FT–MS to Provide Novel Insight into Complex Samples

Article

Special Issues

LCGC SupplementsSpecial Issues-07-01-2007
Volume 0
Issue 0
Pages: 18–26

More than 20 years passed after the introduction of Fourier transform–ion cyclotron resonance mass spectrometry (FT-MS) before advancements in electronics and computer technology enabled the development of practical, high-performance instruments. Modern analytical FT-MS instruments rely on sophisticated electronic circuitry and powerful computer software to achieve the dramatic resolving power and mass accuracy typical for the instrumentation. Here, the power of modern hybrid FT-MS instrumentation is discussed by demonstrating the capability of this instrumentation for selected applications such as the analysis of crude oil, intact protein, and fragile noncovalent complex samples.

More than 20 years passed after the introduction of Fourier transform–ion cyclotron resonance mass spectrometry (FT-MS) before advancements in electronics and computer technology enabled the development of practical, high-performance instruments. Modern analytical FT-MS instruments rely on sophisticated electronic circuitry and powerful computer software to achieve the dramatic resolving power and mass accuracy typical for the instrumentation. Here, the power of modern hybrid FT-MS instrumentation is discussed by demonstrating the capability of this instrumentation for selected applications such as the analysis of crude oil, intact protein, and fragile noncovalent complex samples.

The principles of Fourier transform–ion cyclotron resonance mass spectrometry (FT-MS) were introduced by Melvin B. Comisarow and Alan G. Marshall in 1974 (1,2). More than 20 years passed, however, before advancements in electronics and computer technology enabled the development of practical, high-performance instruments. Modern analytical FT-MS instruments rely on sophisticated electronic circuitry and powerful computer software to achieve dramatic improvements in the sensitivity and resolving power of MS.

Currently, FT-MS is considered the ultimate MS technique because it offers the highest resolving power and the best mass accuracy available in MS instruments. FT-MS is seen by many as the method of choice for tackling the most difficult analytical problems involving peptides, proteins, carbohydrates, and complex mixtures such as petroleum derivatives. Despite the reputation of ultrahigh performance, however, FT-MS instruments have an equally strong reputation for being expensive instruments that are difficult to use and maintain. Fortunately, this has been changing in recent years as a result of the increasing demand for user-friendly and reliable FT-MS instruments.

In addition to the increasing reliability, the footprints of the instruments are decreasing in size. This is mainly because of new magnet technology that provides FT-MS systems with smaller and lighter magnets. The modern magnets also have better active shielding. These recent advancements reduce the site requirements, making the instrumentation more accessible. Furthermore, the latest fashion in magnet technology, maintenance-free designs, further increases the user-friendliness of FT-MS systems. Although several maintenance-free designs exist, they share a main feature: These magnets do not require frequent cryogenic fills and are thus maintenance-free from the user's point of view.

Although FT-MS instrumentation has developed noticeably during the last 20 years with recent advances in electronics and computer technology, the rapid change is expected to continue. Magnet manufacturers are developing more compact magnets and the electronic components are increasing in power and efficiency while decreasing in size, thus allowing instrument manufacturers to work toward developing an affordable, benchtop FT-MS system that includes an ultrahigh-field superconducting magnet.

To illustrate the capabilities of the current hybrid electrospray FT-MS instruments, several distinct applications will be discussed in more detail. First, the benefit of high resolution in accurate studies of isotopic patterns and in obtaining chemical information of dense complex samples such as crude oil extracts will be discussed. Second, the advantage of applying the external ion isolation capability of hybrid FT-MS instruments to large intact protein analysis will be demonstrated along with the electron-capture dissociation (ECD) technique. The ECD technique has proven to be invaluable for gaining better understanding of the structure and function of proteins. The last application will discuss the ability to investigate the temperature dependency of the dissociation pathways and thermal stability of fragile noncovalent complexes.

High Resolution of FT-MS

One of the great advantages of FT-MS over other MS techniques is its ultrahigh resolution, which increases with the magnetic field strength applied. This, combined with the large dynamic range of the instruments, allows for detailed studies of molecular compounds. Isotopic patterns, for example, can be studied with great accuracy. Figure 1 illustrates a spectrum acquired for a dilution of caffeine. The analysis was performed with a Varian 901-MS electrospray QFT instrument(Varian, Inc., Palo Alto, California) with a 9.4-T superconducting magnet. The resolution of >1,000,000 was obtained with ease under the instrument's default conditions. As seen in Figure 1, the low abundance 13C (13CC7H10N4O2) and 15N (C8H10N315NO2) isotopes are resolved for caffeine (C8H10N4O2).

Figure 1

The high resolution of FT-MS instruments enables users to obtain additional and more distinct chemical information about the sample in shorter time than with other MS techniques. This capability is essential in providing further insight into several distinct applications, such as analysis of crude oil extracts.

High-Resolution Analysis of Heavy Crude Oil Extracts by Electrospray Ionization FT-MS

Oils — complex mixtures of aliphatic, olefinic, and polyaromatic hydrocarbons — are difficult to characterize. For example, difficulties arise from the analyte's tendency to fragment or undergo gas-phase reactions during ionization. Furthermore, the density of compounds in a single sample creates additional requirements for the data analysis and the instrumentation. Previous studies have shown that thousands of chemically distinct compounds can be identified in a single sample of heavy crude oil (3,4). Therefore, high resolution, accurate mass determination, and wide dynamic range of FT-MS are requirements for understanding the complete chemical composition of petroleum derivatives.

Data Analysis — Positive Ion Mode

Figure 2 shows the mass spectrum for the analysis of heavy crude oil in the positive ion mode. Mass errors of less than 0.7 ppm on average were achieved using external calibration. Accurate mass measurement combined with high resolving power greatly simplified the interpretation of the complex data. Here, samples of heavy crude oil were diluted to 1 mg/mL in 3:17:0.02 toluene–acetonitrile–acetic acid for positive ion analysis. A Varian 901-MS hybrid QFT instrument with a 9.4-T superconducting magnet and a nanoelectrospray source was applied. Samples were infused directly using a flow rate of 0.1 μL/min. Several mass spectra were summed to enhance the intensity of lower abundance species.

Figure 2

Figure 3a shows an expansion of the mass spectrum in Figure 2 over the range from m/z 290 to m/z 300. Here, the average resolving power is approximately 297,000. Due to the nature of the petroleum derivatives, the m/z spacing between the species can be as small as 3.3 mDa. The small region highlighted around m/z 296.2374 contains a total of six closely spaced peaks that would be demanding to detect with any other MS technique. An example of this is shown in Figure 3b, which illustrates the same 10-Da region of the mass scale with a resolution of only 5,000. This resolving power is typical for a lower performance mass spectrometer such as an ion trap or quadrupole instrument.

Figure 3

Whereas the resolving power of FT-MS is required to separate the individual species, the accurate mass measurement of FT-MS allows for unambiguous identification of the compounds. Figure 4 shows an expansion of the mass spectrum in Figure 3a over the region from m/z 296.14 to m/z 296.24. A total of six peaks were detected with an average mass accuracy of less than 1 ppm.

Figure 4

The "Elemental Composition Calculator" feature of the Omega data acquisition suite was used to determine the molecular formula for each peak. For example, the peak at m/z 296.1468 corresponds to a sulfur-containing species with the elemental formula of C19H22NS+. The presence of nitrogen is characteristic of petroleum samples analyzed in positive ion mode. In addition, considerable interest lies in the very low abundance peaks in the mass range from m/z 296.20 to m/z 296.22. These peaks provide valuable information, such as the type of contaminants present and the geographical origin of the sample.

Data Analysis — Kendrick Mass Visualization

The high complexity of the data provided by this type of analysis makes it challenging to interpret the acquired data quickly and efficiently in an illustrative and informative way. Kendrick mass analysis was developed to deal with the many issues arising when computing, storing, and comparing large amounts of mass data (5). This method scales all the measured m/z values relative to the m/z of CH2, which is assumed to be exactly 14.0000. In the original work, this scaling allowed the conversion of the elemental composition database from single elemental entries to series entries. By applying the Kendrick mass scaling, a database of organic compounds with masses up to 600 Da can be reduced from over 1.5 million to less than 100,000 entries.

Another advantage of the Kendrick mass scale is that the adjacent ions in a homologous series are separated by 14.0000 Da. Furthermore, they have the exact same mass defect, referred to as the Kendrick mass defect (KMD). In a plot of the KMDs as a function of nominal Kendrick mass, referred to as the Kendrick map, the homologous series are aligned horizontally according to the KMD of the series. The horizontal alignments can be seen in Figure 5, where the data from Figure 2 are analyzed and visualized in an automated fashion using the Varian Kendrick/van Krevelen map tool. The different symbols correspond to different hydrocarbon series. The contaminants of the sample are easily detected in the map. These single data points are plotted distinctly outside the lines of the series.

Figure 5

For convenience, the homologous hydrocarbon series can be labeled by the type and the class of the series. The elemental compositions of these compounds is Cc Hh Nn Oo Ss, where the lowercase c, h, n, o, and s are integers and h = 2c - Z. Z defines the type or the "equivalent carbon number" of the series. The class is determined by the elements in the compound, excluding carbon and hydrogen. As an example, CcH2c-13N is abbreviated as -13N, while CcH2c-29NO is abbreviated as -29NO. All theoretically possible hydrocarbon series can be seen in the "View Theoretical Kendrick Mass Defects" view of the Kendrick map tool.

The existence or lack of specific oil series provides essential information about the sample, such as the contaminants present or the geographical origin of the samples. The "View Detected Kendrick Mass Defects" feature, shown in Figure 6, lists all the detected series. The view shows the detected series labels on the left, while the window on the right illustrates the distribution of the members in the series over the mass range. The series -13N is highlighted as an example.

Figure 6

Data Analysis — Negative Ion Mode

As discussed, the basic nitrogen compounds in crude oil extract can be identified with positive ion analysis. Further information about the sample and more detailed analysis of the oxygen-containing compounds can be obtained by performing a complementary negative ion analysis. In addition, negative ion analysis yields information about the acid composition, in particular, the presence of napthathenic acids.

For negative ion analysis, samples were suspended in 3:17:0.1 toluene–methanol–ammonium hydroxide for a 1 mg/mL concentration. Samples were infused directly using a nanospray source, and the analysis was performed with similar conditions as the previously discussed positive mode analysis. Rapid switching between the positive and negative ion mode is performed by the data acquisition software, and the instrument is adjusted to the new polarity in less than 1 s.

Figure 7 shows an example of a crude oil extract analyzed in the negative ion mode. Similar to the positive mode, all of the detected species are singly charged. However, in the negative ion mode, the acid molecules have a tendency to form noncovalent complexes. Furthermore, this ionization mode allows for selective ionization. For example, it is possible to selectively ionize the carboxylic acids in the negative ion mode. The inset to Figure 7 illustrates a close-up of the broadband mass spectrum shown in Figure 7. The complexity of the data even in the negative mode can be observed. In addition, the accuracy of the mass measurement and obtained peak assignments are shown in the inset of Figure 7.

Figure 7

The current discussion demonstrates how the ultrahigh resolution, accurate mass determination, and wide dynamic range of FT-MS are required to resolve and identify the species in a complex crude oil extract. To fully characterize the petroleum extract samples, it is recommended to apply both positive and negative ion modes as complementary analysis techniques. Rapid switching between the complementary positive and negative ion modes allows for efficient characterization of both the nitrogen-containing and the oxygen-containing species. Furthermore, analyses performed using the atmospheric-pressure chemical and photoionization techniques will provide additional chemical information about the composition of the samples.

Efficient Intact Protein Analysis Using Hybrid FT-MS

Novel insight into the character of large proteins or protein complexes often can be obtained by studying their high-resolution mass spectra. The ability to identify the amount and nature of posttranslational modifications (PTM) or the ability to form a suggestion for a possible bonding pattern of an intact protein complex provides a better understanding of the structure and function of the protein.

Performing analysis of large intact proteins on an FT-MS instrument is, however, a demanding task. The difficulty is due mainly to the intrinsic effects, such as the space-charge effects or constructive–destructive interference of the FT-MS system. These effects can cause distortions in the collected transients, resulting in lower resolution and lower signal-to-noise ratio (S/N) in the mass spectra. The current hybrid FT-MS instruments provide an external ion isolation capability that can be utilized to reduce the intrinsic effects and to obtain higher resolution mass spectra for large intact proteins.

In the "improved method," the external multipole region, or ion trap, is utilized to isolate a single charge state of a large intact protein before excitation and detection. This results in higher resolution spectra and enhanced S/N. An additional advantage obtained is a more rapid acquisition speed. The high resolution is obtained acquiring a single mass spectrum while conventionally, the classical method of signal averaging over several spectra is applied. To perform a comparison of the conventional method and the improved method, both of the investigated proteins also were subjected to a more classical analysis.

Figure 8

Intact Protein Analysis with External Ion Isolation

Figure 8 shows a conventional mass spectrum of carbonic anhydrase II. Carbonic anhydrase II was suspended in 50:50:2 water–methanol–acetic acid to a 1 μM concentration and injected at a 1-μL/min rate through an electrospray ionization source. The Varian 901-MS instrument with a 7-T superconducting magnet was applied. In the classical analysis mode, 50 scans were collected and the signal averaged to obtain isotopically resolved envelopes. An approximate resolution of 91,000 was obtained for each of the charge states present in the spectrum. Applying the improved method, the approximate resolution of 91,000 for carbonic anhydrase can be obtained with a single scan, as shown in Figure 9. The close-ups of Figures 8 and 9 illustrate how comparable these two methods are. The improved method, however, has the advantage of an acquisition time of approximately 2% of that of the conventional one.

Figure 9

Figure 10 shows the application of the improved method to study the +47 charge state of intact bovine serum albumin (BSA). The intact BSA was suspended in 50:50:2 water–methanol–acetic acid to a 1 μM concentration and injected at a 1-μL/min rate through an electrospray ionization source. Again, applying the external ion isolation before the ion accumulation, the charge state is resolved isotopically by acquiring a single scan. All of the acquired data can be analyzed in an automated mode with the PeakHunter software.

Figure 10

This method utilizes the external ion isolation capability of hybrid FT-MS instruments. The intrinsic effects causing distortions in the transient can be reduced by applying the external ion isolation. With the improved method, higher resolution and enhanced S/N are obtained in a fraction of the analysis time when compared with the conventional method of signal averaging. This improved method also allows for more efficient fragmentation of the species of interest. Top-down sequencing can be performed with any of the in-cell fragmentation techniques to locate the PTM sites of the protein.

ECD for PTM Studies

The ability to identify the amount and nature of PTMs provides a better understanding of the structure and function of the protein. ECD has proven to be invaluable for top-down proteomics, specifically for the analysis of PTMs of proteins and peptides. Since the initial discovery of ECD in 1998 by Zubarev and colleagues (6), it has made a tremendous impact in driving MS to the forefront of structural analysis of biological molecules. In addition to being invaluable for top-down proteomics, ECD has become an important method for distinguishing isomeric or enantiomeric structures and analyzing the protein conformers in vacuo.

Although the mechanism of ECD is not yet thoroughly understood, it is thought to involve a recombination of an electron with protons to create very active hydrogen radical sites. The electrons captured by the peptide ion recombine with the protons in the ion. The energy released from the recombination causes the peptide ions to fragment. The fragmentation occurs primarily at the N-Cα bond, yielding an extensive series of c- and z-type fragments. The advantage of applying ECD fragmentation to the PTM studies is the nonselectivity of the process. The side-chains remain intact and the labile PTMs remain attached to the backbone during this nonergodic process.

There are two main reasons that FT-MS is an optimal instrument for ECD studies. First, FT-MS has the unique ability to obtain a very minimal difference between the energy of the multiply protonated ions and emitted electrons trapped in the ion cyclotron resonance cell. This maximizes the efficiency of the electron capture. Also, the ECD efficiency is highest for electron energies < 1 eV. Second, although the fragmentation proceeds at a very high rate, several milliseconds are required to ensure the electron capture by most precursor ions. The required trapping time exceeds the storage capability of most other types of mass spectrometers. Continuous efforts are made to create these conditions in other types of mass spectrometers, but currently ECD is exclusively an FT-MS technique.

Figure 11 shows an ECD spectrum of the +4 charge state of melittin. A 0.5 μM solution of honeybee venom melittin (C131H229N39O31; MW of 2846.46) was prepared in 50:50:2 water–methanol–acetic acid and infused at a flow rate of 2 μL/min via an electrospray source into a Varian 901-MS QFT-7T instrument. The injection corresponds to approximately 30 fmol of melittin per spectrum. The quadruply charged parent ion, occurring with an m/z value of 712 Da, was isolated with the quadrupole mass filter and fragmented using the Rapid-Fire ECD module with a 50-ms electron beam pulse. A single scan was acquired to obtain the spectrum shown in Figure 11. An advantage of the Rapid-Fire ECD module is that the infrared multiphoton dissociation (IRMPD) fragmentation could be performed simultaneously with ECD to yield complimentary b- and y-type fragments. The complexity of the ECD MS-MS spectrum is intensified because of the overlapping charge states of the fragments. An example of the overlapping c8 and z122+ fragments is shown in Figure 12, which is a close-up of the spectrum shown in Figure 11.

Figure 11

The acquired data are analyzed using the PeakHunter software. The de novo sequencer in PeakHunter is able to assign even very complex overlapping spectra accurately, as shown in Figures 11 and 12. The average mass error for the identified fragments was 0.98 ppm for this experiment.

Figure 12

PTMs such as phosphorylation, glycosylation, and sulfation usually are lost in collision-induced dissociation and IRMPD experiments. The rationale is that these groups are cleaved off easily because they form weak covalent bonds with the amino acid side chains. This is unfortunate because phosphorylation is a ubiquitous regulatory mechanism. It can result in the activation or termination of many important cellular events, including cell signaling growth and differentiation. Cancer, inflammatory diseases, metabolic disorders, and neurological diseases are among those in which protein phosphorylation plays an important role.

One of the remarkable features of ECD is that PTMs are retained in fragments. Figure 13 demonstrates this for phosphorylated angiotensin II, Asp-Arg-Val-pTyr-Ile-His-Pro-Phe. The doubly charged peptide at m/z 563.758 was selected by the rf/dc quadrupole and injected into the analyzer cell. The electron beam from the dispenser cathode was fired for 50 ms, and 20 scans were accumulated. The peak at m/z 1126.509 results from capture of an electron by the doubly charged ion. It is called a charge-reduction peak. Related to this is a fragment 81 Da lower at m/z 1045.464, which results from loss of H2PO3. The sequence at the top of Figure 13 shows that the arginine (R) residue second from the N-terminus holds the positive charge and accounts for an abundant series of c-type fragments. The site of phosphorylation is revealed by the large gap of 243.03 Da between c3 and c4: m/z 631.260 – m/z 388.230 = 243.03.

Figure 13

This agrees exactly with the expected mass for phosphorylated tyrosine. The phosphate group adds HPO3 (79.9663 Da) to the tyrosine side chain. The other gaps match up well with the expected amino acids. Notice that c6 is missing in the sequence, which is consistent with the proline at position 7.

Although ECD is a relatively new technique, it already has become an important tool in fragmentation studies performed with FT-MS instruments. The advantage of this rapid process is that it cleaves the proteins and peptides nonselectively, leaving the side-chains intact. This makes ECD invaluable for structural studies, including the determination of isomeric or enantiomeric structures of peptides and analysis of the PTM states of proteins.

Characterization of Self-Assembling Metallosupramolecular Complexes by Electrospray FT-MS

Self-assembly is a fascinating strategy for synthesis of complex species such as supramolecular assemblies (7–10). Novel applications in the fields of sensing, catalysis, and electrochemical or photochemical switches have been enabled recently by introducing functionalities such as redox activity or magnetic or luminescent properties into the assemblies. Introduction of metal ions into the supramolecular complexes makes these assemblies more versatile and more controllable with respect to their geometries. This is due to the higher directionality of a metal-ligand bond relative to a noncovalent bond and a larger selection of sometimes even redox-switchable coordination geometries. In addition, the stronger bonding interactions make these assemblies more durable as building blocks for larger systems. Yet another advantage is that the assemblies can be studied under more dilute conditions than many hydrogen-bonded architectures.

The MS studies of metallosupramolecular assemblies are not straightforward. In vacuo, the charge repulsion of the multiply charged species can overcome the bonding interactions in the complexes.

Typical prototypes of metallosupramolecular complexes are self-assembling squares that consist of bipyridine (bipy) edge ligands and Pt-diphenyphosphinopropane (dppp) complexes at the corners. These Pt-dppp squares can exist in equilibrium with triangular analogues, because the formation of triangles is entropically favored. Here, each sample was diluted in acetone to obtain an approximate 100 μM concentration for ESI analysis in the positive ion mode. All the discussed data are collected with a Varian 901-MS hybrid QFT-9T instrument. A typical flow rate applied was 0.5 μL/min with an accumulation time of 100 ms per spectrum. The IRMPD technique was applied to fragment selected species.

Figure 14 illustrates a typical mass spectrum obtained for an m/z range of 200–2500. The assignments of the peaks in Figure 14 are illustrated by a simplified notification explained in the insert of Figure 14. The Pt-dppp corners of the molecular structures are illustrated as circles, while the bipy edge ligands are illustrated as lines. TfO marks the attached CF3SO3 anion. As shown in Figure 14, both triangular and square geometries are detected as intact assemblies.

Figure 14

Expansions of the mass spectrum in Figure 14 are shown in Figures 15a and 15b. Complex areas of the mass spectrum in which there are overlapping distributions of different species are interrogated and identified automatically. This demonstrates that crucial information about the sample would be lost without the ultrahigh resolution, great sensitivity, and large dynamic range of an FT-MS instrument. The close-up shown in Figure 15c illustrates the advantage of the high resolution for detection of the higher mass species.

Figure 15

Isolation and IRMPD

The fragmentation pattern and dissociation pathway of selected complexes provides important structural information about the components in the sample. The species shown here consists of one corner, one side, and one TfO unit. In the full scan shown in Figure 14, this specific species is detected with an m/z value of 756 Da.

As mentioned, the IRMPD technique was applied to fragment this species. Although the power of the laser was kept constant, the amount of fragmentation was controlled by varying the length of the irradiation time. Figure 16 shows the effect of the time of the irradiation on the amount of fragmentation. Fragmentation of the complex at m/z 755 begins by a relatively easy loss of 149.95 Da. This corresponds to the loss of a TfO+ or CF3SO3H unit. Figures 16a and 16b show that the fragmentation of the daughter ions does not seem to occur before all the parent ions are consumed. Figures 16c and 16d illustrate the effect of lengthening the irradiation time even further. In addition to the length of the irradiation time, the fragmentation can be controlled by changing the power of the irradiation beam.

Figure 16

FT-MS is an advantageous technique to study typical prototypes of self-assembling metallosupramolecular complexes. Despite the fragile nature of the assemblies, the intact species such as the triangles and squares were detected. In addition, the isolation and fragmentation of the species can be performed in order to study the structural details and dissociation pathways of the complexes. The dissociation of a selected species using IRMPD was illustrated as an example. Furthermore, by controlling the temperature of the ion cyclotron resonance cell, the temperature dependency of the dissociation pathways and the stability of the assembly can be investigated.

Conclusions

Modern hybrid FT-MS instruments currently are considered the ultimate mass spectrometers. This is mainly because they offer the highest resolving power and the best mass accuracy provided by any MS instruments. The range of applications studied with FT-MS instrumentation has been increasing with user-friendliness and reliability, and FT-MS is seen by many as the method of choice for tackling the most difficult analytical problems involving peptides, proteins, carbohydrates, and complex mixtures such as petroleum derivatives.

The current discussion addresses several distinct applications in which the power of the hybrid FT-MS is demonstrated to provide either novel insight into the system or more chemical information about the system in shorter time. The benefit of ultrahigh resolution of FT-MS in studies of dense and complex samples was addressed. In addition, the advantage of the high resolution in intact protein studies was demonstrated. A method for rapid acquisition of high-resolution mass spectra and an exclusive fragmentation technique for detecting PTMs were discussed. Last, the application of the FT-MS for studies of fragile noncovalent complexes was addressed and the ability to investigate the temperature dependency of ion-ion reactions, dissociation pathways, or thermal stability noncovalent complexes was mentioned.

Although FT-MS instrumentation has developed rapidly in recent years, this trend is expected to continue with upcoming advances in electronics and superconducting magnet technology. These advances will provide means for further increasing the user-friendliness and reliability of the instruments. Furthermore, this allows instrument manufacturers to proceed with the dream of developing an affordable benchtop FT-MS.

Katianna A. Pihakari is with Varian FTMS Systems, Lake Forest, California.

References

(1) M.B. Comisarow and A.G. Marshall, Chem. Phys. Lett. 25, 182–283 (1974).

(2) M.B. Comisarow and A.G. Marshall, Chem. Phys. Lett. 25, 489–490 (1974).

(3) K. Qian and W.K. Robbins, Energy & Fuels 15, 1505–1511 (2001).

(4) K. Qian, R.P. Rodgers, C.L. Hendrickson, M.R. Emmett, and A.G. Marshall, Energy & Fuels 15, 492–498 (2001).

(5) E. Kendrick, Anal. Chem. 13, 2146–2154 (1963).

(6) R.A. Zubarev, Curr. Opin. Biotechnol. 15, 12–16 (2004).

(7) C.A. Schalley, T. Müller, P. Linnartz, M. Witt, M. Schäfer, and A. Lützen, Chem.—Eur. J. 8, 3538–3551 (2002).

(8) M. Engeser, A. Rang, M. Ferrer, A. Gutiérrez, H. tarik Baytekin, and C.A. Schalley, Int. J. Mass Spectrom. 255–256, 185–194 (2006).

(9) M. Mäkinen and P. Vainiotalo, Rapid Commun. Mass Spectrom. 18, 673–677 (2004).

(10) A. Sautter, B.K. Kaletas, D.G. Schmid, R. Dobrawa, M. Zimine, G. Jung, I.H. van Stokkum, L. De Cola, R.M. Williams, and F. Wurthner, J. Am. Chem. Soc. 127, 6719–6729 (2005).

Related Videos
Robert Kennedy
John McLean | Image Credit: © Aaron Acevedo
Related Content