Bruce Quimby Application Scientist Agilent Technologies October 28, 2009 Page 1
Tools For Improving GC/MS Analysis Retention Time Locking (RTL) Method Translation Synchronous SIM/Scan Column effluent split to an NPD Column backflushing Deconvolution Reporting Software (DRS) Page 2
Retention Time Locking Get precisely the same retention times: After cutting or replacing column On all instruments in a lab and other labs (method sharing) Greatly reduces method maintenance for updating: peak recognition (RT) windows SIM group start times integration events Reduces data review time Makes management of large screening databases much easier Simple procedure restores all RTs to original calibrated values Process moves RTs back to calibrated values, does not change calibrated value to current (different) RT Look for new targets in archived data, since RTs always the same Page 3
How Does RTL Work? For each method: A set of five runs of retention time vs inlet pressure for a single locking compound is collected (only once per method) The calibration is stored with the method When locking a new instrument, the locking compound is run and its RT is measured RTL software calculates new inlet pressure that makes the RT of the locking compound and all analytes precisely match that of the original method U pdating RTs of individual cal compounds and timed events unnecessary If a table of hundreds of RTs is collected under RTL conditions, anyone anywhere can lock to the same conditions and get the same RTs. This is basis of Agilent screening databases for Forensic Tox compounds, pesticides, etc. Page 4
RTL Example RTL keeps the retention times of all analytes typically within 0.030 min absolute. Locking is a simple procedure using 1 compound. After pressure adjustment, all compounds fall within their RT recognition window. 4.296 min. Initial run 4.72 psi Original locked method 4.064 min. 4.297 min. Trim 1 meter 4.72 psi Relock 4.42 psi Column trimmed, locking compound run at original pressure (RT is too short) RTL calculates new inlet pressure to restore RTs Page 5
Agilent s Method Translation Software A technique invented by Agilent that calculates the inlet pressure and oven temperature program for changing: Analysis speed (example: 4 x faster) Column dimensions (with same phase ratio) Carrier gas type (example: Changing from He to H 2 ) Column outlet pressure (MSD vs NPD vs CFT Splitter) Determines the required conditions with a calculator instead of trial and error GC runs in the laboratory Key benefits: Relative elution order is maintained. Retention times with new method are accurately predicted Substantial saving in method development time, especially for screening methods with large numbers of analytes. Page 6
Speed Scaling Constant Pressure RTL Methods Allows method to be adjusted to optimal speed on a specific system. Use MTL software to calculate changes in pressure and temperature ramps corresponding to desired speed gain Generate new RTL calibration for the new method Peaks will elute in same relative order as before Lock system using target t compound to: (original target time) (speed gain) Now all retention ti times in new method will be equal to: original RT speed gain Page 7
1-5 ng Test Mix: DB-5MS, 1x MSD Direct 10C/min and 1.8 ml/min initial flow 1 Amphetamine 15 Lorazepam 2 Phentermine 16 Diazepam 3 Methamphetamine 17 Hydrocodone 4 Nicotine 18 Tetrahydrocannabinol 5 Methylenedioxyamphetamine(MDA) 19 Oxycodone 6 Methylenedioxymethamphetamine(MDMA) 20 Temazepam 7 Methylenedioxyethylamphetamine 21 Flunitrazepam 8 Meperidine 22 Diacetylmorphine 9 Phencyclidine 23 Nitrazepam 17 10 Methadone 24 Clonazepam 11 Cocaine 9 25 Alprazolam 12 SKF-525a (RTL Compound) 26 Lysergide (LSD) 13 Oxazepam 10 27 Strychnine 16 14 Codeine 18 28 Trazodone 1 2 3 4 5 6 7 8 11 12 13 14 15 22 21 19 25 27 20 23 24 26 28 4 8 12 16 20 24 28 TIC: 25mix_340C_2.D\data.ms Page 8
DB-5MS: Comparison of Different Speeds on MSD 4X method RTs are precisely the same as 1X divided by 4 1 x 30 m, 10C/min and 1.7 ml/min: 120V oven and Standard Turbo 2x 4 8 12 16 20 24 28 15 m, 20C/min and 1.2 ml/min: 120V oven and Standard Turbo 3 x 2 4 6 8 10 12 14 15 m, 30C/min and 2.7 ml/min: 240V oven and Perf Turbo 4 x 1 2 3 4 5 6 7 8 9 15 m, 40C/min and 5.9 ml/min: 240V oven and Perf Turbo 1 2 3 4 5 6 7 Page 9
Why Use Agilent s Synchronous SIM/Scan? Get SIM and Scan Data in a single run which saves time - SIM = maximum sensitivity for target compounds - Scan = best confirmation of targets and identification of unknowns - For most methods, changing to SIM/Scan results in little, if any, degadation in signal-to-noise noise compared to SIM only or Scan only modes Page 10
How Does Synchronous SIM/Scan Work? Chromatographic Peak SIM data points Scan data points SIM and Scan data points are alternately collected At end of run, two separate data signals are constructed SIM Scan SIM Scan SIM Scan SIM Scan Scan time SIM Scan Each signal can be processed like SIM or Scan collected separately Page 11
Why Collect NPD and MSD Simultaneously? NPD gives sensitive nitrogen selective detection: - Identity confirmation (if it is a nitrogen drug, it should respond on NPD) - Faster data review, alternate to MSD for quantitation - Highlights nitrogen drugs that are not in target list - Splitter used also allows backflushing and changing columns without venting MSD Page 12
Example of Synchronous SIM/Scan/NPD Fentanyl Ion 245 Scan S/N pk-pk = 8 Ion 146 Scan Ion 189 Scan Ion 245 SIM S/N pk-pk = 77 Ion 146 SIM Ion 189 SIM S/N NPD pk-pk 42 Real whole blood tox sample: Screen for 278 drugs in Scan mode. Screen for low level targets in SIM mode Fentanyl found by SIM at trace level because of 10 fold better S/N NPD trace shows nitrogen peak at same RT as fentanyl in MS data 5.52 5.60 5.68 5.76 Page 13
How Does the Splitter Work? Liquid Injector XEPC Vent NPD New splitter technology Column 7890A GC 5975C MSD Column effluent split between MSD and NPD Reliable, inert, easy to use two way splitter with solvent venting Auxillary EPC control allows no vent column changing and post run backflushing Page 14
Why Use Agilent s Column Backflushing? Backflushing removes heavy material from column after elution time of last analyte: - Prevents heavy matrix compounds from eluting in later runs - Reduces column trimming and detector maintenance - Significantly reduces postrun bakeout times (example: 2 min backflush vs 10 min bakeout) Whole blood tox sample TIC Heavies past end of acquisition 1 2 3 4 5 6 7 8 Page 15
How Does Backflushing Work? Split Vent Trap S/S Inlet 25 psi During GC Run Aux EPC 4 psi Flow Splitter or QuickSwap Column MSD Auxillary EPC provides constant pressure makeup gas to post column device. During run, makeup flow is low (~ 1 ml/min) Split Vent Trap S/S Inlet 1 psi After GC Run Aux EPC is programmed Flow Column Aux EPC 45 psi Splitter or QuickSwap MSD to high pressure and inlet to low pressure. Reversed flow quickly sweeps heavies out to split vent trap Page 16
Why Use Deconvolution Reporting Software? Deconvolution of mass spectra removes/reduces interferences from chromatographically overlapped peaks: Better identification and confirmation of analytes in high matrix samples Reduces data review time, especially for large screening methods Reduces false positives and false negatives in dirty samples Identification based on matching entire spectrum cleaned of interferences against library. Much more reliable than target/qualifier ratio method. Page 17
How Does Deconvolution Work? Ions with the same abundance vs time profile are grouped together to create spectra cleaned of interferences from overlapped peaks Three Overlapped Chromatographic Peaks TIC Interference 1 Interference 2 Target Time Deconvolution Interference 1 m/z Interference 2 Time Time Time Target m/z Page 18
AMDIS Looks at Apex Retention Time and Peak Shape 50 75 16 0 17 0 18 5 28 0 31 0 Ions whose EICs have the: 1. same apex RT 2. same shape are grouped to form cleaned (extracted) spectra Extracted Ion Chromatograms (EIC) 160 wrong shape 50 These have same 170 shape and same apex RT 280 185 wrong rt 75 wrong rt 310 wrong rt Page 19
Ions With Same RT and Shape Are Grouped and Called Components (Spectra) 50 Extracted spectrum cleaned of interferences 17 0 This spectrum is compared to msl for target hits. Component Peak heights of Extracted Ion EICs give Chromatograms relative (EIC) abundance values for 50 spectra 170 280 28 0 Page 20
Example:Spectral Confirmation of Carisoprodol in Blood Extract With High Level of Fatty Acids Abundance 200000 Average of 3.564 to 3.601 min.: 120C_1MIN_GB20_VIAL2.D\DATA.MS 54.9 Apex spectrum 180000 #1 NIST Hit is Oleic acid. Carisoprodol is not in top 100 hits Deconvoluted Spectrum Carisoprodol is #1 Hit in NIST05a m/z--> 160000 140000 120000 100000 80000 60000 40000 20000 100 50 0 50 100 96.9 128.9 158.0 Spectrum at Carisoprodol RT, not deconvoluted 185.0 213.0 241.1 284.1 75.8 261.9 311.9 340.0 0 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 44 43 58 55 71 86 83 97 104 97 114 132 149 114 127 144 158 158 Deconvoluted Spectrum at Carisoprodol RT 184 184 205 223 202 217 Match = 80 245 257 288 327 260 245 Library Reference Spectrum of Carisoprodol 40 70 100 130 160 190 220 250 280 310 340 Page 21
Steps in DRS Process Software goes through entire GC/MS scan file and extracts all components (spectra). Each component is automatically searched against the target spectral library. If a component has both: a spectral match factor > than user defined minimum a retention time matching the table RT within a user defined limit It is reported as present. Data reviewer inspects the results and confirms compounds present. Those compounds are then quanted and report generated. Page 22
Example: Design 725 Compound Fast Tox Screening System Retention Time Locking (RTL)- use Agilent s New G1674AA Forensic Toxicology RTL Library Method Translation Run 4x faster version (40C/min) of method Synchronous SIM/Scan and Split to NPD Collect 3 signals Scan,SIM, SIM and NPD in one short run Column backflushing Saves bakeout time, gives cleaner baselines, less maintenance Deconvolution Reporting Software (DRS) Biggest time saver. Data reviewed in ~10 min instead of ~1 hour/sample Page 23
Fast Tox Screening System Post column device provides: A) 1.4:1 MSD:NPD effluent split B) Solvent venting C) Backflushing Liquid D) No vent column changing Injector XEPC 3.8 psig Solvent Vent NPD S/S Inlet 14.9 psig (nom) 7890A GC Column 10mX025mmidX025umDB5ms 0.25 mm X 0.25 DB-5ms 5975C MSD Page 24
Method Parameters for DB-5ms, 4X, NPD/MSD Split on 7890 Ramp 'C/min 'C Hold min MSD Agilent 5975C Initial 100 0.25 Solvent Delay 0.5 min Ramp 1 40 325 1.25 Acquisition Mode SIM/Scan Runtime 7.13 min Scan Range 40 to 570 Postrun 325 'C/min for 0.5 min Threshold 0 Sampling 1 Inlet Split/Splitless p Quad Temp 180 'C Temp 280 'C Source Temp 300 'C Mode Splitless, Constant Pressure Transfer Line 300 'C Pressure 14.90 (adj to lock) Tune Gain Normalized 1X Purge Flow 50 ml/min Purge time 04 0.4 min Splitter 2-way splitter w/ solvent vent Column DB-5ms part # (custom) Pressure 3.8 psig MSD Restictor 0.694 m x 0.15 mm id 10m x 0.25 mm id x 0.25 um film NPD Restictor 0.361 m x 0.15 mm id Initial Flow 252mL/min 2.52 Split ratio 1.4:1 MSD:NPD Outlet Splitter Vent time range 0-0.75 min Outlet Pressure 3.8 psig Backflush Time 0.5 min Injection volume 1 ul Backflush Temp 325 RT locked to Proadifen at 4.285 min Backflush Press 76 Page 25
SIM Groups Acquired With SIM/Scan SIM Group Peak Name RT min Tgt Q1 Q2 1 Amphetamine 0.600 44 91 65 2 Methamphetamine 0.700 58 91 65 3 Methylenedioxyamphetamine(MDA) 1.319 136 135 51 4 Methylenedioxymethamphetamine(MDMA) 1.431 58 135 77 4 Ecgonine Methyl Ester 1.481 94 82 96 4 Ethylecgonine 1.482 94 82 96 5 Meperidine 1.884 246 218 247 6 Ketamine 2.092 180 182 209 6 Phencyclidine 2.166 243 242 200 6 Tramadol 2.259 58 263 59 7 Methadone 2.578 72 57 165 7 Dextromethorphan 2.597 271 212 270 8 Cocaine 2.695 182 82 94 8 Cocaethylene 2.783 196 82 94 9 Diazepam 3.065 258 286 257 9 Tetrahydrocannabinol 3.111 299 300 231 9 6-Acetyl-Morphine 3.182 268 327 328 10 Oxycodone 3.201 315 230 115 10 Temazepam 3.281 271 273 272 10 Diacetylmorphine 3.328 310 268 327 10 Fentanyl 3.451 245 146 189 11 Zolpidem 3.555 235 236 219 11 Clonazepam-M (amino-) 3.622 285 258 286 12 Alprazolam 3.753 308 279 280 12 Zaleplon 3.797 305 263 248 13 Zopiclone 3.937 112 99 139 13 Lysergide (LSD) 4.000 323 324 222 Page 26
Screening A Real Blood Extract Blood extracts are very complex. There are more than 400 detectable peaks in this sample. The problem is to find the compounds of toxicological interest. Fast method: 9.75 min injection to injection (including backflush) Scan TIC SIM TIC NPD 1 2 3 4 5 6 7 Page 27
Target Compounds Identified in Sample By DRS 1 Nicotine 2 Nicotinamide 3 Carisoprodol artifact 4 Cotinine 5 Meprobamate 6 Caffeine 7 Carisoprodol 8 Methadone 9 Cyheptamide (ISTD) 10 Oxycodone 11 Cholesterol 7 9 11 1 2 3 8 5 4 6 10 Scan TIC 3 4 7 2 1 5 6 8 9 10 11 NPD 1 2 3 4 5 6 7 Page 28
Deconvolution of Spectra Eliminates Review Bottleneck For Large Screening Methods Conventional quant approach (~ 1hour / sample review time): Must examine every peak with response at target ion (367 here) Takes ~ 1 hour to review, delete false positives, quant real ones DRS approach (~ 10 min / sample review time) Gives list of compounds present in 1-2 minutes List has many fewer compounds to review. (12 instead of 367) Veryfew false positives and false negatives In the current example, all compounds identifiable by QEDIT or Screener approach were found immediately with deconvolution Can find compounds not identifiable by conventional approach when severe overlap with matrix exists Page 29
DRS Report From Scan Data of Sample Page 30
Caffeine: Is It There? All 3 qualifiers have problems. It would be good to have more info like a spectral match > 60 to confirm ID, since qualifiers are questionable. Interference Caffeine? TIC Scan Q1 and Q2 have interference overlap Q3 has very low S/N NPD confirms nitrogen at caffeine RT 3.00 3.05 3.10 194 Scan 109 Scan 82 Scan 67 Scan NPD Page 31
Caffeine Spectrum Has Interference Problem With Matrix Compound 100 Match = 51 194 Caffeine Apex Spectrum (no deconvolution) 50 0 50 100 43 55 67 71 82 91 42 55 67 82 109 124 94 122 109 137 154 165 136 165 207 221 235 250260 273 284 Target Library Reference Spectrum of Caffeine 194 38 58 78 98 118 138 158 178 198 218 238 258 278 Page 32
Caffeine Component Searched Against Target Library Deconvolution improves spectral match quality by removing interfering compound spectrum Match = 70 100 194 Component Found at RT of Caffeine 50 0 50 100 42 55 55 67 72 82 67 82 94 97 109 109 122 136 165 Target Library Reference Spectrum of Caffeine 194 38 58 78 98 118 138 158 178 198 218 238 258 278 Page 33
Methadone: Is It There? Scan and NPD Data SIM gives high h S/N 72 Scan confirmation of ratios, but 57 still doubtful 57 Scan 165 Scan NPD confirms N containing peak at Methadone RT NPD 3.8 3.9 4.0 Page 34
Methadone Spectrum Has Interference Problem With Octadecanoic Acid in Matrix Match = 42 100 50 72 Methadone Apex Spectrum (no deconvolution) 0 43 55 42 57 77 83 91 97 111 129 143 115 128 151 165 185 199 223 165 178 191 241 253 264 284 236 294 50 100 Target Library Reference Spectrum of Methadone 72 38 58 78 98 118 138 158 178 198 218 238 258 278 298 318 Page 35
Methadone Deconvoluted Spectrum Searched Against Target Library Deconvolution significantly improves spectral match quality by removing interfering octadecanoic acid spectrum 100 50 72 Match = 80 Component found at RT of Methadone 0 42 42 56 57 77 85 91 115 130 143 115 128 151 165 165 178 195 223 178 189 236 294 294 50 Target Library Reference Spectrum of Methadone 100 72 38 58 78 98 118 138 158 178 198 218 238 258 278 298 318 Page 36
Alprazolam: Is It There? Deconvolution and conventional quant are both limited by signal-to-noise ratio of scan data. As S/N drops, ions disappear from deconvoluted spectrum 50 Deconvoluted Spectrum Forward Match 57.5 100 204 Reverse 57.5 Second Hit in NIST 05a 273 279 308 0 50 100 293 39 51 63 190 116 219 253 89 151 163 102 137 177 245 77 273 308 204 279 38 58 78 98 118 138 158 178 198 218 238 258 278 298 318 338 Page 37
Alprazolam at Limit of Detectability in Scan Mode Deconvolution and conventional quant are both limited by signal-to-noise ratio of scan data Both deconvolution and standard quant are at the detection limitit here 279Scan 204 Scan 308 Scan 203 Scan SIM and NPD data aid in confirming very low level of Alprazolam 5.55 5.60 5.70 5.80 279 SIM 308 SIM 280 SIM NPD Page 38
Time Savings Conventional Method: Scan run: 35 min SIM run: 35 min NPD run: 35 min Data Review: ~1 hour Total: 165 min/sample Fast Tox Method: Scan, SIM, and NPD: 9.75 min Data Review: ~10 min Total: 19.75 min/sample Page 39
Summary Retention Time Locking (RTL)- the same RTs from now on Method Translation speed scale to new column sizes, carrier gases, analysis speeds Synchronous SIM/Scan and Split to NPD Collect 3 signals Scan,SIM, SIM and NPD in one run Column backflushing Saves bakeout time, gives cleaner baselines, less maintenance Deconvolution Reporting Software (DRS) Data review time reduced while quality is improved Page 40
Get started quickly Application kit Columns, consumables & checkout mix Application Note and Quick start guide DVD with method parameters 729 compound DRS database for fast screening Video tutorials with step-by-step instructions Pre- configured and pretested for your application On-site application checkout Page 41 Page Feb. 2009 41 FY09 Solution Plan Agilent Restricted
Thanks For Attending Wow! Did you see that? Page 42