Overview of Methods and Considerations for Handling Complex Samples


LCGC North America

LCGC North AmericaLCGC North America-04-01-2020
Volume 38
Issue 4

When working with complex samples, we need effective approaches to deal with matrix interferences. Here, we outline methods of sample preparation, on-line sample treatment, and instrument tools that can help. We also provide examples of applications and guidance for how to evaluate the best option for your complex sample.

Complex samples can be challenging, and there are many caveats that need to be considered to mitigate matrix effects. Matrix interferences can be detrimental to analysis, and the contents of a sample need to be evaluated to determine what could affect analysis. Approaches need to be tailored to remove or work around matrix interferences, to produce reproducible, reliable, and accurate methods of analysis. Here, methods of sample preparation, on-line sample treatment, and instrumental tools are outlined, and several examples of applications are discussed. Some general guidance is also outlined on how to evaluate the best option for your complex sample, and how to move forward.

There are many challenges related to complex sample matrices that analytical chemists have to overcome. These challenges can be met by skillful sample preparation, on-line sample clean-up, sample pre-treatment, or the use of instrumental tools. Sample preparation is an evident place to start, and includes methods such as solid-phase extraction (SPE), liquid-liquid extraction (LLE), salting-out, derivatization, filtration, and centrifugation, among many others. Multiple approaches can also be combined, but doing so can quickly become cumbersome for large sample sets. Online sample cleanup can be a welcomed alternative to relieve some of the manual steps and allow automation from the instrument of choice, but is not always a viable option. Lastly, the instrument’s abilities should not be undervalued. Multiple reaction monitoring (MRM) transitions can be very useful to gain specificity when using a triple quadrupole mass spectrometer, but can sometimes fall short when analyzing similar compounds that do not produce unique MRM transitions. Instrumental tools that allow deconvolutions are also possible when chromatography falls short, as in the case of vacuum ultraviolet spectroscopic detection for gas chromatography. In addition to this, capabilities like post-run spectral filters are a valuable tool to highlight certain classes of compounds in a convoluted complex matrix when analyzing by gas chromatography–vacuum ultraviolet spectroscopy (GC-VUV). All of these options for sample preparation, clean-up, and analysis should be taken into consideration when dealing with complex sample matrices, in order to retrieve meaningful, reliable, and reproducible data for the determination of target analytes.

Consider Your Sample

Before deciding on a technique to move forward with method development, it is important to consider what is in your sample, and what could be a possible interference. Consideration needs to be given to whether the analyte can be analyzed by GC, and, if not in native form, if it can be derivatized to be made more amenable to GC analysis. Derivatization can be a useful technique, but unless it can be automated, it is best to be avoided if the sample set is large, to save time and sanity. GC should not be written off too quickly, though, especially when dealing with complex samples. Headspace sampling can be a terrific technique paired with GC to save time on the front end during sample preparation, and, in many cases, no other sample clean-up is necessary. The measurement of ethanol content in blood samples is a great example of this technique and requires no clean-up of the matrix prior to injection (1). There are very few volatile substances that can be in a blood sample that will interfere with the measurement of ethanol, and if an appropriate column chemistry is used in conjunction with headspace sampling, no matrix clean-up is required.

Liquid chromatography (LC) is imperative for samples with higher molecular weight analytes, those that require extensive derivatization, or those that are otherwise non-volatile or thermally labile; these properties make them less amenable to GC analysis. Special attention needs to be given to cleanup of the sample matrix to ensure that the LC column is not compromised, that the lines do not get clogged, or that the system is not dirtied by the complex sample. Precipitation is also a concern for LC, so compatibility between the mobile-phase solvents, mobile-phase additives, and samples should be taken into consideration. Supercritical fluid chromatography (SFC) continues to re-emerge for use over a wide range of sample and analyte types, as many instrument manufacturers continue to improve their offerings; SFC is not covered in significant detail in this work, but it is acknowledged to be an interesting technique to bridge the gap between GC- and LC-amenable analytes. In recent work, online supercritical fluid extraction (SFE) was paired with supercritical fluid chromatography-mass spectrometry (SFE-SFC-MS), to analyze polycyclic aromatic hydrocarbons (PAHs) in soil (2). The coupling of these two techniques into one instrument and method allows for minimization of sample preparation, and reduction in loss of sample and sample contamination (3).

Food, environmental, and biological samples are some examples of matrices that can be especially tricky. In the case of food analysis, the USDA Food Composition Databases are a great tool to use when evaluating possible interferences (4). They can provide a general idea of what is expected to be in a sample, as far as components with different physicochemical properties, such as fats, carbohydrates, and proteins. Environmental samples can be challenging because of their non-uniformity, and will need tailored methods to mitigate any interferences, while still achieving as much consistency between sample sets as possible. Biological samples, and in some cases food samples, are plagued with large biomolecules and proteins that can greatly hinder analysis and make instrumentation dirty. In order to move forward, it is best to get a handle on what analytes to target and how to exclude or work around interferences. 

Matrix Interferences and the Effect They Can Have on Analysis

Interferences can occur within the sample matrix, and affect the sample analysis in a number of different ways. Matrix effects can mask, suppress, augment, or make imprecise sample signal measurements. This can occur chromatographically, as in the case of coelution, or during ionization, in the case of mass spectrometric detection, and result in highly variable or unreliable data. To correct for matrix effects encountered during electrospray ionization, the use of stable isotopically labeled internal standards is recommended. This is so that the internal standard is nearly perfectly coeluted with the analyte of interest, experiences the same ionization suppression, or enhancement as the analyte, and, thus, can be more effectively used to correct analyte response.

An example of this can be seen in previous research where estrogens were detected in phosphate-buffered saline-bovine serum albumin, gelded horse serum, and mouse serum (5). Internal standards were used to compensate for any fluctuation during the sample preparation procedure and ionization, but because deuterated internal standards were used, a deuterium isotope effect was observed, resulting in slightly different retention times between the internal standard and target analytes (5). The drawbacks of using an isotopically labeled internal standard can be availability and cost. Often, isotopically labeled internal standards are not readily available or can put a dent in the budget. If your laboratory is equipped, it might be advantageous to synthesize the internal standards. In previous research, a laboratory targeting the quantitation of lysosphingolipid bases was able to synthesize their own carbon-13 (13C) sphingoid bases to use as internal standards for their analysis by ultrahigh performance liquid chromatography electrospray ionization tandem mass spectrometry (UHPLC-ESI-MS/MS) (6).

It is worth noting that, when choosing an internal standard, it is important to find an internal standard that is physicochemically similar to the target analyte, but structurally unique, not present in the samples, and coeluted with your analyte but has unique MS transitions. Nitrogen-15 (15N) and 13C labeled internal standards are often preferred over deuterated standards, to eliminate deuterium isotope effects (7). Deuterium isotope effects, in terms of altered chromatographic retention, will be exacerbated the longer the analyte, and its deuterated internal standard are retained in the column, especially in reversed-phase LC mode.


One concern of sample interferences can also be reactivity, especially in the case of reactive analytes. This can happen when the contents of the sample react with target analytes, and is often not reproducible, and can hurt precision. The best way to alleviate this would be to remove the interference that is reacting, but this can be problematic, because it is not always clear what is reacting, especially in extremely complex or unknown samples. A specific case of this can be seen in the detection of formaldehyde, an extremely reactive analyte, in a sample of shale core and produced water. The formaldehyde originated from a resin-coating applied to a proppant, a hydraulic fracturing additive used in unconventional oil and gas extraction (8). The leaching of formaldehyde from the resin coating on the proppants was tested at laboratory simulated subsurface conditions by heating to subsurface temperatures and with the addition of the shale core and produced water, two components that these additives are likely to contact during hydraulic fracturing (8). When the shale core and produced water were added to the sample matrix, the concentration and sometimes precision in measuring the formaldehyde was diminished, likely due to competing reactions taking place from the matrix (8). An example of the results for these experiments can be seen in Figure 1, where the proppants were left to soak in different matrices for 20 h, and then analyzed for their formaldehyde content (8).  These data were obtained with the combined use of derivatization, to “trap” the reactive formaldehyde, in combination with headspace-gas chromatography-mass spectrometry (HS-GC-MS) analysis. In this case, better precision of the HS-GC-MS method was observed compared to a traditional LC-based approach, because all of the reaction chemistry, derivatization, and sampling could be performed in a sealed vial; this limited loss of the volatile analyte (8).

Figure 1: Four different resin-coated proppants (2 phenol-formaldehyde [PF1 and PF2] and 2 polyurethane [PU1 and PU2]) tested for derivatized formaldehyde leaching after 20 h  of soaking in water, produced water inorganic (PWI), untreated produced water (PW), or produced water with added shale core (PW + Shale). Each were tested either at room temperature or heated to 200 °F (93 °C). The produced water with added shale core matrix returned lower quantities of the derivatized formaldehyde leaching from the proppants, likely due to competing reactions in the matrix with the derivatization of the formaldehyde (8).


Sample Preparation

Ideally, sample preparation should be kept to a minimum to streamline sample throughput, but that is not always a viable option when handling complex samples. There are numerous techniques to choose from that can be implemented; the choices depend on the nature of your sample matrix and analyte.

Solid-phase extraction (SPE) is a sample preparation technique that can be of use in preconcentrating samples, removing interferences, or desalinating samples. This can be especially useful in aqueous environmental matrices, as in the case of detecting nonsteroidal anti-inflammatory drugs (NSAIDs) in drinking water, surface water, and wastewater, where the analytes are present in low concentrations (9). The setup usually consists of a manifold and cartridges that are used to trap and elute analytes. Large volumes of an aqueous sample can be loaded onto a cartridge, and eluted in a smaller volume, to preconcentrate the analyte. The system can use positive or negative pressure, and a variety of sorbents are available from which to choose.

Solid-phase microextraction (SPME) can be used to extract volatiles and non-volatiles from a liquid or gas matrix. SPME consists of a fiber coated with a stationary phase, liquid polymer, or both, on the end of a plunger of a syringe or needle (10). SPME can be used to sample from liquid by direct immersion or by headspace sampling. This method of sampling is ideal for offsite sample collection, because it is easily transported to and from the site and back to the laboratory for analysis. An example of the portable capabilities of SPME comes from a recent research article describing a method that was developed to sample a plant’s volatile organic compound profiles. This analysis was performed using SPME on-site and subsequent GC–MS analysis in the laboratory (11). SPME can also be used as a sample cleanup technique. An example of this technique can be seen in a method developed for detecting short chain fatty acids, which are the end products of intestinal bacterial fermentation, from an in vitro gastrointestinal model  (12). The model is made up of a very complex chemical composition, but by sampling using SPME, no additional extraction was needed, and only a simple sample treatment was used (12). Both SPE and SPME techniques require special apparatuses including cartridges, manifolds, and fibers that are available from manufacturers and can be somewhat costly. However, the selective extraction attainable through the use of these techniques can be very effective for eliminating unwanted matrix interferences prior to analysis.

Salting out can also be a useful technique for sample preparation. The addition of salts can help reduce the hydration of target analytes and make them more amenable to extraction. Salting out can be used to remove solid particles, fats, waxes, and even DNA from a sample. It can be used in combination with headspace or liquid-phase extraction techniques.  Salting out is an integral part of the popular method referred to as QuEChERS (quick, easy, cheap, effective, rugged, safe), which also often includes dispersive solid-phase extraction (dSPE) (13). QuEChERS has been implemented in many different applications that would otherwise require more extensive cleanup of the matrix. One such application is the detection of pesticides in food samples, often a very complex sample matrix. In previous work, pesticides were detected in different fruit and vegetable samples using QuEChERS sample preparation followed by analysis with liquid chromatography–tandem mass spectrometry (LC–MS/MS) (14). 

Salting out can also be advantageous during sampling when using the headspace technique coupled to gas chromatography, to facilitate partitioning of analytes into the gas phase. This can be performed with the addition of traditional salts (NaCl) or even ionic liquids (15). Liquid-liquid extraction (LLE) can also use salting out, called salting out-assisted LLE (SALLE), and is similar to QuEChERS. In one example from previous research, SALLE was implemented to extract oxytocin in plasma samples (16). These analytes are often difficult, due to their extremely low concentrations and the interferences present in the plasma. By using SALLE, the methods were able to overcome these challenges and obtain lower limits of detection.

Dispersive solid-phase extraction (dSPE) is another sample preparation clean-up technique if water, polar, non-polar, or pigmentations need to be removed from the matrix. As the name implies, in dSPE, solid-phase extraction particles are dispersed in the sample (rather than being used in a column format) and removed by centrifugation. In previous work using QuEChERS and dSPE as a sample cleanup technique, nicotine and its metabolites were detected in catfish, tuna, salmon, and tilapia (17). It was found that the different types of fish required different dSPE components to optimize the methods, but with the optimized methods minimal or no matrix effects were present. The judicious use of salts to aid phase partitioning of desired analytes can be a simple and economical alternative to purchasing additional sample preparation materials.

Various other simple and cost-effective sample preparation techniques can also be performed in the laboratory. Liquid-liquid extraction (LLE) is typically performed with two immiscible solvents, and can be used to extract certain analytes based on their relative solubilities in the two solvents. In one study, LLE was used to target chlorophenols from wastewater (18). In this example, hydrophobic ionic liquids were tested for their extraction capabilities of the targeted class compounds from the water layer. One aspect of using LLE as a sample preparation technique that needs to be considered is the difficulty OF performing a multiclass compound extraction. This can be challenging, due to the differing degrees of solubility from compound class to compound class.

Filtering and centrifugation are very important steps in the sample preparation process, especially for LC analysis. Centrifugation can help remove solids or small particulates from a sample, and ensure that an autosampler and system do not become compromised. According to an old adage, “if you can dissolve it, you can HPLC it,” but, if you can’t dissolve it, it is going to clog your system.  When trying to remove proteins, precipitation is a quick and easy tool to use. Typically, a chilled organic solvent, such as acetonitrile or acetone, can facilitate the precipitation of proteins, due to their limited solubility in these solvents. Finally, when targeting trace analytes from a complex matrix, concentration by dehydration can be used by drying down the sample with a stream of N2 and regenerating with a solvent. This process also has the ability to be automated. Overall, many of these methods require materials that most laboratories are likely to already have (centrifuge, filters, glassware, gases, solvents), so they can be quite

Derivatizing the sample is sometimes a necessity. Derivatization is not just a technique to facilitate GC analysis. It can also be performed for LC to make a small molecule (>100 m/z) larger and more amenable for MS, or to give a molecule a UV- or fluorescence-active moiety. (19,20) In previous research, dansyl chloride was used to derivatize estrone, 17a-estradiol, 17b-estradiol, and estriol for their determination in human cerebrospinal fluid (20). By this method, no offline extraction or cleanup was necessary and levels in the pg/mL were able to be detected using LC–MS/MS (20).


Derivatizing can also be useful when trying to differentiate chiral compounds. One useful application of this is in the differentiation of amphetamines. D-methamphetamine is a Schedule II controlled substance whereas the enantiomer, L-methamphetamine, is the active ingredient in many over-the-counter medicines in the United States (21). A method was developed for methamphetamine and amphetamine chiral quantitation in blood plasma. Chiral differentiation of these compounds was achieved by the addition of Marfey’s reagent to derivatize these compounds into distinguishable analytes (diastereomers) by LC–MS/MS (21). If chiral columns are not available to differentiate chiral compounds from each other, chiral derivatization is a straightforward way to be able to differentiate compounds when necessary.

Three examples of common methods of derivatization techniques can be seen in Figure 2. When choosing which sample preparation technique is best for your complex matrix, it is important to consider your specific target analytes and the interferences you want to rid your sample of, in addition to how much time you want to spend handling each sample. Although derivatization is often referred to in a negative context, as a time-consuming step to justify an alternate approach, it is sometimes actually quite straightforward.  Many derivatization methods have been developed to be fast and essentially quantitative for the conversion of target analytes to facilitate either GC or LC analysis.

Figure 2: Examples of common methods of derivatization for different methods of analysis: (a) silylation performed on a carboxyl group to make it less polar and more volatile for analysis on GC–MS; (b) derivatization performed on formaldehyde by reacting it with 2,4-dinitrophenylhydrazine to add a chromophore to the molecule before subsequent analysis by HPLC-UV; and (c) dansyl chloride derivatization on estradiol for analysis by LC–MS/MS (8,20,22).


Online Sample Treatment

There are various other techniques that can be implemented online. Moving sample preparation and sample treatment online can be advantageous, because online sample treatment requires less manual sample handling, and provides increased recovery, improved limits of detection, less human error, and reduced exposure of compounds to the environment, which can be useful for analytes that are photosensitive or reactive to oxygen. One common method of online sample treatment is online SPE, which can be implemented to clean and preconcentrate target analytes in an automated fashion (9). Other forms of sample preparation that can be automated include the use of a continuous stirred tank reactor (CSTR) for on-line sample dilution. A CSTR contains a reservoir that allows fluid to travel through the apparatus. This device can be used to continuously dilute a sample injected into it. In previous research, a CSTR was used in the investigation of native carbohydrates to study the electrospray response factors by LC–MS/MS (23). By using the CSTR apparatus, analyte response data for a large range of analyte concentrations were able to be obtained with each single injection.

Restricted access media (RAM) can be particularly useful when dealing with complex matrices, especially when targeting small molecules and trying to rid the sample of large biomolecule interferences, prior to LC analysis (24–30). RAM columns work on a similar principle to size exclusion chromatography. The outer surface of the stationary phase has a non-retentive and size-restrictive layer, while the inner pores of the support material have a bonded group, like C4, C8, C18, and so on. Only the small molecules can access and interact with the inner-pore phase to be retained on the column, while the large molecules are largely unretained and washed to waste. This can be especially useful when looking at small molecules in complex matrices, such as whole blood or plasma, where there is an abundance of protein interferences. This is a technique that can be put in-line, in the flow path before the analytical column, so that the extraction process is completely automated. This can provide a welcomed alternative to extensive sample preparation and sample handling.

An example of this methodology was used in the quantification of lipid mediators in skeletal muscles using the RAM column coupled to LC–MS/MS (28). This technique requires the use of one or two high-pressure valves in the column oven, and some extensive LC programming for loading, eluting, and washing parameters. An example of LC settings and the valve setup can be seen in
Figure 3. (29).

Figure 3: Online restricted access media (RAM)-LC settings for mobile phase concentration, flow rate, valve positions, and valve diagrams. From 0 to 6 min the sample is being loaded by pumps C + D onto the RAM column, while the analytical column is being equilibrated with pumps A + B. The analytes are back eluted from the RAM column by pumps A + B from 6 to 9 min and sent to the analytical column. The valves switch and pumps A + B perform the analytical separation from 9-14 min, while pumps C + D wash the RAM column. Finally, at 20.5 min the valves are switched to their starting position and the RAM and analytical columns are equilibrated for the next injection (29).

When dealing with trace analysis from complex samples, the RAM column can be loaded with large sample injections without affecting the peak shape. In addition to large sample volumes, the RAM can also be loaded with multiple injections to allow for ultratrace analysis to be performed. In previous research, bisphenol A was able to be detected in human saliva samples by using a RAM column in combination with LC–MS/MS (30). Parts per trillion levels of bisphenol A were detectable by performing multiple injections on the RAM column to concentrate the analyte and remove unwanted large biomolecules (30).

Instrumental Tools

Many instrumental tools can be very valuable and can save time and effort in sample preparation and method development. Instrumental tools can include things like multiple reaction monitoring (MRM), spectral filters, and programs that allow for deconvolution of coeluted signals. In the case of mass spectrometry, precursor-to-product ion MRM transitions allow for high sensitivity and specificity from complex matrices, as long as unique transitions can be acquired. Although useful for the differentiation of most compounds, this approach is not always an option when analyzing isomeric compounds. In previous work on cannabinoids, a number of isomers were analyzed and it was found that some of the compounds had common fragmentation pathways when using MRMs in GC–MS/MS (22). In order to differentiate analytes with common fragmentation pathways, the compounds had to be chromatographically separated. This was partially achieved by silylating the cannabinoids; however, some potential interferences still existed. It was still necessary to monitor less sensitive secondary and tertiary precursor-to-product ion transitions to provide adequate specificity.


In complex matrices, there is likely to be a case where the target analytes are isomeric, are not resolved chromatographically, and do not produce unique MRM transitions when analyzed by the mass spectrometer. These can still be accurately quantified using instrumental tools with the right detector. The vacuum ultraviolet (VUV) detector measures in the 120 to 240 nm range, where virtually all compounds have unique, compound-specific absorbance spectra (31,32). The VUV can be used to differentiate isomeric forms of compounds, such as in the case of illicit drugs. In previous research four classes of compounds were chosen that are listed in the Dutch drug legislation, as well as some of their uncontrolled isomers (33). These compounds were then analyzed and were able to be differentiated by GC-VUV analysis.

Coeluted compounds, including isomers, can be distinguished and quantitatively deconvoluted using a VUV detector. Since Beer’s law is additive, the overlapping VUV absorption signals arising from coelution can be easily deconvoluted (31,34). A straightforward least squares approach can be used to discern individual component contributions to the overlapping signals (34). An example of deconvolution can be seen in Figure 4, where coeluted isomers of dimethylnaphthalenes have been separated into their respective contributions to the coeluted peak.

Figure 4: Example of isomers of dimethylnapthalene (DMN) coeluted and measured in the 200–220 nm range (blue). Deconvolutions were performed into the respective concentrations for each isomer (green and orange) using manufacturer software. The deconvolved individual contributions in this case are somewhat noisy because of the inherent similarity of DMN spectra. Even so, these isomers can be accurately deconvolved up to a 100:1 relative abundance (34).

In order to use this tool, the analytes have to first be amenable to GC, which can complicate things if the analytes are not volatile or thermally stable. The similarity of spectra between the coeluted compounds is a governing factor of how accurately the compounds can be deconvoluted. The more distinct the absorbance spectra, the easier the compounds are to deconvolute over a wider range of relative abundance; the more similar the spectra are, the more difficult they are to differentiate.

Spectral filters can also be a great tool in conjunction with VUV for complex samples when certain classes of compounds are of interest. Post-run digital spectral filters can be applied to samples that are too complex to perform deconvolutions on because what compounds are coeluted with target analytes may not be known. In the VUV range, certain classes of compounds absorb more intensely in different wavelength ranges. Saturated compounds absorb in the 125–160 nm range, whereas unsaturated compounds absorb in the 170–240 nm range (34). This information can be used to build and apply a spectral filter to apply to a chromatogram post-run. In the same set of experiments performed for coeluted isomers of dimethylnapthalenes, spectral filters were used on samples of diesel fuel and jet fuel, which are extremely complex, to show where naphthalene, methylnaphthalenes, dimethylnaphthalenes, and trimethylnaphthalenes elute (34). This was accomplished by experimentally determining where naphthalene class compounds absorb strongly (210–220 nm) and then applying a digital filter in that range to selectively identify where the naphthalenes are eluted without performing deconvolutions (34).


How to Evaluate Which Is the Best Option to Use for Your Analysis

A good place to start when deciding what is best to try for method development in complex samples is to first take into account what you are targeting and what you want to remove. Knowing your sample and the possible interferences can facilitate at least an educated guess as to what types of sample handling should be implemented first and can save a lot of guess work, time, and money. Excessive fats, waxes, and proteins can dirty instruments, so special attention needs to be given to remove these from the matrix. Next, decide how removal of problematic compounds can be performed. Does the sample need extensive clean-up on the front-end to remove matrix interferences or can the use of an instrumental tool or on-line sample treatment mitigate unsavory effects on sample analysis, while maintaining reproducible results?

A flow-chart of the topics covered in this article for sample preparation, on-line sample treatment, and instrumental tools discussed in this article can be seen summarized in Figure 5. When starting with a complex sample, the first consideration is which method to choose, LC or GC. Once LC or GC is chosen, there are a number of different sample preparation and automated sample handling choices that are amenable for each technique. Finally, depending on the detector chosen, available tools may aid in the analysis of the complex sample and offer additional tools that can help with complex sample analysis. This flow chart is by no means exhaustive in its content, but can be a good starting point when considering the available options.

Figure 5: Flow chart of a selection of common available options for treating complex samples, which are highlighted in this article. The process begins by deciding whether the target analytes are more amenable to GC or LC analysis.


When starting a new project, it can be overwhelming to try to anticipate the problems your newly assigned complex matrix may give you, but there are many solutions that can be implemented to save time and money. After the sample matrix is evaluated and a method of analysis is chosen (GC or LC), the type of sample clean-up appropriate for the matrix can be determined. If the analytes are going to be analyzed via GC, headspace is a great option due to the limited need for sample clean-up; SPME sampling in solution or in the headspace might be the next best choice to consider. If using LC and excessive proteins are not a concern, a simple dilute, filter or centrifuge, and shoot with the right solvent might just do the trick. Ultratrace analysis in a complex matrix can be accomplished with the use of a RAM column and multiple injections. The most important thing when trying to develop a method for a complex matrix is to start with literature and make the best educated guess you can, and to not be afraid to try something new. In the end, the proof of performance through method validation is key to acquiring reliable, reproducible, and accurate data.


  1. L. Zamengo, G. Tedeschi, G. Frison, C. Griffoni, D. Ponzin, and A.W. Jones, Forensic Sci. Int.295, 213–218 (2019).
  2. A.P. Wicker, D.D. Carlton, K. Tanaka, M. Nishimura, V. Chen, T. Ogura,W. Hedgepeth, and K.A. Schug, J. Chromatogr. B1086, 82–88 (2018).
  3. M. Zoccali, P. Donato, and L. Mondello, TrAC116, 158–165 (2019).
  4. USDA Food Composition Databases (United States Department of Agriculture Research Service, 2019). https://ndb.nal.usda.gov/ndb/search/list (accessed June 27, 2019).
  5. H. P. Nguyen, L. Li, I. S. Nethrapalli, N. Guo, D. Toran-Allerand, D.E. Harrison, C.M. Astle, and K.A. Schug, J. Sep. Sci.34, 1781–1787 (2011).
  6. M. Mirzaian, P. Wisse, M.J. Ferraz, A.R.A. Marques, P. Gaspar, S.V. Oussoren, K. Kytidou, J.D.C. Codee, G. van der Marel, H.S. Overkleeft, and J.M. Aerts, Clinica Chimica Acta466, 178–184 (2017).
  7. R. Grant, Everything you wanted to know about internal standards but were too afraid to ask (2019). https://glc2.workcast.com/clusterSVCFS1/NAS/OnDemand/10994/1775713496937128/Media/10994_20190530100329002_edit1.mp4?mc_cid=0a08c656b8&mc_eid=64db5d8f3e (accessed June 27, 2019).
  8. J. Schenk, D.D. Carlton, J. Smuts, J. Cochran, L. Shear, T. Hanna, D. Durham, C. Cooper, and K.A. Schug, Environ. Sci.: Processes Impacts 21, 214–233 (2019).
  9. Z. Marta, B. Bobaly, J. Fekete, B. Magda, T. Imre, and P.T. Szabo, J. Pharm. Biomed. Anal.160, 99–108 (2018).
  10. A. Spietelun, M. Pilarcyzyk, A. Kloskowski, and J. Namiesnik, Chem. Soc. Rev.39, 4524–4537 (2010).
  11. A.G. Fung, M.S. Yamaguchi, M.M. McCartney, A.A. Askenov, A. Pasamontes, and C.E. Davis, Microchem. J.146, 407–413 (2019).
  12. C. Douny, S. Dufourny, F. Brose, P. Verachetert, P. Rondia, S. LeBrun, M. Marzorati, N. Everaert, V. Delcenserie, and M. Scippo, J. Chrom. B1124, 188–196 (2019).
  13. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, and F.J. Schenck, J. AOAC Int.86(2), 412–431 (2003).
  14. J. M. Montiel-Leon, S. Vo Duy, G. Munoz, M. Verner, H. Hendawi Moya, M. Amyot, and S. Sauve, Food Control104, 74–82 (2019).
  15. E. Varona-Torres, D.D. Carlton Jr., Z.L. Hildenbrand, and K.A. Schug, Anal. Chim. Acta1021, 41–50 (2018).
  16. D. Liu, X. Han, X. Liu, M. Cheng, M. He, G. Rainer, H. Gao, and X. Zhang, J. Pharm. Biomed. Anal. 173, 62–67 (2019).
  17. Y.W. Chang, H.P. Nguyen, M. Chang, S.R. Burket, B.W. Brooks, and K.A. Schug, J. Sep. Sci. 38, 2414–2422 (2015).
  18. R. Sulaiman, I. Adeyemi, S.R. Abraham, S.W. Hasan, and I.M. AlNashef, J. Mol. Liq.294, 111680 (2019).
  19. Y. Baghdady and K.A. Schug, J. Sep. Sci.39, 102–114 (2016).
  20. H. Fan, B. Papouskova, K. Lemr, J.G. Wigginton, and K.A. Schug, J. Sep. Sci. 37, 2010–2017 (2014).
  21. M.N. Newmeyer, M. Concheiro, and M.A. Huestis, J. Chrom. A.1358, 68–74 (2014).
  22. A. Leghissa, Z. L. Hildenbrand, F.W. Foss Jr., and K.A. Schug, J. Sep. Sci.41, 459–468 (2018).
  23. J. Thacker and K.A. Schug, J. Am.Soc. Mass Spectrom.30, 439–447 (2019.
  24. S. H. Yang, H. Fan, R.J. Classon, and K.A. Schug, J. Sep. Sci.36, 2922–2938 (2013).
  25. B. Papouskova, H. Fan, K. Lemr, and K.A. Schug, J. Sep. Sci.37, 2192–2199 (2014).
  26. Y.Z. Baghdady and K.A. Schug, J. Sep. Sci.39, 4183–4191 (2016).
  27. H. Dipe de Faria, L. Abrao, M. Santos, A. Barbosa, and E. Figueiredo, Anal. Chim. Acta959, 43–65 (2017).
  28. Z. Wang, L. Bian, C. Mo, M. Kukula, K. A. Schug, and M. Brotto, Anal. Chim. Acta984, 151–161 (2017).
  29. H. Fan, B. Papouskova, K. Lemr, J.G. Wigginton, K.A. Schug, J. Sep. Sci.37, 2010–2017 (2014).
  30. S.H. Yang, A.A. Morgan, H.P. Nguyen, H. Moore, B.J. Figard, and K.A. Schug, Environ. Toxicol. Chem. 30,1243–1251 (2011).
  31. K. A. Schug, I. Sawicki, D. D. Carlton, H. Fan, H. McNair, J. P. Nimmo, P. Kroll, J. Smuts, P. Walsh, and D. Harrison, Anal. Chem.85, 8329–8335 (2015).
  32. I. C. Santos and K. A. Schug, J. Sep. Sci.40, 138–151 (2017).
  33. R.F. Kranenburg, A.R. Garcia-Cicourel, C. Kukurin, H. Janssen, P. J. Schoenmakers, and A.C. van Asten, Forensic Sci. Int.302, 109900 (2019).
  34. J. Schenk, J. Mao, J. Smuts, P. Walsh, P. Kroll, and K.A. Schug, Anal. Chim. Acta945, 1–8 (2016).

Jamie L. York received her Ph.D. in chemistry from the University of Texas Arlington in 2019.  She is currently an LC Applications Chemist at Restek Corporation, in Bellefonte, Pennsylvania. Kevin A. Schug is a Full Professor and the Shimadzu Distinguished Professor of Analytical Chemistry in the Department of Chemistry & Biochemistry at the University of Texas Arlington. Direct correspondence to: kschug@uta.edu



Related Content