Mark Taylor, Richard Whitney, John Monti and Denny Miser
The combined technique of gas chromatography-mass spectrometry (GC-MS) provides a powerful tool for separation, identification and quantification of compounds in complex mixtures. Many applications rely on the ability of GC-MS instrumentation to detect and identify minute quantities of compounds. Increasingly sensitive instruments are invaluable to experts in a variety of fields who were previously unable to identify trace compounds in complicated or difficult samples.
Figure 1: Shimadzu Scientific Instruments' GCMS-QP2010 Plus
Instrument manufacturers and users alike frequently equate “best” with “most sensitive” when considering GC-MS instrumentation. Instrument specifications should provide the information consumers need to make an informed decision regarding the suitability of a particular instrument for accurate and reliable measurement of samples likely to be encountered in a real lab.
Instrument manufacturers naturally want to publicize impressive performance specifications in order to convince shoppers to purchase their product. Manufacturers are constantly vying to outdo each other in publishing the highest “X” or the lowest “Y” specifications for their models. The buyer is not only faced with comparing features of different models, but decoding confusing market-specific terminology in order to determine whether the information supplied by the manufacturer is relevant to their purchase decision.
Specifications should assist the consumer in making an appropriate selection of a GC-MS. From the buyer’s standpoint, it can be argued that at the very least, instrument specifications should answer three fundamental questions:
1. How accurate will my measurement be?
2. How repeatable will my measurement be?
3. What is the smallest mass or concentration of analyte that can be accurately and reliably measured? In other words, how sensitive is it?
On the surface, it would appear that these questions can be answered quite simply. For example, manufacturers can state measurement accuracy as a comparison between measured and referenced data of a certified reference standard. Measurement repeatability can be easily stated as the standard deviation or relative standard deviation of measurements made on a certified reference standard. However, the reality is that, very often, the first and only question a customer has in mind when selecting a GC-MS is: “How sensitive is it?”
Defining sensitivity
In reality, defining sensitivity is not straightforward. Unlike the responses to the first two questions, the issue of minimum analyte detectability (i.e., sensitivity) does not lend itself to a quick and easy answer. Equally unfortunate, manufacturers tend to use their sensitivity specification as the platform from which to promote their instruments – to demonstrate that their instrument is superior to competitors’ models.
Most frequently, sensitivity is specified as the “signal-to-noise ratio,” or S/N. Ideally, this numerical value stands as a truthful indicator of the minimum detectable analyte that an instrument can measure and a useful comparator of instrument performance. However, such values alone cannot do this. Additional information is needed, particularly in the details of signal and noise measurements. Exact definitions and measurement conditions for both signal and noise are required in order to give meaning to the numerical S/N value.
For hyphenated chromatographic instruments such as GC-MS, chromatographic conditions influence the final S/N result. So, in order to make sense of the S/N specification for a particular model of GC-MS, we need to know:
•the models of GC and MS used
•the chromatographic conditions employed
•the MS conditions employed
• the identity and source of the analyte
• the definition of “signal”
• the definition of “noise”
The industry standard for sensitivity of GC-MS instruments is based on octafluoronaphthalene (OFN); the sensitivity of an instrument is defined as the signal to noise (S/N) obtained from a 1-µL injection of a solution containing 1 pg/µL OFN. On the surface, it appears that all of the information needed to make an objective evaluation of sensitivity is readily available. However, while GC-MS conditions and analyte identity and amount are objective parameters, the definition of both “signal” and “noise” are left up to the writers of instrument specifications and are often a venue for marketing creativity.
Figure 2: Mass chromatogram peak
The “signal” from a mass spectrometer is a series of mass spectra taken at regular intervals over time. The equipment can display them in a variety of ways, often as “mass chromatograms,” or the display of the signal for a specified mass over time. This is shown graphically in Figure 2. It should be noted that the Gaussian-appearing chromatographic “peak” shown in Figure 2 is actually reconstructed from a series of discrete measurements, ten in the example shown.
It is generally accepted that a minimum of ten points (spectra) across a chromatographic peak are required to obtain accurate measurements. This translates to scan rates of 0.2 to 0.4 sec/scan (2.5-5 spectra/sec) for the most commonly used GC columns. Significant error is introduced when lower sampling frequencies are used, and variation in sampling rates can affect signal and noise.
So the question then becomes how to specify “signal.” Is it the maximum response value of a known quantity of a specified analyte measured under defined conditions? Will it use the peak area only or the sum of the discrete responses over a defined analyte peak? And where shall we specify the response to be used as “noise?” These are important considerations when trying to compare and understand an instrument’s sensitivity.
It is critically important to know that the manufacturer’s conditions for calculating the S/N specification are representative of conditions employed in day-to-day operations. All conditions used for measuring S/N, including any noise-suppression algorithms, should be the same for both the S/N calculations and for all other routine analyses.
Achieving sensitivity
There are a number of ways to optimize the sensitivity of GC-MS. When sensitivity is defined in terms of S/N, increasing sensitivity is achieved by increasing signal or decreasing noise, or by a combination of both. Because S/N is a simple ratio, factors affecting signal or noise have equivalent influence on sensitivity (S/N).
Intensity
Optimizing the transfer of analyte from the injector through all of the components of the GC-MS system to the detector can increase the signal. Optimizing the response of the detector to obtain adequate signal without excessive noise is a complementary approach to increasing signal.
Many components of the GC-MS system can contribute to background, or chemical, noise. In addition, noise can originate from the sample itself, from the injection technique or from the type of chromatographic column. Electronic and detector noise are primarily dependent upon instrument design and manufacturing. Significant noise can also emanate from the electrical power source.
Alternate data acquisition modes (SIM and recently introduced Scan/SIM techniques) have dramatic effects on GC-MS sensitivity. In a similar manner, alternate ionization techniques (chemical ionization, negative ion chemical ionization) can enhance sensitivity, especially for selected analytes. The present discussion of sensitivity is limited to the full-scan electron impact mode of operation and applies primarily to single-quadruple instruments.
Increasing Signal
Optimizing the signal can be considered in three phases: maximizing the injection, or transfer, to the GC column; optimizing the GC column for maximum sensitivity and adjusting the detector (MS) conditions.
GC injection
Efficient transfer of sample to the GC column is dependent upon optimizing the injector conditions. Minor inefficiencies in injector operation become major obstacles when trace quantities of analyte are injected. The selection of injection-port liner, split- and septum-purge flow rates, syringe needle penetration, column positioning in the injector and other parameters affect the efficient transfer of analyte to the GC column. The use of alternate injector types can enhance the transfer of analytes to the GC column.
GC column selection
For most GC-MS applications, the GC typically employs relatively long narrow-bore capillary columns (0.18 to 0.32 mm OD; 15 to 75 M long). Sensitivity increases as the GC column diameter decreases, because narrower columns give taller, narrower peaks. If noise remains constant, S/N is greater with tall, narrow chromatographic peaks.
Optimizing the MS signal
Signal generation in the MS consists of three processes:
• creation of ions (from the ion source)
• transmission of ions (through the lenses and mass filter)
• detection of ions (with the electron multiplier)
The combined efficiency of these processes results in optimizing signal intensity. In some cases, inefficiencies in these processes also affect noise. These efficiencies are affected by instrument design and condition (cleanliness, age), as well as operating parameters. As a result, instrument design accounts significantly for differences in sensitivities between different models of GC-MS instruments.
Ionization and ion transmission
Sensitivity is affected considerably by ion source condition (cleanliness) and optimum tuning. Ion source operating parameters have a significant effect on sensitivity because these parameters control ion production.
In recent years, various design improvements in mass spectrometer sources have resulted in significant improvements in sensitivity compared to previous designs. Filament construction and placement is one such design improvement that results in enhanced sensitivity. Filament shielding can ensure efficient ion transport to the detector while protecting thermally labile compounds. Thus, a high-efficiency ion source provides more uniform temperature control for increased sensitivity.
The vacuum system significantly affects ion transmission. Differentially pumped vacuum systems, in some mass spectrometer designs, maintain a higher level of vacuum in the mass analyzer region. An increased vacuum results in longer mean-free paths for ions, which allows for more efficient ion transmission through the mass filter. This, in turn, enables greater sensitivity. Most differentially pumped instrument designs employ two separate turbo molecular pumps or oil-diffusion pumps. One new design utilizes a differential split-flow turbo molecular pump with two sets of rotors, which operate as two separate pumps in a single unit.
Detectors
Electron multipliers detect and convert ions to a signal. When impacted by a charged particle (ion), the surface of the multiplier emits several electrons in a process called secondary emission. This process repeats several times to give up to a million electrons for each ion impact on the electron multiplier surface. The gain, or signal amplification, is determined by the voltage applied across the entire electron—several hundred to several thousand volts.
Most GC-MS instruments employ two types of electron multipliers: continuous dynode types and discrete dynode types. Discrete dynode electron multipliers typically operate at lower voltages, so they show slightly less noise than continuous dynode types. Typically, sensitivity of the electron multiplier can be optimized to obtain a maximum signal with minimum noise.
Decreasing noise
Challenges and solutions in the GC that affect the S/N ratio directly affect the sensitivity of the GC-MS system as a whole. Lowering the background signal originating from the GC and improving the chromatographic resolution can improve GC sensitivity. As the quantity of a specific analyte to be detected is decreased, the effects of minor interferences on the ability to detect the specific analyte become an increasingly significant problem.
Carrier gas
Achieving the highest possible purity in the carrier gas helps to decrease noise, specifically “chemical noise.” For example, carbon dioxide gives a background signal at m/z 44 that increases chemical noise. Oxygen degrades column phase, which results in increased bleed (i.e., background signal at m/z 207). Using pure carrier gas and eliminating contamination and leaks in carrier gas lines are important factors affecting instrument sensitivity as well as overall chromatographic performance. Use of high-purity gases with appropriate filters can significantly lower the chromatographic background signal and therefore increase sensitivity.
Injection ports
Chromatographers face the continuing challenge of assuring complete sample volatilization and transfer of sample to the GC. Use of appropriate injection-port operating parameters and maintenance procedures ensure optimum transfer of sample to the GC column.
Background noise often arises from siloxanes from the GC septa, glass wool used in injection port liners, deactivation of injection port liners and other sources. These peaks are commonly seen in normal GC-MS backgrounds. Use of low-bleed septa, preconditioning of septa and liners and use of injection-port septum purge can minimize this source of contamination.
GC columns
Another common source of background noise originates from the GC column stationary phases. Traces of oxygen in the carrier gas can degrade the liquid phase. To minimize this problem, the GC column, injection port and transfer line temperatures should never exceed the maximum-rated temperature of the liquid phase of the capillary column. Thin liquid phases result in lower column bleed and decreased chemical noise. In addition, several proprietary low-bleed column phases, designed specifically for GC-MS use, have been introduced in recent years.
Electronic and vibrational noise
Electronic noise is determined largely by the instrument design and manufacture. Minimizing overall electronic noise is a major consideration in instrument design and selection of electronic components. Noise is frequently minimized by supplying “clean” or conditioned electrical power for instrument operation. In addition, minimizing vibration from motors and other devices, such as mechanical pumps, is an important consideration in minimizing overall instrument noise.
Evaluating sensitivity specifications
Manufacturers go to great lengths to present their products in the most positive light. These presentations often include technically confusing arguments. In the absence of industry-standardized test procedures (as in automobile crash-safety tests), it is unlikely that vendors will take it upon themselves to standardize conditions used to evaluate sensitivity. Indeed, optimum conditions for one model may not be appropriate for another.
So, with all the considerations affecting sensitivity mentioned above, how can the instrument user objectively evaluate manufacturer specifications for sensitivity? The answer(s) are contained in the details of the S/N determination and “fine print” qualifiers applied to the specifications.
At a minimum, users should expect manufacturers to supply basic information regarding their S/N measurement, which should correspond to “normal” operating parameters. The following questions should be considered when evaluating instrument specifications:
• Are reasonably common chromatographic conditions used (GC column, liner, injector, etc.) Columns with dimensions about 0.25 mm x 15-30 M and a splitless injection are common.
• Would the data acquisition parameters used for the S/N measurement reasonably apply to real-world applications? Scan intervals of 0.2 to 0.5/sec are appropriate for columns of 0.25 mm diameter. Scan ranges should not be unusually small (<100 amu). Unit mass resolution (bandwidth) is typical for normal operation, and serious deviations should be challenged.
• Is noise evaluated objectively? If unusual noise-suppression algorithms or amplifier “zeroing out” of noise is evident, the answer is probably “no.” Threshold or minimum areas set to zero or turned off are indicators that noise is measured properly.
• Finally, does the S/N specification represent an optimum, average or minimum value for units of the model in question? A good test is how frequently the S/N test is applied. If the test is applied to each instrument (unit) during every installation, the S/N specification is probably a realistic minimum representation of instrument sensitivity.
Conclusion
The combined benefits of GC-MS analysis have made the technique an extremely powerful tool for the analytical chemist. The evolution of the GC and MS in combination provides a high degree of qualitative and quantitative accuracy in substance identification. Sensitivity is now more easily achieved with these easy-to-use and affordable instruments.
While instrument sensitivity can be optimized by proper selection of instrument operating conditions and consumables, it is fundamentally related to the mass spectrometer design and manufacture (specific model). When evaluating instrument-sensitivity specifications, it is important to note the exact method by which signal and noise are measured in order to put the sensitivity specification in proper perspective.
