EPA Methods List with Links

US EPA Method 14 - Determination Of Fluoride Emissions From Potroom Roof Monitors For Primary Aluminum Plants

NOTE: This method does not include all of the specifications (e.g., equipment and supplies) and procedures (e.g., sampling and analytical) essential to its performance. Some material is incorporated by reference from other methods in this part. Therefore, to obtain reliable results, persons using this method should have a thorough knowledge of at least the following additional test methods: Method 1, Method 2, Method 3, Method 5, Method 13A, and Method 13B. 1.0 Scope and Application.

1.1 Analytes.

Analyte CAS No. Total fluorides as Fluorine 7782-41-4 Sensitivity Not determined

1.2 Applicability.

This method is applicable for the determination of fluoride emissions from roof monitors at primary aluminum reduction plant potroom groups.

1.3 Data Quality Objectives.

Adherence to the requirements of this method will enhance the quality of the data obtained from air pollutant sampling methods. 2.0 Summary of Method.

2.1 Gaseous and particulate fluoride roof monitor

emissions are drawn into a permanent sampling manifold through several large probe nozzles. The sample is transported from the sampling manifold to ground level through a duct. The fluoride content of the gas in the duct is determined using either Method 13A or Method 13B. Effluent velocity and volumetric flow rate are determined using anemometers located in the roof monitor.

3.0 Definitions.

Potroom means a building unit which houses a group of electrolytic cells in which aluminum is produced. Potroom group means an uncontrolled potroom, a potroom which is controlled individually, or a group of potrooms or potroom segments ducted to a common control system. Roof monitor means that portion of the roof of a potroom where gases not captured at the cell exit from the potroom. 4.0 Interferences. Same as Section 4.0 of either Method 13A or Method 13B, with the addition of the following:

4.1 Magnetic Field Effects.

Anemometer readings can be affected by potroom magnetic field effects. Section 6.1 provides for minimization of this interference through proper shielding or encasement of anemometer components.

5.0 Safety.

5.1 Disclaimer.

This method may involve hazardous materials, operations, and equipment. This test method may not address all of the safety problems associated with its use. It is the responsibility of the user of this test method to establish appropriate safety and health practices and to determine the applicability of regulatory limitations prior to performing this test method.

5.2 Corrosive Reagents.

Same as Section 5.2 of either Method 13A or Method 13B. 6.0 Equipment and Supplies.

Same as Section 6.0 of either Method 13A or Method 13B, as applicable, with the addition of the following:

6.1 Velocity Measurement Apparatus.

6.1.1 Anemometer Specifications.

Propeller anemometers, or equivalent. Each anemometer shall meet the following specifications: Its propeller shall be made of polystyrene, or similar material of uniform density.

To ensure uniformity of performance among propellers, it is desirable that all propellers be made from the same mold. The propeller shall be properly balanced, to optimize performance. When the anemometer is mounted horizontally,

its threshold velocity shall not exceed 15 m/min (50 ft/min). The measurement range of the anemometer shall extend to at least 600 m/min (2,000 ft/min). The anemometer shall be able to withstand prolonged exposure to dusty and corrosive environments; one way of achieving this is to purge the bearings of the anemometer continuously with filtered air during operation. All anemometer components shall be properly shielded or encased, such that the performance of the anemometer is uninfluenced by potroom magnetic field effects. A known relationship shall exist between the electrical output signal from the anemometer generator and the propeller shaft rpm (see Section 10.2.1). Anemometers having other types of output signals (e.g., optical) may be used, subject to the approval of the Administrator. If other types of anemometers are used, there must be a known relationship between output signal and shaft rpm (see Section 10.2.2). Each anemometer shall be equipped with a suitable readout system (see Section 6.1.3).

6.1.2 Anemometer Installation Requirements. Single, Isolated Potroom.

If the affected facility consists of a single, isolated potroom (or potroom segment), install at least one anemometer for every 85 m (280 ft) of roof monitor length. If the length of the roof monitor divided by 85 m (280 ft) is not a whole number, round the fraction to the nearest whole number to determine the number of anemometers needed. For monitors that are less than 130 m (430 ft) in length, use at least two anemometers. Divide the monitor cross-section into as many equal areas as anemometers, and locate an anemometer at the centroid of each equal area. See exception in Section Two or More Potrooms.

If the affected facility consists of two or more potrooms (or potroom segments) ducted to a common control device, install anemometers in each potroom (or segment) that contains a sampling manifold. Install at least one anemometer for every 85 m (280 ft) of roof monitor length of the potroom (or segment). If the potroom (or segment) length divided by 85 m (280 ft) is not a whole number, round the fraction to the nearest whole number to determine the number of anemometers needed. If the potroom (or segment) length is less than 130 m (430 ft), use at least two anemometers. Divide the potroom (or segment) monitor cross-section into as many equal areas as anemometers, and locate an anemometer at the centroid of each equal area. See exception in Section Placement of Anemometer at the Center of

Manifold. At least one anemometer shall be installed in the immediate vicinity [i.e., within 10 m (33 ft)] of the center of the manifold (see Section 6.2.1). For its placement in relation to the width of the monitor, there are two alternatives. The first is to make a velocity traverse of the width of the roof monitor where an anemometer is to be placed and install the anemometer at a point of average velocity along this traverse. The traverse may be made with any suitable low velocity measuring device, and shall be made during normal process operating conditions. The second alternative is to install the anemometer half-way across the width of the roof monitor. In this latter case, the velocity traverse need not be conducted.

6.1.3 Recorders.

Recorders that are equipped with suitable auxiliary equipment (e.g., transducers) for converting the output signal from each anemometer to a continuous recording of air flow velocity or to an integrated measure of volumetric flowrate shall be used. A suitable recorder is one that allows the output signal from the propeller anemometer to be read to within 1 percent when the velocity is between 100 and 120 m/min (330 and 390 ft/min). For the purpose of recording velocity, "continuous" shall mean one readout per 15-minute or shorter time interval. A constant amount of time shall elapse between readings. Volumetric flow rate may be determined by an electrical count of anemometer revolutions. The recorders or counters shall permit identification of the velocities or flowrates measured by each individual anemometer.

6.1.4 pitot Tube.

Standard-type pitot tube, as described in Section 6.7 of Method 2, and having a coefficient of 0.99 ± 0.01.

6.1.5 pitot Tube (Optional).

Isolated, Type S pitot, as described in Section 6.1 of Method 2, and having a known coefficient, determined as outlined in Section 4.1 of Method 2.

6.1.6 Differential Pressure Gauge.

Inclined manometer, or equivalent, as described in Section 6.1.2 of Method 2.

6.2 Roof Monitor Air Sampling System.

6.2.1 Manifold System and Ductwork.

A minimum of one manifold system shall be installed for each potroom group. The manifold system and ductwork shall meet the following specifications: The manifold system and connecting duct shall be permanently installed to draw an air sample from the roof monitor to ground level. A typical installation of a duct for drawing a sample from a roof monitor to ground level is shown in Figure 14-1 in Section 17.0. A plan of a manifold system that is located in a roof monitor is shown in Figure 14-2. These drawings represent a typical installation for a generalized roof monitor. The dimensions on these figures may be altered slightly to make the manifold system fit into a particular roof monitor, but the general configuration shall be followed. There shall be eight probe nozzles, each having a diameter of 0.40 to 0.50 m. The length of the manifold system from the first probe nozzle to the eighth shall be 35 m (115 ft) or eight percent of the length of the potroom (or potroom segment) roof monitor, whichever is greater. Deviation from this requirement is subject to the approval of the Administrator. The duct leading from the roof monitor manifold system shall be round with a diameter of 0.30 to 0.40 m (1.0 to 1.3 ft). All connections in the ductwork shall be leak-free. As shown in Figure 14-2, each of the sample legs of the manifold shall have a device, such as a blast gate or valve, to enable adjustment of the flow into each sample probe nozzle. The manifold system shall be located in the immediate vicinity of one of the propeller anemometers (see Section and as close as possible to the midsection of the potroom (or potroom segment). Avoid locating the manifold system near the end of a potroom or in a section where the aluminum reduction pot arrangement is not typical of the rest of the potroom (or potroom segment). The sample probe nozzles shall be centered in the throat of the roof monitor (see Figure 14-1). All sample-exposed surfaces within the probe nozzles, manifold, and sample duct shall be constructed with 316 stainless steel. Alternatively, aluminum may be used if a new ductwork is conditioned with fluoride-laden roof monitor air for a period of six weeks before initial testing. Other materials of construction may be used if it is demonstrated through comparative testing, to the satisfaction of the Administrator, that there is no loss of fluorides in the system. Two sample ports shall be located in a vertical section of the duct between the roof monitor and the exhaust fan (see Section 6.2.2). The sample ports shall be at least 10 duct diameters downstream and three diameters upstream from any flow disturbance such as a bend or contraction. The two sample ports shall be situated 90E apart. One of the sample ports shall be situated so that the duct can be traversed in the plane of the nearest upstream duct bend. 6.2.2 Exhaust Fan. An industrial fan or blower shall be attached to the sample duct at ground level (see Figure 14-1). This exhaust fan shall have a capacity such that a large enough volume of air can be pulled through the ductwork to maintain an isokinetic sampling rate in all the sample probe nozzles for all flow rates normally encountered in the roof monitor. The exhaust fan volumetric flow rate shall be adjustable so that the roof monitor gases can be drawn isokinetically into the sample probe nozzles. This control of flow may be achieved by a damper on the inlet to the exhauster or by any other workable method.

6.3 temperature Measurement Apparatus.

To monitor and record the temperature of the roof monitor effluent gas, and consisting of the following:

6.3.1 Temperature Sensor.

A Temperature Sensor shall be installed in the roof monitor near the sample duct. The Temperature Sensor shall conform to the specifications outlined in Method 2, Section 6.3.

6.3.2 Signal Transducer.

Transducer, to change the Temperature Sensor voltage output to a temperature readout.

6.3.3 thermocouple Wire.

To reach from roof monitor to signal transducer and recorder.

6.3.4 Recorder.

Suitable recorder to monitor the output from the thermocouple signal transducer.

7.0 Reagents and Standards.

Same as Section 7.0 of either Method 13A or Method 13B, as applicable.

8.0 Sample Collection, Preservation, Storage, and Transport.

8.1 Roof Monitor Velocity Determination.

8.1.1 Velocity Estimate(s) for Setting Isokinetic flow.

To assist in setting isokinetic flow in the manifold sample probe nozzles, the anticipated average velocity in the section of the roof monitor containing the sampling manifold shall be estimated before each test run. Any convenient means to make this estimate may be used (e.g., the velocity indicated by the anemometer in the section of the roof monitor containing the sampling manifold may be continuously monitored during the 24-hour period before the test run). If there is question as to whether a single estimate of average velocity is adequate for an entire test run (e.g., if velocities are anticipated to be significantly different during different potroom operations), the test run may be divided into two or more "sub-runs," and a different estimated average velocity may be used for each sub-run (see Section 8.4.2).

8.1.2 Velocity Determination During a Test Run.

During the actual test run, record the velocity or volumetric flowrate readings of each propeller anemometer in the roof monitor. Readings shall be taken from each anemometer at equal time intervals of 15 minutes or less (or continuously).

8.2 temperature Recording.

Record the temperature of the roof monitor effluent gases at least once every 2 hours during the test run.

8.3 Pretest Ductwork Conditioning.

During the 24- hour period immediately preceding the test run, turn on the exhaust fan, and draw roof monitor air through the manifold system and ductwork. Adjust the fan to draw a volumetric flow through the duct such that the velocity of gas entering the manifold probe nozzles approximates the average velocity of the air exiting the roof monitor in the vicinity of the sampling manifold.

8.4 Manifold Isokinetic Sample Rate Adjustment(s).

8.4.1 Initial Adjustment.

Before the test run (or first sub-run, if applicable; see Sections 8.1.1 and 8.4.2), adjust the fan such that air enters the manifold sample probe nozzles at a velocity equal to the appropriate estimated average velocity determined under Section 8.1.1. Use Equation 14-1 (Section 12.2.2) to determine the correct stream velocity needed in the duct at the sampling location, in order for sample gas to be drawn isokinetically into the manifold probe nozzles. Next, verify that the correct stream velocity has been achieved, by performing a pitot tube traverse of the sample duct (using either a standard or Type S pitot tube ); use the procedure outlined in Method 2.

8.4.2 Adjustments During Run.

If the test run is divided into two or more "sub-runs" (see Section 8.1.1), additional isokinetic rate adjustment(s) may become necessary during the run. Any such adjustment shall be made just before the start of a sub-run, using the procedure outlined in Section 8.4.1 above.

NOTE: Isokinetic rate adjustments are not permissible during a sub-run.

8.5 Pretest Preparation, Preliminary Determinations, Preparation of Sampling Train, Leak-Check Procedures, Sampling Train Operation, and Sample Recovery.

Same as Method 13A, Sections 8.1 through 8.6, with the exception of the following:

8.5.1 A single train shall be used for the entire sampling run.

Alternatively, if two or more sub-runs are performed, a separate train may be used for each sub-run; note, however, that if this option is chosen, the area of the sampling probe nozzle shall be the same (±2 percent) for each train. If the test run is divided into sub-runs, a complete traverse of the duct shall be performed during each sub-run.

8.5.2 Time Per Run.

Each test run shall last 8 hours or more; if more than one run is to be performed, all runs shall be of approximately the same (±10 percent) length. If questions exist as to the representativeness of an 8-hour test, a longer period should be selected. Conduct each run during a period when all normal operations are performed underneath the sampling manifold. For most recently- constructed plants, 24 hours are required for all potroom operations and events to occur in the area beneath the sampling manifold. During the test period, all pots in the potroom group shall be operated such that emissions are representative of normal operating conditions in the potroom group. 9.0 Quality Control.

9.1 Miscellaneous Quality Control Measures.

Quality Control Section Measure


Ensure accurate measurement of gas flow rate in duct and of sample volume

Ensure accurate and precise measurement of roof monitor effluent gas temperature and flow rate

Minimize negative effects of used acid

8.0, Sampling equipment 10.0 leak-check and

calibration 10.3, Initial and periodic



performance checks of roof monitor effluent gas characterization apparatus

Interference/recovery efficiency check during distillation

9.2 Volume metering System Checks.

Same as Method 5, Section 9.2.

10.0 Calibration and Standardization.

Same as Section 10.0 of either Method 13A or Method 13B, as applicable, with the addition of the following:

10.1 Manifold Intake probe nozzles.

The manifold intake probe nozzles shall be calibrated when the manifold system is installed or, alternatively, the manifold may be preassembled and the probe nozzles calibrated on the ground prior to installation. The following procedures shall be observed:

10.1.1 Adjust the exhaust fan to draw a volumetric flow rate (refer to Equation 14-1) such that the entrance velocity into each manifold probe nozzle approximates the average effluent velocity in the roof monitor.

10.1.2 Measure the velocity of the air entering each probe nozzle by inserting a standard pitot tube into a 2.5 cm or less diameter hole (see Figure 14-2) located in the manifold between each blast gate (or valve) and probe nozzle. Note that a standard pitot tube is used, rather than a type S, to eliminate possible velocity measurement errors due to cross-section blockage in the small (0.13 m diameter) manifold leg ducts. The pitot tube tip shall be positioned at the center of each manifold leg duct. Take care to ensure that there is no leakage around the pitot tube, which could affect the indicated velocity in the manifold leg.

10.1.3 If the velocity of air being drawn into each probe nozzle is not the same, open or close each blast gate (or valve) until the velocity in each probe nozzle is the same. Fasten each blast gate (or valve) so that it will remain in position, and close the pitot port holes.

10.2 Initial calibration of Propeller Anemometers.

10.2.1 Anemometers that meet the specifications outlined in Section 6.1.1 need not be calibrated, provided that a reference performance curve relating anemometer signal output to air velocity (covering the velocity range of interest) is available from the manufacturer. If a reference performance curve is not available from the manufacturer, such a curve shall be generated For the purpose of this method, a "reference" performance curve is defined as one that has been derived from primary standard calibration data, with the anemometer mounted vertically. "Primary standard" data are obtainable by: (a) direct calibration of one or more of the anemometers by the National Institute of Standards and Technology (NIST); (b) NIST-traceable calibration; or (c) calibration by direct measurement of fundamental parameters such as length and time (e.g., by moving the anemometers through still air at measured rates of speed, and recording the output signals).

10.2.2 Anemometers having output signals other than electrical (e.g., optical) may be used, subject to the approval of the Administrator. If other types of anemometers are used, a reference performance curve shall be generated, using procedures subject to the approval of the Administrator.

10.2.3 The reference performance curve shall be derived from at least the following three points: 60 ± 15, 900 ± 100, and 1800 ± 100 rpm.

10.3 Initial Performance Checks.

Conduct these checks within 60 days before the first performance test.

10.3.1 Anemometers.

A performance-check shall be conducted as outlined in Sections through Alternatively, any other suitable method that takes into account the signal output, propeller condition, and threshold velocity of the anemometer may be used, subject to the approval of the Administrator. Check the signal output of the anemometer by using an accurate rpm generator (see Figure 14-3) or synchronous motors to spin the propeller shaft at each of the three rpm settings described in Section 10.2.3, and measuring the output signal at each setting. If, at each setting, the output signal is within 5 percent of the manufacturer's value, the anemometer can be used. If the anemometer performance is unsatisfactory, the anemometer shall either be replaced or repaired. Check the propeller condition, by visually inspecting the propeller, making note of any significant damage or warpage; damaged or deformed propellers shall be replaced. Check the anemometer threshold velocity as follows: With the anemometer mounted as shown in Figure 14-4(A), fasten a known weight (a straight-pin will suffice) to the anemometer propeller at a fixed distance from the center of the propeller shaft. This will generate a known torque; for example, a 0.1-g weight, placed 10 cm from the center of the shaft, will generate a torque of 1.0 g-cm. If the known torque causes the propeller to rotate downward, approximately 90E [see Figure 14-4(B)], then the known torque is greater than or equal to the starting torque; if the propeller fails to rotate approximately 90E, the known torque is less than the starting torque. By trying different combinations of weight and distance, the starting torque of a particular anemometer can be satisfactorily estimated. Once an estimate of the starting torque has been obtained, the threshold velocity of the anemometer (for horizontal mounting) can be estimated from a graph such as Figure 14-5 (obtained from the manufacturer). If the horizontal threshold velocity is acceptable [<15 m/min (50 ft/min), when this technique is used], the anemometer can be used. If the threshold velocity of an anemometer is found to be unacceptably high, the anemometer shall either be replaced or repaired.

10.3.2 Recorders and Counters.

Check the calibration of each recorder and counter (see Section 6.1.2) at a minimum of three points, approximately spanning the expected range of velocities. Use the calibration procedures recommended by the manufacturer, or other suitable procedures (subject to the approval of the Administrator). If a recorder or counter is found to be out of calibration by an average amount greater than 5 percent for the three calibration points, replace or repair the system; otherwise, the system can be used.

10.3.3 temperature Measurement Apparatus.

Check the calibration of the temperature Measurement Apparatus, using the procedures outlined in Section 10.3 of Method 2, at temperatures of 0, 100, and 150 EC (32, 212, and 302 EF). If the calibration is off by more than 5 EC (9 EF) at any of the temperatures, repair or replace the apparatus; otherwise, the apparatus can be used.

10.4 Periodic Performance Checks.

Repeat the procedures outlined in Section 10.3 no more than 12 months after the initial performance checks. If the above systems pass the performance checks (i.e., if no repair or replacement of any component is necessary), continue with the performance checks on a 12-month interval basis. However, if any of the above systems fail the performance checks, repair or replace the system(s) that failed, and conduct the periodic performance checks on a 3-month interval basis, until sufficient information (to the satisfaction of the Administrator) is obtained to establish a modified performance check schedule and calculation procedure.

NOTE: If any of the above systems fails the 12-month periodic performance checks, the data for the past year need not be recalculated. 11.0 Analytical Procedures. 13B. Same as Section 11.0 of either Method 13A or Method

12.0 Data Analysis and Calculations.

Same as Section 12.0 of either Method 13A or Method 13B, as applicable, with the following additions and exceptions:

12.1 Nomenclature A = Roof monitor open area, m2 (ft2). Bws = Water vapor in the gas stream, portion by volume.

Cs =

Dd = Dn =

Average fluoride concentration in roof monitor air, mg F/dscm (gr/dscf). Diameter of duct at sampling location, m (ft). Diameter of a roof monitor manifold probe nozzle, m (ft).

F = Emission Rate multiplication factor, dimensionless.

Ft =

Md = Prm =

Total fluoride mass collected during a particular sub-run (from Equation 13A-1 of Method 13A or Equation 13B-1 of Method 13B), mg F- (gr F-).

Mole fraction of dry gas, dimensionless. Pressure in the roof monitor; equal to barometric pressure for this application.

Qsd =

Trm = vd =

vm =

vmt =

Average volumetric flow from roof monitor at standard conditions on a dry basis, m3/min. Average roof monitor temperature (from Section 8.2), EC (EF). Desired velocity in duct at sampling location, m/sec.

Anticipated average velocity (from Section 8.1.1) in sampling duct, m/sec. Arithmetic mean roof monitor effluent gas velocity, m/sec.

vs = Actual average velocity in the sampling duct (from Equation 2-9 of Method 2 and data obtained from Method 13A or 13B), m/sec.

12.2 Isokinetic Sampling Check.

12.2.1 Calculate the arithmetic mean of the roof monitor effluent gas velocity readings (vm) as measured by the anemometer in the section of the roof monitor containing the sampling manifold. If two or more sub-runs have been performed, the average velocity for each sub-run may be calculated separately.

12.2.2 Calculate the expected average velocity (vd) in the duct, corresponding to each value of vm obtained under Section 12.2.1, using Equation 14-1.

vd '

8 D n2 v m


Eq. 14-1

where: 8 = number of required manifold probe nozzles. 60 = sec/min. 12.2.3 Calculate the actual average velocity (vs) in

the sampling duct for each run or sub-run according to Equation 2-9 of Method 2, using data obtained during sampling (Section 8.0 of Method 13A).

12.2.4 Express each vs value from Section 12.2.3 as a percentage of the corresponding vd value from Section 12.2.2. If vs is less than or equal to 120 percent of vd, the results are acceptable (note that in cases where the above calculations have been performed for each sub-run, the results are acceptable if the average percentage for all sub-runs is less than or equal to 120 percent). If vs is more than 120 percent of vd, multiply the reported emission rate by the following factor:

F'1% vd

100vs &120

Eq. 14-2

12.3 Average Velocity of Roof Monitor Effluent Gas.

Calculate the arithmetic mean roof monitor effluent gas velocity (vmt) using all the velocity or volumetric flow readings from Section 8.1.2.

12.4 Average temperature of Roof Monitor Effluent Gas.

Calculate the arithmetic mean roof monitor effluent gas temperature (Tm) using all the temperature readings recorded in Section 8.2.

12.5 Concentration of Fluorides in Roof Monitor Effluent Gas.

12.5.1 If a single sampling train was used throughout the run, calculate the average fluoride concentration for the roof monitor using Equation 13A-2 of Method 13A.

12.5.2 If two or more sampling trains were used (i.e., one per sub-run), calculate the average fluoride concentration for the run using Equation 14-3:

C s ' jn i'1

Vm(std) i

Eq. 14-3

Eq. 14-4

F jn t



where: n = Total number of sub-runs. 12.6 Mole Fraction of Dry Gas.

Md '1&Bws

12.7 Average Volumetric flow Rate of Roof Monitor Effluent Gas. Calculate the arithmetic mean volumetric flow rate of the roof monitor effluent gases using Equation 14-5.

13.0 14.0 15.0 16.0

Qsd ' K1vmt Md Pm A Trm

where: K1 = 0.3858 K/mm Hg for metric units,

= 17.64 ER/in. Hg for English units. Method Performance. [Reserved] Pollution Prevention. [Reserved] Waste Management. [Reserved] References.

Eq. 14-5

Same as Section 16.0 of either Method 13A or Method 13B, as applicable, with the addition of the following:

1. Shigehara, R.T. A Guideline for Evaluating Compliance Test Results (Isokinetic Sampling Rate Criterion). U.S. Environmental Protection Agency, Emission Measurement Branch, Research Triangle Park, NC. August 1977. 17.0 Tables, Diagrams, flowcharts, and Validation Data.

Figure 14-1. Roof Monitor Sampling System.

Figure 14-2. Sampling Manifold and probe nozzles.

Figure 14-3. Typical RPM Generator.

Figure 14-4. Check of Anemometer Starting Torque.

Figure 14-5. Typical Curve of Starting Torque vs. Horizontal Threshold Velocity for Propeller