DIY Oil Testing Questions
I. Questions, Motivations, & Best Practices
Public Lab’s Oil Testing Kit program has sket out to develop a low-cost, Do-It-Yourself kit for differentiating oil pollution, building on the DIY spectrometry kit Public Lab has designed and distributed since 2011.
There are many different questions which such a kit might attempt to answer, and it's important to address them carefully and individually. Asking questions is how we started Public Lab, and questions are at the heart of Public Lab's process of research and development. We've collected many common questions here -- many, but not all of which, we can now answer. Here, we have tried to order each group of questions with increasing difficulty or complexity.
Questions about finding pollution
Imagine you find (as many of us did) what looks like tar on the beach after an oil spill:
A tar ball, photo by @eustatic
How do I tell if it's oil or something else, like a piece of a tire or asphalt, or mud?
The technique we’re using, called steady-state ultraviolet fluorescence spectroscopy, has been shown to be able to distinguish between different weights of oil -- such as crude oil versus motor oil or diesel -- in a laboratory under certain conditions (e.g. Pantoja et al, 2011). The kit we’re developing also attempts to differentiate categories of oil, without laboratory facilities and instrumentation. To do this, there are many unknowns remaining; including:
- the effects of weathering
- false positives
There are also other materials which fluoresce, such as various types of organic matter. Fluorescence alone cannot be used as evidence of petroleum: rotting vegetation or even olive oil or beer will fluoresce, but because the spectrum of each will be different, the theory is that a spectrometer will help you tell plant matter apart from a petroleum sample. However, as of January 2016, this has yet to be clearly demonstrated in our community, and we hope that the refinement of the testing kit itself will make such tests easier to perform.
It’s important to note that, while different kinds of oils have been distinguished by a given user using their specific instrument, not every user has been able to do so, and the results have not always been consistent among different spectrometers. Thus, as of January 2016, we are still working to improve the reproducibility of our method, which is fundamental to being able to distinguish grades of oils, or oils from non-oil fluorescing materials (see Testing Your Hypothesis in Workshop 1).
How do I tell if it's oil from the spill in question?
Matching an individual source of oil, such as crude specifically from the Deepwater Horizon spill, is known as “spectral fingerprinting,” and according to the scientific literature, the technique we are using (“steady state ultraviolet fluorescence spectroscopy”) does not produce spectra that are unique and specific enough to distinguish samples from one source from another if they are very similar -- say, to distinguish between two different crude oils.
However, if there is only one source of crude oil in an area, and you wish to collect evidence that a sample is from that source, as opposed to being motor oil or diesel, or some other pollutant, then this is the test we are attempting to reproduce with a DIY kit. (See previous question)
What does oil pollution mean for my health?
This is a complex question. Oil pollution can have a variety of impacts on a person’s health, or the health of an ecosystem. In addition to birds and other wildlife getting covered by an oil slick which can result in hypothermia or suffocation, oil is comprised of several toxic compounds that can have severe consequences. Polycyclic aromatic hydrocarbons (PAHs) in oils can impair reproduction, physiological and neurological development in marine organisms, and some PAHs are likely to be carcinogenic in humans (e.g. Pelletier et al, 1995, Samanta et al, 2002). Other volatile components of oil may induce headaches, nausea, and respiratory difficulty during even short exposures (e.g. Solomon and Janssen, 2010). For more information about studies investigation human health effects of exposure to oil, please see this ATSDR reference and this NIOSH reference, both hosted by the national Centers for Disease Control.
Also see the Hazards section of the Oil Testing Kit page.
Oil residue on a beach following the BP Oil Spill; photo by Shannon Dosemagen (@shannon)
What do I do if I see oil?
Before handling any oil, make sure that you are equipped to do so safely. Oils contain toxic and carcinogenic components, and several volatile organic compounds, so be sure to wear gloves and stay in well-ventilated areas. See health effects, above.
Contact local response agencies
When you see what seem to be spilled oil or tar balls, the first people to contact will be the environmental response team for the county, parish, or municipality where you observe the suspected oil. In many counties, the response team will be part of the Environmental Engineering Department or Ecology Department. If you are unsure who to call, often a detailed Internet search (e.g. “King County Washington report oil leak") can lead you to the proper website and phone number. If an Internet search does not yield the necessary information, calling the county government’s main office may help you find the appropriate department to contact. When you do get in touch with the environmental response team, if possible, you will want to provide the following information:
- what types of environment are or might be impacted (e.g. crop land, river, ocean, etc)
- what does it look and smell like?
- take a picture
- can you see the source of the oil?
- does it appear as though more is spilling or leaking?
- can you estimate how much oil is spilled?
- is there a particular activity occurring that is causing the spill or impacting it?
- are there other hazardous materials there?
- is there risk of fire?
- include: photographer name, date, time, site location, site description, number the photo, and record this in a logbook or other place that could be referenced
If there is any immediate safety risk, such as potential for a fire, contact the local fire department and police department at once.
If the oil you observe is in marine waters, call the US Coast Guard National Response Center.
Collect a sample for laboratory analysis
In addition to calling and reporting the oil you observe to a local or national agency, if it is possible to do so safely, you could collect a sample of the oil. If you are interested in having the oil tested to determine the type (and potential source of the oil) by a laboratory, it can be helpful to decide which laboratory you will send the sample to before collecting that sample. To find an analytical laboratory, you can either call the county government environmental department and ask their recommendations (as they often need to out-source their lab work), or do a series of Internet searches to find an appropriate accredited lab in your area. You can also email the main discussion list at Public Lab, where thousands of community members from around the world may be able to help you find the appropriate resources. (Likewise, if you can help, please offer assistance to others who ask.)
When you find a laboratory to use, it will be important to learn the collection techniques that are required for the method they are planning to use. Since most methods require very specific bottle preparation (generally including a final step of baking glassware to 450 Celsius for four hours, since by that temperature most organic matter should have combusted), you might want to ask to see if they will send you an appropriate collection container. Note that improper pre-cleaning of sampling containers might be the primary reason that many citizen-collected samples are thrown away by official labs, so cleaning bottles according to specific methods, or obtaining cleaned bottles from laboratories, is important. Different methods may require specific sample collection techniques, but here we include some general sample collection principles to help you prepare.
For samples that you want to analyze for organic compounds, such as oils, you generally want to use amber glass vials -- note that you do NOT want to use plastic bottles (see storage, below), as those will contaminate the sample, and it is best to use a UV filtering glass (such as most amber glass vials) because several compounds in oils are sensitive to UV. If you do not have amber glass bottles, plain glass, such as a mason jar, will work if you keep the jars in a dark box out of the sunlight -- but be aware that many official guidelines require amber glass, so if you're planning to submit a sample to a lab, this is important.
If you are collecting straight oil or oily solids (e.g. sand, soil), the collection container should be large enough for a minimum of 5 grams of sample plus 10 mL water, which will likely be a 40 mL vial, or larger. The analytical method the laboratory will use will indicate whether a PTFE-lined solid cap is sufficient or if you need a cap with a septa (used when the sample vial cannot be opened, and is directly pierced by the analytical instrument).
When collecting solid oil or oily solids samples, you want to avoid using any plastic scoops, and rather, use a stainless steel scoop or spatula that you have rinsed with water and wiped with isopropyl alcohol. When collecting water samples with an oil sheen or emulsion, you want to collect the sample upstream from any disturbance you have caused, with the bottle open to the current in a river, and fill the sample container completely such that there is no headspace. (see collecting sheens, below)
Samples should be stored cold, on ice, attempting to reach 4 Celsius (39 Fahrenheit), or colder if the analytical methods requires it. For samples that must be stored colder than 4 Celsius, you can purchase dry ice. If you purchase dry ice, be careful when handling it, as it can burn bare skin and be a respiratory irritant. Samples should be transported from the field in a cooler and transferred into a refrigerator or freezer of the appropriate temperature as soon as possible.
Suspected oil samples, kept in a cooler; photo by Scott Eustis (@eustatic)
When collecting a sample for third-party laboratory analysis, to submit to a local government agency, or to analyze yourself, it is important to establish a Chain of Custody of the sample to ensure there is a responsible party for it. Information to include in the Chain of Custody are:
- Name and contact information of sample collector
- Sample code (if applicable)
- Date and time of collection
- Location of collection
- Type of media (e.g. soil, sediment, water)
- Environmental conditions
- Collection method used
- How it is stored or transported
- Type of preservation (if applicable)
- Analyses requested
Analyze your own sample
In addition to collecting oil samples to provide to local agencies or independent laboratories, you could collect samples and scan them using a Public Lab DIY spectrometer and oil testing kit it to see if your samples resemble known types of oil (such as crude vs. motor oil). To do this, follow [instructions in Workshops 2, 3, and 4(/wiki/oil-testing-workshops).
Discerning if your samples are oil, and potentially what kind of oil they are, can be useful for a variety of reasons (please see Section IV: Data and Action. One reason is to help local government officials to respond to the oil pollution through action channels more quickly, and potentially contain or mitigate some of the environmental impacts of the spill or leak. While analyzing your own samples will not lead to diagnostic evidence (see What kind of oil testing data is useful in talking to regulators? To lawyers?), it can demonstrate a likelihood that a sample is or is not oil, and be an excellent visual component of community education or advocacy.
Scanning suspected oil samples; photo by Scott Eustis (@eustatic)
There are questions about different types of tests related to oil concentration or identification:
What tests exist, and what are available to me?
Here we’ll discuss a few different tests described in the scientific literature, as well as used in lab analysis of field samples -- but they are not accessible or affordable to people outside those spheres. These descriptions involve some technical language, which we have avoided in other parts of this document; however, part of the reason for Public Lab’s Oil Testing Kit program is to address this very issue, and we feel it’s important to discuss the status quo of oil testing even if it’s not easy to do so for non-scientists.
Currently the most common methods used for “environmental forensics” related to oils involve the use of gas chromatography in tandem with mass spectrometry (GC/MS). Gas chromatography is a method that separates volatile (gaseous) compounds based on certain characteristics, such as their vapor pressure and hydrophobicity, causing different compounds to “elute”, or come out of the gas chromatography column, at different times. As the different molecules elute, they then enter a mass spectrometer, which can determine the masses of the molecules that entered (actually mass to charge ratios). By pairing these two techniques, it’s possible for a laboratory to identify what kinds of molecules they are. GC/MS is used for identification of a lot of the individual components of oils, such as alkanes, sterane, triterpane, and other polycyclic aromatic hydrocarbons (PAHs). Molecules like sterane and triterpane are often called “biomarkers” because they are molecules derived from the living organisms that created the original source rock and its oil. GC/MS is generally conducted after extensive sample preparation using organic solvents, and requires a laboratory setting, but the results are widely recognized. At Louisiana State University, Dr. Ed Overton has used a combination of qualitative and statistical techniques to discern oil sources from GC/MS spectra (e.g. Henry et al, 1993).
Gas chromatography followed by flame ionization detection (GC/FID) is also used to analyze organic molecules, such as components of oils. GC/FID also starts with gas chromatography (described in the preceding paragraph), and the molecules elute to meet a hydrogen flame, which can pyrolyze the molecules to form ions, and create a current running between two electrodes. The current measured is related to the concentration of the molecules eluted. Like GC/MS, GC/FID requires considerable sample preparations involving hazardous chemicals and is performed in laboratory settings.
Fluorescence spectroscopy is also commonly used to analyze oils and other organic compounds with conjugated carbon ring structures. Spectroscopy in general involves observing the relationship between matter and light. Fluorescence spectroscopy involves irradiating a molecule with light that has high enough energy to excite an electron to another orbital, and then observing the light that is emitted when that electron relaxes (fluoresces) back to its normal state. Many organic molecules are excited by ultraviolet light (~200-400 nm), and then fluoresce in the visible range (~400-700 nm). When using a single excitation energy, emissions peaks can be broad and sometimes difficult to distinguish different organic compounds. More recently, people have used excitation-emission matrices -- using a variety of excitation energies and observing the emissions energies and intensities for each -- adding a dimension to assist in molecule identification. Fluorescence spectroscopy usually requires less sample preparation, is not destructive to the sample during analysis, and is generally less expensive than GC/MS or GC/FID.
Infrared spectroscopy can also be used to identify features of oil, helping to determine oil categories and weathering or wear. Infrared spectroscopy involves irradiating molecules with energy in the infrared part of the light spectrum (typically outside the ~400-800 nanometer range of our device, and often around 1600 nanometers), which is lower energy than ultraviolet or visible light, and observing the energies that are absorbed by the molecules. The energies that are absorbed correspond to rotational and vibrational energies of chemical bonds, which relate to the bond strength and atoms present. This technique allows you to discern different “functional groups” which are the most reactive regions of a molecule, and are different in refined versus crude products, and impacted by weathering. Like fluorescence spectroscopy, infrared spectroscopy is less expensive and less difficult than the GC/MS or GC/FID methods.
How much do different tests cost?
According to the Surfrider group, which sent many oil samples to different labs over the course of the BP oil spill response, “When we were doing our oil study, we ran batches of ten samples using the 8272 modified solids GCMS method @ $295/sample so, $2995 a set. This was with the understanding that we would be running a lot of samples, so may not reflect the usual pricing.” This is how the per-sample cost broke down from the lab analysis for a lab contracted to run samples in batches:
- $340 for PAHs including alkylated homologues.
- $1000 for petroleum biomarkers.
- $160 for high resolution GC/FID.
- $2 for sample disposal.
- $700 for data interpretation and reporting.
Total cost per sample would be $2202.00
Costs for oil testing vary widely, to as much as $2000 a sample for dispersant testing (according to Surfrider) and as as much as $10,000 for testing with PACE labs according to Scott Eustis (@eustatic) of the Gulf Restoration Network.
What can each test tell me?
Gas chromatography - Mass spectrometry (GC/MS), described in the question What tests exist, and what are available to me, can elucidate the different organic molecules present in a sample. Oils are mixtures of many different kinds of organic molecules, originating from when the oil was actually living organisms millions of years ago. Petroleum biomarkers are molecules that are generally not as volatile or reactive as many other molecules, and can provide a record of the type of oil and its source rock. However, even biomarkers are prone to react during weathering processes, so the European Committee for Standardization (CEN) promotes the use of “diagnostic ratios” of biomarkers rather than pure concentrations of certain biomarkers. Technical Reports for these tests can be found here. Diagnostic ratios of petroleum biomarkers are the best “fingerprinting” of oil types and sources that are currently available, and at best, can distinguish oils from different wells.
Certain organic molecules can be identified using GC/FID (described in a previous question), but not as selectively as GC/MS biomarker or PAHs analysis.
The downsides of any of the GC methods are the financial costs, length of time for sample preparation and analysis, and the highly skilled nature of the work requiring professional training and experience. Spectroscopy is much less expensive, less time-consuming, and has lower skill-thresholds necessary for operation. Some commercial fluorescence and infrared spectrometers can even be taken into the field for on-site analysis.
Fluorescence spectroscopy is useful for distinguishing oils by grade or API weight, but not fingerprinting an oil source explicitly -- so, not for distinguishing between crude oils from different wells, for example. Fluorescence spectroscopy also can be impacted by weathering and sample preparation or dilution, which is discussed in some following questions. Infrared spectroscopy is useful for identifying the functional groups of various molecules within the oil sample, which can be indicative of both oil type and the extent of weathering or wear it has undergone, but is generally not used to identify a specific source of oil.
Some remote hyperspectral imaging, using visible and near-infrared wavelengths, can be useful in rapidly assessing the extent of an oil spill from satellite or aircraft. However, these techniques cannot be used to discern the type or source of oil.
Preparing samples of crude oil and diesel for DIY testing; photo by Cindy Regalado (@Cindy_ExCites)
What can Public Lab's tests tell me, and with what confidence?
As of January 2016, Public Lab’s Oil Testing Kit is a prototype undergoing further development; we are working to assess data it produces and both improve the kit and methods, as well as evaluate their present utility for analyzing samples.
The tests currently being conducted are with known samples of motor and crude oil, as well as diesel. We asked members of the Oil Testing Kit Beta Program to scan five known, labeled samples of different types of oils, and one unlabeled sample which contains the same kind of oil as one of the five knowns. This was, in effect, a test of the test, in multiple dimensions -- to see if people could assemble the kit, use the software, differentiate spectra of the known samples, and successfully identify the unlabeled sample. The data quality of spectra differed among users, likely due to construction, optical alignment, and incident light intensity differences. Likely due to these factors, the spectra of known samples were substantially different among users, such that inter-user comparisons are currently not sufficient for categorizing an unknown oil sample. However, we focused on tests on a single device; for this reason we shipped reference samples of oil to each tester so that they could be scanned on the same device.
We are improving the hardware design to make inter-user results more consistent, but our primary goal is to pilot testing against references on a single instrument. That is, we hope our test will enable people to tell whether an unknown sample is crude, motor oil, or another type of oil. For the above tests, spectra obtained by individual users on a single spectrometer were consistent and reproducible for many users, and for those who correctly assembled and calibrated their devices, the spectra of at least crude oil and diesel were distinguishable, and were in some cases distinguishable from intermediate grades. The spectra of our “intermediate grades” (5W30, 20W50, and 80W90) were not substantially different enough to be considered distinguishable.
Public Lab fellow Yagiz Sutcu (@ygstcu) will be publishing an in-depth report of these results in coming weeks. (January 2016)
User tests have demonstrated a substantial effect of oil concentration on their fluorescence spectra, larger than would be expected based on published literature. The observed differences between different concentrations of the same oil (diluted in mineral oil) are actually larger than the observed differences between different types of undiluted oils, and so it is very important to know the concentration of oil in the sample if you want to discern what type of oil it is. This may be difficult with field-collected oil samples that may be mixed with substrates such as sand or soil, and is an area of active research in the Public Lab community. As of January 2016, we are exploring the possibility of equalizing opacity of samples in order to correct for concentration. For more on the effects of dilution, see https://publiclab.org/tag/dilution.
As of winter 2016, fluorescence tests of unknown samples have begun as well. With the understanding that sample concentration will have an impact on fluorescence spectra, the current “unknown” tests are simply for presence or absence of fluorescent signal. Unknown samples that do fluoresce will be compared with a series of known oil samples diluted to approximately the same concentration as the unknown sample. Since the unknown sample is likely to have sand or soil agglomerated with it, simple sample weight to solution volume ratios will not be sufficient to accurately calculate true sample concentration, so these comparisons to known samples of approximately the same concentration are not very robust, but are a useful step toward classification of an unknown sample.
For the most recent information on this work, see https://publiclab.org/tag/oil-testing-kit
What kind of oil testing data is useful in talking to regulators? To lawyers?
The federal Water Quality Recommended Guidelines state that waters should be “virtually free from oil and grease”, so county or municipality officials should respond to reports of oil sheens or oil dumping based on citizen reports. Photographic evidence enhances the citizen reports, particularly because the situation could change between the time the report was made and when the officials are able to respond. The more data that is provided, including data obtained using a credible DIY oil testing kit, could potentially expedite the local response or other actions taken. However, note that the process of response does not necessarily include identification of a source or a responsible party. In situations where an individual or group seek a financial penalty on a company for their environmental pollution, that group may need to sue the company under the Clean Water Act.
For marine oil spills, the majority of reporting is visual, often photographic or video from aircrafts (Fingas and Brown, 2007). However, visual techniques are only appropriate for documenting or reporting the existence of an oil spill, but not for determining the type or source of the oil. Since 1972 the U.S. Coast Guard (USCG) has been tasked with developing and implementing techniques for identifying oils and hazardous substances, and the USCG has historically used a combination of spectroscopy and thin-layer chromatography (TLC) in the field and gas chromatography in the laboratory if necessary (Clow, 1977). According to the 2009 American Petroleum Institute Report, only the largest ~10% of all spills are investigated (see the report.
When the USCG or state or local authorities do investigate oil spills, it is common to use ASTM methods D3328 or D5739, which are qualitative methods that use GC/FID and GC/MS respectively (see what kinds of tests, above), and compare samples with oil standards to observe similarities and differences. The methods are widely acknowledged as subjective, however, so there has been pressure to use more quantitative methods, such as using petroleum biomarkers. The European Committee for Standardization (CEN) supports the use of diagnostic ratios of petroleum biomarkers for oil spill investigations, and in the U.S., petroleum biomarker analyses by GC/MS in an accredited lab are used for large spill source identification investigations. However, in smaller cases where GC/MS analyses have not been conducted, compelling visual evidence demonstrating the leaking or spilling source, may be utilized.
Has such data has been used to effect change? How, when, and where?
Petroleum biomarkers and diagnostic biomarker ratios (see what kinds of tests, above) have been used frequently since the Exxon Valdez Oil Spill in Prince William Sound (PWS) in 1989. While much of the oil pollution in PWS was due to the massive Exxon oil spill, petroleum biomarkers were used to evaluate tarballs along the shorelines, and researchers discovered that many of the tarballs actually originated from a source in California, demonstrating the persistent and long-reaching effects of marine oil spills (Wang et al, 2007 and references therein). Petroleum biomarkers were used while investigating a destructive building fire, and it was discovered that the building had caught fire due to ignited leaked bunker oil, which also leaked into the nearby river (Wang et al, 1999). It’s important to note that petroleum biomarkers and biomarker ratios can be indicative of a correlation but do not necessarily provide definitive proof of an oil source; to provide more compelling proof of a source, multiple types of analyses may be necessary.
Fluorescence spectroscopy, on which we have been basing Public Lab’s prototype Oil Testing Kit, has been used to investigate oil spills and their sources as well. After the Hebei Spirit Oil Spill of 2007, several methods were used to monitor the oil spill cleanup effectiveness, including a portable fluorimeter for rapid analysis of oil in pore waters (water between sediment grains). This allowed researchers to monitor oil concentrations in a large geographic spill area with substantial measurement density over several months and monitor oil attenuation (Yim et al, 2012).
Recently, fast-scanning synchronous fluorescence spectroscopy (see What can each test tell me?, above) has been recommended as a screening tool for oil spill source identification in the Turkish Straits, through which more than 55,000 ships travel each year, where fluorescence spectra are rapidly obtained, compared with suspected sources, and any probable matches are then evaluated by laboratory techniques (Karakoc et al, 2015). This fast-scanning screening can make illegal oil discharge or spill investigations more rapid and efficient.
Visual photographic evidence has been used to prompt investigations for decades, and is still the most common evidence included in reports of oil spills (Fingas and Brown, 2007 and references therein).
Set of scans of crude oil by @eustatic; https://spectralworkbench.org/sets/2928
Can results from the Public Lab Oil Testing Kit be used as evidence?
As of January 2016, there remained too many unknowns -- including the effects of dilution, weathering/aging, and device construction variability -- for our DIY data to be considered persuasive enough to be used as evidence yet. That is, spectra obtained using your DIY spectrometer can be informative, but may not be persuasive to your intended audience. For example, regulators will probably not recognize data from a DIY kit or other un-approved methods as diagnostic (see glossary). However, these low-cost methods can provide useful, if not diagnostic, information. With the Public Lab Oil Testing Kit, observing fluorescence which resembles that of a reference sample of known petroleum -- is a probable indicator of petroleum compounds, and can be useful for motivating communities speaking out against oil pollution, informing neighbors about potential hazards in their environment, or prompting further investigation by enforcement or remediation agencies.
Efforts to prompt further official investigation can be greatly enhanced if you can demonstrate a correlation between your results and results of officially recognized tests. If official tests are conducted, try to collect samples at the same time and from the same location, and compare your data with the official results (see co-location, in next question). If you can speak with the local or state agency who is responding to the spill, they may be willing to collect a split sample for you (i.e. take a sub-sample of their sample to provide to you), especially since they will already be going out to collect samples. If any real-time data is collected in the field by officials, try to collect data simultaneously for a co-located comparison test. Read about co-located tests below. Results from these sorts of comparison tests can better inform you and the Public Lab community about the capabilities and limitations of the Oil Testing Kit, and positive results can facilitate greater impact and recognition of your unofficial tests.
What are the advantages of co-locating tests?
By conducting tests on samples collected at the same place and time, or especially on subsamples of the same “split” sample as officially recognized tests, it’s possible to evaluate if a relationship exists between your tests and those which carry much stronger legal and regulatory weight. Of course, if your own tests do not agree with the official tests, it may be difficult to establish credibility for your own testing -- but if they do, this correlation can provide added credibility for your testing.
If a correlation is present between official results and your unofficial results, it may be possible to improve the spatial coverage of testing. That is, if the official tests only cover 10 spots over a 10 mile-long riverbank, but those ten tests have high agreement with a lower-cost, less-recognized test, it may be possible to use the latter test on 100 spots over the same riverbank, and “fill in” data between the official tests, for a more detailed look at pollutant distribution. This approach of filling in data gaps (also called “Data Fusion Approach”) is part of the plan for US Environmental Protection Agency’s (EPA) Next Generation air monitoring programs, likely to be followed by other Next Generation monitoring of land and water contamination.
It is important to note that the usefulness of data from co-location tests, as with any data, will depend on how persuasive the data is to the intended audience. Any interpretation of data must be rooted from the data quality -- the accuracy, precision, resolution, and reproducibility of the tests (see Testing Your Hypothesis in Workshop 1). Beyond that, effective communication of those results is important. For technical audiences, communicating the data quality on a mathematical basis and demonstrating proper statistical analyses will be useful to create a platform for transparent and earnest discussions. For audiences with less technical backgrounds, while data interpretations will, of course, still be rooted in data quality, communication of results may be more effective with more visual demonstration and outlining of analogous situations.
For all audiences, very direct communication of data limitations is as important as assertions of data capabilities. This can often be a difficult balance to achieve, but if you start by clearly explaining the objectives of the test (e.g. to distinguish between different classes of oils rather than an explicit identification of an oil spill source), then explaining the scope of your results follows suit.
There are questions about how to collect and store the samples. Also see Spectrometry Sampling.
How much do I need to collect? Solid vs. liquid?
As of January 2016, there is work being done by Public Lab Fellow Matej Vakula to collect useable sample volumes from sources such as sheens on the surface of water (see sheens), and oils soaked into absorbent plastic pads (see this research note). These techniques aside, we currently recommend 2-10 drops of pure oil diluted in a 2.5 milliliter sample container filled with pure mineral oil (which does not fluoresce). If the samples are very transparent, dilution may not be necessary. As of January 2016, there are remaining questions about the effects of dilution on tests; see dilution below.
For solid samples, it depends on the concentration; dissolving the solid samples in a small amount of mineral oil and allowing dirt and sand to settle out can result in a pale yellow solution which can be tested; see dilution, below.
Can I touch it, and with what?
Do not touch suspected pollutants with bare hands; use gloves and other protective gear, and do not leave them unsealed except in well-ventilated areas. Oil pollutants contain volatile components which can be harmful to breathe, and carcinogenic substances which are dangerous to touch.
What does oil pollution look like?
This depends on where and in what state you find it. Oil may be a liquid sheen on the surface of water, a dried, dirt-like residue in brown, black, or orange, or it may be dissolved in water, and not easy to see. The photos below show a few examples of oil found after a spill:
Tar ball on beach after BP Oil Spill; photo CC-BY Louisana Bucket Brigade
Oil residue on rocks after the BP Oil Spill; photo CC-BY-NC-SA Cesar Harada (@cesarharada)
Can it be mixed with sand, dirt, vegetation, etc?
Samples collected in the field can often be mixed with grit, sand, vegetation, etc. If you’re dissolving a solid sample, like a piece of tar, in mineral oil or another solvent, these may settle out over a few minutes or hours, and so it can be helpful to dissolve in one container, then eye-drop the transparent solution off the top into another container for scanning.
Vegetation or plant or organic matter (like peaty soils) will fluoresce, and so should be kept out of oil samples (see false positives.
What happens if I dilute my sample or don’t know my sample concentration?
As of January 2016, there is ongoing research in our community on whether and by how much dilution affects the color of fluorescence (and specifically the position of the highest point in the spectral graph), which would potentially affect the results of the Oil Testing Kit. If dilution proves problematic, we may be able to correct for this effect using absorption spectroscopy to equalize the opacity of samples. Read the latest on this research at https://publiclab.org/tag/dilution
Dilution may be necessary, however, to get a spectrum from some very dark samples, like concentrated crude oil, due to issues like quenching. See quenching, below.
What is fluorescence quenching?
Quenching is essentially the dimming of emitted light from a sample as the light passes through the sample itself to escape the sample container -- especially if your sample is too concentrated. When we illuminate an oil sample with ultraviolet light, we provide energy that excites an electron into another electron orbital. Fluorescence is the light (energy) that is emitted as the electron relaxes down to its normal or “ground” state. Quenching is the loss of energy to another molecule, rather than energy being lost as a visible light emission. So, when quenching occurs, we see less fluorescence.
Quenching can occur through several mechanisms. A common quenching mechanism involves the loss of energy through collision with another molecule. Quenching is quite common in solutions containing oxidizing agents or electronegative elements (such as halogens), which can gain electrons into their molecular or atomic orbitals. Another type of quenching commonly observed is “Forster Resonance Energy Transfer” (FRET), in which the energy of one molecule’s fluorescence excites another molecule (the emission energy of one molecule is the excitation energy for another molecule in the sample). FRET broadens and lowers the intensities of the observed fluorescence spectra. An observed reduction in fluorescence intensity (though not technically fluorescence quenching) can also be due to molecular shielding and steric effects hindering the excitation of molecules in the sample. All of these types of fluorescence quenching are more pronounced in more concentrated solutions, and can be be mitigated through sample dilution.
It is important to note that all fluorescence involves the loss of energy, which is why fluorescence emissions are at longer wavelengths than their excitation energies. This phenomenon is known as Stokes shift, and is simply the loss of vibrational energy as the excited electron settles from its maximum absorption excited state to a lower energy excited state prior to fluorescing back to its ground state. Jablonski diagrams offer a visual representation of fluorescence, including vibrational energy loss and potential energy transfers.
Could there be a “false positive” in my oil fluorescence test?
Although various polluting oils fluoresce, one reason we can’t rely on fluorescence in itself as an indication of oil pollution is that several organic compounds fluoresce in the UV-Visible range. Generally, most compounds with carbon ring structures can be excited by UV light and fluoresce in the visible spectrum. Common materials that fluoresce, and may fluoresce with a similar fluorescence peak shape and energy include, but are not limited to:
- Other oil types
- Vegetation - humic acid
- Organic matter - beer
- Amino acids, bodily fluids like urine -- crime scene
- Vitamin E in mineral oil
- Containers - gel capsules, for instance
Spectrometers (including our DIY Oil Testing Kit) are usually set up to scan samples placed in small containers called cuvettes, which can be made of different types of plastic or glass, depending on what you want to store in them, and what kind of light you want to shine through them.
Ultra micro cuvettes; photo by Jeff Warren (@warren)
In the lab, cuvettes are considered temporary, and often disposable, especially if plastic. See storage, below, for how to keep samples long-term. Some problems we’ve encountered with cuvettes are that some plastic (polystyrene, a common cuvette material) will dissolve if used to store more corrosive samples like diesel, and that cuvettes with square lids will leak if not kept constantly upright.
A good sample container has flat sides, so you can shine light (or a laser) through it without lots of reflections, which could go into your spectrometer and affect your data. It's also good to have the light travel through a consistent amount of the sample -- many cuvettes (traditional spectrometry sample containers) are 1cm x 1cm, so the light always goes through 1cm of the sample. Some cuvettes are designed for smaller amounts of liquid, such as “micro” or “ultra micro” cuvettes -- these taper towards the bottom, so you don’t need as much sample material to scan. This also means you can shine light through less sample material; it’s thinner. For fluorescence tests, like in the Oil Testing Kit, quenching (see above) can happen in a concentrated solution, especially with a long light path encountering more molecules. Thus, making the light path through the sample shorter, some quenching and molecular shielding is lessened, yielding better results.
Several cuvettes in a row, demonstrating the way light should pass through them; photo by Jeff Warren (@warren)
The different plastics cuvettes are made of are often designed to be super transparent in the intended type of light, as well as to be resistant to corrosion or dissolving from the contents of the sample. This email thread discusses materials a bit, and links to the chart below by Brand (http://www.brandtech.com/cuvette_comp.asp) which shows what types of cuvettes are best for which types of samples, as well as which types of light.
Some cuvettes also have surface treatments like a textured or slightly opaque “diffuser” surface, or a fluted surface, on some sides to affect how light enters or exits the cuvette. For the Oil Testing Kit, we’ve found that if you’re using a laser, you should shine it through the completely clear sides, but that diffusers on the surface facing the spectrometer may have fluting or diffusers; as long as enough light enters the spectrometer, it’s OK.
How should I store samples?
Samples for organic analyses, such as oils, should be stored in glass since their components may stick to the walls of plastic bottles or entrain some plastics components into the oil samples. Oils are comprised of lots of different organic molecules, many of which are photo-reactive, meaning they react and can change chemical or physical form when exposed to sunlight. Thus, keeping oil samples in dark containers, such as amber glass is best. In order to minimize photoactivity, samples should be stored in a dark space as well. Several components of oil are volatile, and to minimize loss of those volatile organic compounds (VOCs), samples should be kept cool -- at 4 Celsius (39 Fahrenheit). If you are planning to have an oil sample analyzed by a laboratory, be sure to find out what kind of sample storage requirements they have prior to analysis, especially because often samples are only acceptable for a couple of days without risk of too much volatilization or oxidation, changing the sample chemistry.
Team members of the Gulf Restoration Network who responded to the 2010 BP Oil Spill collected samples of oiled fish and oiled soil where:
- Oiled fish in the field were wrapped in aluminum foil, placed in plastic bags, placed in coolers and overnight-mailed in cooler with dry ice.
- Soil was grabbed with EPA brown glass jars, placed in plastic bags, placed in coolers and overnight-mailed in coolers with or without dry ice.
Please see the EPA’s technical manual on Methods for Collection, Storage, and Manipulation of Sediments for Chemical and Toxicological Analyses for more information. For a quick list of typical storage requirements for different types of samples can be found here, but be sure to check specific requirements your lab may have.
Interestingly, oils for use (such as motor oil) are typically distributed and stored in plastic containers because they are less expensive and less breakable. In the Public Lab Beta Oil Testing Kit, known oil samples were mailing to beta testing program participants in plastic bottles for these reasons, as well as because the suppliers of these reference samples shipped them in plastic bottles in the first place. We have not observed impacts on oil spectra (obtained from Public Lab spectrometers) from being stored in plastic bottles, but we have not analyzed samples consistently over several months yet. We may need to include glass bottles in future reference sample shipments.
Samples in glass jars and transferred into cuvettes for scanning; photo by Scott Eustis (@eustatic)
How long will samples keep?
There is an ongoing discussion among people interested in oil analyses about how long oil samples will retain their chemical (and, importantly, fluorescent) properties without deterioration or contamination. Most analytical procedures require analysis within a few days of sample collection to ensure sample integrity (see this quick reference, but if stored in cold and dark conditions in sealed containers, oil samples may retain their integrity longer. Organic matter, oxygen, and bacteria may all deteriorate components of oil though, so long storage should be avoided. If you need to store your samples for longer periods of time, deep freezing can preserve their chemical integrity longer.
As discussed in the previous question, samples should be stored in glass, UV-blocking (generally amber) containers, tightly sealed (and ideally without headspace), in the cold (generally 4 Celsius). This is to inhibit, or at least limit, photooxidation, bacterial degradation, and volatilization. Samples that are properly stored can be stored for longer periods of time.
Environmental samples are, of course, exposed to natural conditions including sunlight and air, in addition to water, sediment, or soil mixing. Weathering impacts oil fluorescence spectra and their biomarker or other component analyses, and elucidating specific impacts of weathering on these analyses is an active field of study.
How do I prove where and when I collected it?
In environmental data collection, the “chain of custody” refers to the explicit documentation of the condition of the samples and the person(s) responsible for them, from collection through transport and storage, and finally analysis. To start the Chain of Custody, document:
- Name and contact information of sample collector
- Sample code (if applicable)
- Date and time of collection
- Location of collection
- Type of media (e.g. soil, sediment, water)
- Environmental conditions
- Collection method used
- How it is stored or transported
- Type of preservation (if applicable)
- Analyses requested,
If possible, take a photo with a cellular phone to obtain an official time-stamp and GPS coordinates. It is also useful to take supplemental photos featuring distinguishable landmarks.
There are also questions about the specific site you're concerned with:
How could data I collect relate to existing data?
There are other similar situations where some kind of oil testing would be helpful, and we think it's important to address these independently:
Can I test sheens from the surface of water?
Sheens form from the distribution of oil across the surface of water -- oil being lighter than water, it tends to spread out and can appear brown or orange, or even black if thick. But if it’s very spread out, it often forms a rainbow colored, iridescent sheen, which can be so thin that scooping up a small amount can yield only extremely small amounts of oil.
As a result, Public Lab fellow Matej Vakula (@matej) has been investigating Do-It-Yourself techniques for collecting, concentrating, and analyzing surface sheens over the winter of 2015-16. You can read about his work here: https://publiclab.org/tag/sheen; as of January 2016, he has been exploring the use of plastic that preferentially absorbs oil, as well as the freezing of oil/water mixtures, in an attempt to separate water and oil, and to recover enough to be scanned in an Oil Testing Kit.
Oil sheen in the Chandeleur Islands after the BP Oil Spill
Can I tell how old oil is?
Weathering, especially exposure to air, sunlight, and microbes, can affect the composition of oil samples. Many of the components of oils are photo-reactive, meaning they will undergo chemical reactions when exposed to sunlight, and many organic molecules are volatile and will evade from the samples, particularly under warm or hot conditions. Microbes may be able to metabolize certain compounds in oil, again affecting the overall chemical composition of oils. While it is well known that weathering affects oil compounds in these ways, quantifying the extent to which an oil has “weathered” is challenging.
In order to navigate these complications, the European Committee for Standardization (CEN) suggests using petroleum biomarker ratios (see what kinds of tests, above )rather than the concentration of any one biomarker to mitigate the impacts of weathering on “environmental forensics” of oil samples. For methods such as spectroscopy that do not quantify biomarkers, it is challenging to determine the age of an oil sample or the expected impacts of weathering on its fluorescent spectra. We need more research to address this question and how best to navigate the related issues of impacts on spectra. One approach would be to repeatedly analyze a single sample as it ages in a natural, but controlled environment, and document the changes observed over time.
In general, we've attempted to answer questions with a lower threshold first, and many of the above questions remain unanswered or partially unanswered at the time of writing. We continue to address outstanding questions with our research, development, and application, and if you feel you can contribute, please feel free to improve this document and reach out on the Public Lab discussion lists.
Do-It-Yourself; a tool, technique, or method that a person can do without special facilities, highly technical equipment, or formal expertise.
A split sample is when someone who is collecting a sample agrees to take a sub-sample of their sample to provide to you. As both are collected at the same time, they are useful for comparing the results of two different tests.
Also known as fluorescence spectroscopy; a type of spectrometry where samples are illuminated by a light, typically ultraviolet, which causes them to emit (or fluoresce) their own light. This is then measured using a spectrometer for comparison with other samples.
Polycyclic Aromatic Hydrocarbons, a set of carcinogenic compounds found in petroleum pollution.
an analysis technique which examines and compares the intensities of different colors from samples, typically as illuminated by a light source. See fluorescence spectrometry for the type we have been using in Public Lab’s Oil Testing Kit
Volatile Organic Compound; a group of compounds made of carbon, hydrogen, oxygen, and/or nitrogen that have low boiling points, high vapor pressure, and can often vaporize around room temperature.
Gas Chromatography/Mass Spectroscopy; see What tests exist, and what are available to me?
Molecules like sterane and triterpane are often called “petroleum biomarkers” because they are molecules derived from the living organisms that created the original source rock and its oil. Read more in Has such data has been used to effect change? How, when, and where?
A colloquialism in scientific studies is to call something that can uniquely identify a substance a "fingerprint" since human fingerprints are unique identifiers for individual people. In oil testing, "fingerprinting" means identifying the source or specific type of oil it is, based on chemical properties (such as ratios of biomarkers) that are unique to only that source or type.
Also known as API gravity, the American Petroleum Institute (API) grades oil based on their densities. The less dense an oil is, the higher its API weight. If an oil has an API above 10, it will generally float on water; if it has an API below 10, it will sink. Most petroleum products have API gravity of 10-70, but Canadian tar sands oils are more dense than water, with API gravity of ~8. For more information, please see this page.
A diagnostic test is a test or series of tests that can yield enough information to provide an explanation or answer with a high degree of certainty. In the context of oil testing, a diagnostic test generally refers to a test that can indicate a specific source of oil, such as the specific geologic formation from which that oil came. Diagnostic tests are used to investigate oil pollution without a clear source, in order to identify the source of the leak or spill or illegal discharge.
A range of related scientific literature has been collected on this page; here we list those which were cited in the Questions section above.
Clow JC. 1977. The Coast Guard’s Forensic Oil Identification System. International Oil Spill Conference Proceedings: 1977(1): 189-191.
Fingas M, Brown C. 2007. Oil Spill Remote Sensing: A Forensic Approach. In: Wang Z, Stout S, editors. 2007. Oil Spill Environmental Forensics: Fingerprinting and Source Identification. 1st Ed. Elsevier. 419-447.
Gomez-Carracedo MP, Andrade-Garda JM, Prada-Rodriguez D. 2015. Differentiation of weathered oils using infrared indexes and self-organizing maps. Fuel. 158: 57-65.
Hansen AB, Daling PS, Faksness LG, Sorheim KR, Kienhuis P, Duus R. 2007. Emerging CEN Methodology for Oil Spill Identification. In: Wang Z, Stout S, editors. 2007. Oil Spill Environmental Forensics: Fingerprinting and Source Identification. 1st Ed. Elsevier. 229-256.
Henry CB, Roberts PO, Overton EB. 1993. Characterization of Chronic sources and Impacts of Tar Along the Louisiana Coast. US Dept of the Interior, Minerals Management Service, Gulf of Mexico OCS Regional Office, New Orleans, LA. OCS Study MMS 93-0046, 64.
Karakoç FT, Atabay H, Tolun L, Kuzyaka E. 2015. Fast scanning of illegal oil discharges for forensic identification: a case study of Turkish coasts. Environmental Monitoring & Assessment. 187: 211.
Pantoja PA, Lopez-Gejo J, Le Roux GAC, Quina FH, Nascimento CAO. 2011. Prediction of Crude Oil Properties and Chemical Composition by Means of Steady-State and Time-Resolved Fluorescence. Energy & Fuels. 25(8): 3598-3604.
Pelletier MC, Burgess RM, Ho KT, Kuhn A, McKinney RA, Ryba SA. 1997. Phototoxicity of individual polycyclic aromatic hydrocarbons and petroleum to marine invertebrate larvae and juveniles. Environmental Toxicology and Chemistry. 16(10): 2190-2199.
Samanta S, Singh OV, Jain RK. 2002. Polycyclic aromatic hydrocarbons: environmental pollution and bioremediation. Trends in Biotechnology 20: 243-248.
Schädle T, Pejcic B, Myers M, Mizaikoff B. 2014. Fingerprinting Oils in Water via Their Dissolved VOC Pattern Using Mid-Infrared Sensors. Analytical Chemistry. 86(19): 9512-9517.
Solomon GM, Janssen S. 2010. Health Effects of the Gulf Oil Spill. Journal of American Medical Association. 304(10): 1118-1119.
Staniloae D, Petrescu B, and Patroescu C. 201. Pattern recognition based software for oil spill identification by gas chromatography and IR spectroscopy. Environmental Forensics. 2. 363-366.
Wang ZD, Fingas M, Landriault M, Sigouin L, Grenon S, Zhang D. 1999. Source identification of an unknown spilled oil from Quebec (1998) by unique biomarker diagnostic ratios of “source-specific marker” compounds. Environmental Technology. 20. 851-862.
Wang Z, Stout S, editors. 2007. Oil Spill Environmental Forensics: Fingerprinting and Source Identification. 1st Ed. Elsevier.
Yim UH, Kim M, Ha SY, Kim S, Shim WJ. 2012. Oil Spill Environmental Forensics: the Hebei Spirit Oil Spill Case. Environmental Science & Technology. 46: 6431−6437.
Continue to Working with Communities, part II of this document.