In this installment, we discuss radical ion chemistry as an increasingly important area of mass spectrometry (MS) development and the application to bioanalysis. At the current stage, most of the research is performed by a small set of academic groups. Given the unique capability offered by radical MS compared to the traditional studies on even-electron ions of analytes, it is likely that these types of fundamental studies will attract more attention and even be commercialized in the near future. 

Radicals are atoms, molecules, or ions that contain one or more unpaired valence electrons or an open electron shell. They are, generally, highly reactive because of their need to convert themselves to more stable, even-electron species. Well-controlled radical chemistry can provide the means for conducting some of the most difficult and delicate chemical transformations, like converting RNA to DNA by ribonucleotide reductase, an enzyme with a radical as its catalysis center. Radical chemistry can be coupled with mass spectrometry (MS) to tackle traditionally challenging problems such as sequencing disulfide proteins and determining C=C location in lipids. Here, we discuss this increasingly important area of MS development and its application to bioanalysis.  

MS has been established as a powerful tool for qualitative and quantitative chemical analysis. Its principle of analysis involves at least three crucial steps for extracting as much information as possible from an analyte: forming ions from the analyte molecule of interest, performing mass-to-charge ratio (

 (MS-MS), a process in which ions are allowed to interact with neutral molecules, electrons, ions, electromagnetic waves, and so on. 

During the past several decades, these three steps have reflected significant advances. The development of various soft-ionization methods permits the transformation of almost any molecule or chemical entity from its original physical state to gas-phase ions with the least change to its structure. (Nevertheless, it is frequently necessary to add a charge carrier to the analyte, facilitating MS analysis.) 

Several types of mass analyzers, the component of mass spectrometers that separate ions according to their mass-to-charge ratios, have been developed and commercialized. Instruments with time-of-flight (TOF) analyzers offer a fast speed of analysis. Fourier transform ion-cyclotron resonance (FT-ICR) and orbital ion trap instruments offer high-performance mass measurement and achieve a resolution in millions and sub-parts-per-million mass accuracy. Accelerator MS (AMS) instruments offer superb dynamic range (10

) for isotope analysis. Finally, some instruments could be even be made portable, miniaturized for performing field analyses. 

For tandem MS, many activation and dissociation options are available, including collision-induced dissociation (CID) and electron-based excitation or dissociation methods. Clearly, MS has entered into a golden age of advanced development and can be readily applied to solving complicated problems, whether in the areas of drug discovery, cancer research, and environmental monitoring or exploring Mars. 

Some of our most pressing questions about MS concern whether its usefulness has been fully exploited and, if not, what its next major advance will be. Its history shows that innovation in instrumentation and inquiries in new gas-phase ion chemistry often went side-by-side, providing the impetus for additional MS development. We believe the scope of gas-phase ion chemistry is unlimited, that its exploration so far has been only marginal, and that it may therefore give rise to the next major MS development. 

An area of increasing attention is the development of gas-phase, radical-ion chemistry for bioanalysis. Radical ions, which consist of unpaired electrons, derived from biomolecules (peptides, proteins, lipids, and carbohydrates) are significantly less investigated than even-electron ions. The available ionization methods do not lend themselves to directly forming the radical ions of biomolecules, so additional reaction steps are required to effect the transformation from even-electron ions. Several research groups have invested a major effort in developing methods to produce radical ions of biomolecules. Given the coexistence of two reactive functional groups within one ion, a charge and a radical site, the chemistry of radical ions is rich. It differs significantly from the chemistry of even-electron ions. This difference can be exploited to provide structural information that complements the structural information gleaned from even-electron ions. A well-known example is the development of electron-capture dissociation (ECD) and electron-transfer dissociation (ETD). These techniques provide rich structural information for protein analysis that complements the information from CID. Thus, it substantially improves the capability of protein identification and characterization. 

Bio-radicals have been implicated as important intermediates in a wide variety of biochemical processes. At the molecular level, they are associated with oxidative damage to proteins, DNAs, and lipids, which have been shown to be related to aging, neurodegenerative diseases, and cancers. Several classes of enzymes also use radicals for catalysis. For example, ribonucleotide reductase uses a free radical (thiyl radical) for its catalytic activity. Interestingly, the radical storage center is tens of amino acid residues away from the catalytic center. It is not clear, at this point, how the enzyme tightly regulates radical reactivity along this long-range radical transfer process. In general, knowledge of the bio-radical species is constrained by the limited techniques available to analyze these reactive intermediates. Studying the gas-phase chemistry of bio-radicals, thus, provides direct experimental evidence of their intrinsic chemical properties (reactivity, energetics, and structure), which can help explain the fate, including intra-molecular and intermolecular radical transfer, of bio-radical species after they initially form.  

The following discussions provide examples of methods of forming bio-radical ions of different classes of biomolecules and how radical chemistry and MS can generate useful structural information. 

Table I summarizes the major types of amino acid or peptide radical ions that have been studied, together with their methods of formation. They are categorized as follows, according to their underlying mechanisms of radical-ion formation: 					

electronic state change of even-electron species via electron capture, electron transfer, electron detachment, interactions with high-energy photons, or collisions with meta-stable neutrals

intramolecular electron transfer in peptide-metal-ligand complexes 

homolytic cleavage of labile bonds (either by CID or photolysis) 

radical reactions of disulfide-linked peptides in the plume of electrospray ionization (ESI). 

 Table I: Representative list of peptide radical ions studied in gas phase; type of radical ions, methods of formation, and precursors  						

When an electron adds to an even-electron peptide cation (multiply protonated), it results in more hydrogen than the corresponding even-electron species. Subsequent dissociation of the hydrogen-rich radical ions produces N-Cα backbone fragments, as compared with amide-bond cleavages of even-electron ions from vibrational activation (21–23). Analytical advantages of fragmenting hydrogen-rich radical ions include less dependence on peptide sequences and preservation of the labile modification on amino acid residues. ECD (1,24) and ETD (2,19) are the two major methods using fragmentation of the hydrogen-rich radical ions and have been shown to be effective for protein analysis by providing structural information complementary to that available by the CID of even-electron ions. 

Fragmentation of hydrogen-deficient radical ions, which (when compared with even-electron peptide ions) evidence fewer hydrogen atoms, has also been studied. Their dissociation chemistry differs from that of hydrogen-rich species and is sensitive to the nature of radicals (25). Several research groups have contributed significantly to this area by developing unique methods of forming hydrogen-deficient radical ions and studying their gas-phase ion chemistry (Table I). It has been shown that radical-dissociation chemistry can be analytically useful in protein analysis by inducing site-specific fragmentation (26,27) differentiating structural isomers or stereoisomers (17,28), and probing labile post-translational modifications (PTMs) (29).  

An alternative approach to generating the peptide radical ions used by our group involves a radical addition to a peptide ion, to form peptide radical ions, as shown in equation 1. Except for proton-transfer reactions, other possible reaction routes, such as electron-transfer and functional-group abstraction by radical species, could lead to the formation of peptide radical ions: 

This approach also gives rise to the formation of hydrogen-deficient radical ions, which facilitates choosing the proper radical-reaction chemistry for target peptides, thus conferring the flexibility to produce different types of radicals (such as carbon-centered versus heteroatom centered radicals) with highly specific radical sites. Moreover, because the charging process of the peptide ions can be separated from the radical formation step, this approach makes it possible to manipulate the charge properties of radical ions — the charge states, charge carriers (proton or metal ions), and charge polarities. 

The capability of forming site-specific radicals is essential for studying radical migration in biomolecule systems, whereas manipulation of the charge property is useful in investigating its role in peptide radical chemistry.   

Our group recently applied radical reactions in the nanoelectrospray ionization (nanoESI) plume region for the formation of sulfur-centered peptide radical ions. The experimental setup involves a radical-inducing source: an atmospheric-pressure, low-temperature, helium-plasma source or a low-pressure Hg lamp in the proximity of a nanoESI plume (Figure 1). A high number density of OH radicals can be generated and subsequently reacted with disulfide-linked peptides entrained in the spray plume. Given the high reactivity of the disulfide bond toward radical species, cleavage at the bond is highly favorable. This preference results in the well-controlled formation of sulfur-centered radicals such as the cysteine sulfinyl radical (-SO•), thiyl radical (-S•), and perthiyl radical (-SS•). 

Figure 1: Radical reactions in the nanoESI plume involving disulfide-linked peptides to form sulfur-centered peptide radical ions.			

These peptide radical ions can be mass isolated in a mass spectrometer for further investigation. Note that these three types of radical species have also been detected in enzymatic systems, though little chemistry is known because of their transient nature in condense phases. With their formation in the gas phase, the intrinsic chemical properties lend themselves to extensive study.  

Ion and molecule reactions in a quadrupole ion trap show that the reactivity of the three types of sulfur radicals follows the order of thiyl (RS•) > perthiyl (RSS•) > sulfinyl (RSO•). Although the cysteine sulfinyl radical is quite inert under conventional bimolecular reaction conditions, collisional activation facilitates intramolecular reactions. These reactions reveal novel chemistry, because the supplied external energy overcomes the energy barrier. Two newly discovered reaction channels involve the exchange of the cysteine sulfinyl radical (SO•Cys), with either a free thiol (-SH) or a disulfide bond (S-S) within a peptide ion system (reactions shown in Figure 2). Note that, in a protein system, the exchange of the sulfinyl radical with a disulfide bond leads to disulfide-bond scrambling. These reaction phenomena are significant, for they may affect how a radical intermediate migrates in a protein system and attacks other functional groups, resulting in conformational or functional changes. In addition, the reactions suggest that the 

 — damage done by forming a sulfinyl radical — can potentially be repaired by reacting with a nearby thiol group (Figure 2a).  

Figure 2: Intramolecular cysteine sulfinyl radical (SO
•Cys) transfer with (a) a free thiol and (b) a disulfide bond.			

Disulfide bond formation is a critical PTM, which is highly relevant to the structural integrity of a protein and its biological function. An MS method targeting an intact, disulfide peptide or protein is desirable because it can provide linkage information about disulfide bonds and rich sequence information. It is worth noting that cleavage of the disulfide bond within an even-electron peptide ion is typically a higher-energy process under collisional activation, as compared with cleavage of the peptide backbone (amide bond) to produce protein-sequence ions. To produce a fragment ion from the peptide backbone region within a disulfide bond loop requires the cleaving of at least two bonds. Thus, few sequence fragment ions form under CID conditions, making sequence information difficult to obtain. The routine procedure in proteomics, therefore, involves disulfide reduction and alkylation, to open the disulfide bonds. Indeed, the sequence information obtained from CID-based tandem MS can be significantly improved, but only at the cost of losing information about disulfide-bond connecting patterns. 

Unlike the chemistry of even-electron ions, when a radical is introduced into the peptide or protein system, the disulfide bond presents itself as an easy target. For instance, its cleavage is a favorable process where radical intermediates are produced, in ECD, ETD, electron-detached dissociation (EDD), and UV-photodissociation. Among those means of dissociation, ETD is increasingly applied for disulfide peptide or protein characterization. Facile cleavage at the disulfide bond and separation of chains originally connected by disulfide bridges have been reported from ETD of interchain, disulfide-linked peptides. For peptides consisting of intrachain disulfide bonds, ETD can produce rich-sequence information even from backbone regions covered by the disulfide bone loop. This facility of EDT is the result of radical cascades in which N-Cα bond cleavage from ETD leads to the formation of 

• ions, and the carbon-centered radical 

• can further react and cleave the disulfide bone to release 

 + 33 types of sequence ions, all of which are useful for peptide identification. 

Radical reactions in the ESI plume (setup shown in Figure 1) also efficiently cleaves the disulfide bond. One approach to achieving enhanced sequence information for the disulfide peptide combines radical reactions in the ESI plume with subsequent MS

-CID, to fragment peptide radical ions, thus obtaining the traditional 

 type fragment ions (Figure 3). We applied this method to a series of naturally occurring peptides that contained one or multiple disulfide bonds. In doing so, we obtained significantly enhanced sequence information, as compared with CID of the intact disulfide peptide ions. Using insulin, a polypeptide containing three disulfide bonds, as an example, rich-sequence information corresponding to 75% of backbone cleavages is obtained from radical reactions in ESI and subsequent CID, which compares favorably to the reported ECD and ETD results (~60%). Moreover, this result is significantly more favorable than the 26% backbone cleavages obtained from CID of intact, disulfide, peptide ions. 

Figure 3: Coupling radical reactions in the ESI plume region with a subsequent MS-MS scan of the disulfide-bonded, cleaved-peptide, radical ions, to improve sequence information.			

Lipids serve in numerous biological functions such as structural binding blocks, energy storage, and signaling transduction. The degree of a lipid's unsaturation (the number of C=C bonds) as well as the positions of the C=C bonds determine its overall structure and therefore its biochemical and biophysical properties. For instance, 

-3 polyunsaturated fatty acids (PUFAs) (also called omega-3 fatty acids, in which 3 is the double-bond position counted from 

, the terminal methyl group) are essential for the functional development of the brain and retina. By contrast, no such effects have been observed for 

-6 PUFAs. The efforts to determine double-bond positions in lipids, however, are not trivial, largely because of a need to distinguish the correct structure from a large number of possible double-bond-position isomers. In the field of lipidomics, therefore, MS is the analytical method of choice, because of its high sensitivity, its selectivity, and its ability to provide detailed structural information. Tandem MS of even-electron lipid ions based on collisional activation is useful in providing the head group and acyl-chain composition information. Yet because cleavages at the C-C and C=C positions require significantly higher energies than other possible fragmentation channels, such as neutral or charged head-group losses and fragmenation at the ester bond, obtaining C=C position information cannot be achieved directly from low-energy collsional activation. When the collsion energy is high enough (that is, ~ keV from sector instruments or a TOF-TOF instrument), charge-remote fragmenation is possible. Thus, the C-C bonds adjacent to the C=C bonds are cleaved, so the location of C=C is inferable. Compared with the brute-force method of simply increasing the level of collision energy, reactions targeting the unique chemistry of C=C has been used, including ozonolysis, methoxylation, olefin cross-metathesis. When coupled with MS analysis (on-line or off-line), these chemistry types provide information about the location of C=C. 

Another unique property of C=C is its susceptibility to radical attack, which prompts investigation by radical-involved MS analysis. For example, Blanksby and colleagues explored radical-directed dissociation (RDD) of lipids by attaching a photo-caged radical precursor to the target analyte (30). They found that irradiation of the precursor in the mass spectrometer releases the previously caged radical, and CID of the nascent radical and lipid complex ion yields fragmentation information. Moreover, analyzing the fragment ions and fragmentation patterns can lead to ascertaining the number of unsaturation and the position of double bonds. 

Our group recently explored coupling a Paternò–Büchi (P-B) reaction of C=C in lipids in the ESI plume with subsequent tandem MS, for fast and sensitive determination of the C=C location (31). The P-B reaction is a classic [2+2] photochemical reaction. In organic synthesis, it is widely used to form compounds containing an oxetane ring. The reaction mechanism involves UV excitation of the carbonyl group within an aldehyde or ketone to a ketyl diradical, which subsequently reacts with the C=C in an olefin. Depending on the relative positions of the carbonyl and the C=C bond, two position isomers of the oxtanes can be formed, as shown in Figure 4. When energy (such as heat) is applied to the P-B reaction products, retro P-B reactions proceed through two possible pathways. One route leads to the original reactants. The other leads to cleavage of the C-C bond at the initial C=C bond position and the C-O bond of the initial carbonyl group, forming a new olefin and ketone or aldehyde. The latter pathway is of special interest in this study because the reaction products carry the C=C bond's position information.  			

Figure 4: PaternÃ²âBÃ¼chi (P-B) reaction between ketone, aldehyde, and olefin in lipids together with retro P-B reactions, to form diagnostic ions for the determination of C=C location.			

The P-B reaction can be simply executed in the naonESI plume region by means of the setup shown in Figure 1, a UV lamp with the irradiation wavelength centered at 254 nm. The lipid analyte is dissolved in 1:1 (water:acetone), where acetone functions as the P-B reaction reagent. A reasonable reaction yield of 40% is possible for different classes of lipids, and no reaction selectivity toward specific C=C locations or configurations is found. The P-B reaction product is stable and can be mass isolated and subjected to CID. Figure 5 shows an example of the P-B reaction of oleic acid (C18:1 9Z) prepared at 10 μM in 1:1 water–acetone, with 1% ammonium hydroxide, for nanoESI. Abundant P-B reaction product (

 339.3), because of the addition of acetone to the C=C, was observed upon UV irradiation. MS

-CID of the P-B reaction product promotes the retro P-B reactions and provides a pair of "diagnostic ions" (

 171 and 197) with a mass difference of 26 Da, for confident determination of the C=C bond location. The structures of the diagnostic ions are shown in Figure 5d. The method is sensitive (~femtomolar [fmol] for fatty acid)

and can be readily coupled with "shotgun" lipidomics, for complex lipid-mixture analysis. An in-solution P-B reaction has been developed to couple with on-line liquid chromatographic separation methods.  

Figure 5: PaternÃ²âBÃ¼chi (P-B) reaction between oleic acid and acetone in the nanoESI plume upon UV irradiation. MS1 spectra of oleic acid (a) before UV irradiation, (b) upon UV irradiation, in which the P-B reaction product is located at m/z 339.3, and (c) MS2-CID of the P-B reaction product. The diagnostic ions (m/z 171.1 and 197.2) for the C=C location are observed. (d) Structures of the diagnostics ions (structures 3 and 5). The structures explain the origin of the 26-Da mass difference. 			

In general, the P-B reaction method for lipid C=C location determination has analytical advantages. They include simple experimental setup for reactions, no need to modify the mass spectrometer, easy-to-interpret mass spectra, and inexpensive derivatizing reagents. These characteristics should make this method accessible and attractive to many laboratories. 

Saccharides are highly susceptible to decomposition by free-radical processes. In consideration of this fact, Beauchamp and colleagues (32) proposed using biomimetic synthetic reagent and free-radical chemistry to analyze saccharides. By introducing a radical at the reducing end of the saccharides, both glycosidic bond cleavage and cross-ring cleavages are observed upon CID of radical-precursor ions. Radical-driven fragmentation pathways are proposed as responsible for fragments arising from both types of cleavages. This method for the analysis of saccharides, based on free radicals, also represents a novel method for understanding the reactions between free radicals and saccharides. 

Radical-involved chemical and biochemical reactions are attracting increased interest in the area of MS. This interest is driven by the distinct chemistry of radicals, versus even-electron species, and the potential for new analytic capabilities. This chemistry can be successfully applied for the structural analysis of proteins or peptides, lipids, and saccharides. The extent of structural information obtained from MS analysis of these molecules can be greatly improved relative to conventional methods. The development of radical MS approaches and strategies has only just started. We anticipate its continued growth, both in fundamental ion chemistry and method and instrument development. We are confident that radical MS can serve as one of the new forces advancing the future development of MS in chemical analysis. 

(1) R.A. Zubarev, N.L. Kelleher, and F.W. McLafferty, 

(2) J.E. Syka, J.J. Coon, M.J. Schroeder, J. Shabanowitz, and D.F. Hunt, 

(3) S.L. Cook, O.L. Collin, and G.P. Jackson, 

(4) I.K. Chu, C.F. Rodriguez, F. Rodriguez, A.C. Hopkinson, and K.W.M. Siu, 

Articles

MS – The Practical Art

The chemistry of radicals - which is quite different from that of even-electron species - opens up new analytical capabilities.

Radical Mass Spectrometry as a New Frontier for Bioanalysis

Ultraviolet (UV) detectors are the most common liquid chromatography (LC) detector, and perhaps the most reliable ones. But they are not without problems.

I recently received an e-mail question from a reader complaining of a leaking detector cell in the ultraviolet (UV) detector attached to his liquid chromatography (LC) system. This seems like a good opportunity to address that specific question, other UV detector problems, and some of the more recent advances in detector design. Although the present discussion centers on UV detectors, many aspects will apply to other optical detectors, such as fluorescence or refractive index detectors. 

First, let's take a look at how the typical detector cell is constructed and we'll be able to see where leakage can occur. After we have that information, correcting the problem should be straightforward. 

For UV detectors designed to operate with conventional LC systems with upper pressure limits of 6000 psi (400 bar), the construction shown in Figure 1 is common. The cell itself is made by drilling a 1-mm hole through a 10-mm-long block of stainless steel. To contain the liquid, a quartz window is attached to each end of the cell and a seal is formed with a polymeric gasket. The mobile phase needs to pass through the cell, so a provision is made for this by drilling small-diameter (for example, ≤0.125 mm i.d.) holes to connect the outside world with the cell cavity. Most commonly, the entrance and exit holes are on opposite sides and opposite ends of the cell so that the flow path is Z-shaped. This allows efficient washout and clearance of bubbles, should they enter the cell. UV light then passes through the cell from one end to the other. When sample peaks pass through the cell, some of the UV light is absorbed, and a photodiode measures this absorbance by the change in the intensity of light passing through the cell. 

 Figure 1: Schematic of a conventional UV detector cell. See text for details.			

The amount of light passing through the mobile phase at steady state (no sample present) is affected by the refractive index of the mobile phase. Any change in refractive index will result in more or less light making it through the cell, and the baseline will drift or exhibit other disturbances. To help mitigate temperature-related refractive-index disturbances, a heat exchanger is usually fitted to the inlet of the cell. Most commonly this is a stainless steel capillary wrapped around the body of the cell and covered with a heat-conducting material so that the fluid entering the light path is thermally stabilized, and temperature-induced background disturbances are minimized. 

Another potential problem is the presence of air bubbles in the flow cell. Even though the mobile phase is usually degassed before use, when the mobile phase leaves the column, it moves from a high-pressure region to a near-atmospheric pressure region. When this happens, any residual air present tends to outgas from the mobile phase and form physical bubbles. When air bubbles enter the flow cell, they disrupt the light path and result in a noise spike or false peak in the chromatogram. Usually, bubbles continue through the flow cell and clear by themselves, but tiny microbubbles sometimes become lodged in the corners of the cell that are less well swept. These bubbles can "bounce" in the flow stream and cause additional baseline problems. To help avoid bubble problems in the cell, pressure can be applied to the cell outlet so that the internal pressure in the cell is sufficient to keep the bubbles in solution. Restricting the flow at the outlet of the flow cell can be somewhat delicate — we want to have enough pressure to keep any bubbles in solution until they leave the cell, but we don't want so much pressure that the cell leaks. The most common practice to provide back pressure on the cell is to use a piece of capillary tubing as the waste line. A narrow capillary will restrict the mobile-phase flow and create pressure upstream. The pressure thus created is dependent on the mobile-phase viscosity, temperature, and flow rate. Because pressure will increase with flow rate, there is a risk that increased pump flow, such as might be used to flush a column, will create sufficient pressure to exceed the flow-cell pressure limits (cell pressure limits should be listed in the "specifications" section of the detector manual, but typically are in the 150-psi or 10-bar region). A better solution to provide back pressure on the cell is to use a back-pressure restrictor that generates a constant pressure. These restrictors are constructed like a spring-loaded check valve that opens when the pressure exceeds a set value less than the detector cell maximum. Back-pressure restrictors can be purchased from most suppliers of LC tubing and fittings. 

Now that we know how the UV detector cell is constructed, let's look at why problems with cell leakage occur, where they can occur, and how to correct them. The simplest case of leakage is a problem with the tube fitting connecting the column to the detector or detector to waste. The location of leaks at the fittings should be obvious, with fluid dripping from the fitting, although leaks elsewhere in the detector can drain to the fitting before dripping off. Because replacement of the tubing attached to the detector may be difficult or impossible, be very careful when correcting suspected fittings leaks. Support the fitting so that the tubing isn't bent during tightening. Simply tighten the connecting fitting with a wrench or with your fingers, as appropriate. If the leak is not stopped with one-half turn tightening of the fitting, there may be additional problems, so don't continue tightening or you may cause permanent damage. At this point loosen the fitting, rinse it with clean solvent, and try again, or simply replace the fitting with a new one. If the connections are made with finger-tightened fittings, I suggest that you set the flow rate to zero, loosen the nut, push the tubing firmly into the connection, and re-tighten it. If you try to tighten finger-tightened fittings with the flow on, the tubing can slip in the fitting, creating a small gap that can cause unwanted peak broadening. 

If attempts to correct leaks by tightening the connecting fittings are not successful, it is likely that the gasket forming a seal between the quartz window and the cell body is leaking. Sometimes gasket leaks will occur if the cell is accidently over-pressured, then will correct themselves when the system returns to acceptable pressures. More commonly, however, once the gasket leaks, it continues to leak. Replacement of the cell gaskets may or may not be a user-serviceable operation. Consult the detector operator's manual to see if this is possible. If the cell is user-serviceable, you should be able to buy a repair kit from the detector manufacturer. Instructions on seal replacement should be included with the repair kit or listed in the operator's manual. I recommend replacing both the gaskets and the quartz windows at the same time. Over many hours of use, the quartz windows can deteriorate from exposure to UV light, and organic materials from the sample can degrade or polymerize and coat the windows. If the cell is not user-serviceable, if you feel this is beyond your capabilities, or if you have a service contract on the LC system, call the manufacturer's service technician to make the repair. An intermediate option is to remove the flow cell — usually a simple task — and ship it to the manufacturer for repair. 

A much less common problem that can cause leakage is if the heat exchanger leading into the flow cell becomes blocked. The symptom will be excessive back pressure and often leakage at the fitting connecting the column to the cell. Sometimes you can clear such blockages by reversing the flow through the heat exchanger. Because of the pressure sensitivity of the detector cell itself, it is not a good idea to try to back-flush the cell using the LC pump because cell pressure is likely to be exceeded. A better technique is to disconnect both the inlet and outlet connections to the cell and connect a syringe to the inlet of the cell with a high-pressure fitting. Then pull solvent through the cell in the reverse direction of normal flow. This may pull out anything blocking the inlet, but will not apply any pressure to the cell. Do not use the syringe to try to push solvent through the cell because it is too easy to exceed the cell pressure limits. If these attempts don't clear the heat exchanger, call a service technician or send the cell in for factory service. 

What about cleaning the detector cell? In my opinion, the detector cell falls in the category of "if it ain't broke, don't fix it." That is, it is much easier to cause damage to the cell than it might appear, so well-intentioned cleaning procedures may cause problems instead of preventing them. You should flush the column after each batch of samples to remove strongly retained materials. These flushing conditions typically involve a high concentration of organic solvent for reversed-phase methods, which is likely to remove anything that might remain in the detector cell, as well. As mentioned above, over time sample components may polymerize or otherwise degrade and coat the cell windows, but for most methods this is not a concern. There are detector cleaning procedures, many of which include flushing with dilute nitric acid. If you decide to follow one of these procedures, consult the detector operator's manual for specific directions, and be sure to wear appropriate protective clothing (for example, a lab coat, eye protection, and gloves). Once again, if there is no apparent problem, I do not think that it is a good idea to perform cell cleaning procedures other than the normal flushing that takes place when the rest of the system is flushed with strong solvents. 

With the advent of ultrahigh-pressure liquid chromatography (UHPLC), where pressures of up to 19,000 psi (1300 bar) may be encountered, the conventional flow cell described above and in Figure 1 will not work very well. It is likely that the pressure after the mobile phase leaves the column may still be too high for the pressure limits of a conventional cell, so a higher-pressure design is needed. In addition to this, UHPLC columns generate peaks that typically are much narrower than from conventional columns, so extracolumn band broadening caused by the detector may severely compromise the performance of an UHPLC system. 

A rule of thumb regarding the relationship between peak width and detector cell volume is that the cell should be less than approximately 10% of the peak volume. Under typical analysis conditions, a 150 mm × 4.6 mm column packed with 5-µm diameter (

) particles will generate peak volumes of approximately 100 µL when the retention factor (

) is 1. Another popular conventional LC column configuration is 100 mm × 4.6 mm packed with 3-µm particles, which will give approximately 65-µL peaks with 

 = 1. You can see that these would suggest cell volumes of no more than 6.5–10 µL. So, the 8-µL cell of the conventional 1 mm i.d. × 10 mm long cell is unlikely to contribute unacceptable band broadening, especially if 

 values of more than 1 are used. On the other hand, a 50 mm × 2.1 mm, 2-µm 

 column used with UHPLC will give peaks at 

 = 1 of about 7 µL in volume, so cells with a volume less than 1 µL will be required to avoid unwanted band spreading. 

We could reduce the length of a conventional flow cell to reduce the volume, but this also reduces the sensitivity, because fewer sample molecules will be in the optical path at any given time. Reducing the diameter from 1 mm to 0.5 or 0.25 mm i.d. would reduce the volume from 8 µL to 2 µL or 0.5 µL, respectively. At first glance this seems like a good idea, but there are other problems that are encountered. Any light that enters the flow cell that is not parallel to the cell walls will strike the walls of the cell and tend to be scattered, changing the amount of light that makes it through the cell to the light-sensing photodiode. This can happen if the light is not perfectly collimated or if the refractive index of the mobile phase changes, such as during a gradient or because of temperature or pressure changes. With flow cell diameters of <1 mm, it is difficult to avoid unwanted light scattering within the cell. The net effect is that narrowing a conventional flow cell will not give adequate detector performance. 

The innovation that allows UV detection to work with UHPLC is the total internal reflectance, or light-pipe, design shown in Figure 2. In this design, the inside of the flow cell is coated with a reflective material, such as Teflon TF, that efficiently reflects any light rays that contact it. An alternative design uses an uncoated fused-silica capillary, which reflects light from the external surface of the capillary. In either case, the result is that rather than scattering the light, as with a conventional stainless-steel surface, the light bounces off the reflective surface as it goes through the cell and all the light makes it to the other end where it can be measured with the photodiode. Now the cell diameter is no longer a limitation in design, so a 0.25-mm i.d. cell is practical. Thus, commercial instruments now offer flow cells of <0.5 µL in volume that will work with UHPLC. Because all the incident light makes it through the flow cell, the refractive index–related problems of conventional cells are mitigated. A reduced sensitivity to temperature effects means that the UHPLC flow cell can have a smaller heat exchanger (or none at all), eliminating another potential source of band broadening. 

Figure 2: Schematic of a total internal reflectance (or light-pipe) UV detector cell, as used in UHPLC applications. See text for details.			

The only complaint I've heard about the light-pipe detector cell design is that the internal coating is somewhat fragile. I've heard reports of damage to the coating when the cell was allowed to dry out and heat up with exposure to the UV lamp. Then when the flow is turned back on, the coating can be washed off, blocking or otherwise ruining the flow cell. Some instruments protect against this by automatically putting a shutter between the lamp and the cell when a run is not underway. If your system uses an internally coated cell and is not protected in this manner, when you turn off the instrument you should make sure the detector lamp also is turned off. 

We've looked at the design of the UV detector cell for both conventional and UHPLC applications. With care, both designs can be operated with few problems. With conventional LC systems, the advances in column technology, such as the increasing dependence on 3-µm 

 particles and now the growing popularity of core–shell particles, mean that the conventional 1 mm i.d. × 10 mm, 8-µL flow cell is marginal, at best, in terms of extracolumn band broadening effects. For this reason, I predict that the next generation of conventional LC systems designed to operate up to 6000 psi (400 bar) will incorporate light-pipe flow cell designs. 

"LC Troubleshooting" Editor John Dolan has been writing "LC Troubleshooting" for 

for more than 30 years. One of the industry's most respected professionals, John is currently the Vice President of and a principal instructor for LC Resources in Walnut Creek, California. He is also a member of 

's editorial advisory board. Direct correspondence about this column via e-mail to 

LC Troubleshooting

To solve problems with UV detectors, it helps to understand a few things about how the detector cells are designed.

UV Detector Problems

Sample preparation has often been viewed as the bottleneck in analytical procedures. Surveys have shown that time is typically the most frequent problem area for sample preparation procedures and that analysts can easily spend a majority of the total analysis time on sample preparation. While newly developed extraction techniques address time, modern chromatography advances are also moving toward faster separations. Based on these considerations, what is high-throughput sample preparation? Do modern extraction methods adequately address the issue of time? How can we address the analytical need for speed?

The self-proclaimed world's greatest rock and roll band famously sang, "You can't always get what you want." It's easy to apply the Rolling Stones admonition to analytical sample preparation because surveys (1–3) have consistently shown that time, along with cost and solvent use, is among the most significant desires of analysts. Traditional sample preparation methods are often the rate-limiting step in the overall sample analysis process. We can envision three approaches for addressing the desire for high-throughput sample preparation: parallel sample processing, automation, and improvements in the process kinetics. 

Evolution of High-Throughput Analysis from Microplates

Over the past several years, we've heard increasing calls for high-throughput sample preparation. But what is high-throughput sample preparation? It many cases, high-throughput stems from the high-throughput screening approach to combinatorial chemistry. This approach is the most established, based on the 96-well plate format. These microplates have been established for two to three decades. Key considerations for the acceptance of this approach are standardization of microplate dimensions and attributes by the Society for Biomolecular Screening and development of ancillary devices like repeating pipettes, vacuum manifolds, and more. Countless vendors are involved. Sample preparation and analysis using microplates includes liquid–liquid extraction (LLE), protein precipitation, solid-phase extraction (SPE), fluorescence, matrix-assisted laser desorption–ionization mass spectrometry (MALDI-MS), and separations. 

				Wells (4) presented a list of some of the uses of filtration microplates in pharmaceutical development, including									

clarification of acid, base, or organic digests of plant materials in medicinal chemistry,

solution-phase synthesis in combinational chemistry using resin scavenger media,

membrane-based proteolytic digestion before MALDI-MS,

filtration of plasma or serum samples in combination with direct injection techniques,

filtration of reconstituted extracts, and

solid-supported LLE, SPE, or exclusion chromatography.

As we can see, the development of 96-well plates and the major applications of this approach are in bioanalysis of liquid samples. This is because the transfer of liquid samples from analytical operation to another is somewhat straightforward. Recently, Scoffin (5) reviewed liquid transfer in high-throughput systems. These liquid-handling systems operate either by air displacement or positive displacement and can reliably work with microliter volumes of volatile or viscous liquids. Other combinations of liquid handling and sorbent-based extraction, including solid-phase microextraction (SPME), automated disposable pipette extraction, and microextraction by packed sorbent have also been used in high-throughput approaches (6,7). By combining these solvent-free, sorbent-based extractions with syringe-needle or pipette configurations, direct extraction and liquid transfer is accommodated. 

Turbulent Flow Chromatography in Sample Preparation

In recent years, turbulent flow chromatography (TFC) has emerged as a high-speed extraction alternative for liquid samples. When flow rates increase significantly, the flow pattern shifts from a laminar, bullet-shaped profile to a turbulent profile. An example of this flat, turbulent flow profile is shown in Figure 1. Turbulent flow minimizes zone spreading and, most importantly for sample preparation purposes, facilitates radial diffusion (that is, mass transfer) to the stationary phase. Two years ago, sample preparation applications of TFC were discussed in this column (8). To achieve turbulent flow conditions, large, porous particles are combined with high flow rates. In general, 30–60 µm particles packed in 0.5–1.0 mm i.d. columns with flow rates greater than 1 mL/min and up to 6 mL/min achieve these conditions. Dramatic reductions in time are observed, while maintaining accuracy, precision, detection limits, and other figures of merit. Low-molecular-weight compounds readily diffuse into the porous particles, whereas larger compounds are less retained and are eluted to waste. Proteins are reported to be more selectively removed from biosamples using TFC than restricted-access media (RAM) or solid-phase extraction, as illustrated in Figure 2 (9). Thus, small molecule analytes in samples like whole blood, serum, plasma, or urine can be delivered to the injection port when coupled to high performance liquid chromatography (HPLC) in an on-line manner using column switching. Although minimal sample handling is necessary, resulting in fast analysis, the high flow rates necessary to achieve turbulent flow result in a large solvent consumption. Observed matrix effects require the use of standard addition. Recent reviews (10–13) have illustrated the use of TFC for biofluids, foods, and environmental samples. 

Figure 1: Example of turbulent flow profile.			

So far, we've focused the discussion of high-throughput sample preparation on liquid samples. The reasons are two-fold, as previously mentioned. Liquid handling is easier than manipulation of solids and the microplate technology drove the development of high-throughput methods. How can we make the analysis of solid samples high-throughput? Do the more recently developed extraction techniques take us in that direction? 

Figure 2: Comparison of turbulent flow chromatography (TFC), restricted-access media (RAM), and solid-phase extraction (SPE) cleanup columns as a function of the molecular weight excluded during sample cleanup. Adapted from reference 9.			

During the time when 30–60 min chromatography runs were standard, Soxhlet extractions took 6–48 h. This bottleneck was even further exacerbated when spectroscopy or other rapid analytical procedures were used. During the past 20 years or so, several automated techniques (including supercritical fluid extraction [SFE], microwave-assisted extraction [MAE], and pressured liquid extraction [PLE]) have been developed, which provide the hope for high-throughput sample analysis. In addition to automation of the extraction procedure, these methods are characterized by very rapid extraction times compared with the traditional methods. However, simultaneously to the development of these techniques, advances in sensors, mass spectrometry, and chromatography have rendered even these modern technologies to remain the rate-limiting part of an analytical procedure. Sample preparation that takes 20 min, while several-fold faster than Soxhlet or shake-flask methods, is still slow compared with 5 min analyses. In the days of 6–48 h Soxhlet extraction and 30–60 min chromatographic analysis, the ratio of sample preparation time to chromatography run time was 6–96:1. However, it is common to perform, for example, 12 Soxhlet extractions simultaneously, so the ratio becomes 0.5–8:1. SFE, PLE, and MAE extractions commonly take 20 min, but have we really gained anything? It is now routine to perform chromatography with 10–30 min run times or faster. So the ratio remains 0.5–6:1. 

Figure 3: Relative percent of oil extracted from 725-Âµm coffee particles as a function of time (min) using various extraction processes. Adapted from reference 15.			

The fundamental difference between extracting solid samples and liquid samples is the role of diffusion. Diffusion in liquids at room temperature is on the order of 10

/s, while diffusion in solids at the same temperature is on the order of 10

/s. But what about those solid extraction methods that operate at elevated temperatures? We have conducted a systematic comparison of SFE, MAE, PLE, and sonication extractions to examine the rate-limiting role of diffusion in these extraction methods. We extracted the oils from ground, roast coffee using these methods and determined the diffusion coefficients using the hot-ball model (14). Extractions were performed using dichloromethane at 100 °C under suboptimal conditions, which would allow the determination of the diffusivity. Soxhlet extraction is performed at some elevated temperature near, but below, the boiling point of 39.8 °C, while the modified Soxhlet approach includes leaching in boiling solvent as the main portion of the extraction. Ultrasonic extractions are performed in open-vessels, so the bulk solvent cannot exceed the boiling point, but on a molecular level temperatures during cavitation can reach 5000 K. PLE and MAE in this work were performed at 100 °C. Figure 3 displays the kinetic curves from these extractions (15). When the hot-ball model is applied to this data, the diffusion coefficients shown in Table I are obtained. The results show that analyte diffusion values, under similar extraction conditions applied across each method, were within the same order of magnitude. It seems unlikely that temperatures greater than those used in these techniques will be routinely applied for analytical extractions. Thus, when developing high-throughput extraction systems, the kinetics of extraction will not differentiate these alternative extraction procedures. PLE, MAE, and SFE seem to be the preferred approaches for developing high-throughput processing, but advantages with each technique will arise because of instrumental configurations. 

Table I: Average diffusion coefficients obtained from the extraction of oils from ground, roast coffee			

Is High-Throughput Sample Preparation Necessary? 

As discussed, the major approaches to achieve high-throughput sample preparation stem from parallel sample processing, automation, and the kinetics of the sample processing procedure. The success observed in the high-throughput sample preparation of liquids comes, in part, from automation and parallel processing. The microplates process 96 or more samples simultaneously. The instrumental approaches to extracting solids are also configured for automation and parallel processing. The kinetics of extraction is fundamentally limited by practical considerations. So perhaps the best approaches for high-throughput analysis will result from the integration of methods. Just as turbulent flow chromatography is coupled to liquid chromatography (LC), thermal desorption is coupled with gas chromatography (GC), and 96-well microplates are coupled to MALDI-MS, similar methodologies combining rapid sample and processing parallel approaches will begin to address this need. However, the analyst must also keep in mind what is truly desired. High-throughput sample preparation is not the ultimate goal, but rather high-throughput analysis. Chemometrics and other statistical approaches can glean the desired information from minimal analysis, and dilute-and-shoot or direct analysis mass spectrometry eliminate the need for sample preparation entirely, which is perhaps the fastest method of all. 

High Throughput Sample Preparation Methods and Automation Strategies

 (Elsevier, Amsterdam, The Netherlands, 2003), pp. 199–254. 

(6) J. Pereira, C.L. Silva, R. Perestrelo, J. Goncalves, V. Alves, and J.S. Camara, 

(8) J.L. Herman, T. Edge, and R.E. Majors, 

(9) O. Nunez, H. Gallart-Ayala, C.P.B. Martins, and P. Lucci, 

(10) J. Pan, C. Zhang, Z. Zhang, and G. Li, 

(11) O. Nunez, H. Gallart-Ayala, C.P.B. Martins, P. Lucci, and R. Busquets, 

(14) K.D. Bartle, A.A. Clifford, S.B. Hawthorne, J.J. Langenfeld, D.J. Miller, and R. Robinson, 

(15) J.L. Driver and D.E. Raynie, in preparation (2014). 

is an Associate Research Professor at South Dakota State University. His research interests include green chemistry, alternative solvents, sample preparation, high resolution chromatography, and bioprocessing in supercritical fluids. He earned his PhD in 1990 at Brigham Young University under the direction of Milton L. Lee. 

"Sample Prep Perspectives" Editor Ronald E. Majors is an analytical consultant and is a member of 

's editorial advisory board. Direct correspondence about this column to 

Sample Prep Perspectives

In sample preparation, what does "high-throughput" really mean?

Exploring the Possibilities of High-Throughput Sample Preparation

The following questions were submitted by ChromAcademy members as part of our "HPLC Troubleshooting Master Class" series, which focused on problems in the modern chromatography laboratory. 

Can reversing a modern reversed-phase HPLC column really damage it?

Columns are packed using a constant-pressure rig with a solvent packing slurry forced through the column blank at higher pressure than the intended application limit, with the outlet frit and fitting in place, but the inlet end directly connected to the packing rig reservoir. Although manufacturers do a great job in ensuring the packing is homogenous from inlet to outlet, there is always the possibility the material at the outlet is more densely packed, and so to avoid settling of the packed bed and a void formation, one should observe the "directionality" of the column. 

When column efficiency declines or peaks begin to split or shoulder, users sometimes wish to reverse the column direction to flush strongly adsorbed material from the inlet frit or packing material at the "front" of the column. This is possible using modern high performance liquid chromatography (HPLC) columns without too much risk, other than the packed bed "settling" in the reversed direction — resulting in the same problems all over again. Remember to disconnect your detector from the column when flushing to avoid contamination. 

If you inject a large-volume sample that is different from the starting mobile phase in a gradient can you get partitioning of analyte between the injection slug and the mobile phase?

Yes! As the slug of analyte moves toward the column it will begin to partition at the eluent or diluent interface; however, this mixing is often incomplete, especially because the distance and volume between the injector and column in modern HPLC systems is so small. At the head of the column, especially in gradient analysis, all analytes come to a standstill. Those analyte molecules associated with the most highly eluting solvent (usually, and incorrectly, the sample diluent) will partition down the bed somewhat, giving two "discrete" regions of analyte at the column head — those associated mainly with eluent and those associated mainly with the diluent as they entered the column — leading to a broad and often misshapen peak. 

One should match both the eluotropic and ionic strength of the sample diluent with the initial mobile-phase composition. Where this is not possible (typically because of solubility issues), one should use the lowest injection volume possible to allow effective mixing before, and at the head of, the column, which will help to mitigate some of the effects mentioned above. If solubility is an issue, use a suitable solvent to dissolve the sample initially, but then reduce the amount of organic solvent in any subsequent dilution steps. 

I'm having problems transferring a method between a binary and quaternary ultrahigh-pressure liquid chromatography (UHPLC) system. I've calculated and adjusted for the difference in gradient dwell time but still cannot reproduce the chromatography.

In this problem, even though the system dwell volumes have been catered for by altering the gradient start time (either starting the gradient before injection or inserting an isocratic hold at the start of the gradient), the separation selectivity still cannot be reproduced. This is a problem that we are seeing more frequently, and it stems from the differences in the way the gradients are formed and delivered and the inherent pump volumes associated with solvent mixing. It is important to note that we have also seen this problem when transferring methods between different quaternary HPLC systems — so the issue is not only associated with method transfer between high- and low-pressure mixing systems. 

UHPLC systems have markedly different technologies to mix the required gradient composition on-line and, as well as having different gradient composition delay characteristics, there will be differences in the "slope and wash out" characteristics of the gradient delivery profiles as shown in Figure 1. 

Figure 1: Gradient profile determination for two UHPLC systems.			

Unless one can compensate for the differences in the gradient delivery characteristics, than successful method translation may be difficult. Although some UHPLC systems are equipped with software that can simulate differences from certain manufacturers equipment, not all systems are similarly equipped. 

Whilst not ideal, we have recently adopted the following approximation to adjust the gradient profile between systems, after the gradient slope or wash out volume (washout time × flow rate) has been characterized: New segment time = Original segment time × (Original system washout volume/New system washout volume) 

Adjusting the gradient segment times as outlined above will adjust the slope of the gradient delivered by the new system. When combined with a similar adjustment to account for dwell volume differences at the beginning of the analysis, this approach can help to more closely match the solvent composition at the head of the analytical column at any given time within the gradient. 

The Essentials

A series of real life problems submitted by ChromAcademy members

Troubleshooting Real HPLC Problems

A method was developed using high performance liquid chromatography with tandem mass spectrometry detection for the analysis of phthalate metabolites in urine samples. The urine samples were treated using solid-phase extraction columns. Multiple reaction monitoring models were used for quantitative detection with high sensitivity and selectivity. In this work, a fast, sensitive, and accurate quantitative method is provided to detect toxic phthalate esters in urine and assess environmental toxicants.  

Phthalate esters are widely used as plasticizers in manufacturing. Phthalate esters are also used as raw materials of pesticide carriers, paint, adhesives, cosmetics, lubricants, and so on (1). Because phthalate esters are noncovalently bound to the plastic and maintain relatively independent chemical properties, they can leach out of these products and regularly release into the environment as time goes on (2,3). Globally, more than six million tons of phthalates are used each year (4). Presently, phthalate esters have been detected in many different areas in the ecological environment of the world's major industrial countries; they have become one of the most popular global pollutants. This has severely threatened the environment and human health, despite the fact that phthalate esters play an important role in the production and life of mankind (5). 

Diet, skin, and inhalation are considered the major exposure routes for phthalate esters (6). When phthalate esters enter the human body they are rapidly metabolized and hydrolyzed to the corresponding phthalate monoesters before absorption, which are the biologically active molecules and recognized testicular toxicant (7,8), and then further oxidized into various metabolites (9). General metabolic pathways of phthalates in humans are shown in Figure 1 (10). Most phthalate esters and their metabolites are excreted in urine or feces. The total concentrations of phthalate metabolites in urine are normally used as biomarkers of individual intake of phthalate esters (11). The monoesters contain a free reactive carboxylic acid that can conjugate with α-D-glucuronic acid to produce more hydrophilic compounds. Phthalate esters, which mimic estrogen, have been reported to affect multiple biochemical processes as environmental hormones in humans and wildlife. It is one of the major reasons for male reproductive problems (12–14). Excessive exposure to phthalate esters can increase the risk of breast cancer for women. Many phthalate esters can damage the reproductive systems of offspring as well (15–18). For instance, studies showed that some toxicologically active metabolites, such as MEHP, MBPA, and MBzP, are capable of crossing rodent placenta, which causes adverse effects on the development of the male reproductive system and induces early embryonic death, respectively (19,20). Animal experiments have shown that the toxicities of some monoesters are greater than those of corresponding diesters (21).  

Figure 1: General metabolic pathways of phthalates in humans. R = H; phthalic acid = PA, C2H5; monoethyl phthalate = MEP, (CH2)3CH3; monobutyl phthalate = MBP, CH2C6H5; monobenzyl phthalate = MBzP, CH2CH (C2H5) (CH2)3CH3; mono (2-ethylhexyl) phthalate = MEHP.			

Phthalate monoesters are considered to be valuable biomarkers for exposure to phthalate esters. Measuring phthalate monoesters investigates the contamination status, absorption, and metabolism of phthalate esters in the body better than phthalate diesters. Several developed methods have been used to quantify phthalate monoesters, including gas chromatography–mass spectrometry (GC–MS) (22,23). Calafat developed a sensitive method for measuring 13 phthalate metabolites in breast milk using isotope dilution high performance liquid chromatography (HPLC) with negative ion electrospray ionization tandem mass spectrometry (24). This method showed good reproducibility and accuracy. Liquid chromatography–tandem mass spectrometry (LC–MS-MS) methods with high selectivity and sensitivity show an advantage compared to GC–MS methods because of simple sample pretreatment without derivatization. 

Some methods, such as on-line solid-phase extraction (SPE) coupled to HPLC–MS-MS, are considered the most advanced method to investigate phthalate esters and their metabolites, but the high cost limited their widespread application. To develop a sensitive, selective, low cost, and efficient analytical method to quantify these chemicals appears essential for understanding how human health might be affected by exposure to phthalates under general laboratory conditions. In this article, we report a method for the quantitative detection of four phthalate monoesters including MEP, MBP, MBzP, and MEHP simultaneously in human urine samples. The method involves the enzymatic dissociation of conjugated monoesters, SPE, and separation and detection using HPLC with negative ion electrospray ionization tandem mass spectrometry. The method was used to analyze human urine samples collected in the same area for evaluating potential phthalate ester exposure. 

PA, MEP, MBP, MBzP, MEHP (>99.9%), and their isotope-labeled internal standards (>99.9%) were purchased from AccuStandard, Inc. Methanol, acetonitrile, and ethyl acetate were purchased from Sigma Aldrich Laboratories, Inc. β-glucuronidase (

-K12) was purchased from Roche Biomedical. Ammonium acetate (1 M, pH 6.5), formic acid solution (0.1 M), acetic acid solution (0.1%, v/v), and phosphate buffer solution (pH 2.0) were prepared. Water was purified using a Milli-Q gradient A10 system (Millipore). SPE columns were purchased from Varian, Inc. (Nexus ABS Elut, 60 mg/3 mL). Human urine samples (pooled from 84 individuals) were collected in one area and stored at -80 °C. In this study, plastic equipment was not used in any sampling or experimental processes to avoid contamination, and all glass apparatus were washed with a chromic acid solution and rinsed with deionized water and methanol before drying. 

Chromatographic separations were performed on a Waters Quattro Micro LC–MS-MS system (Waters Corporation) with a 150 mm × 2.1 mm Kromasil 100-5C

 column (Akzo Nobel) set at 40 °C. Mobile-phase A was 0.1% acetic acid in water, and mobile-phase B was 0.1% acetic in acetonitrile. The mobile-phase gradient is shown in Table I. The sample injection volume was 20 µL. The flow rate was 0.3 mL/min.  

The MS detector was used in the negative electrospray ionization (ESI) mode. The optimum operation conditions were as follows: ionization voltage: -2.85 kV; cone voltage: 20 V; ion source temperature: 100 °C; drying gas temperature: 300 °C; electron multiplier voltage: 650 V; and collision energy: 16–19 eV. A multiple reaction monitoring (MRM) model was used. The detection channels are shown in Table II.  

Table II: Linearity of phthalate ester analytes			

Urine samples were thawed and vortexed homogeneously. Each 950-µL urine sample was transferred into a glass tube and mixed with 50 µL of internal standard mixture solution (1 µg/mL 

-MEHP). Then, 5 µL of β-glucuronidase (200 U/mL) and 245 µL ammonium acetate buffer (1 M, pH 6.5) were added to the tube and vortexed in turn. Next, the samples were incubated at 37 °C for 90 min in a shaking water bath and treated with an SPE column. The SPE cartridges were conditioned with 1 mL of methanol, 1 mL of acetonitrile, and 1 mL of phosphate buffer solution (pH 2.0) added successively. Then, 1 mL of urine sample was diluted with 1 mL of phosphate buffer solution (pH 2.0) and added to the SPE column. The cartridges were then washed with 2 mL of formic acid solution (0.1 M) and 1 mL of water. The cartridges were dried under negative pressure. The target analytes were eluted sequentially with 1 mL of acetonitrile and 1 mL of ethyl acetate. The eluent was collected together and concentrated to dry under nitrogen at 55 °C. The dry residue was reconstituted with 200 µL of 1:9 (v/v) acetonitrile–water for the next step of the analysis. 

The analytical method was validated to demonstrate the linear range, recovery, limit of detection (LOD), accuracy, and precision. All samples were prepared as described above using blank (or spiked) urine samples. Urine samples with a series of concentrations (2.5, 5, 10, 50, 100, 200, 500, 1000, and 2000 ng/mL standard mixture solutions, including PA, MEP, MBP, MBzP, and MEHP) were prepared to investigate the linear range, recovery, and LOD. The LOD and limit of quantification (LOQ) were calculated using signal-to-noise of 3 and 10 times based on the blank samples, respectively. Samples spiked at three concentration levels (3.91, 7.81, and 15.63 ng/mL) were used to determine the accuracy of the method. Three repetitions at concentrations of 5, 10, and 50 ng/mL were used to determine the intraday and interday precision in one day and the successive three days, respectively.  

In this study, 84 individual human urine samples were collected in the same area and stored at -80 °C until analysis. The pretreatment and HPLC–MS-MS methods used to analyze these human urine samples were described above. 

To obtain specific, accurate, and sensitive results for quantification during method development, the MS conditions were first optimized by direct infusion. Collision energies were optimized to give the greatest fragmentation with the maximum response for each target analyte. The range of the collision energy was fine tuned from 13–16 eV to 16–19 eV because too much collision energy can cause a loss of valuable stable fragment peaks. The best response of each phthalate ester was about 10

  for MS-MS spectra (see Figure 2). The parent ion and daughter ion for each phthalate in the MRM detection mode was chosen and is shown in Table II.  

Figure 2: MS and MS-MS optimization of the representative target analyte: MBzP.			

The chromatographic conditions were optimized using the instrumentation and column described in the Experimental section of this article. The ratio between the two phases and gradients was investigated using a 30-min separation time and a solvent gradient ranging from 94% mobile-phase A to the final concentration shown in Table I. The target analytes were significantly separated. However, detection development time should be as short as possible while still meeting the resolution required. Because these phthalate esters (including PA) have a polar group such as -COOH, the proportion of organic phase (mobile-phase B) was increased to elute the chromatographic peaks faster from the C18 column. The whole elution time was shortened to 12 min, the initial ratio of the water phase was dropped to 70% with the same gradient procedure, and the five target analytes were separated significantly (Figure 3a and 3b). The chromatographic peaks of MBP and MBzP were overlapped when the water phase was below 70%.  

Figure 3: Optimization of chromatographic conditions: (a) 30-min and (b) 12-min elution times.			

The calibration curves were obtained using standard solutions of five phthalates with concentrations ranging from 5 ng/mL to 2000 ng/mL. The correlation coefficients are shown in Table II. The results show good correlation (the correlation coefficient values were greater than 0.99) between the peak areas and the concentrations of the phthalates. The LODs of the five phthalates were 0.85–5.33 ng/mL and the LOQs were 2.82–17.76 ng/mL. The intra- and interday precision of the method are summarized in Table III. The RSDs of those five phthalate esters ranged from 1.74% to 14.24%. Samples spiked at levels 3.91, 7.81, and 15.63 ng/mL were analyzed three times each. The method validation studies for spiked samples indicated that the present method provides good recoveries and reasonable precision for phthalates at these three levels. As can be seen from Table IV, the recoveries were 81.84% and 125.32% with RSDs between 0.07% and 10.20%. These results demonstrate that the method developed has acceptable precision. 			

Figure 4: Representative LCâMS-MS chromatograms; the blue chromatogram is a nonspiked human urine sample; the red chromatogram is a spiked human urine sample with 200 ng/mL of the five mixture standards.			

The LC–MS-MS method described above was applied to the analysis of 84 human urine samples collected in one area (Figure 4). Table V presents the summary of four phthalate monoester metabolites and PA in these human urine samples. Overall, the detection rates of PA, MEP, MBP, MBzP, and MEHP were 100%, 32%, 99%, 1%, and 21%, respectively. The detection mean concentrations of PA, MEP, MBP, MBzP, and MEHP in all urine samples were 5.31, 16.45, 96.76, 0.07, and 3.83 ng/mL, respectively. The detection concentration ranges of PA, MEP, MBP, MBzP, and MEHP in all urine samples were 1.4–31.3, 0–398.7, 0–663.7, 0–3.1, and 0–184.1 ng/mL, respectively. The results showed that the subjects were exposed to some phthalate esters, especially MEP, MBP, and MEHP with high concentrations in that region. This study confirmed detection levels of four phthalate metabolites and PA in human urine samples. 

Table III: The intra- and interday precision of the method			

We developed an LC–MS-MS method for simultaneously analyzing phthalate ester metabolites in human urine. The method has proven to be sensitive, accurate, and precise with a relatively short chromatographic run time of 12 min. The results demonstrate that this method can be of value in assessing human exposure to phthalates or other environmental toxicants in environmental toxicology studies.  

Table IV: Recoveries of five phthalate esters in human urine samples			

This work was supported by funding from the National Natural Science Foundation of China (21107141, 81273478), the Chinese Scientific and Technological Major Special Project (2012ZX09301003-001-010), and the Major Infectious Disease Special Projects (2013ZX10003002-006). Jianhua Cheng and Haijing Li contributed equally to this work. 

Table V: Concentrations of five phthalate esters detected in human urine samples (n = 84)			

(2) S.Y. Yuan, C. Liu, C.S. Liao, and B.V. Chang, 

(3) H. Faraji, A. Mirzaie, and S. Waqif-Husain, 

(8) R. Habert, V. Muczynski, A. Lehraiki, R. Lambrot, C. Lecureuil, C. Levacher, H. Coffigny, C. Pairault, D. Moison, R. Frydman, and V. Rouiller-Fabre, 

Philos. Trans. R. Soc. Lond. B Biol. Sci. 

(10) E. Samandar, M.J. Silva, J.A. Reidy, L.L. Needham, and A.M. Calafat, 

Peer-Reviewed Articles

This fast, sensitive method provides accurate quantitation.

Analysis of Four Phthalate Monoesters in Human Urine Using Liquid Chromatography–Tandem Mass Spectrometry

Columns packed with superficially porous particles (SPPs) have created considerable excitement over the last few years. Indeed, this column technology manifests the advantages of fully porous material (loading capacity, retention) and some beneficial properties of nonporous particles (kinetic performance). This review provides an updated overview of the theory behind the success of SPP technology, trends, benefits, and limitations. It also summarizes the latest developments of sub-2-µm SPPs and instrumental constraints associated with their use. Finally, it describes several applications to illustrate the performance and the universal applicability of these newly engineered particles.

				Superficially porous particles (SPPs) (also called core–shell, fused-core shell, partially porous, pellicular, or solid-core) are made of a solid, nonporous silica core surrounded by a porous shell layer with similar properties to those of the fully porous materials used in conventional high performance liquid chromatography (HPLC) columns. The "fused-core" terminology was originally introduced by Jack Kirkland to describe the manufacturing procedure that "fuses" a porous silica layer onto a solid silica particle (1). 

The very high efficiency achieved on columns packed with sub-3-µm SPPs, combined with convenient operating conditions (modest back pressures and the ability to use conventional HPLC instruments) has generated significant interest in the chromatographic community and widespread applications in many fields (2). Columns packed with sub-3-µm SPPs rival the efficiency of columns packed with sub-2-µm fully porous particles, but the former generate only half the back pressure. As a result, practitioners can use such columns on regular HPLC equipment, leading to the initial interest in these materials and their successful application. Moreover, further performance improvement is possible with very fine SPPs (1.3–1.7 µm), though the use of ultralow-dispersion ultrahigh-pressure liquid chromatography (UHPLC) systems (for example, extracolumn peak variance σ

) is mandatory. Faster analysis and higher efficiency is always desirable in liquid chromatography (LC), particularly to pharmaceutical scientists or researchers in life science wishing to attain higher productivity in the laboratory or more accurate analysis (higher resolution) of very complex samples. Reducing analysis time while maintaining resolution requires high kinetic performance (more separation power per unit time) using smaller particles, better particle morphology, or both. 

The initial intent of applying SPPs was to enhance kinetic performance in the analysis of large biomolecules such as therapeutic proteins. The rationale behind this concept was to improve column efficiency by shortening the diffusion path that molecules have to travel and, thus, improve their mass transfer kinetics (3,4). Shell-type (pellicular) particles were first developed by Horváth and colleagues in the late 1960s for the analysis of large molecules in ion-exchange mode (5). Shortly afterward, Kirkland demonstrated that superficially porous particles (pellicular materials) with 30–40 µm diameters could provide much better separations than totally porous ones (6). In 2001, a new column for fast protein or peptide analysis was introduced that was packed with a 5-µm SPP with a shell thickness of 0.25 µm.  

In 2007, a revolution started with the commercialization of a new generation of sub-3-µm SPPs adapted for separation of small and large molecules (7) by Advanced Materials Technology. This material possesses a 1.7-µm solid core covered by a 0.5-µm-thick shell of porous silica. It combined the advantages of both fully porous and nonporous particles. In particular, this improved particle design solved the problem of the low loading capacity of early pellicular particles because an approximately 75% volume fraction of these particles is still porous. Since then, many providers commercialized SPPs with particle sizes ranging between 1.3 and 5 µm. Table I provides a current list of commercially available columns packed with SPP materials, their pertinent properties, and available bonded phases.  

Table I: A current list of HPLC columns made with commercially available SPP materials including the newest sub-2-Âµm SPPs			

Column Performance: The Impact of Shell Thickness

In 1956, the famous van Deemter equation was proposed to describe column performance (plate height) as a function of linear velocity (8). Since then, several plate height and rate models have been derived for LC by numerous researchers. Knox suggested a three-term equation to describe the dependency of the theoretical plate height (

) of a column as a function of linear velocity (

 are constants, determined by the magnitude of band broadening because of eddy dispersion, longitudinal diffusion, and mass transfer resistance, respectively. In practice, the 

 term is the sum of contributions from resistance to both trans-particle mass transfer (

The application of reduced parameters is common in engineering to compare the performance of columns in normalized units. It is useful to convert 

An updated overview of the development of SPP technology, including benefits, limitations, and the theory behind its success

LCGC North America-06-01-2014

Radical Mass Spectrometry as a New Frontier for Bioanalysis

UV Detector Problems

Exploring the Possibilities of High-Throughput Sample Preparation

Troubleshooting Real HPLC Problems

Analysis of Four Phthalate Monoesters in Human Urine Using Liquid Chromatography–Tandem Mass Spectrometry

Superficially Porous Particles: Perspectives, Practices, and Trends

The 2014 Emerging Leader in Chromatography: Andre de Villiers

Vol 32 No 6 LCGC North America June 2014 Regular Issue PDF