An instrumentation perspective on reaction monitoring by ambient mass spectrometry

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

An instrumentation perspective on reaction monitoring by ambient mass spectrometry Xiaoxiao Ma, Sichun Zhang, Xinrong Zhang The elucidation of reaction mechanisms in inorganic and organic chemistry, and more recently in supramolecular chemistry, has attracted great interest. Better understanding of these mechanisms would enable people to work more wisely and innovatively in designing new reactions and synthesizing novel chemicals and materials. In addition, monitoring both substrates and reaction products during the reaction process aids the study of reaction dynamics, which will be helpful in optimizing reaction conditions. In comparison with other reaction-monitoring techniques [e.g., electron-spin resonance spectroscopy, ultraviolet spectroscopy and chemically-induced dynamic nuclear polarization], mass spectrometry is widely used due to its high sensitivity, specificity and scanning speed. In this review, we focus on ambient ion sources that have undergone rapid development in the past decade and greatly facilitate direct, rapid analysis of samples with little or no sample preparation. These ion sources, including but not limited to desorption electrospray ionization, extractive electrospray ionization, low-temperature plasma probe, electrospray-assisted laser desorption/ionization, and desorption/ionization on porous silicon, offer varied reaction times (several milliseconds to several minutes) and are controlled by corresponding experimental parameters. We discuss examples of different ion sources coupled with MS for reaction monitoring and highlight some of their unique applications in bioorganic and organic chemical reactions. We believe that, with this abundant toolbox of ion sources as sampling methods, we can be more confident in choosing the most effective technique to address problems of interest in reaction monitoring. ª 2012 Elsevier Ltd. All rights reserved. Keywords: Ambient mass spectrometry; Chemical reaction; Desorption electrospray ionization (DESI); Desorption/ionization on porous silicon (DIOS); Direct analysis; Electrospray-assisted laser desorption/ionization (ELDI); Extractive electrospray ionization (EESI); Ion source; Low-temperature plasma (LTP) probe; Reaction monitoring

1. Introduction Xiaoxiao Ma, Sichun Zhang, Xinrong Zhang* Beijing Key Laboratory of Microanalytical Methods and Instrumentation, and Department of Chemistry, Tsinghua University, Beijing 100084, PR China

*

Corresponding author. Tel.: +86 10 6278 7678; Fax: +86 10 6278 1690; E-mail: [email protected]. edu.cn

50

Most chemical reactions proceed via reactive intermediates through a complicated sequence of reaction steps. Better understanding of the details of these steps would enable chemists to control chemical reactions more efficiently and to find the optimal reaction conditions. The invention of electrospray ionization (ESI) has greatly expanded the applications of mass spectrometry (MS) in inorganic, organometallic and bioorganic chemistry. The massto-charge ratios (m/z) of ions and their corresponding isotopic patterns acquired by ESI-MS in the monitoring process are valuable information for intermediate structure analysis and the investigation of the mechanisms of organic reactions. Through study of the evolution of critical

reactive intermediate ions in the reacting process, detailed information on reaction pathways can be acquired. So far, a number of chemical and physical techniques have been employed to study reactive intermediates in organic chemical reactions. Electron-spin resonance (ESR) spectroscopy [1–5], ultraviolet/visible (UV/Vis) spectroscopy [6–8] and chemically-induced dynamic nuclear polarization (CIDNP) [9–11] have been used for the direct detection of radical intermediates. However, in general, these reaction-monitoring methods cannot simultaneously detect substrates, intermediates, and reaction products from reaction solutions. For example, to use UV/Vis spectroscopy for the monitoring of one species, a chromophore needs to be present in it in order to be detected. As a widely

0165-9936/$ - see front matter ª 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.trac.2011.12.004

Trends in Analytical Chemistry, Vol. 35, 2012

used technique, ESR can be used to detect directly only species with unpaired electrons (e.g., transition metals and free radicals). By contrast, MS is a more universal technique, since it directly measures the molecular weight of species of interest, and, by choosing appropriate ionization methods, most chemical compounds and free radicals can be ionized in a non-labeling approach. Conventional ionization methods [e.g., electron ionization (EI) and chemical ionization (CI)] are widely used for the detection of thermally-stable compounds with low boiling points. To be analyzed by CI/EI, the analytes must be vaporized to be ionized. Thermally-labile compounds (e.g., proteins) cannot be analyzed by EI/CI. EI-MS experiments are typically performed under 70 eV electronic energy. Since the ionization potential of common organic compounds is about 10 eV, under electron impact, the analyte molecule will dissociate through several channels to form fragment ions. Compared with EI, CI is softer because the ionizing species offer lower ionization potential and thus induce less fragmentation. EI and CI have been used in process control to monitor chemical reactions for decades. In common, the systems under study in most cases involve a complex mixture of compounds, so they favor soft ion sources, since components can be identified with high confidence and interpretation of mass spectra facilitated. However, though EI and CI are widely used in many fields, progress was greatly hindered by large biomolecules being difficult to ionize, due to their volatility and decomposition on heating. This problem was circumvented following the invention of ESI and matrix-assisted laser desorption/ionization (MALDI). The most prominent feature of ESI and MALDI is that they are ‘‘soft’’ ion sources that are capable of generating quasi-molecular ions of large biomolecules up to several hundred kDa and induces almost no molecular fragmentation. Fast atom bombardment (FAB) was once successfully used for the analysis of peptides, nucleotides and proteins with molecular weights up to 10,000 Da, and applications of flow FAB in reaction monitoring have been reported. But FAB gave way to ESI and MALDI after their invention. Another convenient feature of ESI is that it can be easily interfaced with liquid chromatography (LC) to analyze compounds eluted at different times. Still, due to its ‘‘soft’’ ionization, ESI is suitable for the simultaneous monitoring of reactants, reaction products, and shortlived reaction intermediates in the solution phase. Traditionally, reaction monitoring by ESI-MS is done off-line. An aliquot of reaction solution is taken from an ongoing reaction system, after which the reaction is stopped (e.g., diluted) to make a solution suitable for MS analysis. The reaction solution is diluted because MS is very sensitive. Without dilution, high concentrations of reactants and products would pollute the instrument. Off-line monitoring by ESI-MS can be efficient in reaction

Trends

monitoring. However, where very fast reaction kinetics and short-lived reaction intermediates are involved, off-line ESI-MS monitoring (time resolution: 0.21 min) becomes ineffective, and high time-resolution techniques are required. Following the invention of DESI in 2002 [12,13], a variety of ambient ion sources have been developed, including desorption/ionization on porous silicon (DIOS) [14–16], electrospray-assisted laser desorption/ionization (ELDI) [17–19], extractive ESI (EESI) [20–24], and low-temperature plasma (LTP) probe [25–29]. As well as vacuum no longer being required, another important aspect of ambient ion sources is that it is not necessary to introduce samples into a closed region or vacuum, as in MALDI and SIMS. This feature offers researchers great flexibility in the coupling or integration of reaction systems to the ion source of MS for on-line reaction monitoring, or even control of reaction kinetics. Although ambient ion sources are used for both sampling and ionization, reaction monitoring by them differs from monitoring reactions occurring inside the mass spectrometer where a high vacuum condition is present (e.g., in the ion trap). Fig. 1, which summarizes the number of publications related to reaction monitoring by MS from 2000 to 2010, shows the rapid development of reaction monitoring with the growth in ambient ion sources. It is very clear that there has been a steady increase in the number of publications since 2002, which is consistent with the invention of ambient ion sources [e.g., DESI and direct analysis in real time (DART)]. Reaction monitoring also has important applications in supramolecular chemistry. It enables researchers to gain insights into the dynamics of self-assembling processes. In this work, we overview the currently available reaction-monitoring methods, categorized according to the ion sources coupled to MS. In particular, we highlight examples of reaction monitoring representative of the unique applications of some ambient ion sources. We introduce the reaction-monitoring techniques by MS based on the approach to sample introduction. In both on-line and off-line sections, we describe ambient reaction-monitoring techniques in chronological order. For ion sources that can be implemented in either approach, we introduce them separately in both sections. Meanwhile, some variants of ESI have also been devised for specific research interests [e.g., cold spray ionization (ESI) was proposed for the study of non-covalent weak interactions, while high-pressure ESI was invented for reactions taking place in high-pressure conditions]. These ion sources, derived from the ‘‘primary’’ ESI, are also categorized based on their on-line or off-line characteristics (Table 1). Organic reactions take place after all the starting reactants are introduced into the reaction system and mixed together. The first step to use MS for the monitoring of ongoing chemical reactions is sampling from http://www.elsevier.com/locate/trac

51

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

Figure 1. The number of publications related to reaction monitoring by mass spectrometry from 2000 to 2010. The search string used in the ‘‘topic’’ field was ‘‘(reaction monitor* OR reaction intermediate* OR reaction mechanism OR real-time monitor* OR on-line monitor) AND mass spec*’’. Data was retrieved from ISI web of science.

the reaction system. Sampling can be performed intermittently or continuously, based on which reaction monitoring can be divided into the off-line and on-line versions. The samples collected are then delivered to the MS, ionized in the ion source and analyzed by the mass analyzer. Since the sample is directly extracted from the system where the reaction proceeds, the MS result is then representative of the state of the reaction system at the time of sampling and can be used to study reaction kinetics and process-control applications.

2. Off-line reaction monitoring Off-line reaction monitoring is suitable for the kinetic study of slow reactions (more than several tens of minutes) and long-lived reaction intermediates (life time >1 min). In this section, we describe reaction-monitoring examples through the use of CI/EI, FAB, ESI, DIOS, DART and MIMS. 2.1. Electrospray ionization Fig. 2 shows an ESI needle that is being operated in the positive-ion mode. As a versatile soft ionization technique, ESI has found wide applications in both off-line and on-line reaction monitoring. Those applications that are performed in off-line require extraction of small amounts of mixture from the system of an ongoing reaction at a series of time points. Because ESI is especially suitable for the detection of small molecules, highmolecular-weight biomolecules (e.g., biological polymers, proteins and peptides), and even relatively unstable non-covalent complexes, it is very useful in studies of

52

http://www.elsevier.com/locate/trac

interactions between proteins and drugs [45–47], proteins and aptamers [48], and antigen-antibody combinations [49,50]. The above examples have already demonstrated that ESI-MS can be used to monitor a variety of chemical reactions. However, besides chemical reactions, during which covalent bonds are formed, there are many other weak, non-covalent interactions between molecules (e.g., hydrogen bonding and p-p stacking). These weak forces are the driving force responsible for the association of molecules or assembly of them into regular structures. Researchers have now realized the critical importance of understanding how these processes occur {e.g., real-time observation of the self-assembly of hybrid polyoxometalates [51] and tracking crown-ether motion along a oligolysine chain [52], among many other early reports [53–56]}. However, complexes formed due to weak intermolecular interactions are very labile, and some will even dissociate due to the relatively harsh conditions in conventional ESI-MS. In view of this, Yumaguchi [57–61] and co-workers proposed the concept of cold-spray ionization (CSI), a variant of ESI operated at ca. 80–10C to allow ready characterization of labile organic species formed via non-covalent interactions. Fig. 3 shows the three typical CSI-MS set-ups. So far, CSI has been used to study labile organometallic compounds, including hostguest complexes, the multiple-link interlocking complex, the box-type complex, Grignard reagents and other organometallic compounds, and biomolecules (e.g., proline aggregation, nucleosides, and hyper-stranded DNAs). These advances have been reviewed elsewhere [60] and are not detailed in this review.

Trends in Analytical Chemistry, Vol. 35, 2012

Trends

Table 1. Characteristics of ambient ion sources for reaction monitoring Ion source

Time resolution

EI CI FAB On-line ESI Off-line ESI MALDI ELDI LTP probe EESI UASI DIOS DART

Off-line Off-line Off-line On-line: several milliseconds Off-line: 0.21 min Off-line On-line On-line On-line: 0.2 s On-line Off-line Off-line

ESI/CI related

Ref. [30] [31] [32,33] [34] [35–37] [38,39] [17] [28] [40,41] [42] [16] [43]

CI ESI ESI ESI related CI related ESI related ESI related CI related

Figure 2. The operation of an electrospray ionization (ESI) needle in positive-ion mode [44].

2.2. Matrix-assisted laser desorption/ionization The fundamental principles of MALDI involve two steps: (1) First is sample preparation. The analyte is dissolved in a solution that contains matrix to form a mixture, which is dried to remove the solvent and to form a ‘‘solid-solution’’ crystal for MALDI analysis [38,62–65]. The matrix should have a strong adsorption towards laser irradiance that serves to form a solid solution of the analytes and to transfer absorbed laser energy to them. (2) Second is laser ablation of the ‘‘solid-solution’’ crystal using a very short, but intense, laser pulse. Laser irradiation causes rapid heating of the crystal due to the accumulation of a large amount of energy in a very short time. As a result, this causes localized sublimation of the matrix crystal and expansion of the matrix and analyte into the gas phase. Though the ionization mechanism for MALDI is still a matter of intense debate, it is widely accepted that proton transfer in the solid phase before desorption or gas-phase proton transfer in the expanding plume of photoionized matrix molecules is involved in the ionization process [66,67].

The number of applications of MALDI in reaction monitoring reported can be divided into two categories. (1) First is the application of MALDI in monitoring solution-phase reactions. Kim and co-workers applied MALDI in monitoring succinylation of collagen, and that proved to be superior to the conventional method using sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDSPAGE). With the rapid development of materials science, newly emerging materials are frequently employed as MALDI matrix due to their high energy-transfer efficiency and analyte pre-concentration capability. More recently, in addition to these benefits, oxidized carbon nanotubes [68] and grapheme [69] were shown to be excellent MALDI matrices, since the interference in the low mass range was eliminated. For example, Jiang and co-workers reported enzyme-reaction monitoring and enzyme-inhibitor screening by MALDI using oxidized carbon nanotubes as the matrix [70]. They monitored the conversion of acetylcholine (ACh) to choline in the presence

http://www.elsevier.com/locate/trac

53

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

Figure 3. Cold-spray ionization (CSI) set-up: (a) prototype; (b) axial type; (c) orthogonal type (for details, see [60]).

of acetylcholinesterase within 1 h and screened acetylcholinesterase inhibitors at the rate of 3.3 s/candidate. Both ACh and choline have a molecular weight of less than 200 and show

54

http://www.elsevier.com/locate/trac

up in the low mass range of the mass spectra, which would be difficult to analyze using a common MALDI matrix. MALDI-MS was also used to study H/D exchange reactions [39].

Trends in Analytical Chemistry, Vol. 35, 2012

Trends

Figure 4. The dual-linker strategy coupled with an ionization tag.

(2) The other category is the application of MALDI in monitoring organic reactions on polymeric supports. Organic synthesis on polymeric supports offers advantages (e.g., the ability to use reagents in large excess and easy isolation of products). However, a serious problem associated with the use of solid supports in organic synthesis is the difficulty in monitoring the progress of reactions and characterization of resin-bound products. In view of this, Fitzgerald and co-workers described a method of derivatizing the beads with a photocleavable linker, and, after cycles of solid-phase peptide synthesis, they were able to monitor the reaction directly by MALDI [71]. But this method failed in the detection of fully protected peptides because of their low ionization efficiency. To solve this problem, Carrasco and co-worker proposed a dual-linker strategy coupled with an ionization tag [72]. The ionization tag was designed to enhance the ionization efficiency of the products. A photocleavable linker, which was cleaved upon laser irradiation, was used to connect the ionization tag with the resin. A chemicallycleavable linker was designed to cleave the products from resins after the reaction is completed. Fig. 4 shows this dual-linker strategy coupled with an ionization tag and the real scheme. Since MALDI analysis is performed under vacuum conditions in the MS and most chemical reactions are in atmospheric/pressurized conditions, reaction monitoring by MALDI is off-line.

2.3. Desorption/ionization on silicon DIOS distinguishes itself from MALDI in that it employs no matrix to transfer the energy of the laser to the analytes [15]. Instead of using a small-molecule matrix, porous silicon is utilized to trap analytes, and also acts as the energy-transfer medium to vaporize and to ionize the trapped analytes upon laser irradiation. Compared with MALDI, there is little interference in DIOS in the low molecular range arising from the matrix, so DIOS is also suitable for the analysis of small molecules. Siuzdak and co-workers, who invented DIOS, reviewed the development of this technique with respect to its application in small-molecule characterization, quantitative analysis, reaction monitoring, protein identification, and functional characterization of proteins [73]. As an example, the conversion of ACh to choline in the presence of acetylcholinesterase was monitored for 250 min. Because DIOS chips were directly placed on a commercial MALDI plate, reaction monitoring was performed in an off-line approach by shooting a laser beam to the sample at different time points. In this way, MS spectrum snapshots representative of the reaction progress were obtained. By extracting the intensities of reactants and/or reaction products from these MS spectra, the changes of their intensities over time (i.e. the conversion of ACh to choline) were reconstructed, as plotted in Fig. 5d. We can see very clearly from Fig. 5d that the conversion of ACh to choline is a slow process, which does not require a high time-resolution monitoring method to study reaction kinetics. It is also worth http://www.elsevier.com/locate/trac

55

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

Figure 5. Configuration of the desorption/ionization on porous silicon (DIOS) chip (a, b, c): (a) on a matrix-assisted laser desorption/ionization (MALDI) plate, four porous silicon plates are placed. Each contains photopatterned spots/grids [15]; (b) silicon-based laser desorption/ionization [15]; (c) cross-section of the porous silicon and the surface functionalities after hydrosilylation; R represents phenyl/alkyl chains [15]; and, (d) plot of the conversion of acetylcholine (ACh) to choline at 25C at an initial concentration of 200 mM substrate and a concentration of 40 pM enzyme [16].

pointing out that DIOS involves a laser, which is shot into the sample repeatedly. Intrinsically, DIOS is an offline reaction-monitoring method, though the time resolution can be improved by shortening the time intervals between two consecutive shots. However, it seems that another ionization method can also be used to monitor the conversion of ACh to choline, so this example does not represent the only application of DIOS in this field.

2.4. Direct analysis in real time The DART ion source was developed in 2005 by Cody et al. [74]. The layout of DART is shown in Fig. 6. The nitrogen gas is introduced into a chamber containing a needle electrode and a grounded electrode. A high voltage of several kilovolts initiates a discharge in the chamber, generating a plasma that contains cations, anions, and metastable ions. To avoid signal loss due to

Scheme 1. Reaction of excited-state helium atoms with atmoshpheric water to form protonated water clusters.

56

http://www.elsevier.com/locate/trac

Trends in Analytical Chemistry, Vol. 35, 2012

Trends

Figure 6. (a) Illustration of the direct analysis in real time (DART) ion source coupled to mass spectrometry (MS); and, (b) gas-ion separator equipped with a vacuum pump to facilitate efficient transfer of ions into MS and maintain stable vacuum conditions [75].

ion-ion combination, after the two electrodes, another set of perforated electrodes are set to repel both cations and anions to allow only neutral species to exit. Before exiting, these neutral species are heated by the heater outside the transfer tube to desorb the analytes more easily. The key processes that occur in a DART ion source include: (1) Desorption: thermal desorption of both liquid and solid analytes is realized through the use of hot gases that are heated by the heater before they impact with the analytes. Inside the hot gases are metastable species, generated from high-voltage discharge and responsible for ionization;

(2) Ionization: if helium gas is used as the discharge gas, long-lived 23S helium atoms in the excited state are formed, with an internal energy of 19.8 eV. The interaction of 23S helium atoms with analyte molecules directly yields radical cations M+Æ, through the well-known Penning process. The mechanism responsible for the positive ions formed in DART includes reaction of excited-state helium atoms with atmospheric water to form protonated water clusters (Scheme 1), and subsequent transfer of protons to the analyte molecules to form [M+H]+ ions. Reports of reaction monitoring using DART are rare. There is one study reporting the monitoring of synthetic

Figure 7. The on-line microreactor used for reagent mixing and subsequent reaction monitoring.

http://www.elsevier.com/locate/trac

57

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

Figure 8. Experimental set-up of stopped-flow electrospray ionization mass spectrometry (ESI-MS) [80]. Notation: S1 and S2 syringes for pulsed sample injection; M mixer; S3 syringe. S3 delivers solvent that pushes reaction mixture through the fused silica capillary (C) to the ESI source after the reaction tube (R) has been filled with fresh reaction mixture; MS: quadrupole mass spectrometer; HS1, HS2: hard stoppers. Arrows indicate the directions of liquid flow.

Figure 9. (A) A capillary mixer with adjustable reaction chamber volume. The high time resolution of this device is realized through adjustment of the reaction chamber between the capillary connected to syringe 1 and the electrospray ionization (ESI) source. By gradually pulling back the capillary, chamber volume is increased from 0 mL and reaction time is also gradually increased from 0 s. (B) Kinetic study of ubiquitin refolding by monitoring its molecular ions of four different charge states.

58

http://www.elsevier.com/locate/trac

Trends in Analytical Chemistry, Vol. 35, 2012

Trends

Figure 10. (a) Design for on-line electrospray ionization (ESI) analysis of reaction under high pressures. 1: high-pressure valve; 2: high-pressure split valve; 3: reactor assembly with fishing tube and sonotrode; 4: syringe for electrospray solvent; 5: ultrasonic piezoelectric transducer with horn; 6: mass spectrometer ion source; 7: fishing tube; 8, sample path and direction. (b) Reactor and MS assembly.

transformation (e.g., N-methylation of indole by using this ion source) [43]. The monitoring process is done off-line by analyzing 2 lL of diluted reaction mixture over 16 h. This off-line feature of DART is essential because it cannot continuously deliver samples to the ion source, since it uses a probe to pick up a small amount of sample for analysis. However, DART has potential for on-line reaction monitoring if an appropriate method for continuous sample introduction can be developed.

3. On-line reaction monitoring To study fast reaction kinetics or to detect short-lived reaction intermediates, on-line methods for MS monitoring are necessary to acquire real-time information from the reaction system. Ion sources serving this purpose include on-line ESI, ELDI, ultrasound-assisted spray ionization (UASI) and the LTP probe, which involve continuous sampling and ionization of a small portion from the system. Practically speaking, all these sources can be used to study reaction kinetics at the millisecond timescale. ESI, ELDI and UASI share some important

features (e.g., they all have the solution under study sprayed), while the LTP probe is a plasma-based ion source. The LTP probe is more suitable for the detection of small non-polar and volatile molecules rather than large biomolecules. Although ESI and EESI are both ‘‘soft’’ ion sources and fall into the same category, the extent of their ‘‘softness’’ differ. As shown in sub-section 3.3, radical species invisible in ESI can be detected in EESI [76]. As such, we finally conclude that EESI is a softer ionization method (compared to ESI) and offers a milder condition for the survival of instable species. For all the ion sources included in this section, an aliquot of solution is continuously transferred from the reaction system to the MS for real-time analysis. Manual transfer of samples from the reaction system, as commonly done for off-line reaction monitoring, is completely eliminated. As a result, on-line MS monitoring techniques can monitor reactions at the millisecond or even sub-millisecond time scale. 3.1. Electrospray ionization To use ESI for on-line and real-time reaction monitoring, Metzger and co-workers devised an on-line microreactor to have two reagents pumped through two separate lines

http://www.elsevier.com/locate/trac

59

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

Figure 11. Liquid electrospray-assisted laser desorption/ionization (ELDI) system [17]. (a) The desorption and ionization of molecules dissolved in liquid during liquid ELDI. (b, c) Photos of the liquid ELDI system displaying (b) the electrospray plume and (c) analyte droplets mixing with the electrospray plume.

of tubing, as shown in Fig. 7. Capillaries and HPLC fittings were used to assemble such a mixer, which offered a reaction time as short as 0.7 s. After mixing, the reaction could be monitored from the ESI spray capillary and, under continuous flow conditions, the mass spectra were acquired. To lengthen the reaction time, they only needed to use a longer ESI spray capillary. Using this device, they detected transient radical cations in electron-transfer-initiated Diels-Alder reactions [77], the catalytically active 14-electron ruthenium intermediate directly from solution [78], and homogeneously catalyzed Ziegler–Natta polymerization of ethene [79], among many other important organic reactions. A prominent advantage of on-line ESI-MS analysis is that the delay between sampling and displaying mass spectra on the MS is eliminated. This allows the user to acquire an immediate profile of the ongoing reaction system, which is especially helpful for monitoring shortlived species and useful in elucidating reaction mecha-

60

http://www.elsevier.com/locate/trac

nisms. This aim is commonly fulfilled by stopped-flow ESI-MS. Shown in Fig. 8 is the experimental set-up of stopped-flow ESI-MS reported by Konermann and coworkers [80]. Briefly, plungers S1 and S2 are operated by computer-controlled drivers to deliver the two reagents to mixer M. When S1 and S2 are in operation, hard stoppers HS1 and HS2 are opened to allow flow of fresh reagent mixture. The injection time of the mixture is 100 ms. Afterwards, HS1 and HS2 are closed and plunger S3 is opened to deliver a 100 lL /min water flow to carry the ‘‘aliquot’’ of reagent mixture to the ESI-MS. The time resolution of this stopped-flow ESI-MS device is 2.5 s (compared with the ms time resolution of stoppedflow optical spectroscopy) [80]. Taking into consideration the particular importance of studying the reaction kinetics of a wide range of chemical and biochemical systems, techniques that improve time resolution are very desirable (i.e. s to ms, or even ls). To reach ms time resolution, Konermann and coworkers proposed a capillary mixer involving a reaction

Trends in Analytical Chemistry, Vol. 35, 2012

Trends

Figure 12. Set-up for ultrasound-assisted spray ionization coupled to mass spectrometry (UASI-MS).

Figure 13. Extractive electrospray ionization (EESI) coupled to mass spectrometry (EESI-MS). The ionizing spray interacts with the sample spray in front of the MS inlet to transfer charges to ionize the analytes the sample spray so that they can subsequently be detected.

chamber with adjustable volume and achieved ms time resolution to study reaction dynamics in the ‘‘kinetic’’ mode. Fig. 9A shows this device [81]. Fig. 9B is a kinetic study of ubiquitin refolding using this reaction-monitoring device by monitoring its protonated molecules in four different charge states. Please note that the x scale (average reaction time) is 0–1.4 s and the ESI-MS intensity points plotted in Fig. 9B It can be easily seen that the time resolution is high (ms scale).

Following such on-line reaction-monitoring techniques based on ESI, Plattner and co-workers explored the possibility of on-line ESI analysis of reactions under high pressures by making some improvements [34]. Fig. 10 shows their experimental set-up. Briefly, CO2 was introduced into the reaction system via higher pressure valve 1 to maintain a high-pressure environment for the reaction to proceed. To record a mass spectrum, valve 2 was opened for 1 min to allow the reaction mixture to be

http://www.elsevier.com/locate/trac

61

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

Figure 14. Set-up for extractive electrospray ionization (EESI) reaction monitoring.

Figure 15. Concept and set-up of extractive electrospray ionization (EESI) coupled to mass spectrometry (EESI-MS) for monitoring chemical reactions in viscous liquid [94].

driven out by the high pressure. The reaction mixture was then mixed with electrospray solvent delivered by syringe 4 to make the solution suitable for ESI-MS analysis. Using this assembly, they monitored the dehydration of hydroxenin monoacetate and the hydrogenation of 5-norbornene-2-carbonitrile using Pd/ CaCO3 as the catalyst. However, the time resolution of this work was unclear. 62

http://www.elsevier.com/locate/trac

3.2. Electrospray-assisted laser desorption/ionization Electrospray-assisted laser desorption/ionization (ELDI) combines the features of ESI and laser desorption to allow direct, sensitive, rapid detection of small organic molecules and large biological molecules [18,19,82–85]. The basic ionization mechanism of ELDI includes release of neutral-analyte molecules from the solid surfaces into the gaseous phase and subsequent ionization by charged

Trends in Analytical Chemistry, Vol. 35, 2012

Trends

Figure 16. Reaction-monitoring procedures using a low-temperature plasma (LTP) probe [28].

species in the ESI plume. Unlike other ambient desorption/ionization techniques, ELDI has desorption and ionization as two separate processes, which offer independent control and optimization over the composition of the sample and the ESI plume. This ionization technique was invented by Shiea et al., who subsequently demonstrated its application by coupling ELDI with thin-layer chromatography (TLC) for the detection of separated chemical compounds without any sample preparation [18]. Afterwards, it was observed that by adding carbon powders into the reaction system, ELDI can be utilized for direct analysis of liquid samples and continuous monitoring of chemical reactions if the surface of the solution is continuously disturbed by the UV laser beam [17]. The role of carbon powders is to transfer the energy of the laser beam to the solvent and analyte molecules, liberating them from the bulk system, and no viscous liquid medium [as used in surface-assisted laser desorption/ionization (SALDI)] is needed. Fig. 11 shows ELDIMS for reaction monitoring. It was demonstrated that ELDI is suitable for the on-line monitoring of chemical reactions, including fast complexation reactions and relatively slow protein enzymolysis [17,86]. However, although the monitoring was continuous, the time resolution of this method was not reported. 3.3. Ultrasound-assisted spray ionization More than a decade ago, ultrasound-assisted ESI [87] was first demonstrated through the use of an ultrasonic transducer for the enhancement of the nebulization efficiency of analytes and then subjecting them to ESI. More recently, an ambient ion source, termed sonic spray ionization (SSI), was invented [88–91]. In SSI, a pneumatic gas (e.g., nitrogen) flowing at a high speed

through the space between two concentric capillaries nebulizes the analyte solution in the inner capillary. Small organic and drug molecules can be successfully detected by SSI. However, for the detection of large molecules (e.g., proteins), a high DC voltage needs to be applied to the analyte solution to increase the charge density therein for the production of charged droplets and generation of multiply-charged protein ions. The most important motive for the design of UASI lies in the fact that both ultrasound-assisted ESI and SSI need postionization for the detection of large biomolecules, so an ion source that eliminates post-ionization would simplify the experimental set-up and make analyses more convenient. Briefly, in UASI, a capillary with a tapered tip is inserted into a solution under ultrasound. The ultrasonicator provides sufficient energy to drive the solution along the capillary tube to the tip, resulting in the generation of sonic spray [92]. In their first publication on UASI-MS, Chen et al. discussed the possible mechanism for charge accumulation and ion generation in the ionization process. Also, it is straightforward to envisage that, since UASI is an ion source that continuously extracts samples from a reservoir, it can be applied for reaction monitoring, especially monitoring reactions under ultrasound, which has two roles: (1) on the one hand, it is part of the UASI-MS system, which is needed for nebulization and ionization; and, (2) on the other hand, it is the driving force in accelerating chemical reactions. There is therefore no compromise in the application of UASI-MS in ultrasound-assisted reactions. In their work, Chen et al. chose a trans-esterification reaction, called a Zemplen reaction, to demonstrate this very interesting concept [42], as shown in Fig. 12.

http://www.elsevier.com/locate/trac

63

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

3.4. Extractive electrospray ionization Fig. 13 illustrates EESI [93], whose name reflects its specific features: by interaction of the ionizing spray (ESI spray) and the sample spray, analytes in the sample can be ‘‘extracted’’ into the ionization plume where charges are transferred to them to ionize them. Compared with ESI, EESI offers long-time signal stability and better tolerance to complex samples without sample preparation. Reaction-monitoring and mechanism elucidation using EESI-MS has also been reported. Zenobis group explored this by direct coupling of a three-necked flask with EESI [40]. As shown in Fig. 13, the reaction mixture was sampled by continuously passing nitrogen gas through the headspace of the flask. In this work, it was the headspace gas (rather than the bulk solution) that was monitored. The utility of the technique was demonstrated by monitoring a Michael addition reaction. However, a major problem was that it may not have faithfully reflected the changes in the bulk solution over time, as the headspace gas is presumed to be in equilibrium with the bulk solution and is rapidly flushed out and renewed. The time resolution of EESI-MS was reported to be no more than 0.2 s in the experimental set-up shown in Fig. 14. Subsequently, the same group explored EESI for monitoring reactions in a highly viscous medium [94]. The particular usefulness of EESI for this purpose rests with the phenomenon called microjetting [95] (Fig. 11), during which the bubbles generated due to the introduction of nitrogen gas burst at the liquid-air interface, resulting in the creation of microdroplets. Since EESI has been reported to analyze aerosols, it should also be applicable in the analysis of these microdroplets, thereby providing rich chemical information about the viscous system. The authors successfully monitored the conversion of fructose to 5-hydroxymethylfurfural at 80C in 1-ethyl-3-methylimidazolium chloride (EMIMCl), which is viscous (Fig. 15). Liquids with a viscosity ranging from a few cP to 300,000 can be rapidly characterized by this simple, but robust, method without any sample preparation. 3.5. Low-temperature plasma probe With the same underlying principle as dielectric barrier discharge ionization (DBDI), the low-temperature plasma (LTP) probe was developed in 2009 as an improved version of DBDI in that it allows direct interaction of the plasma blown out of the discharge tube with the sample to be analyzed, thereby facilitating convenient analysis of samples of any shape and size. So far, the LTP probe has been applied to a large variety of compounds, including explosives [96,97], drug tablets and drugs of abuse [26,98], milk powders adulterated with melamine [99], and non-destructive imaging of works of arts [27] without any sample preparation.

64

http://www.elsevier.com/locate/trac

Taking into consideration the simple design and low temperature, Zhangs group succeeded in applying the LTP probe to direct monitoring of organic reactions in the ambient conditions [28]. Three classic organic reactions were selected as representative cases to illustrate the feasibility of the proposed method (as shown in Fig. 16). Reaction monitoring using an LTP probe may be more suitable for certain kinds of reactions rather than others, depending on the method sensitivity towards a specific category of compounds. Current studies show that the LTP probe generally offers high sensitivity for compounds with a relatively high vapor pressure (e.g., esters and amines). But other experiments also revealed that explosives, the vapor pressures of which are very low, and active ingredients in tablets can also be easily detected. The mechanism responsible for the desorption/ionization in the LTP probe is still vague and needs more supporting data to study it. We believe that, with better understanding of this mechanism, a whole category of chemical reactions can be effectively monitored by this simple, yet convenient, LTP probe in the ambient environment. 4. Conclusions and perspectives In this review, we have summarized the applications of different reaction-monitoring methods using a variety of ion sources from the conventional ESI to the more recently emerged DIOS and the LTP probe. Existing studies have demonstrated the great usefulness of conventional ion sources in reaction monitoring, especially in reaction kinetics studies and interception of reaction intermediates for the elucidation of reaction mechanisms, and it is believed that ESI will continue to play an important role in this fast-expanding field. However, reaction monitoring also benefits from the rapid development of ambient ion sources or novel sample-introduction methods. For example, because of the interesting microjetting phenomenon, the real-time monitoring of reactions in viscous liquids was, for the first time, successfully carried out by EESI without any sample pretreatment. The obstacles frequently encountered in ESI upon analysis of unusual systems can therefore be tackled by using novel ion sources. The objective of this review was to give a brief introduction of the currently available ion sources and to show their particular applications. As demonstrated, the large family of MS reaction-monitoring techniques offers much more flexibility and convenience in effectively addressing the problem of concern. We believe that the rapid advancement of ion sources, combined with sample-introduction techniques, will continue to drive the development of this very important field.

Trends in Analytical Chemistry, Vol. 35, 2012

Acknowledgements This work was funded by the National Natural Science Foundation of China (No. 21027013), the National High Technology Research and Development Program of China (No. 2009AA03Z321), and the Initiative Scientific Research Program of Tsinghua University. The authors also acknowledge Prof. Huanwen Chen from East China Institute of Technology for his helpful suggestions on the manuscript. References [1] C. Felby, B.R. Nielsen, P.O. Olesen, L.H. Skibsted, Appl. Microbiol. Biotechnol. 48 (1997) 459. [2] B.D. Flockhart, K.J. Ivin, R.C. Pink, B.D. Sharma, J. Chem. Soc., Chem. Commun. (1971) 339. [3] R.B. Ingalls, L.A. Wall, J. Chem. Phys. 35 (1961) 370. [4] B. Kraeutler, C.D. Jaeger, A.J. Bard, J. Am. Chem. Soc. 100 (1978) 4903. [5] R. Leardini, A. Tundo, G. Zanardi, G.F. Pedulli, Tetrahedron 39 (1983) 2715. [6] H. Beckers, H. Willner, M.E. Jacox, Chemphyschem 10 (2009) 706. [7] L.A. Chen, K. Sung, Org. Lett. 11 (2009) 3370. [8] F. Marchetti G. Pampaloni, C. Pinzino, J. Organomet. Chem. 696 (2011) 1294. [9] M. Goez, J. Rozwadowski, B. Marciniak, J. Am. Chem. Soc. 118 (1996) 2882. [10] M. Lehnig, Arch. Biochem. Biophys. 368 (1999) 303. [11] O.B. Morozova, S.E. Korchak, R.Z. Sagdeev, A.V. Yurkovskaya, J. Phys. Chem. A 109 (2005) 10459. [12] R.G. Cooks, Z. Ouyang, Z. Takats, J.M. Wiseman, Science (Washington, DC) 311 (2006) 1566. [13] Z. Takats, J.M. Wiseman, B. Gologan, R.G. Cooks, Science (Washington, DC) 306 (2004) 471. [14] Z.X. Shen, J.J. Thomas, C. Averbuj, K.M. Broo, M. Engelhard, J.E. Crowell, M.G. Finn, G. Siuzdak, Anal. Chem. 73 (2001) 612. [15] J. Wei, J.M. Buriak, G. Siuzdak, Nature (London) 399 (1999) 243. [16] J.J. Thomas, Z.X. Shen, J.E. Crowell, M.G. Finn, G. Siuzdak, Proc. Natl. Acad. Sci. USA 98 (2001) 4932. [17] C.Y. Cheng, C.H. Yuan, S.C. Cheng, M.Z. Huang, H.C. Chang, T.L. Cheng, C.S. Yeh, J. Shiea, Anal. Chem. 80 (2008) 7699. [18] S.Y. Lin, M.Z. Huang, H.C. Chang, J. Shiea, Anal. Chem. 79 (2007) 8789. [19] I.X. Peng, R.R.O. Loo, J. Shiea, J.A. Loo, Anal. Chem. 80 (2008) 6995. [20] H. Chen, S. Yang, A. Wortmann, R. Zenobi, Angew. Chem., Int. Ed. Engl. 46 (2007) 7591. [21] H.W. Chen, A. Wortmann, W.H. Zhang, R. Zenobi, Angew. Chem., Int. Ed. Engl. 46 (2007) 580. [22] K. Chingin, H.W. Chen, G. Gamez, L. Zhu, R. Zenobi, Anal. Chem. 81 (2009) 2414. [23] J.H. Ding, H.W. Gu, S.P. Yang, M. Li, J.Q. Li, H.W. Chen, Anal. Chem. 81 (2009) 8632. [24] W.S. Law, R. Wang, B. Hu, C. Berchtold, L. Meier, H.W. Chen, R. Zenobi, Anal. Chem. 82 (2010) 4494. [25] J.D. Harper, N.A. Charipar, C.C. Mulligan, X.R. Zhang, R.G. Cooks, Z. Ouyang, Anal. Chem. 80 (2008) 9097. [26] Y.Y. Liu, Z.Q. Lin, S.C. Zhang, C.D. Yang, X.R. Zhang, Anal. Bioanal. Chem. 395 (2009) 591. [27] Y.Y. Liu, X.X. Ma, Z.Q. Lin, M.J. He, G.J. Han, C.D. Yang, Z. Xing, S.C. Zhang, X.R. Zhang, Angew. Chem., Int. Ed. Engl. 49 (2010) 4435. [28] X.X. Ma, S.C. Zhang, Z.Q. Lin, Y.Y. Liu, Z. Xing, C.D. Yang, X.R. Zhang, Analyst (Cambridge, UK) 134 (2009) 1863.

Trends [29] Z. Xing, J.A. Wang, G.J. Han, B. Kuermaiti, S.C. Zhang, X.R. Zhang, Anal. Chem. 82 (2010) 5872. [30] S.L. Koontz, M.B. Denton, Int. J. Mass Spectrom. Ion Processes 37 (1981) 227. [31] M.S.B. Munson, F.H. Field, J. Am. Chem. Soc. 88 (1966) 2621. [32] L.A. Smith, R.M. Caprioli, Biomed. Mass Spectrom. 11 (1984) 392. [33] H.T. Kalinoski, U. Hacksell, D.F. Barofsky, E. Barofsky, G.D. Daves, J. Am. Chem. Soc. 107 (1985) 6476. [34] R.A. Paz-Schmidt, W. Bonrath, D.A. Plattner, Anal. Chem. 81 (2009) 3665. [35] F.M. Romero, R. Ziessel, A. DupontGervais, A. VanDorsselaer, Chem. Commun. (1996) 551. [36] R.B. Cole, Abstr. Pap. Am. Chem. Soc. 209 (1995) 115. [37] U.A. Mirza, B.T. Chait, H.M. Lander, J. Biol. Chem. 270 (1995) 17185. [38] R.C. Beavis, T. Chaudhary, B.T. Chait, Org. Mass Spectrom. 27 (1992) 156. [39] N.B. Kasper, D. Rand, M.M. Nedertoft, T.J.D. Jørgensen, Anal. Chem. (2011). [40] L. Zhu, G. Gamez, H.W. Chen, H.X. Huang, K. Chingin, R. Zenobi, Rapid Commun. Mass Spectrom. 22 (2008) 2993. [41] C.A. Marquez, H.Y. Wang, F. Fabbretti, J.O. Metzger, J. Am. Chem. Soc. 130 (2008) 17208. [42] T.Y. Chen, C.S. Chao, K.K.T. Mong, Y.C. Chen, Chem. Commun. 46 (2010) 8347. [43] C. Petucci, J. Diffendal, D. Kaufman, B. Mekonnen, G. Terefenko, B. Musselman, Anal. Chem. 79 (2007) 5064. [44] P. Kebarle, U.H. Verkerk, Mass Spectrom. Rev. 28 (2009) 898. [45] C.S. Allardyce, P.J. Dyson, J. Coffey, N. Johnson, Rapid Commun. Mass Spectrom. 16 (2002) 933. [46] F.M. Han, C.H. Qi, Y. Chen, Chin. J. Anal. Chem. 35 (2007) 993. [47] M.L. Shen, K.L. Johnson, D.C. Mays, J.J. Lipsky, S. Naylor, Rapid Commun. Mass Spectrom. 14 (2000) 918. [48] M. Vairamani, M.L. Gross, J. Am. Chem. Soc. 125 (2003) 42. [49] M. Adamczyk, J.C. Gebler, C.A. Harrington, A.F. Sequeira, Eur. Mass Spectrom. 5 (1999) 165. [50] M.A. Tito, J. Miller, N. Walker, K.F. Griffin, E.D. Williamson, D. Despeyroux-Hill, R.W. Titball, C.V. Robinson, Biophys. J. 81 (2001) 3503. [51] E.F. Wilson, H.N. Miras, M.H. Rosnes, L. Cronin, Angew. Chem., Int. Ed. Engl. 50 (2011) 3720. [52] D.P. Weimann, H.D.F. Winkler, J.A. Falenski, B. Koksch, C.A. Schalley, Nat. Chem. 1 (2009) 573. [53] B. Qiu, J. Liu, Z. Qin, G.B. Wang, H. Luo, Chem. Commun. (2009) 2863. [54] T.S.B. Baul, C. Masharing, R. Willem, M. Biesemans, M. Holcapek, R. Jirasko, A. Linden, J. Organomet. Chem. 690 (2005) 3080. [55] W. Jiang, A. Schafer, P.C. Mohr, C.A. Schalley, J. Am. Chem. Soc. 132 (2010) 2309. [56] R. Franski, B. Gierczyk, Rapid Commun. Mass Spectrom. 19 (2005) 585. [57] S. Sakamoto, K. Yamaguchi, Angew. Chem., Int. Ed. Engl. 42 (2003) 905. [58] K. Yamaguchi, Bunseki Kagaku 59 (2010) 985. [59] S. Sakamoto, K. Nakatani, I. Saito, K. Yamaguchi, Chem. Commun. (2003) 788. [60] K. Yamaguchi, J. Mass Spectrom. 38 (2003) 473. [61] S. Sakamoto, K. Yamaguchi, Tetrahedron Lett. 44 (2003) 3341. [62] K. Strupat, M. Karas, F. Hillenkamp, Int. J. Mass Spectrom. Ion Processes 111 (1991) 89. [63] H. Rehulkova, J. Chalupova, M. Sebela, P. Rehulka, J. Mass Spectrom. 45 (2010) 104. [64] C.D. Calvano, S. Carulli, F. Palmisano, Rapid Commun. Mass Spectrom. 23 (2009) 1659. [65] T. Kobayashi, H. Kawai, T. Suzuki, T. Kawanishi, T. Hayakawa, Rapid Commun. Mass Spectrom. 18 (2004) 1156.

http://www.elsevier.com/locate/trac

65

Trends

Trends in Analytical Chemistry, Vol. 35, 2012

[66] F.L. Brancia, M. Stener, A. Magistrato, J. Am. Soc. Mass Spectrom. 20 (2009) 1327. [67] K. Breuker, R. Knochenmuss, R. Zenobi, J. Am. Soc. Mass Spectrom. 10 (1999) 1111. [68] X.L. Dong, J.S. Cheng, J.H. Li, Y.S. Wang, Anal. Chem. 82 (2010) 6208. [69] X.Z. Zhou, Y.Y. Wei, Q.Y. He, F. Boey, Q.C. Zhang, H. Zhang, Chem. Commun. 46 (2010) 6974. [70] L.G. Hu, G.B. Jiang, S.Y. Xu, C.S. Pan, H.F. Zou, J. Am. Soc. Mass Spectrom. 17 (2006) 1616. [71] M.C. Fitzgerald, K. Harris, C.G. Shevlin, G. Siuzdak, Bioorg. Med. Chem. Lett. 6 (1996) 979. [72] M.R. Carrasco, M.C. Fitzgerald, Y. Oda, S.B.H. Kent, Tetrahedron Lett. 38 (1997) 6331. [73] W.G. Lewis, Z.X. Shen, M.G. Finn, G. Siuzdak, Int. J. Mass Spectrom. 226 (2003) 107. [74] R.B. Cody, J.A. Laramee, H.D. Durst, Anal. Chem. 77 (2005) 2297. [75] J. Hajslova, T. Cajka, L. Vaclavik, Trends Anal. Chem. 30 (2011) 204. [76] H.W. Chen, S.P. Yang, M. Li, B. Hu, J.Q. Li, J. Wang, Angew. Chem., Int. Ed. Engl. 49 (2010) 3053. [77] S. Furmeier, J.O. Metzger, J. Am. Chem. Soc. 126 (2004) 14485. [78] H.Y. Wang, J.O. Metzger, Organometallics 27 (2008) 2761. [79] L.S. Santos, J.O. Metzger, Angew. Chem., Int. Ed. Engl. 45 (2006) 977. [80] B.M. Kolakowski, D.A. Simmons, L. Konermann, Rapid Commun. Mass Spectrom. 14 (2000) 772. [81] D.J. Wilson, L. Konermann, Anal. Chem. 75 (2003) 6408. [82] I.X. Peng, R.R.O. Loo, E. Margalith, M.W. Little, J.A. Loo, Analyst (Cambridge, UK) 135 (2010) 767. [83] M.Z. Huang, H.J. Hsu, L.Y. Lee, J.Y. Jeng, L.T. Shiea, J. Proteome Res. 5 (2006) 1107. [84] M.Z. Huang, H.J. Hsu, C.I. Wu, S.Y. Lin, Y.L. Ma, T.L. Cheng, J. Shiea, Rapid Commun. Mass Spectrom. 21 (2007) 1767.

66

http://www.elsevier.com/locate/trac

[85] I.X. Peng, J. Shiea, R.R.O. Loo, J.A. Loo, Rapid Commun. Mass Spectrom. 21 (2007) 2541. [86] J. Shiea, C.H. Yuan, M.Z. Huang, S.C. Cheng, Y.L. Ma, W.L. Tseng, H.C. Chang, W.C. Hung, Anal. Chem. 80 (2008) 4845. [87] J.F. Banks, S. Shen, C.M. Whitehouse, J.B. Fenn, Anal. Chem. 66 (1994) 406. [88] L.S. Eberlin, P.V. Abdelnur, A. Passero, G.F. de Sa, R.J. Daroda, V. de Souza, M.N. Eberlin, Analyst (Cambridge, UK) 134 (2009) 1652. [89] R. Haddad, R. Sparrapan, T. Kotiaho, M.N. Eberlin, Anal. Chem. 80 (2008) 898. [90] A. Hirabayashi, M. Sakairi, H. Koizumi, Anal. Chem. 66 (1994) 4557. [91] V.G. Santos, T. Regiani, F.F.G. Dias, W. Romao, J.L.P. Jara, C.F. Klitzke, F. Coelho, M.N. Eberlins, Anal. Chem. 83 (2011) 1375. [92] T.Y. Chen, J.Y. Lin, J.Y. Chen, Y.C. Chen, J. Am. Soc. Mass. Spectrom. 21 (2010) 1547. [93] H.W. Chen, A. Venter, R.G. Cooks, Chem. Commun. (2006) 2042. [94] W.S. Law, H.W. Chen, J.H. Ding, S.P. Yang, L. Zhu, G. Gamez, K. Chingin, Y.L. Ren, R. Zenobi, Angew. Chem., Int. Ed. Engl. 48 (2009) 8277. [95] M.D.L. de Castro, F. Priego-Capote, Trends Anal. Chem. 26 (2007) 154. [96] Y. Zhang, X.X. Ma, S.C. Zhang, C.D. Yang, Z. Ouyang, X.R. Zhang, Analyst (Cambridge, UK) 134 (2009) 176. [97] J.F. Garcia-Reyes, J.D. Harper, G.A. Salazar, N.A. Charipar, Z. Ouyang, R.G. Cooks, Anal. Chem. 83 (2011) 1084. [98] A.U. Jackson, J.F. Garcia-Reyes, J.D. Harper, J.S. Wiley, A. MolinaDiaz, Z. Ouyang, R.G. Cooks, Analyst (Cambridge, UK) 135 (2010) 927. [99] G.M. Huang, W. Xu, M.A. Visbal-Onufrak, Z. Ouyang, R.G. Cooks, Analyst (Cambridge, UK) 135 (2010) 705.