1. Scope
    1.1 The purpose of this guide is to provide a procedure to be used to perform a mass balance calculation of trace metals entering and leaving a combustion device. Such a combustion device may be an incinerator, an industrial furnace or steam generation boiler or a heat transfer media heater (such as a direct fired hot oil heater). A mass balance calculation aids in determining the quality of the analyses of the input and output streams of a combustion device. Due to analytical imprecision, it may not be possible to demonstrate that the input mass of a specific trace metal is equal to the output mass of this metal. To demonstrate "closure" in a mass balance it must be shown that the mass input plus or minus the analytical imprecision overlaps the range of values of the mass output plus or minus the analytical imprecision. By utilizing the appropriate sampling and analytical methods, this guide could also be used to determine the mass balance of non-metallic elements.
    1.2 The units may be expressed in any format provided they are used consistently for both the input and output values. Input and output values may be expressed as mass per unit time, however the units must be consistent for both input and output values; e.g., grams per second or grams per hour, the units of time must be the same for both input and output values.
    1.3 This guide does not purport to address all of the safety concerns, if any, associated with its use. It is the responsibility of the user of this guide to establish appropriate safety and health practices and determine the applicability of regulatory limitations prior to use.

2. Reference Documents
    2.1 These are covered in Section 7, Test Methods.

3. Terminology
    3.1 Description of terms specific to this guide:
    3.1.1 Trace metals -- any metal consti-tuent that is less than 1% by weight in any one of the device input or output streams.
    3.1.2 Combustion device -- any device that is intended to convert organic based fuels into energy by oxidation. Generally for the purpose of this guide, this is limited to combustion devices such as incinerators, industrial furnaces, direct fired steam generation units or direct fired heat transfer media heaters.
    3.1.3 Mass balance -- the sum of the inputs into a process (in this case the combustion device) of a specific metal are compared with the sum of the outputs from the device of that same metal.
    3.1.4 Closure -- when the total mass input of a metal plus or minus the analytical imprecision overlaps the ranges of values of the total mass output of that same metal plus or minus the analytical imprecision.
    3.1.5 Analytical imprecision -- each analytical method has a determinable level of imprecision. Generally this is stated as a percentage of the measured analytical value. As an example, SW-846 Method 0060 typically has a QA/QC data quality goal of +/-25% of the measured value. The achievement of this goal is demonstrated by the QA procedures during the execution of the analytical method.

4. Summary of Practice
    4.1 The person wishing to perform a trace metals mass balance on a combustion device selects the appropriate sampling and analytical methods which will measure the targeted trace metals in all of the combustion device inputs and all of the combustion device outputs. A careful accounting of the mass in each of the device inputs and outputs must be made over a selected time period. This is accomplished by selecting the appropriate sampling point for each input and output stream and an appropriate sampling frequency based on the knowledge of the process, and by measuring the mass input and output of each of these streams during the sampling period. The analytical data and the mass input and mass output data are used to calculate a range of total mass input and total mass output for each of the targeted metals. A comparison of the range of input versus the range of output values for each metal will determine whether a balance has been achieved.

5. Significance of Use
    5.1 A demonstration of closure in a mass balance would be indicative of a set of analyses and input and output stream rate measurements that adequately characterize the concentration of the targeted trace metals present in each of the combustion device's input and output streams. A failure to demonstrate closure would be indicative of a failure to characterize the metal concentration in one of the input or output streams or the failure to adequately measure the rate of an input or output stream or, possibly, the omission or misidentification of a stream. Lack of closure can also occur if a metal is retained and "builds up" within the system and the testing did not allow sufficient time for the system to reach equilibrium. (For certain cement kiln designs and some metals, equilibrium may never be reached. Under those circumstances a different approach is required to determine the lonng term emission rate and metal balance.)
    5.2 A failure to characterize the concentration of metal in an input or output stream may be indicative of an inadequate level of precision in the sampling and/or analytical methods. It may also be that the mass input or output rates have not been adequately measured during the sample period. Either of these indications would require an examination of the sampling and analytical method and their execution, sampling frequency and process input/output measurements and controls. The successful demonstration of closure of the mass balance ultimately rests on the achievable accuracy of the analyses and input/output stream measurements. If input and output streams cannot be accurately metered, particularly if these streams exhibit wide variability in flow rate over the test period, closure of the mass balance is unlikely. Steady state operating conditions are generally required for the test period. This test period should not closely follow a period of non-steady state operation. Processes with highly erratic feedrates, process cycling or highly erratic trace metals concentrations in the input/outputs may require an elaborate sampling and analytical plan to achieve closure of the mass balance.

6. Procedure
    6.1 A person knowledgeable of the process should examine each process input and output stream and determine the following:
    6.1.1 The accuracy of the measurement of the stream flow rate. The accuracy of this measurement should be as good as possible over the test period, but should at least be no worse than +/-10% of actual.
    6.1.2 The expected variability of the trace metals concentration in the various input and outputs must be considered when selecting a sample frequency for that stream. More variability in the trace metals concentration will require more frequent sampling. Alternately, if the stream is highly variable in its trace metals concentration and these variable concentrations can be isolated to discreet volumes, it is possible to sample and analyze these volumes separately; either prior to the test if this is an input or subsequent to the test if it is an output.
    6.1.3 If very low levels of a trace metal are expected for a stream, there should be a consideration of collecting a larger sample than normal or utilizing an analytical procedure that achieves a lower detection limit. Either of these considerations may affect the sampling method or sampling location.
    6.1.4 Some input or output streams should be sampled on an advanced or delayed schedule. In some cases the only safe sample point of a feed stream may result in the feed stream entering the device several minutes or more subsequent to sampling. Or, an output stream may represent the inputs fed to it an hour or more previous to its sampling. Sampling times must be adjusted to accommodate such time delays, otherwise the test period will not be characterized by analyses.
    6.1.5 Sample point selection and sampling method must be considered as a part of the overall quality of the performance of the mass balance. In addition to the safety of the person performing the sampling consideration must be given to how representative a sample from that location is of the stream. As an example; is the stream well mixed? Have two or more sub-streams entered the stream prior to the sample point? Is the sample likely to be contaminated during sample retrieval? This can occur due to the stream being very hot or very cold, or being at a location that is dusty. Sampling tools must be appropriate for the location, but not introduce contaminants into the analysis. As an example, a stainless steel sample cup may be the standard sampling tool, but such a tool can contaminate the sample with chromium and/or nickel.
    6.2 After examining all these considerations, a detailed plan is prepared to specify the sample points, the sampling method at each point, the schedule for sample collection at each point, a sample storage and label designation system and a plan to modify the sampling schedule in the event of test delays or interruption.
    6.3 Prior to the test period the sample storage materials and sampling tools must be strategically located. The persons performing the sampling must be trained and a sample coordinator designated. A clean, dry location must be selected for the cataloging and storage of the samples. Invariably, a stack emissions sampling and/or analytical firm must be selected. This firm must understand the QA/QC requirements that they are expected to meet and the importance of communication of run start and stop times and, in the event of an interruption, the start and stop times of any interruption of their sampling due to any cause.
    6.4 On the day of the test, the sampling of the various streams is performed in accordance with the sampling plan determined above. Those streams that must be sampled prior to the test period must be sampled the appropriate time period in advance of the start of the stack gas sampling. This requires coordination with the stack sampling firm. The sampling schedule is keyed to the stack sampling execution. If the stack sampling is interrupted, the input/output stream sampling schedule must be altered accordingly. At the end of the test period after all of the input and output streams have been collected, the properly labeled samples are sent for analysis.
    6.5 It is recommended that the analysis of the samples be periodically monitored. This is usually done by the sample coordinator. The purpose of this monitoring is to ensure that the samples are analyzed prior to their expiration date, that the QA/QC checks have been performed as agreed and to spot check the data for obvious errors such as misdesignation of sample ID and mathematical errors.
    6.6 A sample trace metals mass balance report is attached.

7. Test Method
    7.1 Process Stream Sampling and Analytical Methods - Process streams such as kiln feed, cement dust or incineration fly ash, clinker or bottom ash and fuels are sampled utilizing a "grab" sample method and subjected to an analysis for trace metals utilizing the ASTM E926 Method A.
    7.2 Stack Emissions Sampling and Analytical Methods - Stack emission samples are collected and analyzed utilizing EPA SW-846 Method 0060.

8. Report
    8.1 Once all of the analytical data is compiled, as well as the input/output stream flow rate data, a spreadsheet is constructed. This spreadsheet calculates the mass or each targeted trace metal for each input and output stream.
    8.1.1 Each analysis has a stated or determined precision. Generally this is expressed as +/-XX%. For each metal in each stream, a minimum and maximum rate is calculated by multiplying the analytically determined concentration times the mass input or output rate according to the following method.
    8.1.2 If the analysis is below the detection limit of the analytical procedure, the minimum metal mass rate is 0 (zero). The maximum value is the detection limit concentration times the mass rate of the stream.
    8.1.3 If the analysis is above the detection limit but below the quantitation limit, the minimum metal mass rate is the detection limit times the mass rate of the stream. The maximum metal mass rate is the quantitation limit times the mass rate of the stream.
    8.1.4 If the analysis is above the quantitation limit, the minimum metal mass rate is the declared imprecision percent-age subtracted from 100% and the resultant times the concentration and that value times the mass rate of the stream. (e.g. 75% x conc. x mass rate) The maximum metal mass rate is the declared imprecision percentage added to 100% with the resultant multiplied with the concentration and that value times the mass rate of the stream. (e.g. 125% x conc x mass rate)
    8.1.5 This is repeated across the various metals and the streams to result in a minimum and maximum metal mass for each metal in each input and output stream.
    8.1.6 At this point, it is now possible to create a minimum and maximum input value for each targeted metal by summing the minimum values for each metal in the input streams and the maximum values for each metal in the input streams. Perform a similar summing of the minimum and maximum values for the output streams.
    8.2 For each targeted trace metal, there are now two ranges of values, the mass input ranging from minimum to maximum and the mass output ranging from minimum to maximum. If these ranges overlap when compared, meaning a value of each is within the range of values of the other, closure of the mass balance for that trace metal has been demonstrated.

9. Appendices
    9.1 Appendix "A" - Sample Mass Balance Report