5.0 Acetonitrile/Water (5:95 v/v); 0.05mM TFA ACN/Water (95:5 v/v); 0.05mM TFA -0B (22 min) C_11 C_11 12 354.23 354.23 Retention Time (min) 188 14 494.25 561.25 467.13 467.13 16 368.25 368.25 C_13 C_13 C_14 C_16 18 C_18 UV *) 20 Retention Time (min) MSD SM+14, 5.29 215.25 min SM 1, SM+29, 5.15.min 5.24 min 201.02 229.25 296 355.25 341.25 328 422.25 436.31 3.25 C_04 C_19 490.25 476.33 504.33 354.23 C_11 188 494.25 467.13 561.25 C_13 368.25 C_14 C_16 509.25 382.22 C_18 + 629.25 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 min While the majority of impurities are baseline resolved, 2 unknowns were identified under the SM1 peak with the difference in [M+H]/retention time 14 Daltons/0.14 min (major peak) and 29 Daltons/0.9 min respectively, indicating peak co-elution in all the columns tested. Composite chromatogram simulated by the software (on the right) shows 3 components identified under the peak at 5.2 min detected by MSD (on the lower left). *) Data were also intended to be used for the release method development, so the worst case scenario reaction mixture with multiple impurities was used for the screening of columns under acidic conditions.
Side Products Structures Proposed Structures R H N Structure Confirmation by Independent Synthesis R H N MeI R CH 3 H3C CH 3 N N + R + SM 1+15 Da R NH 2 N SM 1+13 Da R H H N + Further software-assisted modeling/optimization resulted in partial separation of piperidines on ACE PFP column: Ratio of piperidines in the marker mixture: ~ 20/40/40 Ratio of piperidines in the reaction mixture: ~ 7/83/ 6000000 SM 1+14 Da SM 1+29 Da 5000000 4000000 0 0 2 4 6 min R CH 3 N R H 3 C CH 3 N + 3000000 2000000 00000 SM 1 Methyl-SM 1 Di-Methyl- SM 1 Reaction Mix 0 0 20 Markers of piperidines, mix min The retention times of the peaks in the marker mixture match with those of the peaks identified in the reaction mixture. NMR analysis also confirmed the proposed structures. A reliable IPC method was needed with baseline resolution of all starting material-related species and the product.
Further Method Development: Mixed Mode Chromatography Mixed-mode chromatography adds another dimension to the separation of polar ionizable analytes allowing resolution of compounds based on their differences in both hydrophobicity and pka/pkb. mv 200.00 150.00 0.00 50.00 Product-Me - 2.885 piperidine - 5.658 Methyl-piperidine - 7.804 di-methyl piperidine - 9.856 Primesep A column, 4.6x50mm, 5um 1 ml/min; 55/45/0.05 ACN/H20 TFA, Isocratic, CAD Detector 0.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00.00 11.00 12.00 13.00 14.00 15.00 Minutes All compounds of interest were baseline resolved under isocratic conditions on a Primesep A column, enabling further process optimization. The reaction yield was vastly improved by altering the original starting material SM 2 to an ethyl ester, thus eliminating the undesired alkylation. The initial 80% yield was improved to 94% with a mechanistic understanding of the reaction enabled by the analytical investigation.
Summary Peaks that overlap with the starting material can lead to false results during reactions monitoring, which can hinder process optimization. Software assisted peak co-elution check that includes automated processing of LC/MS data can significantly expedite process development and provide knowledge of generated impurities, as described in this work. Findings obtained during AutoChrom-assited method development allowed to increase the yield and provided better understanding of the reaction mechanism. Isocratic HPLC method that employed mixed mode chromatography and CAD was developed and qualified enabling successful process development and DOE study followed by the transfer to an outside vendor.
Conclusions Using AutoChrom software in tandem with Agilent LC/MS : 1. Significantly simplifies Method Screening - runs column, buffer, and solvent screening experiments, automatically finds and tracks all peaks in your samples, and selects the best result. 2. Assists in Method Optimization - guides through data processing and method optimization, helps to avoid the experiments that are unlikely to work, and develops high quality separation methods. 3. Tracks peaks from run-to-run based on UV or MS spectral similarity, finds all trace components to avoid missed peaks. 4. Build a model based on peak tracking. Offer a choice of mathematical equations for more accurate model 5. Manages projects allows to see a clear overview of your experiments, and dig down into the original data when necessary. Experiments are summarized in a peak table as you work. 6. Provides knowledge about process impurities. 7. Assists in peak co-elution check.
Scheme of Automated Method Development for Difficult Separations *