Ultrafast GC/MS 

and Direct Measuring Mass Spectrometry

Environmental Analysis

        Instrument manufacturers and analytical chemists design instruments and develop methods with competing goals in mind; for example, 1) improving measurement precision, accuracy, sensitivity, or selectivity, 2) increasing analysis speed without/while sacrificing data quality, 3) increasing overall laboratory productivity (staff and/or equipment), and 4) lowering the per sample analysis costs. Improvements in instrument hardware and data analysis software concomitant with the remarkable advancements in computer processing speeds and control technologies have provided the tools needed to meet these challenges. High throughput laboratories routinely employ automated systems capable of delivering the sample to the instrument, detecting a wide variety of substances, and in reporting their results. 

        Gas chromatography/mass spectrometry (GC/MS) is the most often used technique for analyzing complex samples. High resolution capillary columns yield separation efficiencies of up to several hundred thousand plates per meter.  Today, high resolution fused silica capillary columns produce the same resolution in ten meters that once required sixty meters or more to accomplish.  While, improvements in MS peak deconvolution techniques, statistical methods of analysis, and automated library searches have significantly enhanced the data reduction and reporting process.

        Despite these advancements, the rate at which high throughput laboratories can process samples is limited by the complexity of the sample matrix. Although selective detectors such as chemiluminescence (CD), flame photometric (FPD), or electron capture (ECD) are used to minimize sample cleanup, they cannot provide unambiguous compound identification. For example, when analyzing polychlorinated biphenyls (PCB), dibenzofurans (PCDF), or dibenzodioxins (PCDD) at least two-column GC separation is required to obtain some degree of confidence in congener identification if an ECD is used, where solute retention is orthogonal on the two columns. If, on the other hand, a single column is used, PCB misidentification can result in as much as 200% overestimation of the total PCB concentration. Whether single- or multi-column GC/ECD is employed, compound selectivity must be achieved through long GC run times.

        Unlike selective detectors, where a single dimension of data is produced, mass spectrometry yields two-dimensional information. The time-band trace (dimension 1) is obtained by summing the signals produced from mass fragments (dimension 2) at some specified time-scan. Positive identification is made when sample constituents are separated during GC and have sufficiently different fragmentation patterns. Results indicate that the most often used mass spectral search routines, e.g., NIST and Wiley data analysis software, produce good agreements between sample and library spectra (50-75% accuracy) when sample spectra are relatively clean. Poorer results (25-50% match) are obtained for two-component or more mixed spectra where peak purity is less than 85%. These findings indicate that pattern recognition algorithms have difficulty unambiguously identifying sample constituents when more than one compound results in the same GC peak.  In addition to long GC run times, most laboratories spend 15 min confirming the MS data analysis for every 30 min of GC analysis time.

        The data analysis software program we developed is based on a new set of mathematical algorithms that deconvolute highly complex sample spectra, see figure.  The algorithms can be applied to instruments that produce characteristic, narrow band signals including MS, NMR, AE, and FTIR. 

        Figure2 depicts a 40 min total ion current (TIC) chromatogram of a standard solution containing a mixture of Aroclor 1248 (> 50 different PCB congeners), 17 polycyclic aromatic hydrocarbons (PAH), and 16 chlorinated pesticides. Ideally, each TIC peak is represented by the sum of the ions produced by a single compound at a specified time. The characteristic narrow band signal is from a tetrachlorinated biphenyl (PCB).  All MS data analysis software can correctly identify and quantify organic compounds when this type of "clean" spectra exists.  

What happens when the same standard mixture is spiked with engine oil and gasoline, and then analyzed by GC/MS in 10 min?  

        The resulting TIC chromatogram shows several ill defined peaks superimposed on a "humptygram," while the mass spectrum reveals fragmentation ions from all organics that hit the MS at time t.  Untangling this signal at a specified scan(time) is analogous to re-constructing 10-50 puzzles/(compounds), where each puzzle/(compound) may contain 10 to 50 pieces/(fragments), and where each piece may be suppressed in color/(fragment signal) relative to the more vibrant pieces/(hydrocarbon fragments). 

 
        Figure 3 shows the results after data analysis.  Compound identity can be shown as a function of each compound's reconstructed ion current chromatogram, by family grouping, or by compound class. The corresponding match between the measured and library relative abundances indicate the certainty of match.  Cl-5, Cl-6, and Cl-7 PCB congeners produced signals of ~ 10², while all other organics contributed     > 10⁶  to the total ion current response, resulting in measurement selectivity of 10⁴.  Ultrafast GC/MS methods have been used to detect a wide variety of volatile organics (EPA method 8260) and semivolatile organics (EPA method 8270) in 5 to 10 min at more than 50 Superfund, Brownfield, RCRA site investigations and cleanups.  For example, a 5 min PCB analysis was used to support a PCB cleanup projects at a gas pipeline distribution center (soil excavation) and for harbor sediment (supercritical fluid extraction). A new method has been developed for the simultaneous analysis of VOCs and SVOCs, when subsurface soil samples are collected by thermal extraction cone penetrometry.

see also dynamic workplans and subsurface detection of organics

Drug Discovery

      Automated, solid phase, parallel synthesis, more commonly known as combinatorial chemistry, is used to make focused libraries of related compounds. These compounds share structural features necessary for binding to pre-selected protein targets. Both small molecules and polypeptides can be quickly made by this method . This process, called split-and-pool synthesis, results in structurally complex and diverse libraries of compounds. Beads are put into reaction vessels (wells), with each well receiving a unique set of reagents. The beads are subsequently split, with cycles of pooling, re-splitting, with further chemistry resulting in thousands of new compounds spatially segregated in wells on unique beads. Split-pool synthesis has the potential of producing one new compound per bead. Encoding methods, which are analogous to the genetic code, have been developed that record the chemical history of the synthesis. Encoding allows the structures of compounds selected in assay screens to be inferred. The literature indicates that the potential for discovering new drugs increases with increased diversity and the ability to effect specific biological pathways. Combinatorial synthesis offers the means to identify proteins that can serve as targets for therapeutic intervention and provide leads to optimize pharmacokinetic and pharmacodynamic properties.

The Problem Small molecule, combinatorial synthesis, can be carried out on 500-µm polystyrene beads. A cleavable linker can be put on the bead onto which the small molecule is synthesized. Different tagging compounds can be used in each step to evaluate synthesis progress. The presence/absence of a particular tag indicates the success or failure of the step. Because the bead can be washed after each step, it may be possible to determine intermediates as well as the final synthetic product.

Assume that ten 96-well plates are used in the synthesis. Also assume your lab is automated to efficiently analyze well contents and that the tags consist of the following compounds: Cl_1-5Benzene, Br_1-5Benzene, Cl_1-9Bipenyl, and Br_1-9Bipenyl, with an oxygen atom (ether) between the aromatic ring and, C₃, hydrocarbon chain, see Figure 1. Another set of tags can also be used, where the length of the hydrocarbon chain increases by one carbon unit per tag. Design an analytical method that will help identify whether the final product was produced as desired. How might you determine the identity of the intermediates (assume the tag also inserts into the small molecule)? Choose an appropriate tagging system consistent with the method you develop.