Multidimensional Gas Chromatography: Past, Present & Future

Transcription

1 4 th MDC Workshop: TORONTO Multidimensional Gas Chromatography: Past, Present & Future Philip Marriott School of Chemistry, Monash University Discovery Outstanding Researcher Award: Australian Research Council Australian Centre for Research on Separation Science

2 Thanks to My Monash ACROSS Group; all our collaborators across many continents Agilent Technologies Sigma-Aldrich Supelco Defence Science Technology Organisation Other research sponsors: ARC; Linkage Project supporters; Industry supporters

3 Has GC GC spurred new developments ~ renaissance ~ in MDGC? If GC GC represented a new and exciting development in GC, and showed how enhanced separations benefits analysis, how should GC companies respond? The easiest is to use a method that is compatible with their software, but gives improved separation = MDGC

4 Landmarks in GC GC John Phillips pioneering work LECO agrees to commercialise GC GC-TOFMS ZOEX GC Image GC GC-qMS Cryogenic Modulation Diaphragm Modulation Flow Modulation The GC GC Symposium series ~ now 10 th.. And every one of you for believing

5 Landmarks in GC GC Retention indices Retention prediction / modelling Chemometrics Novel interconversion processes Application of new detection systems Optimisation Modulation ratio and sampling theories.

6 Definitions of MDGC / GC GC Giddings: two principal kinds of multidimensional systems (1) continuous two-dimensional (2D) operation (simultaneous zone displacement), and (2) coupled column assemblies (sequential displacement) [1984; 1987]. Marriott: MDGC the process of selecting a (limited) region or zone of eluted compounds from the end of one GC column, subjecting the zone to a further GC displacement. (2005).. Blumberg and Klee: re-defined MDC as n-dimensional analysis is one that generates n-dimensional displacement information. (2010) None of these definitions implicitly state nor require that the separation of compounds will be improved by MDGC.

7 PJ Marriott; ZY Wu; P Schoenmakers. Nomenclature and Conventions in Comprehensive Multidimensional Chromatography An Update. LC-GC Europe, 25 (2012) the guiding principle (3) is that a 2D separation can be called comprehensive if: 1. Every part of the sample is subjected to two different (independent) separations; 2. Equal percentages (either 100% or lower) of all sample components pass through both columns and eventually reach the detector; and 3. The separation (resolution) obtained in the first dimension is essentially maintained.

8 And that was fine, until along came flow modulated GC GC thanks to John Seeley. Now we had to change the concepts of how we implemented the procedure. Changes to the linear carrier flow rate from 1 D to 2 D now permitted the fast elution required on 2 D, and to achieve adequate retention and resolution, we now have to use a longer and often wider bore column, than in the more common thermal modulation interfaces.

9 The need for NOMENCLATURE Consistency amongst researchers Universal understanding of the meaning of symbols Clarity in definitions and terms Use of 1 X and 2 X for symbols, vs 1X, 2X, or X1, X2 e.g. 1 D and 2 D for first and second dimensions, not 1D and 2D, or D1 and D2 (2D means two-dimensional) 1 t R ; 2 t R, not 1t R ; 2t R nor t R 1, t R 2 We can now define any symbol correctly: 2 I; 2 d f ; 2 w b

10 lingering confusion or inconsistency over the use of the words comprehensive and orthogonal, and of the multiplex sign ( ) as a short notation for comprehensive twodimensional (2D) separations. Do NOT use comprehensive GC GC! Do NOT use GCxGC, GCXGC, GC X GC Do NOT use normal phase column set And especially do NOT use reverse-phase column set

11 Are we at a stage where new-comers into GC GC (and MDGC) do not have an adequate understanding of the technology, fundamentals and processes of these methods? Or maybe where some have little experience with classical capillary GC? Is doing a GC GC analysis because it is new justified if the analysis does not require GC GC, or.. If the method is not properly (or poorly) implemented?

12 High or low frequency Modulation / Sampling in MDGC The difference between classical MDGC and GC GC is the sampling frequency and modes. MDGC conventionally samples from 1 D in a few discrete steps; normally a long 2 D column is used. With or without cooling the oven. GC GC requires faster, and regular sampling, normally at a faster rate that 1 w b. So this could be considered a continuum of technology from MDGC thru GC GC.

13 GC qms (non-modulated) 4.0 Response (x10 5 ) zone Retention time (min)

14 MDGC qms (Modulated) 1.0 / 1.5 min 10 FID (x 10 5 ) M N qms 1 min modl n qms 1.5 min modl n M N Zone 10 = 1.5 min Zones M & N = Retention time (min)

15 MS 2 x 10 7 FID x 10 6 DL MS 2 DS FID Possible to get split peaks; Peaks are SHARP and TALL * Min 2 t 2 R 1.5 Min (b) 4* 6 1(a) Retention time (min) 1(a): (b): : : : : : (B) FID x 10 6 MS x * Retention time (min) 2 t R 1: : : : : : 1.304

16 & complete each heartcut in 2 min Multidimensional sampling strategy Sample 0.2 min, 2 min apart = 10 analyses to cover a full sample

17 No cryo 2 min + cryo

18 Single (12 s) heart-cut; with ( ) & without ( ) cryotrapping. Without: can hardly recognize oxygenates: With: easy to recognize oxygenates

19 Another approach - Single heart-cut, cryotrap, reduce oven temp., release... Cut a Very 0.2 min high section peak capacity from the (~600) 1D GC run Would need ~150 runs to cover the whole 1D GC Single (12 s) heart-cut (see previous slide) PROCESS: (1) cryotrap; (2) cool the oven; (3) then elute on the long 2 D column

20 GCxGC ~ a tutorial 2 D Inj 2 D 1 st Dimension; 1 D Modulator, M 2 nd Dimension; 2 D (very short) Det C A B D 1 D Peaks are sliced (modulated); each slice is introduced into a short 2 D column to provide further separation. Data presented as a 2D plot.

21 Movies of Modulation Operation MODULATION DRIVE UNIT CRYOTRAP MOVEMENT REAL-TIME DATA ACQUISITION as seen on chromatography data system

22 How do we calculate the exact first dimension retention time?.. and is it important? Some commentators say that GC GC loses information on the 1 D column Due to the modulation process, and loss of resolution. But what if we can re-construct the 1 t R value? This will return to the 1 D all of the primary information that we need from the GC process. Reconstruct based on the modulated peak distribution, and the appropriate peak model

23 Modulation Phase (Φ M ) in GC GC Start time In-phase Start time Start time o out-ofphase Start time In-phase

24 Modulation Phase P M = 6 s; 0.1min) Symmetric REAL 2 D PEAKS Adjust time of starting modulation allows phase of modulation to be readily altered 180 o out-of-phase In-phase 3.10 min 3.08 min 3.06 min 3.04 min 3.02 min Same phase 3.00 min

25 A 2s B 3s Modulation period incremented A-E 200 Response (pa) C 4s D 6s Start time 16.03min A range of phases is observed E 9.9s Retention Time (minutes)

26 Response (pa) Modulation Period (P M ) in GC GC ? Retention Time (minutes) 17 Effect of changing modulation period on pulse presentation - This affects apparent resolution in 2D

27 How do we calculate the second dimension retention index in GC GC? Injector deactivated FS 0.2m x 0.25mm first - 2 D column BPX5 (0.95m x 0.1mm x 0.1mm) 2 I 1 longitudinally modulated cryogenic system (van den Dool) temp prog 3 way splitter FID1 FID2 2 I 2 1 I 1 D column SolGel Wax (30m x 0.25mm x 0.25mm) second - 2 D column BP10 (0.95m x 0.1mm x 0.1mm) (Kováts) isothermal

28 Make successive SPME injections of alkanes tr (z+1) 4 C 10 C 11 C 12 C 13 C 15 C 16 C 14 C 17 C C 19 C n-alcohols t=0 injection tr x tr (z) 2 D retention time [s] (BPX5 column) C 9 C 9 -C 10 C 9 -C 11 6 C 9 -C 11 7 C 9 -C 12 8 C 9 -C 13 9 C 13 C 10 -C C 10 -C 15 C 10 -C 16 C 11 -C 17 C 11 -C 18 C 12 -C 19 C 12 -C 20 C 13 -C 20 C 14 -C 20 C 15 -C Slide 28 1 D retention time [min] (SolGel Wax column)

29 2 D polar pa ? 1 0 First retention index dimension 1 C 10 C C C 15 C 16 C 17 C C C C 13 C Slide 29 C a b 17a b 23 21a 22b 7 22a b * A 1? A 2? 1 D nonpolar Second retention index dimension Modulated alkanes t =0 injection

30 How do we calculate the second dimension retention index in MDGC? We have to get alkanes into the 2 D column, and analyse them under identical conditions as the analyte. A number of ways are possible --- see later

31 Detection in GC GC Apply classical GC detectors to the fast peak profiles in GC GC FID Of prime concern: SCD Is the chemistry and physical process of NCD the detection transducer compatible ECD with the fast data acquisition and peak NPD flux needed for a GC GC detector? AED Often the answer is no ~ or at FPD least, not quite.. ~ tailing; broadening; less than anticipated sensitivity

32 Dual Detection NPD (A) & (B) and ECD (C) & (D) Response (pa) (A) 2 tr (s) 5.00 (B) st Dimension Retention time (min) Response (Hz) (C) 2 tr (s) (D) st Dimension Retention time (min) 50.00

33 B) matrix GC GC-NPD D Retention time (s) matrix C) iprodione matrix + D) all D Retention time (min)

34 2 tr sec FPD/P - mode FPD/S - mode Methyl Chlorpyrifos Diazinon Ethyl parathion Malathion Methyl Pirimphos Chlorfenvinphos Fenthion ethyl Methidathion Bromophos ethyl Carbophenothion Good shape ~ same as FID min Broad, Tailing s m in

35 Mass Spectrometry with GC GC Why do we want / need MS? Argument: If a peak s position is equivalent to identity and this is so much more certain in 2D space then do we still need MS? Implies need for / use of authentic standards; (and if only target analysis is wanted, for known compounds, maybe MS is less important) Need to give a name / tentative name to a spot and the skeptic is likely to say OK so you have lots of resolved peaks what are they???

36 Mass Spectrometry with GC GC Slow scanning qms Frysinger Faster scanning qms Marriott, Shellie, Song, ~ reduced mass scanning range Fast acquisition TOFMS Marriott (loan from LECO) Isotope ratio MS - Brenna Accurate mass MS usually TOFMS systems - Patterson Quantification vs qualification of components Use of GC GC-FID integrated with GC GC-MS

37 Mass Spectrometry with GC GC Some history.. Meeting in Park City, Utah ISCC (Milton Lee Symposium) LECO undertook commercialisation of fast TOFMS, and Jan Beens, Rene Vreuls and others (John Dimandja?) tried to convince LECO (Rick Parry) to invest in GC GC, since ALL GC GC users NEEDED fast MS. A focus on 1D GCMS means less of a demand for TOFMS. LECO obviously agreed!

38 Modulation & Modulators in GC GC The Sweeper originally we were told that this was a system that would end the need to innovate different modulation systems Cryogenic systems have proved to have longevity, established the present user-base, and much novel work has been performed using these. Diaphragm modulators Flow modulators See the presentation of Tadeusz Gorecki for more information..

39 Modulation Ratio M R in GC GC Mod Number n M vs Mod Ratio M R n M : Number of modulations per primary dimension peak Schure, et al. suggested ~ 3 4 modulations per peak How do we decide n M?? It varies across peaks in the chromatogram; It varies with phase; It varies with amount injected; It varies with how small a peak we want to include in the Number of Modulated peaks!!!!!!!! So define modulation ratio M R as the ratio of the M R = (4 x 1 w b ) / P M Or more generally M R = (4 / w h ) / P M

40 Response (pa) Response (pa) M R ~ 6.0 M R ~ Response (pa) 40 P M = 4s Retention time (min) M R ~ 3.7 M R ~ Retention time (min) P M = 3s P M = 5s Response (pa) P M = 6s Symmetric REAL 2 D PEAKS Alter P M 1 w b ~ 18 s Retention time (min) Retention time (min) Response (pa) M R ~ 2.5 P M = 7s Retention time (min)

41 Use Modulation Ratio to.. M R = (4 / w h ) / P M Approximately tells us (how many large ) modulated peaks we will obtain. IF we require a M R of 3. If 1 w h = 6 s, then P M ~ 1.6 s; So modulate at ~ s What does this require of the METHOD? If we want no wraparound, then the desired P M above determines the max 2 t R possible. This decides the column ID, L, 2 d f, etc required for the analysis. Note that all these parameters will alter 2 t R, (or 2 k I we know 2 t M ).

42 The Future MDGC & GC GC.. we will continue to investigate more integrated methods of analysis, incorporating GC GC and MDGC, or using a combination of the two approaches. MDGC will see increasingly sophisticated implementation methods, improved prediction of separations, more chemometrics, and GC GC will see more standard methods / protocols. More modulation methods?? Better detectors, more hyphenation approaches.. All areas will benefits for MANY MANY MORE applications!

43 Some of our recent Publications B Mitrevski; PJ Marriott. A novel hybrid comprehensive two-dimensional multi-dimensional gas chromatography method for precise, high resolution characterisation of multicomponent samples. Anal Chem 84 (2012) PJ Marriott; S-T Chin; B Maikhunthod; H-G Schmarr; S Bieri. Multidimensional Gas Chromatography. TrAC, 34 (2012) S-T Chin; GT Eyres; PJ Marriott. System Design for Integrated Comprehensive and Multidimensional Gas Chromatography with Mass Spectrometry and Olfactometry. Anal. Chem. 84 (21) (2012)

44 V Alternate GCxGC & MDGC INJ EPC DET 1 DET 2 VI Simultaneous GCxGC / loop H/C INJ DET 1 DET 2 L M 1 M M 2 DS VII 1DGC / MDGC with loop H/C & DS VIII Simultaneous GCxGC/1DGC & MDGC INJ DET 1 EPC DET 2 CT/SP INJ EPC DET 1 DET 2 SP L M M DS DS

45 IX Dual parallel MDGC INJ EPC DET 1 DET 2 X Alternate GCxGC & MDGC with CT INJ EPC DET 1 DET 2 1 M 1 M 2 2 DS DS XI Selective heart-cutting for semi-prep XII Multiple/repeated H/C screening INJ EPC DET 1 large bore trapping capillary INJ EPC MS DET 2 1 DS 2 1 M 2 DS

46 Development of GC GC/MDGC/FID/O/MS System.. various columns INJ 2 Deans switch column unions LMCS interface cryotraps looped direct sniff port MS detectors

47 Selectable 1D or 2D GC system coupled to 3 simultaneous detections (MS, O / NPD / FPD) K. Sasamoto & N. Ochiai (2010) J. Chromatogr. A 1217, p.2903 (a) 1D GC MS analysis; (b) Heart-cutting; (c) 2D GC MS analysis & 1D GC back flush.

48 MDGC/GC GC O/FID/MS Configuration 2012 Inlet P1 1 D Polar SolGel-WAX 30m*0.32mm*0.50um LMCS CT 2 D short Apolar DB-5ms 1m*0.1mm*0.1um FID CT DS P2 & 2 I values 2 D Long Apolar DB-5ms 30m*0.25mm*0.25um ES P3 MS S-T Chin et al. Anal. Chem. 84 (21) (2012) 9154

49 HYBRID GC GC-MDGC Results 1D GC-MS of JP-5 (TIC) Dominant 57, 71 and 85 m/z GC GC-FID 5% phenyl WAX

50 Two modes: Just Determined DS OFF: (GC GC) by Deans switch DS ON: GC GC-MDGC) 1 D column: SolgelWax (30m x 0.32mm x 0.5µm) 2 D M column: VF5 (5m x 0.15mm x 0.15µm) 3 D L column: Rxi 17Sil (20m x 0.18mm x 0.18µm)

51 FID1, without H/C all cpds Full GC GC run, P M = 30 s (20, 12 s also tested) Cut the compounds in this region (lots of DS events!) Only the cut compounds (FID2) appear; Separation improved.. On FID2, with H/C (wanted cpds) FID1, with H/C (unwanted cpds.) After H/C, on FID1 only MATRIX cpds remain

52 COFFEE volatile sample by using SPME: GC GC FID1; Not heart-cut DS scissors

53 COFFEE volatile sample by using SPME: GC GC FID1; Now heart-cut With Heart-cut of the selected components Targeted components are excised from the 2D plot!

54 Chiral Interconversion Processes Z (C) E E (B) Response (S)Z & (R)Z (S)Z (R)Z (R)E (S)E Response Retention Time (min) (A) Retention Time (min)

55

56 ISCC and GC GC 2013 Sunday, May 12, - Thursday, May 16, 2013 Renaissance Palm Springs Hotel