Sample Preparation in Mass Spectrometry
The Essential Role of Sample Preparation in Trace Quantitative Analysis
- Fig. 1: Schematic of an electron ionization source (A), which shows the production of ions from the GC eluent and their acceleration into the mass spectrometer under high vacuum. Ion current composed from high abundance fragment ions is generated by applying 70 eV electron energy (B). For any given molecule, as you increase electron energy, you will exceed the ionization potential of the analyte (a), begin to form increasing numbers of intact molecular ions in the threshold region (b), then generate significant fragment ions (c), and finally, reach a plateau where abundant fragment ions are generated for a wide array of molecule types (d).
- Fig. 2: Electrospray ionization takes place in an atmospheric pressure spray chamber (A). A capillary held at a high voltage creates a mist of charged droplets that desolvate and release ions into the gas phase. The ions are sampled by the mass spectrometer, the inlet for which is often set orthogonal to the spray axis in order to limit the introduction of neutrals into the high vacuum region. In the droplet (B), analytes compete for a limited number of charged sites at the droplet surface. Based on the Equilibrium Partitioning Model , the droplet can be divided into an electrically neutral bulk interior phase and a charged surface phase. Surface active analytes (ASA+) can easily outcompete non-surface active analytes (ANSA+) and the former will be preferentially observed in mass spectra.
The manipulation of gas phase ions based on their mass-to-charge (m/z) ratio using electric and magnetic fields forms the basis of mass spectrometry (MS). It is a very powerful, widely-used, and versatile technique for quantitative and qualitative analysis, and it is capable of being hyphenated with different separation formats. In this article, a brief discussion of ion formation in some of the common embodiments of LC-MS and GC-MS is given, along with reasons why the removal of interferences by various sample preparation techniques is crucial to ensure highest performance.
Gas chromatography - mass spectrometry (GC-MS) and liquid chromatography - mass spectrometry (LC-MS) are currently the gold standard for analytical determinations when an analyte of interest is present in a complex mixture (or "matrix"). However, these high efficiency techniques often require some additional sample preparation prior to sample introduction. In fact, for optimal work flows that exhibit the highest levels of reproducibility and sensitivity, sample preparation is absolutely essential. An inspiration for this article is the new Virtual Special Issue available in the Journal of Separation Science on "Sample Preparation for Mass Spectrometry," that has been co-edited by the author together with Prof. Michael Lämmerhofer from the University of Tübingen, Germany . Provided here are some general introduction to basics of mass spectrometry and why you need sample preparation. You can then take that information to the JSS website and enjoy the wide variety of detailed and exceptional research articles available for download (see QR code).
Before ions can be sampled and analyzed by a mass spectrometer, they must first be formed and transferred into the gas phase. Depending on the sample type (solid, liquid, or gas), there are a lot of choices to accomplish ion generation. Early studies by J. J. Thompson (the "Father of Mass Spectrometry") around the turn of the 20th century were performed by creating gas discharges in a confined apparatus. The resultant ions could then be accelerated through the application of an electric field and exposed to electric and magnetic fields, which could deflect and bend, respectively, the ion paths to degrees dependent on the m/z ratio of a particular ion.
An excellent and concise history of mass spectrometry can be found in the 2002 book by Grayson .
For most of the first half of the 20th century, mass spectrometry was performed by ion physicists. In the 1950s, with the coupling of MS to GC, mass spectrometry began its journey toward its modern place - in the laboratories of analytical chemists and user facilities meant to service and support a massive range of research endeavors. With GC, samples of semi-volatile and volatile analytes present (most often) in volatile solvents can be easily introduced for separation and mass spectral analysis. Various liquid and solid samples can be processed by extracting the analytes of interest into solution for subsequent introduction into a GC-MS or LC-MS instrument. Otherwise, solid samples can even be exposed to other means of ion generation, such as through laser irradiation. In the past decade, the means by which analytes can be converted to gas phase ions has expanded to an astonishing degree with the creation of new ambient ionization techniques .
A discussion of the mechanisms associated with all of the different ion generation techniques would go beyond the scope of this article, so the author would like to focus on a couple of common ones - electron ionization (EI) for GC-MS and electrospray ionization (ESI) for LC-MS. Taken together, these two sources still account for the majority of GC-MS and LC-MS applications reported in the literature. These ion sources are versatile and they do their job well for a wide range of different analyte types.
A schematic for EI is shown in Figure 1A. The effluent from the GC column is placed in close proximity to an EI filament, which bombards analytes with 70 eV electrons. This is a lot of energy, and it causes organic molecules to readily ionize and form fragment ions. An energy of 70 eV is chosen so that across different instrument platforms, for a variety of different analytes which have different ionization potentials, fragmentation is consistent. Because of this, mass spectral libraries containing hundreds of thousands of mass spectra for different compounds are available to aid in identification of signals for unknown compounds. In Figure 1B a plot showing the onset of ionization (generation of ion current) and fragmentation for a generic molecule is given. Initially, ionization occurs as the ionization potential of the molecule is exceeded. The number of ions, and eventually fragment ions, increases as electron energy is increased. A plateau region of consistent ion generation and fragmentation, regardless of molecule type (for the most part), is found once you reach 70 eV.
EI generates abundant fragments and is often referred to as a hard ionization technique. It is also fairly tolerant of matrix effects, in that if multiple constituents are ionized simultaneously, the presence of one would generally not heavily affect ion generation of the other. Of course there are limits to any such statements. For example, chemical ionization (CI) is another GC ionization technique where the EI source is flooded (a large excess relative to the analyte) with a reagent gas (e.g. methane or ammonia), and this vastly alters the ionization of analytes. Matrix effects are not a large problem for EI, but efficient separation of analytes prior to their ionization is still important for qualitative analysis and for confirming appropriate signals are being targeted for quantitative analysis. If multiple co-eluting analytes enter the ion source at the same time, then the mass spectrum observed will be a mix of all of the different ion signals. It would be very difficult to reference such a spectrum with the library to determine the identity of one of the components. Perhaps you might not even know that co-elution was complicating the spectrum. Efficient chromatographic separation is one key to obtaining clean spectra (if all of the compounds in the mixture can be resolved), but sample preparation is still very important for removal of interferences. These interferences might not only co-elute and convolute mass spectra, their decomposition in the hot injection port of the GC could compromise column and injection port liner lifetimes, which in turn can hurt reproducibility and sensitivity. Thus, for GC-MS, sample preparation is just as important for preserving the efficient operation of the GC system as it is for ensuring quality mass spectra are generated.
If we move to ESI, then the situation becomes more complicated. ESI, used commonly as an ionization source for LC-MS is a soft ionization source (analytes experience minimal fragmentation) and is very prone to matrix effects. Shown in Figure 2A is a general ESI source set-up. A mist of highly charged droplets is formed at the tip of a capillary, which is held at a high potential. The electric field at the capillary surface causes charge separation, and droplets are formed as the excess charge tries to traverse the atmospheric pressure source chamber to the inlet of the mass spectrometer. During the process, droplets lose solvent by evaporation, until they are forced to subdivide when the charge repulsion at the droplet surface overcomes the surface tension holding the droplets together. After several generations of subdivision, analytes species that have migrated to the surface of the droplets and acquired charge find it favorable to evaporate from the droplet surface, and eventually become gas phase ions. In Figure 2B is depicted the concept of the Equilibrium Partitioning Model , which well describes the competitive migration of species from the droplet interior to the charged droplet surface. The droplet can be envisioned as having two phases - a bulk electrically-neutral interior phase and a charged surface phase. Surface active species compete for a limited number of charged sites at the droplet surface. Measured ion intensity is proportional to droplet surface (not total droplet) concentration of an analyte. What this means is that if multiple species are in an ESI droplet together with your analyte of interest, the competition for droplet surface sites might limit the signal of the analyte of interest (incidentally, it is also possible that matrix effects can also increase signal intensity).
Limiting matrix effects in ESI is absolutely essential for sensitive and reproducible quantitative analysis by LC-MS. Plasma and serum contain abundant levels of phospholipids, and phospholipids are inherently surface active compounds. Thus, they can effectively outcompete your analyte of interest if they are in the same ESI droplet (the same can happen if you use soap to clean glassware, and there is residual soap present in your solutions; do not use detergents to clean glassware that will contain solutions for MS analysis!). Sample preparation in biomonitoring is absolutely essential to limit the co-elution of high abundance salts, lipids, and proteins with your analyte. Of course, the use of an internal standard can improve precision and account for signal suppression, but a) signal can still be lost, which compromises overall method sensitivity and b) a stable isotopically-labelled version of the analyte as an internal standard (Is it available? How much does it cost?) is absolutely essential so that it can experience exactly the same ESI droplet environment as the analyte of interest. Think about what else is in your sample, and then devise an appropriate sample preparation scheme to limit the introduction of matrix components into the LC-MS that might compromise performance.
Implications for Sample Preparation
The many ways to perform sample preparation can hardly be recounted in a single monograph. Simple and common approaches for treating liquid samples, particularly aqueous environmental or biological samples, include liquid-liquid extraction, solid phase extraction, and protein precipitation. For solid samples, solvent extraction or Soxhlet extraction are common. For gas samples, either the adsorption of components on a solid sorbent or direct collection in a canister are common. These all have different advantages and limitations with respect to time, cost, preconcentration factor, recovery, etc. Furthermore, different sample preparation approaches can be performed in tandem. On- versus off-line formats can be considered to alter the amount of human intervention needed. Overall, it is important to remember that while sample preparation is critical to optimizing precision, accuracy, and sensitivity for a method, it is also the step that often introduces the most error into an analysis. Thus, the more sample preparation steps that you incorporate, the worse will be the precision of the overall method. This is why when trace quantitative analysis methods are reported in the literature, they should include reporting of extensive method validation. This allows the end-user to easily compare the performance, advantages, and limitations of any work-flow. It is often improvement in these parameters, particularly if better sensitivity can be reached in shorter time with fewer costs, which distinguishes a new method and makes it popular.
In summary, mass spectrometry is an extremely versatile technique, but as the complexity of a sample increases, more effort needs to be placed into segregating the components of the mixture using sample preparation and high efficiency separation prior to mass analysis. When you look into the Special Virtual Issue in JSS, think about how the innovations in sample preparation have enabled new performance metrics for different sample types and analytes. This is an area of research where more innovation will always be welcome. There will always be new challenges to address.
 Grayson M. A.: Chemical Heritage Press, Philadelphia, PA. (2002)
 Monge M.E. et al.: Chem. Rev. 113, 2269-2308 (2013)
 Enke C.G.: Anal. Chem., 69, 4885-4893 (1997)