Section 80.65(e)(1) of the regulations states that "[e]ach refiner or importer shall determine the value of each of the [reformulated gasoline] properties for each batch of reformulated gasoline it produces or imports prior to the gasoline leaving the refinery or import facility, by collecting and analyzing a representative sample of gasoline taken from the batch." "Batch of reformulated gasoline" is defined at § 80.2(gg) as "a quantity of reformulated gasoline which is homogeneous with regard to those properties which are specified for reformulated gasoline certification."
Samples that accurately represent batch properties are necessary in order to determine if RFG standards are being met. Therefore, the first concern of batch sampling is to determine whether or not the tank contents are homogeneous.
Gravity analyses of upper, middle, and lower samples is an appropriate means of establishing tank homogeneity. EPA would consider a tank to be homogeneous where the maximum difference in tested gravities between any two samples from different tank strata is no greater than 0.6 °API, unless there is reason to believe the tank contents are not mixed in spite of such gravity test results. For example, if samples from a storage tank have noticeably different colors, the gasoline in the tank should not be considered homogeneous even if the samples have gravity tests that are within the 0.6 °API range.
If a question remains about whether the contents of a storage tank are fully mixed following gravity testing the party could resolve the homogeneity issue by conducting tests on the upper, middle and lower tank samples for benzene and oxygen. (Tank homogeneity could be established using benzene and oxygen tests on upper, middle and lower tank samples without the need for gravity testing.) EPA would consider a tank to be homogeneous if the maximum difference in benzene tests is 0.10 vol% and the maximum difference in oxygen tests is 0.15 wt%. The benzene and oxygen testing to establish homogeneity (as opposed to certification testing) could use a non-regulatory method such as mid-infrared analysis.
Where it is found that tank contents are not homogeneous, further mixing should be performed before collecting a representative sample for reformulated gasoline analysis.
Product stratification should also be avoided downstream of refiner or importer facilities, because samples must meet the downstream "per gallon" standards, and stratification could result in a portion of the gasoline in a tank being out of compliance with "per gallon" standards. For further discussion of homogeneity, see the Independent Sampling and Testing Section, Question 20 of the July 1, 1994 Question and Answer Document).
Storage tanks should be sampled according to 40 CFR part 80, Appendix D, using the method that will best represent the contents of the tank or batch. EPA expects the refiner, importer, or independent laboratory to use its best professional judgment in determining the procedures that are necessary in order to best represent a given batch within the guidelines of Appendix D.
EPA preference for sampling storage tanks is a "running" or "all-levels" sample collected from an un-confined (no gauge tube) roof port. A "running" or "all levels" sample collected from a perforated gauge tube is the next best choice. In no case should a sample be collected from a solid gauge tube.
EPA prefers to collect "running" samples as opposed to "all-levels" samples for two reasons. First, assuming both "all-levels and "running" samples are collected with uniform lowering and retrieval rates, the "running" procedure achieves better representation of the tank contents than the "all-levels" procedure. This occurs because with the "running" procedure, one half of the sample is collected when lowering the apparatus, and the column sampled is undisturbed at that point. The second reason is that "running" samples are easier to collect than "all-levels" samples because the sample collector is not required to stopper the sample bottle.
If a tank cannot be bottle sampled from the top, then tap sampling is an appropriate substitute. For best representation, a single composite should be collected by proportionally filling the sample container from all available taps. If homogeneity is well documented, the entire sample may be collected from a single tap. If a refinery or importer tank has no roof sampling port or sampling taps, then a pipeline sample is the only other sampling means that is possible. Pipeline sampling is discussed in Question 2 of this section.
In the case of downstream quality assurance sampling from a storage tank which does not have a roof sampling port or taps for sampling, a sample collected from a truck or barge that has just loaded from that tank is marginally acceptable. The truck or barge should be completely empty before loading, and a "running" sample should be collected from the truck or barge compartment.
Appendix D contains general instructions and precautions that must be followed when choosing sampling equipment and containers, and when collecting samples. RVP is the most sensitive reformulated gasoline property, relative to sampling, and therefore precautions to prevent loss of "light ends" must be followed carefully. Also, sampling containers must be clean and rinsed well with the gasoline to be sampled in order that the sample is not contaminated, for example, with trace amounts of heavy metals. When collecting tap samples, the tap and connecting piping must be completely flushed, and the sample container must be bottom filled strictly according to the procedure outlined in Appendix D. Always label the container as soon as possible, and note the location of the sampling point and method of collection.(1/23/95)
This question and answer was posted at Consolidated List of Reformulated Gasoline and Anti-Dumping Questions and Answers: July 1, 1994 through November 10, 1997 (PDF)(333 pp, 18.17 MB, EPA420-R-03-009, July 2003, About PDF)