Recently, I had the opportunity to participate in Colacro XII, a chromatography symposium held in a Latin America country every two years. This year's meeting was held in Florianopolous, Brazil, October 27-30. During the symposium, I perused the large number of applications posters for the sample preparation techniques being used. I was surprised to find many of the investigations of such diverse matrices as natural products, fruits and vegetables, petroleum products, and body fluids used solvent extraction (liquid–liquid extraction, or LLE) as the initial (or the only) sample preparation technique. So, I thought that I would devote an installment of "Sample Prep Perspectives" to this well used technique. In our last comprehensive sample preparation survey, nearly 40% of the respondents reported on the use of LLE as at least one of their routine sample preparation procedures (1).  

Earlier, I wrote on the basics of LLE providing some of the theory (which will not be repeated here), examples of classical LLE including continuous and countercurrent chromatography, and some of newer techniques of the time (2). More recently, I covered approaches for miniaturization of classical LLE (3). In this installment, I will provide some practical hints to those who are considering the use or already are using LLE in hopes of improving your methodology. 

LLE is performed using two immiscible liquids and soluble samples. LLE is useful for separating analytes from interferences by partitioning the sample between these two immiscible liquids or phases. Usually, one phase in LLE will be aqueous (often the denser or heavier phase) and the second phase is an organic solvent (usually the lighter phase). The more hydrophilic compounds prefer the polar aqueous phase, while more hydrophobic compounds will be found mainly in the organic solvent. Analytes extracted into the organic phase are recovered easily by evaporation of the solvent, while analytes extracted into the aqueous phase often can be injected directly onto a reversed-phase high performance liquid chromatography (HPLC) column. Obviously, when the analyte of interest in found in the aqueous phase, evaporation is a much slower process. However, the following discussion assumes that an analyte of interest is concentrated preferentially into the organic phase, but similar approaches are used when the analyte is extracted into an aqueous phase.  

The generic flow diagram of Figure 1 summarizes the steps involved in a typical LLE separation. Because extraction is an equilibrium process with limited efficiency, significant amounts of the analyte can remain in both phases. Chemical equilibria involving changes in pH, ion-pairing, and complexation can be used to enhance analyte recovery and eliminate interferences.  

Figure 1: Classical generic LLE flow chart. 

In its classical and simplest form, LLE is performed conveniently in a separatory funnel where the two immiscible phases are added from the top and, after the extraction process, the heavier phase is drained out the bottom stopcock. However, LLE can be performed in other devices such as beakers, vials, centrifuge tubes, and graduated cylinders. In these cases, the liquids must be removed by some type of pipette and great care must be exercised not to disturb the interface layer. Otherwise, the extraction process may be incomplete. 			

The LLE organic solvent is chosen for the following characteristics: 					

Volatility for easy removal and concentration after extraction.

Compatibility with the HPLC or gas chromatography (GC) detection technique to be used for analysis (for example, avoid solvents that are strongly UV-absorbing for LC and chlorinated solvents for GC using electron-capture detection).

High purity because extracted analytes often are recovered by the evaporation of large volumes of organic solvent; impurities in the solvent can be concentrated as well.

Polarity and hydrogen-bonding properties that enhance recovery of the analytes in the organic phase — that is, increase the value of 

Several approaches can be used to increase the value of 

If the analyte is ionic or ionizable, its 

 can be increased by suppressing its ionization to make it more soluble in the organic phase. The analyte also can be extracted into the organic phase by ion pairing provided that the analyte is ionized and a suitable ion-pair reagent is added to the organic phase.

–"Salting out" can be used to decrease an analyte's concentration in the aqueous phase, by addition of an inert, neutral salt (for example, sodium sulfate) to the aqueous phase. 

Table I provides examples of typical extraction solvents, as well as some unsuitable (water-miscible) extraction solvents. Apart from miscibility considerations, the main selection criteria is the polarity 

' of the solvent in relation to that of the analyte. Maximum 

 occurs when the polarity of the extraction solvent matches that of the analyte. For example, the extraction of a polar analyte from an aqueous sample matrix would be best accomplished with a more polar (large 

') organic solvent. An optimum-polarity organic solvent can be selected conveniently by blending two solvents of different polarity (for example, hexane and chloroform), and measuring 

 versus the composition of the organic phase (4). A solvent mixture that gives the largest value of 

 is then used for the LLE procedure. Further changes in 

 can be achieved, with improvement in the separation of analytes from interferences, by varying organic-solvent selectivity. Solvents from different regions of the solvent-selectivity triangle (5) are expected to provide differences in selectivity. 			

In solvent extraction, the organic species of interest sometimes can be transferred into either phase, depending upon the selected conditions. For example, consider the extraction of an organic basic analyte (a primary amine) from an aqueous solution. If the aqueous phase is buffered at least 1.5 pH units above its p

 value, the analyte will be neutral and prefer the organic phase; more polar interferences will be extracted into the aqueous phase. If the pH of the aqueous solution is lowered (<<p

), so that the analyte is now ionized (protonated), it will be extracted into the aqueous phase, leaving less polar interferences in the organic phase. The principles of acid–base extraction as a function of pH are the same for LLE and HPLC. 

 is unfavorable, additional extractions may be required for improved recovery. In this case, a fresh portion of immiscible solvent is added to extract additional solute and all extracts are combined. Generally, for a given volume of final extracting solvent, multiple extractions are more efficient in removing a solute quantitatively than the use of a single extraction volume. Back extraction can be used to further reduce interferences. For example, consider the previous example of a basic analyte. If the analyte is first extracted at high pH into the organic phase, polar interferences (for example, hydrophilic neutrals, ionized acids) are left behind in the aqueous phase. If a fresh portion of low-pH aqueous buffer is used for the back-extraction of the organic phase, the ionized basic compound is transferred back into the aqueous phase, leaving nonpolar interferences in the organic phase. Thus, a two-step back-extraction allows the removal of both acidic and neutral interferences, whereas a one-step extraction can eliminate one or the other of these interferences, but not both. 

 is very small or the required quantity of sample is large, it becomes impractical to carry out multiple extractions for quantitative recovery of the analyte. Too many extractions are required, and the volume of total extract is too large. Also, if the extraction is slow, a long time might be required for equilibrium to be established. In these cases, continuous LLE can be used, where fresh solvent is recycled continually through the aqueous sample. Continuous extractors using heavier-than-water and lighter-than-water solvents have been described (6).  

For more efficient LLE, a countercurrent distribution apparatus can provide a thousand or more equilibration steps (but with more time and effort). This allows the recovery of analytes having extremely small 

 values; countercurrent distribution also provides a better separation of analytes from interferences. Small-scale laboratory units are available commercially. For further information on these devices, see (7).  

In some cases, LLE can enhance analyte concentration in the extract fraction. Equation 1 shows the fraction of analyte extracted (

 is the concentration of the analyte in the organic phase, 

 is the concentration of the analyte in the aqueous phase, 

Thus, by choosing a smaller volume of organic solvent, the analyte concentration can be increased by the volumetric ratio of organic to aqueous phases (assuming complete extraction into the organic phase). For example, assume 100 mL of aqueous sample, 10 mL of organic solvent, and a large 

 > 1000). The concentration of the analyte in the organic phase will then increase by a factor of 10. For large ratios of aqueous to organic solvents, the organic solvent can be slightly soluble in the aqueous phase. This effect can reduce the volume of the recovered organic solvent and make the calculation of analyte recovery uncertain. This problem can be avoided by presaturating the aqueous solvent with organic solvent.  

 is small, the physical manipulation of two phases becomes more difficult. Microextraction is another form of LLE where extractions are carried out with organic–aqueous ratios of 0.001 to 0.01. In classical microextraction, analyte recovery can suffer compared to conventional LLE, but the analyte concentration in the organic phase is increased greatly and solvent usage is reduced greatly. Such extractions are carried out conveniently in a volumetric flask. In classical microextraction, the organic extraction solvent is chosen to have a density less than that of water, so that the small volume of organic solvent accumulates in the neck of the flask for easy removal. For quantitative analysis, internal standards should be used and extractions of calibration standards carried out.  

A new variation of microextraction, called dispersive liquid–liquid microextraction (DLLME) has been applied successfully to a variety of analyte–matrix pairs (8). The technique is based upon a three-component solvent system. In DLLME, the container is usually a centrifuge tube and the appropriate mixture of immiscible organic extraction solvent that is heavier than water (usually a few microliters, ~8 μL of tetrachloroethylene) and a dispersive solvent (for example, a milliliter or so of acetone) is injected rapidly into an aqueous solvent (approximately 5 mL) with a syringe. The role of the dispersive solvent is to ensure miscibility between the organic phase and the aqueous phase. When the solvents are mixed rapidly, a cloudy solution is formed. This cloudy solution actually consists of fine particles (droplets) of extraction solvent, which is dispersed entirely into the aqueous phase. Because the solution is already finely dispersed, no vigorous shaking is required. Extraction time is almost instantaneous and much faster than solid-phase microextraction (SPME) and liquid-phase microextraction (LPME), which often can require 30 min or longer. Next, the entire mixture is centrifuged (1.5 min at 6000 rpm) and organic extraction solvent fine droplets are sedimented in the bottom of the centrifuge tube and are removed with a microsyringe or micropipette. The organic solution containing the analyte of interest can be injected directly or evaporated and taken up in a more appropriate solvent for the chromatographic technique. In their original work, Rezaree and colleagues (8) were able to achieve 

 values of approximately 600–1100 for polynuclear aromatic hydrocarbons from a water sample with excellent recoveries. 

Recently, Caldas and coworkers (9) were able to show excellent extractions of pesticides (carbofuran, clomazone, and tebuconazole) in aqueous samples using DLLME along with LC with tandem mass spectrometry detection (MS-MS). For their work, they found that a 60-μL portion of extraction solvent (carbon tetrachloride) and 2 mL of dispersive solvent (acetonitrile) for a 5-mL sample of fortified water acidified with phosphoric acid at pH 2.0 gave the best results. Because carbon tetrachloride gave chromatographic problems, they evaporated the solvent to dryness and dissolved the residue in HPLC-grade methanol. The variables in DLLME method development include: choice of dispersion solvent and its volume, choice of extraction solvent and its volume, pH if necessary, and centrifuge speed. Figure 2 shows the results of their optimization studies on extraction solvent and extraction solvent volume (9). For all three pesticides, the linear range was found to be 0.001–1.0 mg/L and the limit of quantitation (LOQ) was 0.02 mg/L. Overall, the DLLME technique is simple and fast, provides good recovery, is low cost, and provides good enrichment factors.  

				Some practical problems associated with classical LLE include					

Analytes bound to high molecular weight compounds (for example, drugs to proteins)

Emulsions are a problem that can occur with certain samples (for example, fatty matrices) and certain solvent conditions. An emulsion represents a dispersion of organic liquid within the aqueous environment inside the separatory funnel and has a lack of boundary between the organic and aqueous phases. If emulsions are not "broken" with a sharp boundary between the aqueous and organic phases, analyte recovery can be affected adversely. Emulsions can be broken by 					

Filtration through phase separation filter paper

Addition of a small amount of different organic solvent

If particulates are present in a sample, adsorption onto the particulates can result in a low recovery of the analyte. In such cases, washing the particulates after filtration with a stronger solvent will recover the adsorbed analyte; this extract should be combined with the analyte phase from LLE. A "stronger" solvent for recovering adsorbed analyte might involve a change in pH, increase in ionic strength, or the use of more-polar organic solvents. 

Compounds normally recovered quantitatively in LLE can bind to proteins when plasma samples are processed, resulting in low recovery. Protein-binding is especially troublesome when measuring drugs and drug metabolites in biological fluids. Techniques for disrupting protein binding in plasma samples include 					

Addition of organic solvent, chaotropic agents, or strong acid

Displacement with a more strongly binding compound

Articles

Sample Prep Perspectives

Columnist Ron Majors discusses some of the practical considerations in the successful application of the popular yet age-old technique of solvent extraction (also known as liquid–liquid extraction, or LLE).

Practical Aspects of Solvent Extraction

Here we are, more than 40 years after the initial exploration of high performance liquid chromatography (HPLC). Many things have changed, some innovations have come and gone, but the reciprocating-piston pump still remains as a key component of the LC system. We've tried pneumatically driven pumps, piston-diaphragm pumps, syringe pumps, and some other ideas, but none have proven to be superior to the reciprocating-piston pump. Yes, some changes have been made, but the basic operation remains the same. This month's "LC Troubleshooting" installment will take a look at the design of these durable pumps and also examine some of the potential weaknesses and how to overcome them. 

The basic design and operation of the reciprocating-piston pump is illustrated by the single-piston pump shown in Figure 1. The key components are a piston, pump seal, pump head, and a couple of check valves. The piston, usually made of sapphire, is driven back and forth in the pump head by a rotating motor. Various means have been conceived to convert the rotary motion of the motor into the bidirectional movement of the piston. Most commonly, this is done by a cam pressing against one end of the piston to push it into the pump head and a spring to push the piston back out. The pump seal is shown in the inset in Figure 1c. It is a polymer ring that fits around the piston, and a small lip forms a liquid-tight seal against the piston with the aid of a spring and liquid pressure. A pair of ruby check valves with sapphire seats are mounted on the top and bottom of the pump head. The check valves control the direction of flow through the pump. On the intake stroke (Figure 1a), the piston is withdrawn, which creates a low pressure area inside the pump head. This allows the outlet check valve to close and the inlet check valve to open, so that mobile phase flows in to fill the pump head. On the delivery stroke (Figure 1b), the piston moves into the pump head, and the inlet check valve is closed as the pressure increases. When the pressure inside the pump head exceeds the pressure in the column, the outlet check valve opens and mobile phase flows to the column. When all is working well, this simple pump design is quite reliable. 

Figure 1: Single-piston reciprocating pump: (a) Pump in intake, or fill cycle; (b) pump in delivery cycle; and (c) partial section of piston showing cut-away portion of piston seal. See text for discussion.			

The weak points in the design are the check valves and the pump seal. The operation of the check valves and some design improvements were discussed in the June 2008 installment of "LC Troubleshooting" (1). Contamination of the check valves can cause them to leak, and under the right circumstances, acetonitrile in the mobile phase can cause the inlet check valve to stick closed. The pump seal forms a seal against the moving piston, and although the sapphire piston is very smooth, the seal eventually will wear out. This wear is accelerated if buffer is allowed to sit in an unused pump, because the liquid behind the seal evaporates and leaves an abrasive layer of buffer crystals. These abrade the seal when the pump is restarted, and can shorten the seal life. Generally the pump seals will last 6–12 months under normal operation if buffers are rinsed from the system before shutdown. 

Bubbles of air trapped in the pump head will cause the pump to under-deliver mobile phase. Bubbles can result when mobile phases are not degassed sufficiently. Fortunately, most of today's LC systems incorporate an inline vacuum degasser to degas the mobile phase automatically. So if the pump is purged of bubbles when it is started, bubble problems in the pump are not a common problem today — they were the bane of the chromatographer in years past. 

The original reciprocating-piston pumps were crude by today's standards. I remember using the Milton–Roy Minipump, one of the standard pumps in the late 1960s and early 1970s, particularly for laboratory-built systems that were common in graduate school laboratories. The flow rate was controlled by adjusting a stop that limited the distance the piston could move. With a single-piston design, the pump spent half the time filling and half the time delivering mobile phase. This resulted in pulses of pressure and flow that were reflected in a fluctuating baseline and short column lifetimes of the then-poorly-packed columns. Huge pulse dampeners, sometimes 50–100 mL in volume, were required to reduce the pulses to an acceptable level. Refinements in the shape of the driving cam and the use of stepper-driven motors helped to minimize pulses with single-piston pumps, but as other aspects of the LC system were improved, single-piston pumps were not sufficiently pulse-free to be acceptable. 

The simple cure for the pulsating nature of the single-piston pump was to operate two such pumps so that when one piston was filling, the other was delivering solvent. This is shown in the diagram in Figure 2a, where the left-hand piston is at the end of its fill stroke and the right-hand piston is at the end of its delivery stroke. The solvent delivered from both pump heads is combined into a single flow-stream. Because one piston is always in the delivery cycle, solvent flows continuously to the column and pulses are minimized or eliminated. In practice, dual-piston pumps are configured so that the pistons are mounted side-by-side in parallel. The dual-piston pump was developed early in the history of HPLC and remains as one of the two most popular designs in use today. 

Figure 2: Two-piston pump designs: (a) Dual-piston pump with left piston in fill cycle and right piston in delivery cycle and (b) accumulator-piston pump with top piston in delivery cycle and bottom piston in fill cycle. Arrows show direction of motion of pistons; pressure and flow profiles are shown at right. See text for discussion.			

A later development in pump design is the accumulator-piston pump shown in Figure 2b. This also is called a tandem-piston pump. In this design, the two pistons deliver solvent at two different flow rates. For example, if the pump of Figure 2b is set to deliver 1 mL/min of solvent to the column, the top piston (shown at the end of the delivery stroke) pumps at 1 mL/min. Meanwhile, the bottom piston fills at 2 mL/min. Next, the top piston fills at 1 mL/min. The 2 mL/min delivered by the bottom piston is split so that 1 mL/min serves to fill the top piston and the other 1 mL/min flows directly to the column. In this manner, solvent always flows to the column at 1 mL/min. 

One purported advantage of the accumulator-piston pump is the reduced number of check valves. In the diagram of Figure 2b, three ball-type check valves are used, as opposed to the four check valves required by the dual-piston pump of Figure 2a. Furthermore, if there is constant flow to the column, supplied by either the top or bottom piston, the outlet check valve is always open, so it is not needed. Some implementations of this pump do not use an outlet check valve — now only two check valves are used. If the inlet check valve is replaced by an active check valve (see discussion in reference 1), only one ball-type check valve remains. This is in contrast to two ball-type check valves if the inlet check valves are replaced by active check valves in the dual-piston pump. In both cases, the number of check valves is reduced for more reliable operation.  

The next advance in pump technology was the addition of the ability to mix two or more solvents online. Before this, isocratic mobile phases (for methods in which the mobile phase composition is constant) were hand-mixed. Gradients (in which the mobile phase composition changes during the run) were formed by use of a crude dilution flask, siphon, or other gradient former. Online mixing comes in two primary forms — high- and low-pressure mixing, illustrated for two-solvent mixing systems in Figure 3.  

Figure 3: Online mixing systems: (a) High-pressure mixing and (b) low-pressure mixing. See text for discussion.			

High-pressure mixers are the simplest in concept (Figure 3a). Two separate pumps are used, with one pump delivering the A-solvent (water or buffer for reversed-phase separations) and the second delivering the B-solvent (generally acetonitrile or methanol). The ratio of the flow rates for the two pumps determines the A/B ratio. Constant flow from both pumps is used for isocratic separations, whereas changing flow rates of the two pumps can form a gradient at constant flow rate. 

Low-pressure mixers came a bit later. In these, two, three, or four solvents are combined at a low-pressure proportioning manifold prior to the pump (Figure 3b for a two-solvent system). A single pump operates at constant flow rate. The solvent mixture is controlled by alternately opening proportioning valves to allow one of the solvents to flow into the proportioning manifold. For isocratic mobile phases, the solvent proportions stay constant, whereas for gradients, the open-close cycles of the various valves is programmed to vary during the run. This design easily can handle four different solvents, whereas the high-pressure mixing requires a separate pump for each solvent, and is limited to binary mixtures unless more pumps are added. It should be obvious that continuing improvements in pump controllers have improved the precision and accuracy of blending solvents online. 

In both designs of online solvent blending, active or static mixers are used to facilitate mixing the mobile phase. All of the various components from the point the solvents are first blended until they reach the head of the column comprise the dwell volume, as shown inside the dashed lines for the systems in Figure 3. This dwell volume is of little concern for isocratic separations. For gradients, however, the dwell volume becomes more important, because it results in a delay between the time the sample is injected and the time the gradient reaches the head of the column. Differences in dwell volume account for one of the primary reasons gradient methods are difficult to transfer from one LC system to another. LC systems designed for conventional separations (for example, using 150 mm × 4.6 mm columns) have dwell volumes in the range of =1.5–4 mL. A more detailed discussion of the effects of dwell volume and how to measure it was presented in an earlier "LC Troubleshooting" column (2). 

There is a delicate balance between having the mixing volume large enough to provide acceptable solvent mixing and having the dwell volume small enough that excessive gradient delay is not encountered. Furthermore, if the mixing volume is too large relative to the gradient volume (gradient run time × flow rate), gradient distortion can occur (3). For a 20-min gradient at 1.5 mL/min, as might be typical for a 150 mm × 4.6 mm column, the gradient volume is 30 mL. For the equivalent gradient with a 50 mm × 2.1 mm column, as is used commonly for LC methods with mass spectrometry detection (LC–MS), a 6-min gradient at 0.3 mL/min (gradient volume = 2 mL) might be used. It has been reported (3) that the mixing volume (which accounts for 25–50% of the dwell volume) should be ≤10% of the gradient volume. This means that for good performance with the newest column technology, such as 50 mm × 2.1 mm columns containing 3-μm particles or 50–100 mm × 1.0 mm columns with sub-3-μm particles, dwell volumes of <1 mL are required. LC systems that accommodate these separation conditions must have small internal volumes, yet provide adequate solvent mixing. 

The simple reciprocating-piston pump that was used in the earliest LC systems remains as the central design element of today's higher performance pumping systems. Dual-piston and accumulator-piston pump designs are most common today, and design refinements allow them to generate highly precise and accurate isocratic or gradient mobile phases. The challenges of smaller-particle columns, pressures of more than 6000 psi (400 bar), and faster gradient runs have continued to keep pump designers busy modifying today's pumps for tomorrow's applications. 

"LC Troubleshooting" Editor John W. Dolan is Vice-President of LC Resources, Walnut Creek, California; and a member of 

's editorial advisory board. Direct correspondence about this column to "LC Troubleshooting," 

, Woodbridge Corporate Plaza, 485 Route 1 South, Building F, First Floor, Iselin, NJ 08830, e-mail 

  For an ongoing discussion of LC trouble-shooting with John Dolan and other chromatographers, visit the Chromatography Forum discussion group at 

(1) J.W. Dolan, LCGC 26(6), 532–538 (2008). 

(2) J.W. Dolan, LCGC 24(5), 458–466 (2006).  

(3) G. Hendriks, J.P. Franke, and D.R.A. Uges, J. Chromatogr., A 1089, 193 (2005).

LC Troubleshooting

This month's "LC Troubleshooting" installment will take a look at the design of these durable pumps and also examine some of the potential weaknesses and how to overcome them.

LC Pumps

Part III concludes the print version of our three-part series, "A Mass Spectrometry Primer," with the glossary to follow in a fourth and final installment. To recap the basis for undertaking this work, technological changes that affect our knowledge, in terms of its depth and our speed of accessing it, spur a couple of observations about print media versus electronic. Anything published must be scholarly because it often serves as a primary resource for certain facts, equations, and other things we cannot, or would not, remember. Paradoxically, however, almost as soon as we commit words to describing or explaining a fast-developing area of technology, their value diminishes as new insights form. And though the marvels of electronic communication are considerable, they nevertheless fail to resolve all of the print's shortcomings as a static medium. For example, increasingly ubiquitous web logs usually are focused narrowly and posted by a single individual. We have yet to benefit from the deeper understanding of our science that interactive conversations could bring.  

As for primers, they abound in various forms and by various authors (I reference some of them here). But this primer, which lives on the internet, differs from all others. It is a self-validating document, continually updated with the comments and suggestions of American and European scientists. I invite your observations on this electronic primer, which you can access by visiting the Waters website at 

 and then clicking Resource Library > Primers. 

Increased, measured mass accuracy and resolution is now a dominant tool for structural characterization in various applications beyond early drug discovery. With their broad reach of specificity and utility, quadrupole time-of-flight (QTOF) instruments, offered by a number of manufacturers today, are replacing other LC–MS technologies. 

Figure 1: The effect of increasing mass accuracy for unambiguous identification of compounds (T.L. Quenzer, J.M. Robinson, B. Bolanios, E. Milgram, and M.J. Greig, Automated accurate mass analysis using FTICR mass spectrometry, Proceedings of the 50th Annual Conference on Mass Spectrometry and Allied Topics, Orlando, Florida, 2002).  

Although there are even higher order instruments a QTOF's high mass accuracy, which falls within a few parts per million of the true, calculated, monoisotopic value, and its high resolution — as much as 10 times higher than a quadrupole instrument's — permits us to determine empirical formulas according to mass defect (where the critical mass value of hydrogen and other atoms present serve as a differentiator). Speciation analysis — discerning the difference between an aldehyde and a sulfide, for example — becomes possible with an increase in mass accuracy above the quadrupole limits to 30 ppm, where the two masses differ by 0.035 Da. Differentiation between the metabolic processes involving methylation is more demanding, however. Adding CH

 produces an increase over the precursor (response for the drug alone) in the measured mass of +14.0157 Da, as compared with a two-stage biotransformation involving hydroxylation (addition of oxygen) followed by oxidation at a double bond (loss of H

), which produces an increase of +13.9792 Da. Yet both measurements, when limited by nominal resolution — a typical quadrupole response — will look like +14 Da. 

See MS - The Practical Art, LCGC (chromatographyonline.com)  

 Low-resolution quadrupole instruments perform well for extremely high mass-accuracy measurements, like those used for analyzing proteins. The masses of proteins generally are defined as "average" values when the isotope peaks are not resolved relative to each other. Average mass is the weighted mean of all the isotopic species in a molecule. The instrumental resolution normally employed on quadrupole instruments broadens the resolved response for a 10-kDa protein by a factor of x1.27. That factor increases significantly as the mass increases (for example, to x2.65 at 100 kDa). However, by reducing the peak width to 

 0.25 (increasing resolution to 4000 resolution) rather than limiting the instrument resolution to 1000 using the typical peak width (

 0.6) improves the situation dramatically. 

In practice, electrospray ionization (ESI)–MS analyses of large molecules produce multiply charged ions. Hence, the widths need to be divided by the number of charges on an ion to give the width on the mass-to-charge ratio scale. For example, a 20-kDa protein with 10 or 20 charges on it will produce isotope envelopes that are 0.9 or 0.45 

When these ions are observed on an instrument set for a significantly lower resolution than that required to resolve the isotopes (say less than 10,000 resolution), a single peak is produced for each charge state. The overall width is determined by combining the instrument peak width with the theoretical width of the isotopic envelope divided by the number of charges on the ion. The instrumental peak width would be determined on the first isotope peak of a low molecular weight compound at the same 

 value as the multiply charged protein peak. 

How much accuracy do we need, or can we realistically achieve, and what are the compromises?

 Consider the requirements for unambiguous characterization from the 

Journal of the American Society for Mass Spectrometry

 author's guidelines (March 2004). For C, H, O, N compositions (C

 response at 118 needs only an error not exceeding 34 ppm to be unambiguous, where a 

 response at 750 requires precision better than 0.018 ppm to eliminate "all extraneous possibilities." 

Comparing precision from instrument to instrument: millimass units (mmu), measurement error (ppm), and resolution

 According to the Accurate Mass Best Practice Guide of the VIMMS Programme, an initiative that forms part of the UK National Measurement System, most instruments used for accurate mass measurements are capable of achieving precision of 10 ppm or better. 

A calculated mass of 118 Da measured by a modern mass spectrometer to within 2 mmu accuracy would display 17 ppm error, sufficient by today's standards for unambiguous determination of a chemical formula of that mass:  

Monoisotopic calculated exact mass = 118 Da

An instrument capable of a response at 750 

, also deficient by 2 mmu, would be accurate to 2.7 ppm. In the first case, the measurement is more than sufficient for unambiguous identification of a chemical formula, according to the published standards of the 

Journal of The American Society for Mass Spectrometry

. But in the latter case, the measurement is insufficiently precise. Only the highest order Fourier-transform ion cyclotron resonance (FTICR) MS can achieve such precision at higher masses. 

A comprehensive method of evaluating instrument mass accuracy measurement capability, which resembles intended use, is to calcualte the root mean square (RMS) error. To illustrate its use, the following is adapted from the mass measurement accuracy specification of a commercial TOF mass spectrometer. 

 "The mass measurement accuracy of the instrument, under normal operating conditions, will be better than a given ppm RMS over the given 

 range, based on a number of consecutive repeat measurements of an analyte peak (of given 

), using a suitable reference peak (of given 

). Analyte and reference peaks must have sufficient intensity and be free of interference from other masses.""

There are some important points and assumptions to be considered: 

1. Assumes an instrument calibration has already been performed with peaks of known mass using a calibration standard. The reference peak is used to account for any variation in the instrument calibration over time and mass measurement accuracy is determined using the analyte peak. 

2. Normal operating conditions — also can include details of chromatographic conditions (for LC–MS performance specifications) and any related MS operating conditions (for example, mass resolution, 

 of interest, or spectral acquisition rate). 

3. Sufficient intensity — assumes that the ion count is not detrimental to characterization of the (mass measurement) accuracy and precision of the instrument in question. Too few ions leads to poor ion statistics and too many ions can lead to detector saturation, both of which result in a greater variation in the standard deviation of repeat measurements and will influence calculation of the RMS error adversely (also relevant to instrument calibration). 

4. Free from interferences — assumes that the mass measurement of the peak of known mass is free from interference by ions of the same or similar mass. Overlapping peaks lead to poor mass measurement accuracy, which is also detrimental to characterizing the accuracy or precision of the instrument properly (also relevant to instrument calibration). 

5. The reference or analyte peak selected should be a good representation of the 

 range, which is relevant to the analysis of a particular sample type. 

The RMS error is calculated using the following relation, where Eppm is the parts-per-million error and 

It is worth noting the RMS error allows some measurements to fall outside the ppm error "window of interest" (for example, 5 ppm RMS). To ensure quality measurements, the conditions described previously must be satisfied (particularly regarding intensity and influence of interferences - balanced ion statistics with clear peak definition in the spectra) over a number of repeat injections. Many reported resolution and mass accuracy numbers that you see are not RMS error numbers, but instead originate from a single selected (favorable) ion. 

Figure 2: Increased filtering or restriction of error in the measurement reduces the possible candidates for a given result.  

It is important to remember in all applications that a weak signal (excessively high resolution) can yield poor ion statistics and can therefore be unusable. Too strong a signal can be equally useless, causing detector saturation. Ideally balanced ion statistics with definition in the spectra is the goal. 

Figure 3: Importance of increased resolution when differentiating closely related masses.  

quadrupole resolution is not sufficient to differentiate the two compounds in the upper figure. 

with a resolving power of ~5000, TOF data clearly show two distinct peaks, which can be mass measured accurately to < 5 ppm.

It is important to appreciate the various interrelated roles played in accurate-mass precision by the shift between definitions of mass and increasing resolution and factors such as peak shape and the need for calibration. If these are not understood clearly and taken into consideration, mass misassignments and other undesirable results can occur. 

Figure 3 shows quadrupole and TOF response where both mass values on the TOF data are within 1 mDa of the exact mass. The two fragments of different compositions are from the same analyte and therefore in the source at the same time. Even the best chromatography won't help in this case so this emphasizes one of the reasons higher resolution is useful especially in the analysis of unknowns. This applies equally well to QTOF product ion data versus product ion data from a triple quad. As an added advantage with this higher degree of resolution the extracted ion current plot of each allows differentiation of the oxygen containing and alkyl containing analytes selectively from the chromatograms where the quadruole data would lack this capability. 

Figure 4: Resolution becomes increasingly important as mass increases to properly determine the monoisotopic and average mass relative to the peak top (S. Carr and R. Annan, Unit 16.1, , in Current Protocols in Protein Science [J. Wiley and Sons, New York, 1996])  

Nominal — Unit mass.  Sulfamethazine Nominal = 278 [C

Average mass — Calculated using all isotopes of each element and their natural abundance. 

Calculated exact mass (monoisotopic) — Determined by summing the masses of the individual isotopes for a given ion. 

Accurate mass (Actually "measured exact mass") —What we do with our instruments. It is the measure of an 

 reported to (typically) three or four decimal places. 

As mass increases, differences between the definitions increase, and peak shape plays a bigger role: Ubiquitin  Nominal     Exact       Avg.  [C

Figure 5: Quantitation example using GCâMS with CI - linear dynamic range: 5 orders. 

The mass spectrum is a display of unique ions present at a specific time in the experiment, whether that duration represents a long-term ablation of a solid sample in the source or the passage of a transient gas chromatography (GC) or LC peak. Software is available from several sources. It is often tailored to specific practices, such as metabolite identification. It can be expeditious, reducing huge volumes of data while highlighting issues the unaided eye might overlook. Software can help us reduce uncertainty if, with properly applied skills, we make use of fundamental chemistry: the electron valence rules for nitrogen-containing compounds, characteristic spectra of halides, rings-and-double bond calculations, and so forth to arrive at what we believe is an unambiguous conclusion. No single software application can answer all inquiries satisfactorily. And so, what really counts is a practitioner's ability to apply well-honed skills and educated judgment. 

MS–The Practical Art

Part III concludes the print version of our three-part series, "A Mass Spectrometry Primer," with the glossary to follow in a fourth and final installment.

A Mass Spectrometry Primer, Part III

Al-Assaf, Saphwan, and Phillips, Glyn O.  Characterization and Prediction of   Emulsification Performance of Acacia Gums.   February, p. 196. 

Anderson, Melanie; Simpson, Richard; Archer,   Laura; and Desai, Rajesh. Quantifying an    Anthrax LF Inhibitor in Human Plasma by   HPLC-MS-MS. February, p. 202. 

Aruda, Wayne O.; Walfish, Steven; and Krull, Ira  S. Review and Optimization of Linearity and   Precision in Quantitative HPLC-ELSD with   Chemometrics. October, p. 1032.  

Balogh, Michael P. MS - The Practical Art.   Column. 

Chen, Jianping; Gu, Ming; Wang, Yongdong;    and Kuehl, Don. Improving the Confidence   of Unknown Compound Identification by    First Responder Mobile GC–MS Laboratories  in Time-Critical Environmental and   Homeland Security Incidents.  September, p. 938. 

Crimi, Christina M., and Snow, Nicholas H.    Analysis of Pharmaceutical Residual Solvents   Using Comprehensive Two-Dimensional Gas   Chromatography. January, p. 62. 

Crowther, Jonathan; Dowling, Jeremy; Hartwick,  Richard; and Ciccone, Bill. Performance    Qualification of HPLC Instrumentation in   Regulated Laboratories. May, p. 464. 

Culbertson, Adam W.; Williams, W. Brent;    McKee, Andrew G.; Zhang, Xiang; March,   Keith L.; Naylor, Stephen; and Valentine,    Stephen J. Inside the Personalized Medicine   Toolbox: GC×GC-Mass Spectrometry for    High-Throughput Profiling of the Human   Plasma Metabolome. June, p. 560. 

Desmet, Gert. "Comparison Techniques for    HPLC Column Performance," in Column    Watch. June, p. 514. 

Deventer, K.; Pozo, O.J.; Van Eenoo, P.; and   Delbeke, F.T. Detection of Doping Agents by  LC–MS and LC–MS-MS. April, p. 376. 

Dowling, Jeremy. See Crowther, Jonathan. 

Ettre, Leslie S. Milestones in Chromatography.   Column. 

Fox, Neal M. See Snoble, Karel A.J.  Garrard, Ian J.; Fisher, Derek; and   Sutherland, Ian A. "Dynamic Extraction: A   High-Speed, High-Capacity Purification    Process That Is Rapidly Scalable," in Sample   Prep Perspectives. May, p. 424. 

Grob, Koni. "Kurt Grob - One of the Fathers of   Capillary GC," in The History of   Chromatography. October, p. 1022. 

Hartwick, Richard. See Crowther, Jonathan. 

Hinshaw, John V. GC Connections. Column. 

Jandera, Pavel. Column Selection for   Two-Dimensional LC×LC. January, p. 72. 

Jennings, Walt. "From Academician to   Entrepeneur - A Convoluted Trek," in The   History of Chromatography. July, p. 626. 

Kemperman, Anthony R. See Snoble, Karel A.J.  Kind, Tobias, and Fiehn, Oliver. "Hardware   and Software Challenges for the Near Future:  Structure Elucidation Concepts via   Hyphenated Chromatographic Techniques,"   in MS - The Practical Art. February, p. 176. 

Kissinger, Candice B., and Kissinger, Peter T.    "Linear Thinking in a Nonlinear World," in   Validation Viewpoint. August, p. 708. 

Kissinger, Peter T. See Kissinger, Candice B. 

Regular Article

An alphabetical listing of all 2008 LCGC articles by author and subject.

2008 Index of Authors and Subjects

This installment of Validation Viewpoint describes how statistically rigorous quality-by-design (QbD) principles can be put into practice to accelerate each phase of liquid chromatography (LC) instrument method development. Here, in Part I, the authors examine the current approaches to column screening in terms of design space coverage — a key element of process knowledge. The second part presents novel data treatments to both accelerate and bring quantitation to the column-screening effort. The third installment will focus on integrating quantitative method robustness estimation into formal method development. Moving robustness estimation upstream into the method development effort is consistent with both FDA and ICH guidances. It also enables the identification of instrument methods that simultaneously meet both mean performance and performance robustness requirements.  

Development of analytical methods for LC instrument systems typically is carried out in two phases. The first phase involves column screening, sometimes referred to as column scouting. Column screening is the experimental work done to identify the analytical column (stationary phase) with the best selectivity in terms of all compounds in the sample that must be resolved adequately. Formal method development, the second phase, involves experimenting with additional instrument parameters believed to affect compound separation strongly. The overall goal of the two phases is identification of the instrument parameter settings that provide optimum chromatographic performance.  

This three-part column describes how statistically rigorous QbD principles can be put into practice to accelerate each phase of LC instrument method development. The column is presented in three parts. In this first part, we address column screening. We examine current method development approaches in terms of design space coverage — a key element of process knowledge. In Part II, we will present QbD data treatments to both accelerate and bring quantitation to the column-screening effort. The final installment will focus on integrating quantitative method robustness estimation into formal method development.  

Traditional LC Method Development Practice

Reversed-phase LC is by far the most widely used LC separation method in the pharmaceutical and biotechnology industries. Reversed-phase LC is therefore the basis of the discussions and examples used in this paper. However, the reader will recognize that the instrumentation, software, and QbD-based methodologies presented here are applicable to other LC approaches such as normal-phase LC and hydrophilic interaction liquid chromatography (HILIC). 

The traditional approach to LC method development is to systematically vary one factor across some experimental range while the level settings of all other controllable factors are held constant. This "one-factor-at-a-time" (OFAT) approach, still widely practiced today, is carried out by selecting one instrument parameter to study while holding all other parameters fixed. The "best" performing level of the study parameter is identified normally by visual inspection of the trial chromatograms; the parameter is then fixed at this level, and a new instrument parameter is selected for the next iteration. The OFAT process is repeated parameter by parameter until an adequately performing instrument method is obtained. 

The OFAT approach is carried out routinely in two informal phases, with a specific set of instrument parameters relegated to each phase. The instrument parameters associated with the first phase are those that were historically "easy to adjust," meaning that new levels of the parameters could be set directly in the instrument method without the need to physically change the instrument configuration. The instrument parameters associated with the second informal phase were those for which changing a level normally required changing the instrument configuration; for example, switching a solvent reservoir or switching to a different analytical column. Table I lists the instrument parameters commonly utilized in the two informal OFAT phases of traditional method development.  

Table I:	Historical phasing of method development workflow			

The obvious limitation of the traditional OFAT approach is that the instrument parameters that often affect selectivity most dramatically, and therefore adequate separation, are the hard-to-adjust parameters relegated to the second informal development phase. This builds inefficiency into the approach as well as risk, as critical time can be spent studying secondary affecting parameters in the first phase without achieving much performance improvement. 

However, modern LC instrument systems can support column and solvent selection valves. These new automation capabilities eliminate the traditional categories of easy-to-adjust and hard-to-adjust. The experimenter is now able to select instrument parameters for study based upon relative potential to affect method performance. A direct consequence of the new automation capabilities is that LC method development protocols in many major pharmaceutical companies have promoted analytical column selection to "poll position" — that is, analytical column screening, also referred to as column scouting, now routinely constitutes the first phase of LC method development. 

Figure 1 is a diagram of the LC method development workflow as it is practiced commonly today. As the diagram indicates, the first phase is analytical column selection. After the "best" column is identified, a second development phase is carried out that addresses the remaining important instrument parameters. The goal of this second phase is to identify the parameter settings that meet all critical method performance criteria, which normally include both compound separation and total assay time. After these goals are met, a separate experiment can be performed to demonstrate the robustness of the resulting method. The robustness experiment is done normally as part of the method validation effort.  

Figure 1: Current method development workflow.			

This three-part series of "Validation Viewpoint" columns will apply the concept of design space in a comparison of current LC column-screening approaches. The FDA defines design space as "The multidimensional combination and interaction of input variables (for example, material attributes) and process parameters that have been demonstrated to provide assurance of quality. Working within the design space is not considered as a change. Movement out of the design space is considered to be a change and normally would initiate a regulatory postapproval change process" (1). 

Screening analytical columns at fixed levels of all other instrument parameters — the OFAT approach — does not provide any process knowledge on how the candidate columns being screened would perform at other levels of instrument parameters that also can greatly affect column selectivity. To see this in terms of design space, consider pH and gradient slope — two parameters that can affect column selectivity strongly. Figure 2 illustrates a two-parameter design space delineated by typical study ranges of these parameters. Here the gradient time is held constant, and so the gradient slope is set by the final percentage of organic solvent.  

Figure 2: Design space - pH and final percentage of organic solvent.			

As shown in Figure 2, the two-parameter design space must be duplicated for each column included in a column-screening study (five columns in this example), because it cannot be assumed that the parameter effects observed with one column will hold for other columns. Process knowledge of these parameters within the defined design space would be obtained by conducting an experimental design that statistically samples the design space by experimentally testing specifically defined pH–slope combinations within the space (1). However, as shown in Figure 3, the OFAT approach only evaluates each column at a single level setting combination of pH and slope, that is, screening the column at only one point within the design space. The OFAT approach therefore does not provide any process knowledge for these parameters within this design space.  

Figure 3: Design space coverage - OFAT column screening.			

The most common alternative to the OFAT approach is the first-principles equation approach. This approach employs predefined equations, also called models, that predict key method-performance metrics such as retention time (

) of a compound as a function of one or a specific pair of LC instrument parameters within a design space. For example, a first-principles equation can predict the retention time of a compound as a function of mobile phase percent organic given reversed-phase LC, a specific column, and a specific organic solvent. 

To understand how a first-principles equation relates to a design space, we must understand the structure of the equation in terms of the design space parameters. The simplest equation is the equation of a straight line relating the linear (straight-line) effect of one study variable to a response. Equation 1 is the general form of a linear equation in one variable: here 

 is a coefficient that defines the slope of the line, and β

, is set to zero). The magnitude and sign of the coefficient define the strength (steepness of the line) and direction of the variable's effect (positive or negative) on the response as it is adjusted from low to high through a defined range. The equation is therefore a model of variable behavior, because the equation can predict values of the response 

 for different input values of the study variable 

Equation 2 is an example of a first-principles linear equation, which can be used to predict the retention time of a compound as a function of the percent organic solvent in the mobile phase. In this equation, 

' is the retention time of the compound relative to the retention time of the void volume. 

 is a constant that depends upon the molecular weight of the compound, Φ is the percent of the organic solvent in the mobile phase, and 

' when Φ is set to zero. The equation therefore predicts values of 

First-principles equations for reversed-phase LC have been developed from historical experiments. Their use involves running one or two injections at different levels of instrument parameters linked to the equation. The data from these injections is used to normalize the equation, that is, improve its predictive accuracy in terms of the current sample compounds. The equation is normalized to the data using linear regression analysis, which updates the values of the equation coefficients β

Figure 4 depicts a design space in two parameters: pH and gradient time, along with two run conditions that can be used to normalize a first-principles equation such as equation 2. As shown in Figure 4, an extremely limited amount of the design space is addressed by the two normalizing runs. The first-principles equation methodology therefore operates under the assumptions that the parameter effect is represented accurately by a straight line, and that the effect is the same across the entire design space. In other words, it assumes that the effect of changing the percentage of organic solvent on the retention time is the same at any level of pH within the specified range. Another way of stating the second assumption is that the approach assumes that pH and % organic do not interact — that is, that their effects are strictly additive.  

Figure 4: Design space sampling - first principles equation normalization runs.			

In fact, first-principles equations that include two instrument parameters, here designated 

, cannot model interaction effects between the parameters. This can be seen by examining the general forms of these equations, which are presented as equations 3 and 4. Equation 3 is a linear equation in two variables; this equation can represent only linear additive effects of the variables. Equation 4 is a pure quadratic equation in two variables. The squared terms [β

] in this equation can represent simple curved variable effects (curvilinear effects). Note that neither equation contains even a simple two-way interaction term of the general form β

However, interaction effects are common among LC instrument parameters, and these effects can cause major deviations from expected behavior within a design space. In fact, the importance of interactions is implicit in the FDA definition of design space ("the multidimensional combination and interaction of input variables"). The assumption of simple additivity in the first-principles equation methodology can therefore impart significant error to the method development work. What is needed is a methodology that enables visualization and quantitation of all variable effects, which might be present in the design space. 

The Classical Design of Experiments Approach

Fortunately, QbD principles can be applied to the task of screening analytical columns to include factors such as pH, gradient slope, and organic solvent type, as these factors are recognized major effectors of column selectivity. In a QbD approach, a statistical experiment design plan (design of experiments, or DOE) (1) would be used to systematically vary multiple study factors in combination through a series of experiment trials that, taken together, can comprehensively explore a multifactor design space. A statistically designed experiment can provide a data set from which interaction effects of the instrument parameters can be identified and quantified along with their linear additive effects, curvilinear effects, and even nonlinear effects.  

This installation of "Validation Viewpoint" describes how statistically rigorous quality-by-design (QbD) principles can be put into practice to accelerate each phase of liquid chromatography (LC) instrument method development.

LCGC North America-12-01-2008

Practical Aspects of Solvent Extraction

LC Pumps

A Mass Spectrometry Primer, Part III

2008 Index of Authors and Subjects

A Quality-by-Design Methodology for Rapid LC Method Development, Part I