Note 76: Determination of the Sensitivity of a CRIMS System

By Eric Butrym and Steven Colby

Presented at ASMS, Orlando, FL., June 1998


Chemical Reaction Interface Mass Spectrometry (CRIMS) is a technique that exploits stable isotope enrichment for enhanced detection of analytes and their by-products. The sensitivity of a CRIMS system depends on its ability to detect low levels of material as well as to distinguish isotopic enrichment from natural isotopic abundance. Many uses for the CRIMS technique have been proposed ranging from drug metabolism analysis to facilitating pollution remediation.1,4

Crims cavity

CRIMS Cavity Inside GC Oven

CRIMS enhances isotope detection by assuring that an overwhelming majority of the element in question (Carbon, for instance) is in the form of one small inorganic species. This is accomplished by combusting the effluent from a chromatographic separation in a plasma, then reacting the resulting atomic species with a strong (usually oxidizing) reagent. In the case of organic compounds reacted with SO2, most of the carbon will be in the form of CO2. 12CO2 is represented by ions of m/z 44 and 13CO2 by m/z 45. Enrichment is detected by deriving an "enrichment chromatogram" from the total ion chromatogram, or by integrating selected ion chromatograms separately and then comparing the relative abundance of mass 45 to mass 44 for a given peak. Enrichment chromatograms are generated by mass calculation software that performs mathematical operations on the signals from each isotope channel. For instance, multiplying the abundance of 12CO2 (m/z 44) by the amount of 13CO2 occurring in nature yields the theoretical natural abundance of the heavy isotope for each data point in the chromatogram. Subtracting this result from the actual m/z 45 signal gives a chromatogram in which peaks are present only when the ratio of 13C to 12C rises above the natural abundance value used in the calculation. Enrichment levels based on careful integration of the individual ion chromatograms yields more reproducible results than integration of the enrichment-only chromatogram and should be used when making critical measurements. More information about the chemistry and application of CRIMS can be found in the literature.1,2

The issue of sensitivity in a CRIMS system is related to the ability to distinguish between naturally occurring levels of heavy isotopes and enriched or elevated amounts. Due to the random nature of the noise in a blank baseline, examining the isotope ratios in the baseline itself has little meaning. The blank baseline contains sources of noise that are independent of any discreet compound and affect all masses equally; electronic background from the detector is one example. The influence of these mass-independent signals is dominant when only a small amount of analyte is present, and small isotopic ratios are buried. Only when there is a sufficient amount of material in the ion source (e.g. a peak) can the correct ratio be expected. For this reason, experiments concerned with determination of exact or very small degrees of enrichment should be carried out using substantial amounts of material.3 Figure 1 illustrates the change in isotope ratio as a peak elutes. Nominal or "baseline" isotope ratios can be determined by repeated analysis of unenriched material, and sensitivity is related to the variation in the nominal value. Generally, if a peak produces an enrichment signal elevation in excess of three standard deviations of the mean nominal value, it can be integrated and this value is considered the limit of detection for enrichment.

In terms of absolute sensitivity to small amounts of material, the methods for CRIMS do not differ from those used in normal chromatography, and simply involve determining where the signal-to-noise ratio is unacceptable or where linearity of response falls off. Such methods would most likely be used when CRIMS is employed as an element-specific detector.1,2


Figure 1

Figure 1. Effect of Eluting Peak On isotopic Ratios. Isotope Ratios (blue) Drop to Appropriate Levels as CO2 (black) enters detector

In the examples that follow, the enrichment detection limit for 13C enriched caffeine was examined using a GC-CRIMS system (SIS Inc.) based on a Hewlett-Packard 5890 Series II GC and a 5971 Mass Selective Detector modified after Song et.al 2 and using SO2 as the reagent gas. Hewlett-Packard G1034C ChemStation software was used for data aquisition and analysis, and enrichment-only chromatograms were generated by ProCalc (ProLab Resources, Inc.).

Seven 1 ml samples of unenriched Caffeine were injected at various concentrations ranging from 226 ng/ml to 565 ng/ml. Variation in the isotope ratios was found to be independent of sample concentration, and the results of the sample were averaged to give the nominal natural abundance. The enrichment-only chromatograms (EOCs) are rarely completely flat, owing to slight variation in the isotopic ratio of the material used, minor chromatographic effects and baseline drift. For unenriched material, the "signal" generated in the EOCs should fall within three standard deviations from the mean baseline value calculated close to the peak. EOCs were evaluated by taking boxcar sums of 9 data points (representing the approximate width of the baseline disturbances) in the vicinity of the caffeine peak. The greatest of these was compared to the sum of 9 other points taken in a clear area of the chromatogram approximately one minute before the peak. The difference is the baseline signal due to unenriched material. As mentioned above, the seven values of the baseline signal were averaged, and the standard deviation is used to define the detection limit. All of the unenriched samples fell within the three standard deviation limit.

Three 1 ml samples containing 2.48 ng of enriched caffeine and 283 ng of unenriched caffeine were injected next. The enrichment-only signal of each of these samples was also within three standard deviations above the mean baseline, indicating that this level is below the enrichment detection limit. Next, three samples containing 12.4 ng enriched caffeine in the same amount of unenriched material were injected. All three samples gave signals greater than the mean-plus-3 S.D. limit and were thus considered within the detection limits of the instrument.


Figure 2

Figure 2 Extracted Ion Chromatograms of m/z 44 (blue), m/z 45(red) and Enrichment-Only (green) for Unenriched Sample

The enriched caffeine used (Cambridge Isotope Labs, Inc.) contained 13C enriched to 99.9% on one of the eight carbons in the molecule. The detected enrichment was therefore approximately 0.65 atom %. In terms of the established nominal isotopic ratio, d13C = 261.6.

Figure 3

Figure 3. Extracted Ion Chromatograms For Sample with High Degree of Enrichment (310 ng 13C caffeine, 283 ng 12C caffeine)

These values were also calculated by integrating the selected ion chromatograms for m/z 44 and 45. The signal for m/z 45 was first multiplied by 100 to allow consistent integration of peaks in both traces, and the ChemStation autointegration feature was used to determine peak areas. Peak areas for the different ions were compared directly to give percent enrichment. This method of calculation gave highly reproducible ratios for the samples tested. Interestingly, the samples with less enrichment, while very tightly grouped by the integration method, gave ratio values that were lower than the unenriched standard material. This may be due to the fact that the enriched samples were run the day after the control samples, and the anomaly may reflect differences in instrument tuning or CRIMS chemistry. However, one unenriched sample was run as a system stability check on the same day and its ratio agreed with those run previously. Data is summarized in Table 1.

Table 1. Summary of Enrichment Sensitivity Data

Sample  12C/13C (ng/ml) Nominal Value(abundance units)   Ratio by Integration (% 13C)
565/0 77   1.33
565/0 165   1.19
377/0 65   1.25
226/0 35   1.23
283/0 99   1.23
283/0 152 Mean : 89.9 1.32
283/0 36 3 x S.D: + 156.2 1.23
    Detection Threshold: 246.1 Mean:1.25
283/2.48 156   1.18
283/2.48 87   1.18
283/2.48 97   1.15
283/12.4 405   1.56
283/12.4 250   1.56
283/12.4 376   1.64


Measurement of the sensitivity of a CRIMS system to isotopic enrichment requires enough material to overcome random baseline noise and display appropriate isotope ratios with unenriched material. GC-CRIMS is very sensitive to slight enrichment of 13C, and results can be obtained with high precision.

Other factors that affect the sensitivity of a CRIMS system range from the obvious steps of optimizing the tuning parameters and eliminating sources of chemical background to more involved tasks of selecting Selected Ion Monitoring parameters for optimum dwell time and scan rates. Correct plasma conditions are also a factor. Inappropriate amounts of reagent gas or overloaded samples can quench the plasma or alter the CRIMS chemistry, resulting in lower sensitivity and unreliable results.


1.    Abramson, F.P. 1994. CRIMS: Chemical Reaction Interface Mass Spectrometry. Mass Spectrometry Reviews. 13, 341-356.

2.    Song, H., Kusmierz, J., and Abramson, F. 1994. Implementation of the Chemical Reaction Interface Mass Spectrometry Technique on a Hewlett-Packard       Mass-Selective Detector. J. Am. Soc. Mass Spectrom. 5, 765-771.

3.    Kusmierz, J.J., and Abramson, F.P., 1993. Improved Measurement of Stable Isotope Ratios in GC-MS using CRIMS. Biol. Mass Spectrom. 22, 537-543.

4.    Using CRIMS to Monitor Bacterial Transport in in situ Bioremediation.  DeFlaun, M., Fuller, M., Thomas, A., Butrym, E., Colby, S. and Onstott, T. Poster presentation given at Pittcon '98 New Orleans, LA March 1-6, 1998.