ASTM D6469 standard guide for microbial contamination in fuels and fuel systems
8. Sampling
8.1 Bottom samples, as described in Practice D4057, provide the best material for evaluating microbial contamination.

8.2 Because sample analyses may be performed by more than one laboratory, good sample chain of custody procedures should be followed (see Guide D4840).

8.3 Both biological and nonbiological deterioration processes continue in a sample during the period between collection and analysis. Ideally, all testing should be accomplished at the sampling site, within a few minutes after a sample is drawn. As this is rarely possible, good practices for preserving and preparing samples for analysis should be following (see Practices D3325 and D3326).

8.4 Samples for pH, alkalinity/acidity, and dissolved oxygen determinations should be tested within 1 h after sampling.

8.5 Samples for microbiological testing should be kept on ice for transport to the laboratory. Tests should be performed within 1 h and no later than 36 h after sampling. Samples stored at higher temperatures, or for longer times, may show the presence of microbial contamination that does not represent actual fuel system conditions.

8.5.1 Samples for microbiological testing should be collected in new, unused containers.

8.5.2 If microbiological tests are not going to be completed within an hour after sample collection, the container should not be more than half-full. This provides adequate headspace to minimize the risk of conditions within the sample container becoming anoxic. Samples to be examined for anaerobic bacteria should be filled completely to maintain oxygen depleted or anoxic conditions.

8.6 Sampling intervals should be set so that there are at least three sets of data obtained during the period between system changes. For example, if microbial loads take six months to exceed criteria levels after biocide treatment, then tests should be performed every 1.5 to 2 months. This provides a compromise between controlling monitoring costs and detecting potential problems before they affect operations.

9. Examination and Testing
9.1 Some analytical methods can be performed in the field under less than optimal conditions, but many others will require the services of a laboratory with specialized equipment.

9.2 Gross Observations:
9.2.1 Gross observations, such as color, odor, clarity, and appearance of the fuel/water interface, are made during routine housekeeping and change over practices. When careful records are kept, they can identify changes in operating practices and environmental conditions that result in increased levels of microbial contamination. Gross observations should be made whenever a sample is drawn from the tank.

9.2.2 Check any accessible tank surfaces for the presence of microbial mats or slime. Their presence is evidence of microbial infestation.

9.2.3 Observe an interface sample that contains both water and fuel (see Fig. 2).

9.2.3.1 Uncontaminated samples contain either no water or only two clear and bright phases. Turbidity in the fuel phase (see Test Methods D4176 and D4860) indicates a significant problem, which might be due to microbial activity, high water content, surfactant contamination, or chemical instability.

9.2.3.2 Emulsified brown to red to black material in either phase indicates the presence of microbes. These colors generally reflect the presence of iron oxide or iron hydroxide, or both. Formation of these precipitates may involve microbial activity or may be the result of nonbiological processes.

9.2.3.3 The presence of a third phase (the rag phase or cuff layer) between the fuel and water suggests a microbial problem, although the rag layer may also be formed nonbiologically due to fuel component polymerization or inorganic precipitate formation, or both.

9.2.3.4 Presence of significant amounts of precipitated material in jars containing tank or pipeline samples further suggests the presence of microbes.

9.2.3.5 Hydrogen sulfide and other atypical (rancid) odors may indicate heavy microbial contamination.

9.2.3.6 Compare the gross properties (see Test Method D4176) or near-bottom fuel (from 5 to 10 cm above fuel-water interface or tank bottom) with those of the bottom sample fuel-phase. Differences in clarity or color indicate that there are differences in fuel chemistry. Microbial activity may be the cause.

9.2.4 Pipe or filter blockage also indicates that severe infestation may be present in a fuel system. Test Method D5452 can be used to predict a fuel's filter plugging tendencies.

9.2.5 The major advantage to gross observations is their speed and simplicity. Their principal disadvantage is that gross changes typically occur late in the biodeterioration process, after significant damage has occurred.

9.3 Physical Testing:
9.3.1 Physical testing requirements for fuels are listed in each product grade specification (see Specifications D396, D910, D975, D1655, D2069, D2880, D3699, D4814, and D6227).

9.3.2 Microbial contamination is most likely to affect filterability (see Test Methods D2068, D2276, D4870, D5452, D6217, and D6426) and oxidation stability (see Test Methods D2274, D3241, and D5304). Severe contamination can eventually affect viscosity (see Test Method D445) and density (see Test Method D1298). Consequently, when these parameters are outside the specification range, microbial contamination should be considered as one of the possible causes.

9.3.3 Physical tests for bottom-water samples are not included in fuel standards. One particularly useful test is determination of suspended solids by filtration (see Test Method 2540 D). This test determines the mass of material that is present in the contaminated water sample.

9.3.4 The measurement of interfacial tension (see Test Methods D1331) is sometimes used as a surrogate for direct surface active by-product (fatty acids and proteins) analysis. The presence of these biosurfactants is indicative of microbial activity. This test method is most useful if baseline data (fresh fuel over sterile water) are available. Both near-bottom fuel and bottom-water should be tested periodically to determine changes in their respective contact angles (interfacial tension).

9.4 Chemical Testing:
9.4.1 Chemical testing should be performed on the fuel phase, water-phase, and filter material, if available.

9.4.2 ASTM fuel product specifications include chemical tests. Microbial contamination may contribute to changes in any chemical property of a fuel. Consequently, a microbial cause should be considered when investigating a fuel's failure to meet chemical specifications. Several additional standard fuel chemistry tests facilitate diagnosis of fuel diodeterioration. In addition, there are a variety of commercially available test kits for many of the parameters discussed below. Typically the test kits provide a simplified means for obtaining data. In many circumstances, data from test kits are sufficient for contamination monitoring and diagnostic purposes. Operators or laboratory personnel intending to substitute a simplified test should confirm that the simplified test meets their specific needs and expectations. Guide E 1326 offers suggestions for evaluating the suitability of nonconventional or nonstandard test methods.

9.4.3 Acid/base (see Test Methods D664, D974, and D3242) number indicates the extent to which the fuel composition has changed through production of fatty acid by-products, carbon dioxide, or other by-products of metabolism or non-biological processes.

9.4.4 Water content (see Test Method D1744 or D3240) provides information related to the amount of water present in the fuel phase. Water is critical to the successful establishment of a microbial population in the fuel phase.

9.4.5 Copper corrosion (see Test Method D130) testing may indicate the presence of anaerobic microbes, sulfate-reducers, and clostridia and other protein fermenters that also produce hydrogen sulfide (H2S). The copper corrosion strip test is rapid and economical and is included in some fuel specifications.

9.4.6 Proteins, lipids, and fatty acids in the fuel and bottom-water are directly proportional to biomass.

9.4.7 Gas chromatography/mass spectroscopy (GC/MS) fingerprinting may reveal compositional changes in the fuel components that can be evaluated using simulated distillation methods and elemental analysis. Certain types of changes are expected in a product contaminated with microbes. These include a diminution in alkanes (normal, iso-, and cyclo-) relative to uncontaminated fuels.

9.4.8 Water phase tests are not included in fuel specifications but provide critical information about fuel system biodeterioration.

9.4.8.1 Bottom-water samples should be measured for pH (see Test Methods D1293). Bottom-water pH is normally alkaline (> pH 7.0). Samples with pH <7.0 are suspect for microbial activity.

9.4.8.2 Acidity or alkalinity (see Test Methods D1067) data are used to compute the aggressiveness index of bottom-water. The alkalinity of uncontaminated bottom-water under middle distillate fuel typically ranges from 75 to 150 mg CaCO3/ L. Under gasoline that has been augmented with amine-based anticorrosion additives, alkalinities > 2500 mg CaCO3 are not unusual. To be most useful, baseline bottom-water alkalinity should be determined using fresh fuel over local tap or freshly collected rainwater, or any readily available fresh water with a known low (< 50 mg CaCO3/L) alkalinity. If total alkalinity falls below half of the baseline value, microbial activity is the most likely cause. Microbial activity is also likely in bottom-water samples with any measurable acidity, as the by-products or microbial metabolism tend to be acidic and readily partition into the water phase. These include the by-products of hydrocarbon metabolism: acetate, butyrate, formate, and others, as well as carbon dioxide.

9.4.8.3 Although not affected directly by microbial growth, hardness (see Test Method D1126) is one of three variables used to compute aggressiveness index. Moreover, hard water (> 200 mg CaCO3/ L) tends to support more robust microbial communities than does soft water.

9.4.8.4 Compute A.I. from pH, alkalinity/acidity, and hardness data.
(a) Water with A.I. > 12 is considered noncorrosive. If A.I. is between 10 and 12, the water is moderately corrosive, and if A.I. <10, it is strongly corrosive. Low bottom-water A.I. is a strong indication of biological activity.

9.4.8.5 Dissolved oxygen (see Test Methods D888) in bottom-water from uncontaminated tanks is typically at 50 to 75 % saturation. Moreover, the oxygen concentration is an uncontaminated sample that is aerated to saturation will remain greater than 50 % saturation for at least 4 h after aeration is discontinued. Consequently, a bottom-water sample with an oxygen concentration less than 50 % of the saturation concentration, or a sample in which the oxygen concentration decreases by more than 50 % in a 4 h period is likely to have a significant bioburden.

9.4.8.6 Metal ion analysis by inductively-coupled argon plasma atomic emission spectroscopy (see Test Method D1976) provides a measure of the extent to which metals from tanks, pipes, and other equipment surfaces have been solubilized. Microbes involved in MIC can directly or indirectly cause solubilization of metals from system components. Consequently, the presence of increased soluble iron over time strongly suggests bacterial activity. Changes in aluminum and manganese (along with iron) are also indicative of corrosion.

9.4.8.7 Nitrogen analysis (see Test Methods D1426 and D3867) is useful because nitrate-reducing bacteria convert nitrate ions to nitrite and ammonia.

9.4.8.8 Sulfate reducers produce hydrogen sulfide from sulfate ions. Increased sulfide concentrations in bottom-waters (see Test Methods D4952) are characteristic of biodeterioration in fuel systems.

9.4.8.9 Fuel components that have partitioned into the water-phase can be analyzed by Test Methods D3328 or Test Method D3414. Biosurfactants increase petroleum emulsification into bottom-water.

9.4.8.10 Phosphorus is often a growth-limiting nutrient in bottom-water. Determine phosphorus in bottom-water samples using Test Methods D515.

9.5 Microbiological Testing:
9.5.1 In contrast to many of the gross observations, physical and chemical tests, there are no generally accepted criteria for acceptable microbial contamination levels. In many applications, any detectable microbes in the fuel-phase trigger corrective action. However, populations of greater than 105 cells/mL may be tolerable in bottom-water. As discussed in Section 10, it is critical to monitor contamination trends and to correlate microbiological data with operational, chemical, and physical data in order to define the role of microbial contamination in specific fuel or system problems

9.5.2 Direct microscopic examination enables the analyst to directly examine all sample phases (fuel, water, filter medium, and so forth). A skilled analyst can assess the relative distribution of different groups of microbes and enumerate total cell numbers (see Test Method D4454). IP Proposed Method BY provides a technique for recovering and observing fungal fragments from fuel samples.

9.5.3 Electron microscopy (Scanning Electron Microscopy (SEM) or Time of Flight SEM) is useful in special circumstances. These have a practical resolution of about 25 nm, compared with about 300 nm for conventional light microscopy, and a depth-of-field at least 300 times greater. However, both methods currently require water to be removed, which may alter surface structures creating artifacts.

9.5.4 Enumeration methods estimate microbial loads from the growth of microorganisms either in liquid or on solid nutrient growth media (see Practice D3870).

9.5.4.1 On solid media, viable titer estimates are computed from the number of colonies that form on the inoculated growth medium. The three most common solid media viable titer methods are the pour-plate, spread-plate, and membrane filtration methods.

9.5.4.2 In liquid media, viable titer estimates are based on turbidity development or color change, or both. Liquid media are better suited for estimating viable titers in fuel-phase samples, because fuel hydrophobicity (tendency to reject water) interferes with sample dispersion across a solid medium's surface.

9.5.4.3 Viable titer methods are most meaningful for detecting subpopulations with specific physiological attributes. Examples include hydrocarbon or additive mineralization and sulfate reduction (see Test Methods D4412).

9.5.4.4 The choice of medium has a large effect upon what microbes are recovered and constitutes a bias introduced into this type of testing. Frequently, microbes that are important components of the ecosystem being studied do not grow in the media used for viable titers.

9.5.4.5 Typically sample volumes from 0.01 to 1.0 mL are used to inoculate solid media, and 1.0 to 5.0 mL are used to inoculate liquid media. Given the nonuniform (heterogeneous) distribution of biomass in fuel systems, it is difficult to get a truly representative sample. Any given sample of water or hydrocarbon may not come from where the bulk of the microbes are growing at the site. Dislodging bacteria attached to equipment surfaces is not easy. Population counts based on planktonic (suspended) microbes generally severely underestimate the total population present at the site.

9.5.4.6 Several suppliers offer solid media on dip-slides and liquid media in pre-measured vials. These devices provide data comparable (precision and accuracy) to viable titers determined by standard laboratory test procedures while offering convenience for field use.

9.5.4.7 Since viable recoveries from fuel are typically less than 1 CFU/mL, samples can be filtered through 0.22 and 0.45-μm filters to concentrate bacteria and fungi respectively. It is useful to filter up to 1000 mL of fuel in order to decrease detection limits from 1 CFU/mL (from the plate count method described in 9.5.4.5) to 1 CFU/L (see IP 385).

9.5.5 Microbial activity tests may be used as an alternative or supplement to viable titer tests. These tests typically require some skill and laboratory equipment to complete but provide a more direct indication of actual or potential biodeterioration.

9.5.5.1 A variety of chemical and potentiometric test methods can be used to measure the rate of molecule production or disappearance. Examples include periodic measurements of dissolved oxygen (see Test Methods D888) to determine the rate of oxygen consumption (see Test Methods D4478) and periodic spectroscopic analyses to measure changes in relative concentrations of fuel components, primarily molecular species.

9.5.6 Cell component measurements provide a fourth approach to quantifying bioburdens. Over the past two decades, numerous test methods have been developed and automated for analyzing clinical samples for specific cell constituents. A growing number of these test methods have been adapted for environmental and industrial testing. Adenosine triphosphate (ATP) (see Test Method D4012), catalase activity, nucleic acid, and protein concentrations are four examples of molecules routinely monitored in industrial systems. Guide E 1326 provides more guidance on test method selection.

10. Data Interpretation
10.1 The critical element for all data interpretation is change measurement. This, in turn, depends on three factors.

10.2 Tests should be selected based on their ability to guide system management decisions. This means that test results should indicate conditions of the system, which if not properly controlled could result in biodeterioration.

NOTE 2 - The mere presence of recoverable microbes in bulk fuel indicates heavy contamination levels. In contrast, values from 1 x 10(3) to 1 x 10(5) CFU/mL in bottom-water may be acceptable if all other fuel and system conditions are normal.

10.3 The precision and accuracy of each test method needs to be known to determine whether observed changes are real or are within the limits of experimental error (see Practice E 177). Once test precision and variation are defined, criteria levels can be assigned. For fuel samples, the criteria listed in the respective ASTM standards can be used. For water and system samples, managers should define criteria that meet their specific needs.

10.4 Different microbiological growth media will yield different results. Therefore, consistency in sampling and test procedure is important for establishing baseline and special-cause variance conditions.