Independent Consultants in Environmental and Forensic Chemistry
Volume 4, Issue 3, Summer 2001
What is the difference between phenol, phenols, and phenolics?
The terms Aphenol,@ Aphenols,@ and Aphenolics@ have caused confusion in environmental analyses for over 20 years. APhenol@ (or hydroxybenzene) is a single organic compound. APhenols@ refers to the class of compounds having a hydroxyl (-OH) group, as well as other substituent groups, on a benzene ring or, prior to the introduction of the list of priority pollutants in the late 1970s, to the colorimetric analyses of phenols. APhenolics@ is the title used for the US Environmental Protection Agency=s (EPA) SW-846 Methods 9065 and 9066 for the colorimetric measurement of numerous known and unknown substances with one or more hydroxyl groups attached to a benzene ring. The organic compound, phenol, is the chemical used to calibrate the colorimetric test for the classes of compounds known as Aphenols@ or Aphenolics.@
The EPA=s list of priority pollutants includes a number of phenols measured in the acid fraction of the base/neutral/acid (BNA) analysis or the semivolatile organic compounds (SVOCs) analysis using EPA Methods 525, 625, 1625, and 8270 and the Contract Laboratory Program Statement of Work (CLP SOW). These phenols include 2,4-dimethylphenol; phenol; p-cresol (4-methylphenol); 2,6-dichlorophenol; 2,3,4,6-tetrachlorophenol; 2-chlorophenol; 2-nitrophenol; 2,4-dichlorophenol; 4-chloro-3-methylphenol; 2,3,6-trichlorophenol; 2,4-dinitrophenol; 4-nitrophenol; 2-methyl-4,6-dinitrophenol; pentachlorophenol; o-cresol (2-methylphenol); 2,4,5-trichlorophenol; and 2,4,6-trichlorophenol. Note that phenol is included in the list of the priority pollutants.
Each of the priority pollutant phenols, including phenol itself and the para substituted phenols, can be found, identified, and quantified by several gas chromatographic/mass spectrometric EPA methods such as 525, 625, 1625, and 8270. The colorimetric analytical procedures entitled Aphenols@ or Aphenolics@ measure many compounds and provide a quantitative result for phenols or phenolics as a class. However, no specific compound can be identified or quantified using the colorimetric method. Consequently, if a sample is analyzed by one of the colorimetric methods, no specific phenolic compound, including phenol itself, can be stated to be present in the sample.
Decaying vegetation and, in particular, wood produces numerous phenols because the benzene with the hydroxyl group is in a major portion of a woody substance called lignin. Lignin is removed from paper pulp made from trees and degrades to form numerous substances including phenols. These phenols include humic and fulvic acids, large molecules of organic matter that are soluble in water, as well as aromatic substances with several hydroxyl groups on the benzene ring. These phenolic compounds can and do form groundwater plumes emanating from swamps and wetlands, and the plumes can contain a variety of phenols. A positive result from colorimetric analysis can only tell you that the hydroxy benzene structure is present. It cannot tell you what specific compounds are present.
What Are Duplicate, Replicate, or Co-located Samples?
Duplicate samples are used to estimate the precision or reproducibility in an analytical system. Laboratory duplicates originate in the laboratory and are used to estimate the laboratory=s ability to reproduce an analytical result. A field duplicate is generated in the field and, consequently, provides a measure of the reproducibility of not only the laboratory=s efforts but also that of the sampler and sampling techniques.
Field duplicates are the Asame@ sample in separate containers. However, the manner in which the duplicates were collected may impact whether they are truly the Asame@ sample. A subtle assumption in collecting field duplicates is that the matrix being sampled is homogeneous, i.e., any one part of the matrix is identical in composition to any other part. While this may be true in many cases for liquid samples, solid samples tend to pose greater chances that this is not true. For soil samples, differences in particle size alone could yield considerably different analytical results for the duplicates. If differences in analytical results for field duplicates arise, the interpreter of the data must know how the samples were collected to fully evaluate the apparent imprecision in the results. The evaluator must know if the containers for both samples were filled simultaneously or sequentially. The manner in which a sample is collected may compromise the composition of the samples. This type of information should be documented in a field log and provided by the sampler.
A common misconception is that field duplicates provide an estimate of the laboratory=s ability to analyze field samples in a consistent manner. While the laboratory=s ability is a factor in producing comparable results from the analysis of field duplicates, the homogeneity of the duplicate samples submitted from the field is a major factor. If field duplicates are significantly different due to either the collection procedure or the homogeneity of the area, comparable results between the two samples may not be achieved regardless of the laboratory=s expertise.
Laboratory duplicates are sample aliquots taken from either the same or identical sample containers containing a field sample. The aliquots are treated in the same manner through the entire analytical procedure from preparation through instrumental measurement. In many cases, samples for laboratory duplicate analyses are provided in separate containers from the field. In effect, these are no different than field duplicates. When the laboratory takes aliquots from the same container, the aliquots should be collected in a manner to ensure that the aliquots are comparable to each other and representative of the total container contents. Collected appropriately, laboratory duplicates can provide a measure of the laboratory=s precision in performing the analytical method.
Although sometimes used synonymously with duplicates, replicates are usually a repeated operation occurring within an analytical procedure. Two or more measurements of the same prepared solution (digestate or extract) constitute replicate, not duplicate, analyses.
Co-located samples are samples which are taken from the same general area. Unless what constitutes Aco-located@ for a particular project is defined, the term is nebulous. Are samples considered co-located if they are within six inches? Twelve inches? Three feet? Is the sampler using Aco-located@ synonymously with Afield duplicate?@
The terminology used to describe duplicate samples is not critically important. What is important is the understanding of how the terms are used with respect to a particular project and how each of the duplicates was obtained.
Improper Practices in Calibrations (Cheating by Any Other Name?)
The purpose of calibrating an instrument is to determine the concentration of an analyte in a sample based on its response. Initial calibration (IC) of an instrument is established by analyzing different concentrations of an analyte and plotting the responses against the concentrations. Responses for unknown concentrations (samples) can then be calculated against this curve.
The IC must be established prior to the analyses of samples and must then be verified periodically thereafter by analyzing a continuing calibration (CC) standard. Method criteria for each IC and CC must be met before a laboratory can analyze samples. Since calibration can be a difficult and time-consuming process, attempts are made to keep the same IC in effect for as long as possible. Analysts have developed some creative ways to meet method criteria for calibration. While some of these practices are not specifically forbidden by the methods, they are technically wrong.
Running multiple analyses of the calibration standards and choosing the responses that provide the best curve is often done. A variation of this practice is to analyze more than five concentrations and choose those concentrations that provide the best curve. The problem is that the discarded data are as equally important as the data retained. If the discarded data do not fit the desired curve, they are telling you something about the accuracy and/or precision of your system. Dropping the highest or lowest concentration standard is technically reasonable as long as the calibration range is modified to match. However, discarding data from within the calibrated range provides a distorted view of the capability of the analytical system to measure accurately or precisely.
A creative way to verify that a calibration has not changed is to insert responses from a current CC standard into the IC that was analyzed days or weeks earlier. If a current CC standard gives a response that is out of criteria, its response is added to the IC data obtained some time before, and the criteria are recalculated. This has the effect of weighting the curve toward any changes occurring in the analytical system so that the calibration will appear to have been maintained for a longer period of time when, in fact, it was not.
By far the most creative and most improper means of establishing and maintaining the calibration is to simply fabricate numbers that fit the desired curve and use them to replace the true responses in the calibration (Adry labbing@).
Reviewing the raw data for a calibration curve is the only way to detect these types of creative processes. Sometimes, even then, they will not be detected.
Are your acceptance limits acceptable?
In the analyses of organic compounds by gas chromatography, surrogate compounds are spiked into every sample. The recoveries of these compounds are monitored to determine if the analytical process is "in control." If you request sample results without a full deliverables data package, you will receive sample results and quality control (QC) summaries but no raw data. The narrative may state that all of the QC results were within acceptance limits or QC criteria were met. With inherent variations in the analytical process, recoveries are rarely 100%, but they should be within acceptance limits. But what exactly are these Aacceptance limits?@
Acceptance limits define ranges into which recoveries of surrogates must fall in order for the analyses to "pass QC." If an analysis does not pass, the sample must be re-analyzed. SW846 methods use a range of 70-130% for default limits until the laboratory establishes their own limits based on the recoveries they normally achieve for spiked compounds. Acceptance limits of 0 - 200+% have been reported by some laboratories, which means that surrogate recoveries within this range would be considered as Apassing QC@ for that laboratory.
Matrix spikes and matrix spike duplicates are field samples which are spiked with target compounds as well as surrogates. The recoveries of those compounds are a measure of the ability of the laboratory to detect and quantify the materials of concern. In some cases, the laboratory=s lower limit of acceptance may be as low as zero percent. This means that a compound can be spiked into a sample matrix by the laboratory, and the laboratory can completely fail to detect the spiked compound. Yet, the result would still be within the laboratory=s QC acceptance limits.
For some methods, laboratory fortified blanks (LFBs) are prepared by spiking contaminants of concern into a clean matrix (de-ionized water or clean sand). These LFBs are used as laboratory control samples (LCS) to monitor system performance. Recoveries of compounds from an LFB or LCS are also compared against laboratory-established acceptance limits. If those limits range from 0-200 % and a 100 part per billion (ppb) spike was not recovered at all (0%), the result would still Apass QC@ and be considered acceptable by the laboratory. Conversely, a result of 200 ppb would also be considered acceptable.
What do the results obtained for surrogate recoveries, matrix spikes, and blank spikes say about the laboratory=s ability to measure contamination in your samples? If a compound is spiked at 100 ppb but cannot be recovered, the implication is that the compound cannot be measured reliably below 100 ppb. If a result of 200 ppb is obtained for a 100 ppb spike, sample results may also be exaggerated, and limits may appear to have been exceeded when, if fact, they were not. Limits of 50-100% may not satisfy the needs of the project although they may satisfy the acceptance criteria of the laboratory. It is up to the data user to determine how much error is acceptable.