Evaluation of ethanol in biological samples such as blood, urine, and saliva is requested frequently for legal purposes such as driving under the influence, postmortem alcohol evaluation, and so forth. (1,2). Ethanol intake also has been associated with violent death (3,4). In addition, recently methanol determination has gained a great deal of importance and has become an interesting scientific area with regard to unintentional oral ingestion by food and beverages causing blindness, coma, and death by respiratory and cardiac arrest (5). 

Methanol and ethanol testing has been done by using blood as the biological material for years. However, there is increasing interest in urine and saliva as a biological specimen because both are obtained in a noninvasive way (6). No significant endogenous ethanol production is observed in urine from healthy patients (7). Furthermore, postmortem ethanol production in urine is less when compared with blood, because negligible quantities of glucose are excreted in urine from individuals having normal carbohydrate metabolism. However, the probability of diabetes must always be kept in mind (8). The concentration of ethanol in saliva accurately reflects blood ethanol concentration as the ratio of the blood flow to the tissue mass of the salivary gland is high (9). 

 Numerous chemical and enzymatic methods have been described to determine volatiles in biological materials (10). Furthermore, chromatographic methods certainly stand out with higher precision (11). For instance, liquid chromatography (LC) still has been used as an analytical technique for the assessment of ethanol in matrixes other than human specimens, as well (12). On the other hand, combining ease of application with precision gas chromatography (GC) equipped with headspace is the most reliable method in use (13–15).

Extraction is the most important step for an analytical procedure for separation of the organic compound from the biological material. Regarding the volatile compounds, such as alcohols, headspace-GC analysis allows direct injection of specimens and is the most common procedure used (14). On the other hand, in this technique, depending upon the sample matrix, the resulting data can vary mildly, which is explained by matrix effect (16). 

In this study, a simultaneous and reliable headspace-GC–flame ionization detection (FID) method was validated to establish methanol and ethanol assessment in blood, urine, and saliva as a routine forensic–toxicological analytic application. In addition, the influence of the specimen matrix onto the chromatographic response was also investigated for each of the biological fluids used for routine analysis. Therefore, the main purpose and the novelty of this article fundamentally depends upon validation and comparison of specimens in this respect. 

-butanol were obtained from Merck (Gernsheim, Germany). Headspace vials were purchased (Agilent Technologies, Palo Alto, California). GC was performed on an HP 6890 series II gas chromatograph (Agilent Technologies) with a flame ionization detector. The system was equipped with an HP 7694 E Headspace Sampler (Agilent Technologies). A 30 m × 320 µm, 0.25-mm 

 HP-Innowax (Agilent Technologies) column was used. Headspace conditions were adjusted according to the parameters described in the study conducted by Wasfi and colleagues (15). Vial equilibrium time was set to 10 min. The sample volume was 1 mL and sample loop was 3 mL in volume. Headspace oven temperature, sample loop temperature, and transfer line temperature were fixed to 70 °C, 80 °C, and 90 °C, respectively, to avoid the possibility of condensation until injection. The GC oven temperature was held constant at 110 °C for 6 min. The GC system had an injection temperature of 180 °C and a detector temperature of 250 °C. 

 All biological specimens were collected by the volunteer laboratory crew of Ankara University Institute of Forensic Medicine. For collection of saliva and urine,  clean polypropylene vessels were used by each volunteer; urine and saliva samples were collected, providing a pool. Blood samples from each volunteer were collected in anticoagulant tubes and pooled prior to analysis. Each pooled specimen was mixed by vortex before use. All specimens were kept at +4 °C until the time of analysis. 

Blank specimens (namely water, blood, urine, and saliva) (800 µL) were placed into clean glass headspace vials containing 100 µL of 800 mg/dL 

-butanol as internal standard (IS). In the next step, 100 µL of ethanol and methanol standards was spiked, yielding concentrations of 50, 100, 200, 300, and 400 mg/dL. The total volume was 1 mL, achieving an IS concentration 10 times diluted to 80 mg/dL. Each vial was sealed with a rubber cap and an aluminum crimp seal immediately after addition of the methanol and ethanol standards. Calibration standards were prepared five times for each specimen at each concentration. The precision within-run and between-run tests were performed with five replicates at concentrations of 50, 200, and 400 mg/dL. For evaluations of recovery, 200 mg/dL was chosen as mid-concentration and 10 replicates of this concentration were tested. 

 To overcome the risks of any artifacts during analysis, the IS was used. Thus, peak areas were measured and calculations were carried out considering peak ratios of analyte to IS. 

The method for each specimen was validated fully, as were previous validation studies (17), to follow criteria in agreement with ICH guidelines (18). 

 The method demonstrated excellent chromatographic specificity with no endogenous interference at the retention times of, methanol, ethanol, and the IS (3.78, 3.61, and 5.30 min, respectively). The representative chromatogram of blood spiked with ethanol and methanol is displayed in Figure 1. Retention times for other specimens observed exactly the same revealing lack of matrix affect on retention time and duration of the analysis. 

Figure 1: Representative chromatogram of  spiked human blood containing 200 mg/dL methanol, 200 mg/dL ethanol and 80  mg/dL n-butanol (IS). Peaks: 1 = methanol, 2 = ethanol, 3 = n-butanol.			

After establishing the chromatographic conditions, separate calibration curves were prepared for each specimen at ethanol and methanol concentrations of 50–400 mg/dL. For each concentration, five individual replicates were injected and linearity was obtained with the correlation coefficients (

) equal to 0.998 and 0.999, respectively, for methanol and ethanol in all biological specimens. 

The recovery tests of this analytical procedure were performed by spiking blank water, blood, urine, and saliva specimens with 200 mg/dL of methanol and ethanol. A recovery rate of over 90% was obtained, and the recovery results are shown as percentages and ranges in Table I.  

 Table I: The results of recovery studies for blood, urine, and saliva spiked with 200 mg/dL methanol and ethanol 			

 Precision, defined as relative standard deviation (RSD), was determined by five individual replicates at three different concentrations (

= 5). Accuracy, defined as % bias, was also determined at the same concentrations (

= 5). Intraday and interday accuracy and precision studies were carried out as well. Samples solutions of biological specimens were freshly prepared on the same day (intra) and on five consecutive days (inter). Detailed data of accuracy and precision studies are tabulated in Table II.  

Table II: Intra- and interday accuracy and precision data of water, blood, urine, and saliva spiked with methanol and ethanol 			

 The limit of detection (LOD) and lowest limit of quantification (LOQ) were determined based on the standard deviation of the response and the slope of the calibration curve, according to ICH guidelines (18) (LOD = 3.3 σ/S, LOQ = 10 σ/S, where σ is the standard deviation of the response and 

 is the slope of the calibration curve). Table I shows LOD and LOQ values of methanol and ethanol calculated for water, blood, urine, and saliva. 

Comparison of Specimens and Matrix Effects

: Regarding different specimens, dissimilarities in resulting data of validation parameters (Table I and Table II) were observed. Intraday studies showed that for methanol and ethanol analysis, urine conduces to most precise results at mid-concentration in contrast to blood. For interday studies, high precision was obtained in saliva as well. Similar to precision data, analysis of both methanol and ethanol in urine and saliva had the highest accurate results for intra- and interday studies (Table II). No matrix effect was presumed for water in this study due to constant homogeneity and no other molecule in its composition other than water itself. On the other hand, with rich protein content, blood certainly has the strongest matrix, affecting the volatile substance to be analyzed in the headspace-GC system. In this respect, detection and quantification limits and recovery data were affected as well. Table I shows that blood has the lowest recovery rate and highest detection-quantification limits. 

In headspace-GC analysis, in addition to the phase transition between liquid and gas phase, volatile analytes are affected only by the physical and chemical properties of the sample. Therefore, the effect of matrix content on analytical signal should be considered in particular (19). As performed in the present study, addition of IS is one of the suggested solutions for minimization of matrix effect (16).   			

A headspace-GC–FID method was established for simultaneous measurement of methanol and ethanol and was validated in terms of reproducibility, sensitivity, accuracy, precision, and detection limits. It was revealed that headspace GC for methanol and ethanol determination in all routinely available biological fluids is applicable as a reference method and also in routine diagnostics and monitoring for forensic–toxicological and analytical purposes. For nonadvanced local toxicology laboratories, the method presented here stands out as simple and easily applicable.  

The influence of the matrix composition on the chromatographic response was investigated for each biological fluid. The matrix effect on headspace analysis of methanol and ethanol was determined in blood in comparison to saliva and urine, and had less protein content. 

The presence of proteins indicated stronger matrix effects in headspace-GC analysis. Matrix effect of blood was determined more in comparison to saliva and urine, which have less protein content. 

This study was supported by grants from the Ankara University Institute of Biotechnology Research Fund, project number 2004-118.  

Görkem Mergen*, Zeliha Kayaalti*, Emrah Dural*, Vugar Aliyev*, Seda Kaya*, Serap Yalçin

, Ayşe Karakuş*, and Tülin Söylemezoğlu*

*Ankara University Institute of Forensic Medicine, Ankara, Turkey. 

(2) D. Zuba, A. Parczewski, and M. Reichenbacher, 

(3) C.L. Winek Jr., C.L. Winek, and W.W. Wahba, 

(4) S.J. Heishman, E.G. Singleton, and D.J. Crouch, 

(7) M. Zilly, P. Langmann, U. Lenker, V. Satzinger, D. Schirmer, and H. Klinker, 

(12) T. Yaritaa, R. Nakajimab, S. Otsukac, T. Iharaa, A. Takatsua, and M. Shibukawa, 

(14) B.S. Martinis, M.A.M. Ruzzene, and C.C.S. Martin, 

(15) I.A. Wasfi, A.H. Al-Awadhi, Z.N. Al-Hatali, F.J. Al-Rayami, and N.A. Al Katheeri, 

(17) G. Mergen, T. Söylemezoglu, I.Ilhan, and Y.B. Dogan, 

Articles

An HS-GC-FID method for the determination of ethanol and methanol content in biological fluids is presented.

Simultaneous Headspace-GC–FID Analysis for Methanol and Ethanol in Blood, Saliva, and Urine: Validation of Method and Comparison of Specimens

Sample preparation and analysis of trace compounds in difficult matrices has always been a challenge. In modern ages, the drive to expand food production to feed an ever-increasing population has led to the increased use of a multitude of toxic pesticides in all countries of the world. Analysts have had to keep up with techniques to measure low levels of these pesticides, sometimes used indiscriminately for a wide variety of crops, including those for which they were not intended. Hence, methods for the trace analysis of multiresidues for multiclasses of pesticides have received a great deal of attention, especially in view of recent scares of contaminants showing up in our food supplies. 

Sample preparation traditionally used liquid–solid extraction followed by liquid–liquid extraction and more recently, solid-phase extraction (SPE) to isolate pesticides of interest. In 2003, building on extraction techniques of the past, chemists working at the U.S. Department of Agriculture introduced a technique that was simple yet effectively isolated trace pesticides from a variety of fruits and vegetables (1). Relying heavily on marked improvements in high sensitivity, high selectivity tandem techniques of gas chromatography–mass spectrometry (GC–MS) and liquid chromatography–mass spectrometry (LC–MS) (and more recently on their MS-MS counterparts), the technique of QuEChERS (quick, easy, cheap, effective, rugged, and safe, and pronounced "Catchers") was developed (see Figure 1). An acetonitrile salting-out extraction of a solid sample in an aqueous environment is followed by liquid–solid extraction (called dispersive SPE, abbreviated d-SPE) to remove a majority of the remaining matrix interferences. The sample is then analyzed by one or more of the tandem techniques. 

The two chemists who, along with their coworkers, developed and piloted the technique through the international validation and accreditation organizations (AOAC International and European Committee Standardization [CEN], respectively) were Drs. Steven J. Lehotay of the USDA in Philadelphia, Pennsylvania, and Michelangelo Anastassiades of Chemisches und Veterinaruntersuchungsamt Stuttgart, Germany, who worked together in the USDA laboratory from 2000 to 2002. The technique has grown in such proportions that it is becoming the standard sample preparation technique for pesticides in fruits and vegetables in a majority of laboratories in the world. In addition, QuEChERS is quickly expanding outside of its traditional applications area to address a wide variety of contaminants, antibiotics, and other pharmaceuticals, drugs of abuse, and so forth in a wider variety of matrices such as meat, fish, whole blood, wine, and even soil. In this installment of "Sample Preparation Perspectives," columnist Ron Majors interviewed Anastassiades and Lehotay about the technique, its successes and challenges and where it might be heading in the future. 

The original QuEChERS publication was in 2003. What would you both consider to be the three biggest improvements in these extraction protocols since the original method? 

 During the conception of QuEChERS, we conducted many experiments to isolate each facet of the extraction and cleanup process for a wide range of representative pesticides and commodities. The original method from the 2003 publication (1) is fine and even advantageous over current modifications, depending upon the analytes and matrices. When we independently conducted validation experiments for hundreds of pesticides in different commodities, we both found that there were about 10 pesticides that showed pH dependence in the recoveries, depending on the matrix (2–4). Buffering at pH = 5 was the best compromise to achieve acceptable recoveries for those few pH-dependent pesticides, and we each introduced buffering during extraction in our preferred way (3–7). This was the first "improvement" (however, buffering also leads to more coextractives in some matrices, so we prefer to call it an "adjustment" rather than an "improvement"). 

Secondly, the addition of C18 along with the primary-secondary amine (PSA) in the d-SPE step definitely helps improve cleanup for some samples, particularly those that contain lipids such as olives (8), and it does no harm in any case. Thus, Lehotay regrets that he did not uniformly include C18 in the AOAC Official Method 2007.01 protocol while in the European Standard Method EN 15662, C18 was made an option. Some chemists employ a freeze-out step to reduce lipid coextractives, but C18 in d-SPE is faster and easier, and has been shown to work equally well in removing lipids, although freezing out also precipitates additional matrix components having limited solubility in QuEChERS extracts (4,9,10).  

Thirdly, the reduction of chlorophyll coextractive from green commodities and carotenoids via the addition of a small amount of graphitized carbon black (GCB) as in the CEN Standard Method 15662 (7) or ChloroFiltr product (UCT, Inc., Bristol, Pennsylvania) constitutes a third prominent adjustment, at the sacrifice of ≈25% reduced recovery of some structurally planar pesticides. Another option of note entails the use of a greater amount of PSA in d-SPE to remove more fatty acid coextractives from cereals and grains (at the expense of ≈20% lower recovery of certain polar pesticides) (11). 

Table I: QuEChERS vendors and web addresses			

What have been the three biggest setbacks or difficulties in advancing the QuEChERS methodology?  

We don't think we've really had any setbacks in advancing QuEChERS concepts, and the number of vendors selling QuEChERS and d-SPE products has continued to increase (see Table I). QuEChERS generated much excitement from the day it was introduced at the 4th European Pesticide Residue Workshop (2002 in Rome), and it has continued to gain popularity since (see Figure 2). For example, Anastassiades has compared QuEChERS with the traditional method extensively used throughout Europe for 15–20 years (12) and calculated an approximate 95% reduction in solvent consumption (10 versus 535 mL), an approximate 95% reduction in consumable costs, and an approximate 90% reduction in time. Looking at the reports of the official EU-proficiency tests for pesticides in fruits and vegetables with 130–150 participants each year, the number of laboratories using QuEChERS has been continually increasing to reach approxmately 40% in 2009. Among private laboratories, where financial aspects weigh even more, the adaptation rate to QuEChERS is reportedly even greater and close to 70%. 

The biggest difficulty with QuEChERS arises from its greatest strength: it is extremely flexible and rugged in the sense that dozens of different small modifications will still give high recoveries for hundreds of pesticides in hundreds of commodities. There is an old expression that "analytical methods are like toothbrushes and everybody prefers to use their own." We can see numerous modifications of the original method or the two official versions of QuEChERS in the literature that are based solely upon personal preferences or limited experimentation. Sometimes modifications are necessary to better adjust the method to a specific application, but in many cases the impression is that the introduction of amendments arises merely from need to do something different to better justify projects or publications. In various cases, modifications actually make the method less rugged, longer, or more expensive to perform. 

We went through this divide ourselves when Anastassiades preferred to use citrate buffering (at a relatively low buffering capacity) (7) while Lehotay preferred acetate buffering (at higher concentration to give greater buffering strength) (3,6). We each have our reasons, but ultimately both methods have been extensively validated for many pesticides in many foods by many laboratories.  

Another difficulty inherent in the flexibility of the method is that it accommodates different sample sizes and different injection volumes and solvents. Therefore, the vendors provide many products to accommodate the many preferences and options that laboratories use throughout the world. Too many details and too many choices also lead to misinformation and confusion among novices in the pesticide analysis field.  

 Perhaps the greatest difficulty that all multiclass, multiresidue methods possess is that certain pesticides (for example, captan, captafol, folpet, and chlorothalonil) don't give high and consistent recoveries with the method. Acidic conditions help to reduce losses of these pesticides, but these analytes degrade readily in nonacidic samples even before extraction. Unfortunately, these troublemaker compounds are not amenable to LC–MS-MS and their GC analysis is notoriously problematic. The GC–MS conditions have to be well controlled for their high-quality detection, which is not always possible in routine monitoring operations. Therefore, QuEChERS faces criticism by some for not being perfect. The use of analyte protectants (APs), another interesting feature introduced in the original QuEChERS paper, improves GC analysis of these troublemakers (1,13,14). 

You both worked on the original methods, yet as you have gone your separate ways, your refinements have led to two additional and different extraction protocols. Currently, there are three popular "standardized" methods: original, AOAC Official Method 2007.01 (attributed to Lehotay and coworkers), and CEN Standard Method EN 15662 (attributed to Anastassiades and coworkers, see Figure 3), along with many homemade modifications. Is there a movement to "harmonize" these methods to provide the users (and the manufacturers who supply kits) a more generic choice? 

According to Lehotay, AOAC International has approved moving forward with another Collaborative Study to "harmonize" the QuEChERS method in which the best of both current official methods are combined into a single method for a variety of foods. This entails the use of acetate buffering during extraction and d-SPE (or disposable pipette extraction, DPX, DPX Laboratories, Columbia, South Carolina) cleanup with 50 mg PSA, 50 mg C18, and 7.5 mg GCB (plus 150 mg MgSO4 drying agent) per milliliter of extract. For cereals and grains, a 2.5- or 5-g sample (plus 10 mL water) is extracted and 150 mg of PSA is used rather than 50 mg. The study protocol still needs to be written and reviewed by the AOAC Int. Chemical Contaminants and Residues in Food Community. 

 Although QuEChERS started out as a method for the extraction of multiclass, multiresidue pesticides from fruits and vegetables, it now is being used for a wide variety of analytes (for example, vet drugs, PAHs, and antibiotics) in a wide variety of matrices (for example, plasma, meat, and soil). Do you think that it is expanding too quickly without the careful interlaboratory testing that you performed during your early work?  

We are glad to see the exploration and expansion of QuEChERS concepts in different applications. In terms of quality of the work, it is a case-by-case basis, which boils down to good science or not, independent of QuEChERS. In all endeavors, people have to critically assess the quality and validity of the information they are given. 

From the beginning, we did not try to patent or control the use of QuEChERS concepts, foremost because we did not think the chemical approach or d-SPE could be patented (after all, Tswett used sorbents in tubes for chromatography prior to the idea to use a column), but also because philosophically, we both feel that QuEChERS would be more widely used if we (or our respective employers) did not become involved in business. Neither of us personally nor our organizations profit monetarily from the QuEChERS name or selling of commercial products.  

What we have done to "protect" QuEChERS is work through the auspices of international official method validation and accreditation organizations (AOAC International and CEN). AOAC Official Method 2007.01 and CEN Standard Method EN 15622 are gold standard methods, and modified versions are acceptable if they are properly validated and fit-for-purpose in the particular application.  

However, we caution that many methods achieve high and consistent recoveries for spiked samples during method development and limited validation experiments, yet the method may not work well in real-world samples in which the analyte–matrix interactions are stronger or routine analysis poses instrument maintenance problems. For example, we are skeptical of methods using QuEChERS for soil analysis because stronger extraction conditions are typically needed, such as pressurized liquid extraction or Soxhlet extraction, rather than just shaking to overcome the strong binding characteristics of the soil–sediments. Pesticide extraction from many foods does not entail strong matrix interactions, and shaking has been demonstrated to be acceptable for incurred as well as spiked samples.  

Along those lines, the original QuEChERS was devoted to trace pesticides, mainly in pesticide laboratories that were very familiar with the techniques of statistical and representative sampling. As QuEChERS finds its way into nonpesticide laboratories that are doing trace analyses but without the wisdom that pesticide analysts have in that area, do you think that QuEChERS might get a "bad rap" because of nonrepresentative samples being analyzed? After all, the result is only as good as the sample supplied. 

Again, representative sampling issues are fundamental considerations with any analysis, not just with QuEChERS. In our original paper (1), subsequent papers, and all presentations we've given about QuEChERS, we emphasize that the original sample collection and homogenization step is integral to the QuEChERS concept due to the smaller subsample size and use of shaking for extraction. In this regard, the $5000 chopper used for sample comminution is more important than the $300,000+ worth of LC–MS and GC–MS instruments used for analyses.  

With pesticide laboratories now encountering large numbers of samples, the question of automation of the entire QuEChERS process naturally arises. In its generally practiced form, QuEChERS is still a manual procedure with lots of shaking and sample manipulations. How do you see the automation of the technique taking hold? 

In our views and experiences, QuEChERS is so fast and easy that, rather than making it quicker and simpler, automation adds to the time and effort for sample preparation, and certainly adds to the expense. A chemist can often complete the procedure by hand for a batch of samples faster than a person can load and unload the tubes in automated systems. Fortunately, we have large enough hands to hold 6–10 extraction tubes (50 mL) at a time, and enjoy the 1-min exercise from the vigorous shaking. However, one of the most frequent complaints from bench chemists is that they don't like the shaking step (despite the suggestion for them to put on some music and do the QuEChERS dance!). Also, some people just don't shake as vigorously as others.  

Therefore, a strong shaker can be a beneficial tool in the laboratory for the extraction step, depending on analyst's capabilities, and commercial shakers are an option for QuEChERS. Anastassiades is currently studying the impact of extraction conditions on the extraction yields of incurred pesticide residues from food commodities. In many cases, 1-min extractions are just fine, but there are also cases where there is a notable increase in the extracted amounts when extraction time is extended. This effect may concern for example residues that have diffused into wax layers. Wax particles obviously need some time to soak for those enclosed residues to become accessible to the extractant. Automated shakers thus become very interesting for QuEChERS applications also from this point of view. 

Otherwise, the d-SPE step does not require vigorous shaking and it only takes 30 s, thus it does not need automation. The semiautomated DPX device can handle 20 samples at a time, which does not slow sample throughput versus manual d-SPE. The use of automated robotics to conduct d-SPE may save time in theory, but the logistics, costs, and added instrument downtime seem daunting for little benefit.  

The most laborious and time-consuming sample processing and sample weighing steps seem beyond the ability to automate, and the little details of entering information about samples and keeping track of them are also of critical importance and require careful human inputs.  

Now that QuEChERS is commonly used for sample preparation, the limiting step for sample throughput involves data analysis — reviewing the results for hundreds of pesticides in the analyses. The data processing software is still not trustworthy enough in our opinion to allow it to generate results without human review. Residue chemical analysis does not currently entail the same degree of high-throughput sample analyses that clinical or pharma tests have employed, but if QuEChERS is used in those type of applications, then its automation becomes more important. Gerstel (Germany, UK, and US) is working with DPX Labs and others to automate QuEChERS. 

Extractions of all kinds are often matrix dependent. An analyte in one matrix can behave well but in another matrix be difficult to extract. Have you seen cases like this in QuEChERS? If so, are there some recommended modifications in the QuEChERS protocols to handle these situations? 

Food samples in most cases do not have pesticide–matrix interactions to affect recovery (certain pesticides in dry and oily samples are some exceptions). However, there are some complex matrices (such as teas, herbs, spices, liver, citrus oil, and the like) that give a tremendous number of coextractives no matter what method is used. It is inherent in multiclass, multiresidue methods that the sample preparation process cannot be so selective to remove many matrix chemical compounds if those chemicals have similar properties as the diverse range of analytes being monitored. Sure, modifications can be made by adjusting solvents, pH, salts, volumes, water content, and cleanup sorbents.  

In terms of pH adjustment, for example, Anastassiades found out that the quantitative extraction of nicotine from mushrooms requires pH values in the range of 10–11 (15,16). On the other hand, the extraction of chlorothalonil from certain commodities, such as onions and leeks, requires pH values close to 2 to minimize losses caused due to interaction with the matrix. Similarly low pH values are needed to achieve good recoveries for pesticides belonging to the nereistoxins class.  

Acidic herbicides such the phenoxyalcanoic acids tend to form covalently bound residues. As these are typically included in legal residue definitions, they have to be released before extraction. This is conveniently achieved by introducing an alkaline hydrolysis step (30 min at pH 12) before reneutralization and normal QuEChERS extraction (15,17).  

Individual pesticide classes can be better isolated through more intensive cleanup, but this inevitably reduces recoveries of other analyte classes and typically leads to fractionations into multiple extracts and multiple injections.  

For such complex commodities, we thus very much rely on high quality chromatography–MS analysis using highly selective and sensitive modern instruments. High sensitivity instruments allow to reduce the "sample concentration" in the final extract (for example, to 0.1 g/mL in the case of spices) and thus also the sample equivalents injected, whereas the high chromatographic and MS selectivity allows us to largely "fade-out" matrix components coeluting with pesticides (for example, terpenes in GC applications). If ruggedness or detection limits are compromised too much, then traditional methods involving gel permeation chromatography (GPC) or fractionation of extracts might be a better option. Such decisions depend on the required data quality objectives that need to be met, and the practical matters of time, cost, labor, and instrumentation in the laboratory.  

We have never claimed that QuEChERS works well for everything, and we only say that it is fast and easy to try. If it is found to work well in the application, then it saves a tremendous amount of method development time and labor. 

QuEChERS extracts can be relatively "dirty." Having selective, sensitive detection such as MS and MS-MS helps provide the analytical selectivity on the back end of the analysis. For labs that can't yet afford these rather expensive detectors and are stuck with flame ionization detection (FID) in GC and UV–diode-array detection (DAD) in LC, do you recommend that they even try to use QuEChERS?  

Even historically, FID, and UV–DAD have not been selective enough for use in pesticide residue analysis. Instead, selective detection methods such as flame photometric detection (FPD), pulsed flame photometric detection (PFPD), nitrogen–phosphorus detection (NPD), electrolytic conductivity detection (ELCD), electron-capture detection (ECD), halogen-specific detection (XSD), and so forth have been used in GC, and postcolumn derivatization fluorescence in LC. Before LC–MS-MS, a wide range of polar pesticides was largely ignored in monitoring programs because the single-analyte methods were too cumbersome in practice. The previous multiclass, multiresidue methods were designed for GC analysis, and QuEChERS combined with LC–MS-MS helped advance pesticide residue monitoring capabilities.  

With LC–MS-MS or GC–MS-MS (or selective ion monitoring [SIM]) analysis, we rarely even notice the coextractives or observe interferences in typical commodities (except with spices and citrus perhaps). Similarly, QuEChERS with selective detectors in GC and LC does not provide much difference in terms of interferences than with prior methods. As described earlier, QuEChERS can be adapted by adjusting extraction and cleanup parameters for a narrower range of analytes.  

GC–MS and especially LC–MS-MS are nowadays considered as indispensable for the effective control of pesticide residues. A large number of pesticides giving rise to toxicology concerns and trade restrictions are essentially only amenable to LC–MS type applications. We think that bodies offering technical and financial assistance to developing countries for the establishment of export-control laboratories are well-advised to help these labs to directly embark into LC–MS–MS rather than to scroll through the entire history of instrumental analysis by starting with basic techniques such as LC–DAD, GC–NPD and GC–ECD. The capital investment might be larger but in the long run LC–MS-MS is clearly worth the money by covering a broader scope of pesticides including most critical ones, thus improving the recognition of export certificates and reducing problems with consignment-rejections.  

Now for some specific questions about the procedures: 

During the initial acetonitrile extraction, a good bit of heat is generated by exothermic hydration during the addition of anhydrous MgSO

 to the samples. In one sense, this heat generation aids the extraction speed or efficiency while, on the other hand, too much heat may lead to a loss of volatile or thermally labile pesticides. In practice, how to you reconcile these conflicting occurrences? Or, isn't it an issue? 

The recoveries have been shown to be fine for hundreds of pesticides and other analytes in the method. The heat causes potential degradation losses of the same few pesticides mentioned earlier, but acidification helps to stabilize these pesticides. When using frozen samples (as advised), the temperatures reached after extraction are very moderate.  

 I've heard that at last count there are over 400 pesticides in various matrices that have been extracted using QuEChERS. Besides the planar ring pesticides that are strongly retained by graphitized carbon black during the d-SPE step, are there any other classes of pesticides that cannot be extracted (or at least have very poor recoveries) using these official QuEChERS methods? If so, what alternatives do the users have for extracting these troublesome ones? 

Sample Prep Perspectives

The inventors of the QuEChERS technique present an update and a look at the future.

QuEChERS, a Sample Preparation Technique that Is "Catching On": An Up-to-Date Interview with the Inventors

Occasionally the chromatogram generated by a liquid chromatography (LC) method looks strange. One of those oddly appearing cases is when the peak doesn't seem to belong with the neighboring peaks in the chromatogram. This month's "LC Troubleshooting" looks at one such case. An example is shown in Figure 1, where the peaks at 1 min and 2.75 min look normal, but there is a broad peak that shows up at approximately 1.5 min. I do not know the chromatographic conditions for this run, but for the present discussion, let's assume that it is a reversed-phase method run on a C18 column. As a general rule, all the peaks in a particular part of the chromatogram, whether it is an isocratic or gradient run, should be about the same width. So when a broad peak appears, it is a sign that something is wrong. 

One of the most likely causes of broad peaks in the presence of normal, narrower ones is late elution — a peak that belongs to a previous run. This is simple to determine by extending the run time until the peak emerges. I've shown this in the pair of simulated chromatograms shown in Figure 2. In Figure 2a, a broad peak appears at 2.2 min, obviously out of place when compared to its neighbors. In Figure 2b, the run time was extended and another peak appears at 7.2 min. This peak has the expected width for its position in the chromatogram. This tells us that the 2.2-min peak comes from the prior injection. 

Let's go back to the broad peak of Figure 1 to consider another way to figure out where the extra peak really belongs. To do this, we need to do a few simple calculations. First, we need to determine the column plate number 

 for the peaks that behave as expected. Recall that the plate number (also called the column efficiency) is calculated as 

 is the width of the peak at the baseline, measured between tangents drawn to the sides of the peak. We can use equation 1 to calculate the plate number for a normal peak in the chromatogram. In this case I chose the third peak of Figure 1. I expanded the chromatogram and did my best to measure the retention time and peak width, based on the supplied time scale: 

 = 0.1 min. Using equation 1 with these values gives 

 = 12,100 for peak 3. I have assumed that all peaks in the chromatogram have the same plate number, which is reasonable for this estimation technique. 

Now that I know the plate number for the a normal peak, I can determine the true retention time for the broad peak. First, I have to rearrange equation 1 to 

From my expanded chromatogram, I have estimated 

 = 1.05 min for the broad peak. Now that I know both 

 = 28.9 min for the broad peak, not at 1.5 min as it appears. This means that it might be frustrating to extend the run time until the peak is eluted, because it doesn't come from the prior chromatogram but from run six injections earlier. It should be noted that using this procedure to estimate the retention of a peak is rather sensitive to small measurement errors — an estimated peak width of 1.0 min gives 

 = 27.5 min. So do your best to estimate the retention times and widths from the chromatogram, but don't expect an exact result. 

So now that we know the retention time of the broad peak, let's consider how we might modify the method to either allow us to quantify the peak if it is a peak of interest, or get rid of it if we are not interested in it. 

If we are interested in quantifying the peak, we should first ask ourselves if an isocratic method is reasonable for a peak that is eluted so long after the prior peak of interest at 

 ≈ 2.75 min. For this, it is useful to calculate the retention factor 

 is the column dead-time. If we assume the peak at 1.0 min in Figure 1 to be the unretained peak at the beginning of the run, we can use this for 

 = 27.9 for the 28.9 min peak. As a general guide, we can use an isocratic separation if the ratio of 

-values for the first and last peak is less than 20. In this case, the ratio is approximately 16, so strictly speaking, an isocratic method is feasible. However, with such a wide gap between the two peaks, a gradient might be a much better choice. If we decide to convert the method to a gradient method, I would start the gradient at the same solvent composition as the current isocratic run, or perhaps with 5% less organic, and run a 5%/min gradient to 100% organic. This will give us an idea of where the last peak is eluted under gradient conditions and we can modify the gradient as required for a more favorable chromatogram.  

An alternative to a linear gradient would be a step-gradient. In this case, I would run the current isocratic conditions until the 2.75-min peak is eluted, then at 3 min, step the mobile phase concentration to a higher value. We can use the rule-of-three to estimate the required mobile phase change. The rule-of-three states that a 10% change in organic will change 

 by approximately threefold, So a 10% increase in organic should reduce 

 = 28/3 = 7. I would try a 10% step first, then refine it a bit, as appropriate, to reduce the last peak to a reasonable retention time. Just make sure to allow enough time after the step change in mobile phase so the baseline stabilizes before elution of the last peak. 

If we decide to stay with an isocratic method, I would start with a 30- or 35-min run under the current isocratic conditions to determine the true retention time of the last peak, then adjust the run time as necessary. This would make a very long run time for a two-peak chromatogram. An alternative approach would be to adjust the run time so that the last peak is eluted during a later injection, but in a region where it does not overlap the 2.75-min peak. For example, increasing the run time from its present 5-min injection cycle to a 10-min cycle, the 28.9-min peak from the first injection would be eluted at 8.9 min in the third injection, but well separated from the 2.75-min peak. The actual run time should be adjusted for best results using this technique. 

If, on the other hand, we are not interested in the broad peak, there are several options to eliminate it. One would be to adjust the isocratic run time, as earlier, to move the peak to an unimportant region of the chromatogram. A second approach would be to use a step-gradient or a steep gradient to flush the peak from the column as soon as the 2.75-min peak was eluted. 

If we are willing to modify the sample preparation process, there are more options to eliminate the broad peak. Reversed-phase chromatography separates samples primarily based upon their polarity, with nonpolar, or hydrophobic, analytes eluted later than their more polar, hydrophilic, counterparts. We know from its long retention time that the broad peak is much less polar than the 2.75-min peak. We can take advantage of this fact to design a sample cleanup method to remove the unwanted peak before injection.  

One option to remove the strongly retained peak would be to use solid-phase extraction (SPE). A simple C18 SPE cartridge should do the job, because of the large difference in polarity between the desired and undesired compounds. Following the initial activation of the cartridge, usually with a milliliter or two of methanol followed by the same volume of water, the sample would be loaded onto the cartridge in a highly aqueous solution. This should trap both the 2.75-min peak and the 28.9-min peak on the cartridge. Next, a series of elution solvents would be tried, for example, 10% methanol, 20% methanol, and so forth. Each fraction would be collected and injected to determine how much methanol was required to elute the 2.75-min peak and leave the other peak on the cartridge. The strongest solvent that elutes the 2.75-min peak and not the late peak is likely to be the best choice for fast sample preparation.  

An alternative to SPE is to use liquid–liquid extraction. In this case, a nonpolar solvent that does not mix with water would be the first choice, for example, methyl 

-butyl ether (MTBE). As a first experiment, I would place 0.5 mL of aqueous sample in a 1.5-mL microcentrifuge tube, add 0.5 mL MTBE, cap, and vortex the sample. If necessary, the tube could be centrifuged briefly to separate the phases. The lower, aqueous phase would then be injected and if we are lucky, the 2.75-min peak will remain in the aqueous phase and the 28.9-min peak will have moved to the MTBE phase. The extraction conditions may need to be adjusted to obtain the desired results. 

What are the validation implications for the changes discussed earlier? Regulatory guidelines, such as the United States Pharmacopoeia (USP), state that it is permissible to adjust a method to meet system suitability, but if the method is modified, some revalidation is required. However, there is no consensus on where adjustment stops and modification starts. We can speculate how each change might fit in the adjustment-modification continuum. In any case, I would consult with the quality assurance group for advice on the official interpretation of the regulatory guidelines by my company, and how these would affect any requirements for revalidation. 

Any of the changes to the method so that the broad peak could be quantified suggest that the broad peak was not an expected part of the original method. In this case, some additional validation would be required to show that the peak could be properly measured. This validation is concerned more with quantification of the peak than modification of the chromatographic conditions.  

Changes to the method that move the broad peak to an unwanted part of the chromatogram or strip the peak from the column after the 2.75-min peak is measured would be more likely to fall in the adjustment category. My conclusion is based upon the fact that whatever is done to the conditions after the desired peak is eluted should have no impact on its quantification. In this case, it should be sufficient to make a few experiments to show that system suitability is met with the new conditions and that the data gathered for the 2.75-min peak are not significantly different from those obtained under the original conditions. Proper documentation of these tests should be made, of course. 

I would consider any alteration of sample pretreatment to be a modification of the method, and it would require some revalidation. 

When a broad peak appears among narrow peaks in a chromatogram, one of the most likely sources of the broad peak is that it originated from a prior injection. We have looked at several options to move the broad peak, either to allow it to be quantified or to remove it from the chromatogram. Some of these changes are simple and would be unlikely to impact the validity of the method, whereas other changes would require some revalidation of the method. 

In many cases, a broad peak comes from the previous injection, and an extended run time will allow it to be eluted in the proper position, as in Figure 2b. In other cases, the true retention is much longer, and other actions are more appropriate. And some samples contain materials that are so strongly retained and are at such low concentration that they do not show up as peaks at all, but rather result in a wavy, undulating baseline as more and more samples are run. When this kind of sample is encountered, flushing the column with a strong solvent after each sample batch might be sufficient to strip these materials from the column and restore a flat baseline for the next batch of samples.  

"LC Troubleshooting" Editor John Dolan has been writing "LC Troubleshooting" for LCGC for over 25 years. One of the industry's most respected professionals, John is currently the Vice President of and a principal instructor for LC Resources, Walnut Creek, California. He is also a member of LCGC'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 troubleshooting with John Dolan and other chromatographers, visit the Chromatography Forum discussion group at 

LC Troubleshooting

John Dolan looks at a case in which the peak doesn't seem to belong with the neighboring peaks in the chromatogram.

Where Did That Peak Come From?

The goal of quantitative chromatography is to separate two or more substances sufficiently so that they can be measured with a desired degree of accuracy and precision. The resolution of a peak pair provides a standardized measurement of the extent of a separation but strictly speaking, resolution is defined only for two Gaussian peaks of equal size. This month's "GC Connections" discusses resolution and the effects that relative peak size have on area measurements. 

The resolution of a pair of adjacent peaks is defined in the chromatographic literature and in standard nomenclatures such as IUPAC (3) or ASTM E355 (4) as follows: 

For a Gaussian shaped peak the width at half-height is related to the width at base by a factor of 1.699: 

It is usually more convenient to calculate resolution from the width at half-height of the second peak using this related formula: 

The width at half-height is more easily measured than the width at base, and it is measured and used by many chromatography data-handling systems in, for example, system suitability calculations. The width of the second peak is used because it generally will be either equal to or slightly greater than the first peak's width and thus, produces a the more conservative result while not requiring measurement of the widths of both peaks.  

Most chromatographers are familiar with the appearance of partially resolved, baseline resolved, and fully resolved peak pairs. Figure 1 shows an ideal pair of generated Gaussian peaks of equal sizes and widths with partial resolution at 

 = 1.0 (Figure 1a), baseline resolution at 

 = 1.5 (Figure 1b), and full resolution at 

 = 2.0 (Figure 1c). In this illustration, resolution increases as the separation between the peaks increases, while the same peak shape is maintained. Keeping the same separation while decreasing the peak widths also would increase their resolution proportionately. Figure 1c illustrates the relationship between the peak width at half-height and base, and it also shows how the width at base is determined from the baseline intersection points of lines drawn tangent to the peak's upward and downward slopes. A common misconception is that the width of a peak at base is the time between when the peak starts to rise above the baseline until it subsides back into the baseline — the peak integration start and stop times — but in fact, the base width of a Gaussian peak is measured at 13.4% of the peak height above the baseline, at the times corresponding to the tangent–baseline intersection points. 

Table I lists the retention times, peak width measurements, and resolution for the three cases shown in Figure 1. 

Table I: Resolution measurements from Figure 1. (a), (b), and (c) refer to the parts of Figure 1.			

At resolution of 1.0 the valley point between the peaks corresponds with the end of the first peak's base width and the start of the second peak's base width. At baseline resolution, where 

 = 1.5, the ending and starting base widths of the first and second peaks, respectively, are separated by 0.5 times the base width. At resolution 2.0, the ending and starting base widths of the first and second peak are each spaced at intervals equal to the base width. 

The identification of specific peaks by their retention times becomes more certain as resolution increases. It is possible to identify peaks on the basis of a partial separation, but the certainty of identification improves with increasing resolution. The detector signal comes closer to baseline between the peaks and the presence of additional components becomes more apparent. A mass-selective detector helps to identify merged peaks on the basis of their mass spectra, but closely eluted peaks often are isomeric or otherwise chemically similar. They may have similar spectra that preclude unambiguous identification by spectral library matching.  

Each component in a group of two or more merged peaks should be identified with separate analyses of the pure components, and the influence of matrix components should be determined through blank injections. Bear in mind, too, that the elution order of close peaks will be influenced by changes in the column temperature conditions: a change in temperature program rate of as little as 2 °C/min can reverse elution order. When faced with the partial resolution of a critical peak pair, or with the presence of a new matrix compound that interferes, make sure that the current separation conditions are optimal and then consider upgrading the column resolving power by installing a longer or narrower-bore column. If all else fails, then try changing to a different stationary phase that better separates the peaks of interest. 

Measurement of a partially resolved peak pair can be difficult to achieve with acceptable levels of accuracy. The simple but unrealistic case of two equal-sized peaks is a good starting point for understanding the effects that colliding peaks have on quantitative analysis. Both the influence of the portion of one peak that extends into another and the effects that merged peak shapes have on peak integration play strongly into the accuracy of quantitative measurement of merged peaks. 

Returning to Figure 1, the degree of area overlap of the coinciding portions of the peaks decreases as resolution increases. At 

 = 1.0 in Figure 1, the contribution of either peak to the other's total area is about 2.3%, as measured from where a vertical dividing line for peak area integration is dropped from the valley point between the two peaks to the baseline. I measured the overlap areas with a measurement tool in my data system for the first generated peak alone by starting at the time of the valley point, which corresponds with the end of the peak's 

 segment, and extending to the end of the peak. In this case, the peaks are symmetrical and identical so the overlap and total areas are the same for each. 

The overlapping areas are highlighted in color in Figure 1a and in the inset in Figure 1b. As the resolution increases to 1.5 — baseline resolution — the overlap between the two peaks drops toward 0.1%, and at a resolution of 2.0 no measurable overlap remains. Table II gives the area measurements from Figure 1. The values for each peak are very close to the known areas as set when the peaks were generated. 

Table II: Peak area measurements from Figure 1. Overlap is the amount of area or percentage of one peak into the other. (a), (b), and (c) refer to the parts of Figure 1.			

In this idealized situation, there is no significant error between the measured and known peak areas, even down at a resolution of 1.0. However, real peaks are not equal in area and width because their amounts, detector response factors, and shapes vary. Also, real-world peaks incorporate uncertainty in their measurement due to injection variability, detector noise, and slight shifts in retention time from run to run. It is instructive to consider what happens when one peak is considerably larger than the other and what the effects imply about the influence of peak resolution on quantitative analysis. 

Increasing the size of the first peak in our example affects both the peak separation and area measurements. The disparity in peak sizes changes the degree of peak overlap, but more significantly, it has a profound effect on quantitative measurements. Figure 2 shows the same three cases as in Figure 1, but the first peak is now 10 times larger. A question arises of whether the classical resolution measurement, which is based upon two equal-sized peaks, is applicable when the peak sizes are significantly different. 

We can use the degree of area overlap between the peak pair as an indicator of the degree of peak resolution, and compare the case of equal-sized peaks to that of unequal peaks. Looking first at  Figure 2a (the case of partial resolution of 1.0 of peaks with a size ratio of 10:1), the first peak's percentage of overlap into the second, as listed in Table III, is similar to that of a pair of equal-sized peaks when there is a simple vertical division between the two peaks at the valley point. In the case of baseline resolution, in Figure 2b, Table III lists the overlap of the larger first peak into the smaller second peak at 0.31% of the second peak's area, while the second peak's overlap into the first is only 0.06%.

Table III: Peak areas and overlap of unequal peaks. Overlap is the amount of area or percentage of one peak into the other. (a), (b), and (c) refer to the parts of Figure 2.			

The overlap for the unequal peaks is small and similar to the degree of overlap of 0.14% for a pair of equal-sized peaks, however, so the definition of the resolution of equal-sized peaks seems to extend to at least at a peak area ratio of 10:1. In the case of resolution equal to 2.0, in Figure 2c, there is no measurable influence of either peak on the other: they are indeed fully resolved and independent. 

Quantitative area measurements in the case of baseline or better resolution of unequal peaks are fairly accurate. The measured and known areas in Table III for cases of baseline resolution (Figure 2b) and full resolution (Figure 2c) compare very well. Going from a 1:1 area ratio to a 10:1 area ratio with a pair of peaks at baseline resolution or better should yield a nearly linear trend between peak size and measured area. At some point, as the area ratio increases above 10:1, however, the degree of influence of the larger peak on the smaller one's measured area will become significant and the linear relationship between peak size and measured area will break down. Greater resolution between the peaks will raise this ceiling on linearity.  

 = 1.0, the story is quite different. Now, there is a much larger overall error in measured area counts with a peak size ratio of 10:1, which was not hinted at when only the similar degree of area overlap for equal and nonequal peak sizes was considered. Not surprisingly, the largest errors occur for the smaller peak. 

The error level depends strongly upon the type of baseline allocation selected for the data system. Figure 3 illustrates three common choices. First is a simple vertical separator dropped from the valley point to the baseline, which we've used throughout this discussion thus far, shown in Figure 3a. This choice produces an error of about +1 % for the larger peak, while the smaller peak comes back with an area that's about 10% smaller than the known amount, as listed in the first line of Table IV. This effect is seen graphically in the inset of Figure 3a, where the relative amount of the second peak that's overlapped into the first is a significant fraction of the second peak's size. 

Table IV: Influence of integration type on areas of partially resolved unequal peaks Rs = 1.0. (a), (b), and (c) refer to the parts of Figure 3.			

A tangent skim baseline choice puts a straight line at the valley point and tangential to its intersection with the declining side of the smaller peak. When a tangent skim is selected, the error in the second peak's area becomes unacceptably huge at –63%. This effect is visualized easily by examination of the inset in Figure 3b, where a very large portion of the second peak is lost into the area of the first. This shift of area also biases the first peak's area upward by 4.5%.  

Finally, using an exponential skim or curve fit off of the first peak reduces the error relative to the tangent skim, but for the second peak the loss of area is still very large at –31%. Looking at the inset in Figure 3c, the curve-fit of the exponential skim to the first peak is fairly good and overestimates the peak size by only 1%. The loss of a significant portion of the second peak is obvious.  

The errors associated with these three different baseline allocation choices depend on the relative sizes of the peaks and their resolution. What I've shown here is just a snapshot of one artificial situation. Chromatography data-handling systems include conditional code that allows users to select boundaries for cases in which different baseline allocations are to be applied, based upon relative peak heights, valley points, and other parameters. This means that the choices for baseline allocation must be made carefully and validated across the entire targeted range of analyte concentrations.  

The resolution of a pair of Gaussian chromatographic peaks provides a useful metric for gauging the ability to achieve reliable qualitative peak identification as well as predicting the accuracy of area measurement. As long as peak resolution is at baseline or better levels, quantitative errors can be very low, even for adjacent peaks at size ratios up to 10 to 1. When the resolution falls below the baseline level, the measurement of peak areas becomes difficult. Careful choices for area integration in data-handling systems can minimize the errors, but must be checked across the entire operating range of a method. Selecting conditions and columns that better resolve critical peak pairs can be a better choice than relying upon the data system's rules of area integration to slice away partially resolved peaks from each other. 

  "GC Connections" editor John V. Hinshaw is senior Research Scientist at Serveron Corp., Hillsboro, Oregon, and a member of LCGC's editorial advisory board. Direct correspondence about this column to LCGC, Advanstar Communications, 485 Rt. 1 S, Bldg F, 1st Floor, Iselin, NJ 08830, or contact the author via 

 (International Union of Pure and Applied Chemistry, Research Triangle Park, North Carolina 27709-3757, USA, 1997). 

(4) "ASTM E355-96 (2007) Standard Practice for Gas Chromatography Terms and Relationships," Annual Book of Standards, Vol. 03.06 (ASTM International, American Society for Testing and Materials, Conshohocken, Pennsylvania). 

GC Connections

This month's "GC Connections" addresses the effects of two overlapping tailing peaks on peak resolution and quantitation.

When Peaks Collide

Pressurized liquid extraction (PLE) was introduced in the mid-1990s as an alternative to tedious classic extraction methods such as Soxhlet extraction or sonication for solid samples (1). It involves extraction with solvents at high temperature and pressure, which is highly beneficial for the extraction efficiency. The elevated temperature and pressure leads to significantly higher capacity of the extraction solvent to dissolve the target analytes and improves the rate of mass transport and effectiveness of sample wetting and matrix penetration, which improves the desorption of analytes from active sites on, and within, the sample particles. Instrumentation for PLE consists of a stainless-steel extraction cell, which is filled with sample and kept at the desired pressure and temperature with the aid of a pump for solvent delivery, a flow restrictor or on–off valve, electronically controlled heaters; and one or more vials for extract collection. PLE may be conducted in two ways: 

Dynamic PLE, where the solvent is continuously pumped through the extraction cell.

Static PLE, where the extraction cell is filled with solvent, pressurized for a specified time, and then drained to the collection vial. It is also possible to combine the two or to perform multiple extraction cycles. Dynamic PLE is suitable for rapidly equilibrating (easily extracted) samples, and, conversely, static PLE is suitable for slowly equilibrating samples. As PLE is primarily used for samples that are difficult to extract, the static PLE approach is more widely used. The majority of PLE applications (about 75%) have been performed in the static extraction mode. 

Such systems can reach temperatures up to 200 °C and pressures up to 200 bar and can accommodate cells of various volumes (that is, 1, 5, 11, 22, or 33 mL [ASE 200] and 34, 66 or 100 mL [ASE 300]). Both systems are produced by Dionex (Sunnyvale, California).

Since its introduction a decade ago, PLE has grown rapidly in popularity, as indicated by an almost exponential increase in the number of scientific publications and conference presentations (Figure 1, left). It has also been adopted by many commercial laboratories and more or less completely replaced Soxhlet and other traditional extraction techniques. 

Figure 1: Overview of the indexed PLE literature (SciFinder, December 2009).			

Over time, there has been a shift in major application areas. Initially, PLE was almost exclusively used for solid environmental samples, primarily soils and sediments, later it was also adopted by researchers working with biotic samples and food control laboratories and, recently, it has gained increasing popularity amongst natural product chemists and laboratories working with product control (Figure 1, right). 

Generally, the PLE extracts are characterized, or analyzed for target (mainly organic) analytes, by various chromatographic techniques, primarily gas chromatography (GC) (75–80%) and liquid chromatography (LC) (10–15%), but also ion chromatography (IC) (approximately 8%) and capillary electrophoresis (CE) (approximately 2%); mass spectrometry (MS) is the most frequent detection technique. The main areas have been summarized in some recent reviews: abiotic environmental (2), food, feed, and biota (3), and natural products (4). 

PLE is very efficient, as mentioned earlier, and generally gives recoveries equal to or better than those obtained using Soxhlet (5). However, the exhaustive nature of PLE is both its largest assets and one of its main drawbacks. Although PLE allows the sample to be (particle) filtered (Figure 2, left), through its flow-through design and sometimes produces extracts that can be analysed directly using highly selective techniques such as liquid chromatography-tandem mass spectrometry (LC–MS-MS), its low selectivity more often leads to extracts rich in coextracted material, which has to be removed before instrumental analysis (6). This process is tedious and costly, severely hampers the sample throughput, negatively influences the quality of analyses, and frequently involves the use of large volumes of harmful organic solvents. Alternative procedures are, therefore, needed. 

Figure 2: Extraction schemes for (from left to right) normal PLE, pre-extraction, sequential PLE, PLE with matrix retainer and matrix solid-phase dispersion (MSPD) PLE.			

				Some authors have used a pre-extraction (pre-PLE, Figure 2) with a weak solvent to remove some interfering (less strongly sorbed) compounds before the extraction of the compounds of interest (7,8). The extract complexity also can be reduced by using a sequential extraction procedure with solvents of increasing solvent strength (9). Such procedures also can be used to assess the strength of matrix–analyte interactions (10).

The selectivity of the extraction or leaching process may be further enhanced through the addition of a matrix retainer to the extraction cell. Use of alumina as a fat retainer was already suggested in 1996 in a Dionex application note (11). Since then, many other adsorbents have been used for the same purpose, such as  Florisil, silica gel, and diol- and cyanopropyl-silica, for example (4). For chemically inert compounds such as PCBs and dioxins our group has recently shown than even better results may be obtained using sulfuric acid–impregnated silica (12). Matrix retainers have also been used for abiotic samples. Peterson and Richter (13) used an ion-exchange resin to retain unwanted ionic interferences during extraction of perchlorate from vegetation samples, Ong and colleagues (14) used silica gel to retain humic matter and other polar co-extractants during PLE of polycyclic aromatic hydrocarbons (PAHs) from soil, and Kim and colleagues (15) used a silica–copper mixture to simultaneously remove polar compounds and sulfur when extracting PAHs from sediment. In the closely related matrix solid-phase dispersion (MSPD) technique the matrix retainer is mixed with the sample before the PLE extraction (16). Both the matrix retainer and MSPD technique can be adjusted to retain particular matrix constituents by choosing an appropriate retainer or dispersion material and by carefully tuning the solvent composition and PLE parameters.  

There are also a few recent applications in which the concept of PLE with in-line sample purification has been taken one step further, by using the adsorbent bed not only for matrix removal but also for analyte fractionation. Lundstedt and colleagues (17) used 2% deactivated silica gel–filled extraction cells (Figure 3, left) to separate parent PAHs from carbonyl-containing PAH oxidation products (oxy-PAHs). The former were extracted with cyclohexane–dichloromethane (9:1) at 120 °C and the latter with cyclohexane–dichloromethane (1:3) at 150 °C. By using a similar approach Poersschmann and colleagues (18) were able to fractionate lipids into two fractions containing neutral lipids (acylglycines, sterol esters, and long-chain free fatty acids) and polar phospholipids. In both cases, analyte classes of different polarity were separated through adsorption chromatography processes. 

Figure 3: Two types of PLE with in-cell fractionation: polarity fractionation using polar adsorbents (left) and shape-selective fractionation using activated carbon (right).			

Another mode of separation, shape selective separation on activated carbon, has proven to be very useful. By using carbon-filled extraction cells, halogenated aromatic compounds such as PCBs and dioxins can be fractionated according to planarity (Figure 3, right). Because the PCBs are less planar than the dioxins, they can be extracted with apolar or moderately polar solvents or solvent mixtures, pass the carbon trap, and be recovered in the first fraction together with the lipids. The planar dioxins, on the other hand, bind tightly to carbon and can only be released by backward elution with an aromatic solvent such as toluene. Depending on the amount of carbon, the PLE settings, and the solvent composition, the dioxin-like non-ortho PCB can be directed to the first or second fraction. Following miniaturized multilayer silica cleanup the toluene fraction is ready for non-ortho PCBs or dioxins analysis by bioassays (19), immunoassays (20), or GC–MS (21). 

The applications described above serves as illustrative examples of what can be accomplished using in-cell purification techniques. However, although the streamlined approaches significantly reduces the analysis time, manual labor and solvent volumes, compared with the traditional Soxhlet extraction and multistep open-column cleanup, there are also limitations of the in-cell clean-up and fraction techniques. These primarily stem from the intrinsic ambiguity in exhaustively extracting analytes from complex samples, efficiently trapping analytes or matrix on an adsorbent, and, in the case of analyte trapping, quantitatively releasing analytes without coextracting large quantities of matrix material. This equation cannot be solved using current commercial hardware for PLE, which houses sample and matrix retainer or adsorbent in the same compartment.  

A flexible modular approach for PLE with in-cell clean-up recently has been developed in which extraction cell segments are added or removed as needed to overcome the previously discussed problems and ensure optimum extraction efficiency and selectivity. 

A modular system using in-house manufactured adapters to couple existing ASE 200 extraction cells (Dionex) recently was developed (Figure 4) (22). The adapters consist of two cylindrical stainless steel guides, with threading identical to the original PLE end-caps, and a PEEK disk with a stainless steel frit pressed into its center. The guides primarily align the extraction cell segments and keep the PEEK disk in place, but they are not tightened hard. Instead, it is the compression unit of the ASE system that presses the extraction cell segments into the PEEK disk and creates the seal. These adapters allow two or three cells to be serially connected. All combinations of 1-, 5-, 11-, and 22-mL cells, except two or three 22-mL cells, are possible. The same company produces a system with a similar design (ASE 350) and it should be possible to manufacture larger adapters to couple its larger cells (34, 66, and 100 mL) to increase the sample capacity.  

Figure 4: Photo of one assembled and one dissembled modular PLE assembly.			

Due to the cell and heater oven designs of other PLE systems on the market, it is currently difficult to transfer this modular approach to ther instrument platforms. Fluid Management Systems (FMS, Watertown, Massachusetts) does, however, offer a similar solution in which PLE is coupled through an evaporative interface to a multicolumn system for subsequent (at-line) cleanup. This system is primarily intended for dioxin analysis. 

The separation of the extraction cells into two or three separate compartments greatly enhances the flexibility and versatility of the PLE with the in-cell purification concept. Modules with sample, matrix retainer, and adsorbents or ion-exchange resins may be added or removed as needed to obtain the desired separation, much like the solid-phase extraction (SPE) cartridges that are frequently used for extraction and purification of aqueous samples (also for purification of extracts of solid samples). The combinations are almost endless, but a few are described below and in Figure 5.   

Figure 5:  Selected modes of operation of modular PLE.			

The "trap and release" mode (Figure 5, top left) is very similar to the procedure described in Figure 3 and the corresponding text, with one major difference: the sample modules can be removed in modular PLE before elution of the trapped material. The "class separation" mode (Figure 5, top right) is closely related to trap and release, but involves sequential extraction of the trap. The final extraction of the most tightly bound analytes is done in the reverse direction, after turning the trap. For samples of moderate complexity it may also be possible to perform extraction and purification based on a "dual mode cleanup" (Figure 5, lower left), by loading a two-module extraction cell with sample and an adsorbent, then extracting the sample, trapping the desired analytes on the adsorbent (adsorbent A), removing the sample module, and back-extract the trapped material through a second adsorbent (adsorbent B). More complex samples can be handled using a three-module "matrix retainer and trap" set-up (Figure 5, lower right), in which the compartments are packed with sample, matrix retainer, and adsorbent, respectively. The samples are extracted and the extracted material passes through a matrix retainer and adsorbent which, once again, is used to trap selected analytes. The sample and matrix retainer modules are then removed and the trap is sequentially extracted. 

Spinnel and colleagues (22) combined a user-friendly sampling technique based on gas adsorption on polyurethane foam plugs (PUFP) and aerosol particle removal using glass fiber filters with modular PLE for cost-efficient extraction and sample preparation of polychlorinated dibenzo-

-dioxins and dibenzofurans (dioxins or PCDD/Fs) in flue gasses from municipal solid waste incinerators. The procedure utilized the trap and release mode (Figure 5) with the sampling material packed in the upstream cell, followed by an activated carbon-filled cell. The dioxins were extracted from the PUFP and the filter using three cycles of 

-heptane–acetone (static time, 5 min; temperature, 150 °C; pressure, 2000 psi). The upper cell and the coupling assembly were removed and the cell with the carbon trap was capped. The trap was inverted and extracted with three cycles of toluene (static time, 7 min; temperature, 175 °C) to back-elute the dioxins.  

The procedure offers clear advantages as compared with normal PLE and selective PLE with in-cell carbon trap (21). Because the carbon adsorbent is very selective to planar compounds, the bulk of matrix is passing through the carbon trap during extraction. Furthermore, the coupled cell assemblies enable the removal of the matrix before back-flushing the carbon trap, thereby preventing extraction of unwanted matrix constituents during the exhaustive toluene extraction of the trap. 

Excellent agreement was obtained between Soxhlet extraction and multistep cleanup and modular PLE followed by a miniaturized open column clean-up (Figure 6). For mono- to tetra-CDD/F the results of the two extraction techniques were statistically equivalent while modular PLE provided superior extraction of highly chlorinated congeners. This analytical strategy was proposed to be used to improve national and international dioxin emission inventories.  

Figure 6: The graph illustrates the relative extraction efficiency of modular PLE in comparison to Soxhlet extraction (modified from reference 23).			

A modification of the popular PLE technique is discussed.

LCGC North America-07-01-2010

Simultaneous Headspace-GC–FID Analysis for Methanol and Ethanol in Blood, Saliva, and Urine: Validation of Method and Comparison of Specimens

QuEChERS, a Sample Preparation Technique that Is "Catching On": An Up-to-Date Interview with the Inventors

Where Did That Peak Come From?

When Peaks Collide

A Modular Approach to Pressurized Liquid Extraction with In-Cell Cleanup

Ion Chromatography: An Overview and Recent Developments

Vol 28 No 7 LCGC North America July 2010 Regular Issue PDF