Environ. Sci. Technol. 2002, 36, 2048-2053
Continuous Real-Time Monitoring of Phosphine Concentrations in Air Using Electrochemical Detectors Interfaced by Radio Telemetry TOMMY G. THORN, JR.,* EDWARD M. CHODYNIECKI, KENNETH W. INGOLD, GERALD A. LONG, CHARLES D. MILLER, EDWARD A. ROBINSON, F. SCOTT COWAN, AND ROBERT L. THOMAS Lorillard Tobacco Co., 200 Kentuck Road, Danville, Virginia 24540
This work involves the novel use of a radio telemetrybased system that continuously monitors phosphine using two different types of electrochemical detectors (ECD/ RT). The ECD/RT units were used to monitor phosphine inside and at varying distances from large tobacco storage warehouses. A master controller unit transferred the data to a personal computer that received and displayed the data. Supervisory control and data acquisition software assimilated the data from each ECD/RT unit, displayed and updated it as new transmissions were received, and stored the data in secure databases. Phosphine concentrations outside five warehouses simultaneously under fumigation and at the facility boundaries were 25%) that are dependent on the type of handoperated sampling pump used, sampling technique, and subjectivity of color intensity. These variables may skew the results obtained from colorimetric tubes. The ECD/RT system eliminates these variables because the electrochemical detectors are automated, provide results via digital readout, and can be calibrated. Data gathered during fumigations performed in 2001 are presented.
Experimental Section This study was conducted in and around large tobacco storage warehouses located in Danville, VA. All structural measurements are approximate. Each corrugated aluminum warehouse measured 91 × 34 × 12 m high at the roof peak. The total warehouse volume was 30752 cubic meters. Two large doors were located at the center of the 91-m sides. Small personnel doors were located at the north and south ends of the warehouses. Roof vents were at the top of the warehouses. Fumigation procedures were performed according to the Fumi-Strip fumigant label (10). All doors and roof vents were sealed with 6-mil polyethylene and secured with tape and adhesive. Magnesium phosphide plates were placed inside the warehouse at a dosage rate of 1 g per cubic meter. Phosphine monitoring was initiated at least 2 h prior to application of the Fumi-Strip fumigant, thus allowing zeroing and calibration of the sensors with no analyte present and the establishment of a baseline response. A limited-access, secure control room (CR) was utilized to house the personal computers and essential stationary 10.1021/es015774+ CCC: $22.00
2002 American Chemical Society Published on Web 04/29/2002
instrumentation. The CR was located a minimum of 91 m from the nearest warehouse being fumigated. Radio communication between personnel in the field performing calibrations and personnel in the CR allowed the confirmation of calibration data observed at the ECD/RT unit and the displayed phosphine concentration in the SCADA software. Constant temperature conditions were maintained inside the CR to ensure proper functioning of instrumentation therein. Analytical Methods. Continuous monitoring during fumigation of large tobacco storage warehouses was conducted using ECD/RT units (Gastronics True Wireless gas transmitters; Gastronics Inc., 23660 Miles Road, Suite 110, Bedford Heights, OH 44128). The ECD/RT units used in this study were constructed of a weatherproof housing which contained a 5-W UHF radio on a licensed frequency (463.925 megahertz), a rechargeable 12 V direct current battery, and a circuit board with an external keypad interface. The ECD/RT units can be recharged using either a solar panel or 115 V alternating current via the circuit board through sealed couplings. An antenna was mounted onto the exterior of the gas transmitter enclosure, which allowed radio transmissions of data from the unit. Two different Sensoric 4-20 milliampere (mA) output ECDs (ATMI GmbH Sensoric Division, Justus-von-Liebig Strasse 22, D-53121 Bonn, Germany) were mounted onto the RT unit. The RT unit was comprised of a weatherproof enclosure with keypad and display and a solar panel. Lowlevel detectors (AsH3 3E1) were used for measuring outside the warehouses. The phosphine concentration alarm levels for the low-level ECD/RT units were set at 0.1 and 0.2 ppm to enable rapid response to levels approaching 0.3 ppm. The concentration above which a data transmission was prompted by the sensors was set to 0.03 ppm. High-level detectors (CO 3E 300) were used for monitoring phosphine in warehouse air. These high-level ECD/RT units were mounted on the outside of the warehouse at either the north or south end, depending on power connection availability. The actual sensor was placed inside the warehouse air via a sealed tube. One ECD/RT unit was used to monitor each of the five warehouses for a total of five separate units. The concentration alarms were set at the maximum level for the high-level detectors. A master controller remote terminal unit (Gastronics, Inc.) was configured to receive the data from each of forty ECD/ RT units. The MC/RTU was connected via an RS-232 connection to a communications port on a desktop personal computer. The PC had supervisory control and data acquisition (SCADA) software (Events Version 5.19; Roper Associates, 9477 Greenback Lane #527, Folsom, CA 95630) and MC/RTU programming software (Version 5.12; Gastronics Inc.) installed. The RTU programming software was used to configure the MC/RTU for receiving and transmitting data to the remote units. The SCADA software was designed to import the MC/ RTU program file so that each data point can be stored in individual databases. Data were archived daily. Phosphine concentration data, expressed as parts per million (ppm), were received from the MC/RTU each minute for each ECD/ RT unit and stored automatically. Also, when a phosphine concentration increase or decrease of greater than 0.03-ppm was detected, data were transmitted to the MC/RTU controller and stored on the PC. This transfer and storage of data allowed for the comprehensive collection of phosphine concentration data from each ECD/RT unit, as well as the simple calculation of phosphine concentration decay data from the five high level units inside each warehouse. Another aspect of the SCADA software is the on-screen display. Data being received by each ECD/RT unit are displayed on various forms appearing on the PC monitor. The information displayed on a PC monitor included alarms, current con-
TABLE 1. Span Data from Three Different ECD/RT Units PH3 concentration observeda sensor date
A
B
C
6/4/01b
0.26 0.26 0.27 0.26 0.29 0.30
0.27 0.26 0.29 0.30 0.27 0.33
0.28 0.28 0.28 0.28 0.30 0.30
6/5/01b 6/8/01b 6/11/01b 6/13/01b 6/15/01c
a Tolerance is ( 0.02 ppm. Values outside the specified range indicate need for calibration. b Calibration standard ) 0.28 ppm PH3/N2. c Calibration standard ) 0.32 ppm PH3/N2.
centration data, battery voltages, and transmission receipts using a time/date stamp. This characteristic of the software facilitates monitoring of all systems from one central location and quick response to communication failures, elevated phosphine levels, and battery failures. The MC/RTU system was equipped with an Autodialer capable of dialing preset telephone numbers when any ECD/ RT unit reaches an alarm condition. Possible alarm conditions include low battery voltage, variation of phosphine concentrations exceeding a set limit, and communication failure at a particular ECD/RT unit. Personnel in the control room verified the readiness of the Autodialer by ensuring that the concentration announced by the Autodialer alarm corresponded to the concentrations observed in the Events software and at ECD/RT unit. If the proper Autodialer response was not given, the source of the problem was investigated and corrected. Validation. Certified gas mixtures of 0.28-ppm and 0.32ppm phosphine in nitrogen (Spectra Gases, Inc., 3434 Route 22 W, Branchburg, NJ 08876) were used to calibrate the lowlevel ECDs. Certified gas mixtures of 771-pm, 262-ppm, 104ppm, 52-ppm, 26.2-ppm, and 10.4-ppm phosphine in nitrogen (Spectra Gases, Inc.) were used for high-level ECD/ RT unit calibration. ECD/RT units were calibrated prior to study initiation. Thereafter, ECD/RT unit readiness was assessed at least once every 72 h until the study was terminated. Phosphine sensor readiness was determined by zeroing the unit and passing a certified phosphine standard over the sensor (spanning) at a rate of 0.5 L/min until the response stabilized (approximately 2 min). The value displayed by the sensor was then compared to the nominal value of the standard being used. The low-level ECD/RT units were calibrated when the observed concentration varied more than 0.02 ppm from the nominal value of the phosphine standard used. The high-level ECD/RT units were calibrated at different levels during the study depending on the concentration inside the warehouse. The calibration range for the high-level ECD/RT units was ( 10% of the nominal value of the phosphine standard. The ECD/RT units were spanned at six different times during fumigation. Table 1 provides data from three different lowlevel units for the six spanning/calibration points during fumigation. These data are typical of data collected from each of 35 sensors used during the study. Each of the data points is within 0.02 ppm of the nominal value of the calibration standard used. These same phosphine standards were used to calibrate field-portable micro gas chromatographs (MGCs) (11). The MGCs (Varian Inc., 2700 Mitchell Dr., Walnut Creek, CA 94598) were equipped with thermal conductivity detectors (TCDs) and PoraPLOT Q columns. These Micro GCs were used to monitor phosphine concentrations inside the five warehouses fumigated during this study. Colorimetric tubes (Dra¨ger-Tubes, phosphine 50/a (15-1000 ppm): phosphine; 0.01/a (0.01−0.3 ppm) Dra¨ger VOL. 36, NO. 9, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2049
FIGURE 1. Distribution diagram showing ECD/RT units around a warehouse being fumigated and around property boundaries. Safety Inc., 101 Technology Dr., Pittsburgh, PA 15275-1057) were used along with the MGC data to compare with data from the ECDs. Air was sampled with the Dra¨ger-Tubes using Dra¨ger Model 31 pumps (Dra¨ger Safety Inc.). Polyethylene tubing (6-mm) was used to facilitate monitoring the internal warehouse phosphine concentrations with Dra¨ger-Tubes. The polyethylene tubing was sealed with tape or a gastight fitting when not in use. The tube extended into the warehouse just enough to allow a representative sample to be pulled through the tube. Historical data from studies conducted at our facility show a 12-h equilibration time between phosphine concentrations in the warehouse air and the concentrations inside the tobacco bales. Therefore, for the purposes of fumigant efficacy, the concentrations inside the bales were
not monitored during this study. Only the warehouse air was monitored. The equilibration time was factored into the fumigation duration to ensure adequate fumigation of the product. Monitoring Studies. The ECD/RT units used for monitoring fugitive emissions of phosphine from warehouses under fumigation were deployed around the Danville facility based on the recommendations of an experienced consultant (12). Figure 1 illustrates a general view of a representative warehouse being fumigated and the distribution of the ECD/ RT units around the property boundaries and the warehouse. The monitors were positioned so that an overall view of phosphine movement around the facility could be determined. This helped to decrease reaction times to elevated levels of phosphine detected. A comparison between the ECD/RT technique and colorimetric tubes was made as part of this study. An ECD/ RT unit located approximately 4 m downwind of a warehouse under fumigation was selected. The sensor was fixed approximately 1 m above the ground. Three low-level colorimetric tubes were used during three different 20-minute monitoring periods. The concentration range for the colorimetric tubes was 0.1-1 ppm. The number of strokes made with the Model 31 pump was 10 (pump volume ) 100 mL). A time interval of 2 min was used between each stroke to allow maximum expansion of the pump bellows. A paper sleeve was used to enclose the body of the colorimetric tube so that ambient sunlight would not discolor the tubes. The weather data collected during this study show winds predominately out of the west with speeds as high as 482 m/minute. The average temperature was 24 °C. The relative humidity ranged from 26 to 100%.
Results and Discussion The phosphine concentration from the colorimetric tube was a single data point read directly from the tube at the end of the sampling period. With colorimetric tube monitoring,
FIGURE 2. Comparison between an ECD/RT unit and a colorimetric tube over a 20-minute period. 2050
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 9, 2002
FIGURE 3. Data from ECD/RT unit A located 3 m (downwind) from a warehouse being fumigated.
FIGURE 4. Data from a single ECD/RT unit and a Micro Gas chromatograph (MGC) monitoring phosphine inside a warehouse under fumigation. sporadic increases in phosphine concentrations are not addressed because they are not detected because colorimetric tubes give an average phosphine concentration over the sampling period. Figure 2 illustrates the variance in data collected over a 20-minute period at a single ECD/RT unit. The average phosphine concentration was 0.13 ppm over the 20-minute sampling interval, which was greater than the
0.05-ppm value obtained during the same 20-minute sampling interval from the colorimetric tube. This shows significant personal exposure when monitoring phosphine concentrations outside fumigated warehouses with colorimetric tubes. Figure 3 illustrates data collected from a single ECD/RT unit during the study. Phosphine concentration data g0.1 VOL. 36, NO. 9, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2051
FIGURE 5. Natural logarithm (ln) plot of decay data (from the highest concentration to just prior to abatement) from a single ECD/RT unit monitoring phosphine inside a warehouse under fumigation.
TABLE 2. Comparison among ECD/RT High-Level Sensor, Micro GC, and Dra1 ger-Tubes Phosphine Concentration Data from a Warehouse under Fumigation ECD/RT micro GC Dra1 ger-Tube data (ppm) data (ppm) data (ppm)
date 6/5/01 6/6/01 6/8/01 6/9/01 6/10/01 6/11/01 6/12/01 6/13/01 6/14/01 6/16/01
593 510 381 318 276 246 215 189 154 106
548 537 349 308 260 235 216 154 145 141
585 498 300 280 175 180 205 120 120 50
av (x) 575 515 343 302 237 220 212 154 140 99
SD % RSD (σ) (σ/x)*100 24 20 41 20 54 35 6 35 18 46
4 4 12 7 23 16 3 22 13 46
ppm prompted response by fumigation personnel to check for leaks around fumigated warehouses. This prompt response to elevated phosphine concentrations can be seen in data collected on 6/6/01. There was an increase in phosphine concentration at the ECD/RT unit approaching 0.2 ppm. Personnel investigated the source of the increase of the fugitive emission. A leak was detected and corrected. Subsequently, the phosphine concentration returned to a concentration below the alarm setpoint. Ten-day data from a high-level sensor monitoring phosphine concentrations inside a warehouse were plotted (Figure 4). The phosphine concentrations were also measured using a micro gas chromatograph (13) and Dra¨ger-Tubes. The data are compared and summarized in Table 2. There was good agreement between at least two of the technologies on a given date. Variances can be attributed to the time of day samples were collected, limitations inherent in the technology, and sampling variance between the personnel acquiring data using the Dra¨ger-Tubes. The ECD/RT system not only provides comprehensive automated real-time monitoring outside warehouses but also allows for internal monitoring of a warehouse under fumi2052
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 9, 2002
gation and obtains a more complete assessment of the phosphine concentration decay rate. The phosphine concentration decay rates allow the assessment of the variance between fumigations (usually once annually). The assumption is made that most of the phosphine is leaving the structure at the rate of phosphine concentration decay and a relatively small amount, if any, is reacting inside the warehouse. The natural logarithm (ln) of selected ECD/RT data was plotted (Figure 5). Specifically, beginning with the highest observed concentration and ending prior to abatement, the data were plotted yielding a negatively sloped line. The slope is taken as the rate of decay of phosphine from that warehouse per day. That number is converted to cubic meters per minute by eq 1:
(
(24slope h/day) 60 min/h
)
(30752 m3/warehouse) ) phosphine decay in m3/min (1)
For the plot given in Figure 5, the slope of the line is -0.1651. So, the phosphine decay rate is 3.52 m3/min (124 cubic feet per minute, cfm) by high-level ECD/RT which was confirmed by a phosphine concentration decay rate of 3.44 m3/min (122 cfm) calculated from data obtained by micro GC (slope ) -0.1614). An increase in the rate of decay could indicate a problem with the warehouse integrity. Repairs can be made during the interim period to improve the gas tightness of a warehouse. Extensive studies in our facilities have determined fumigant decay rate analysis to be the best means of assessing the readiness of warehouses for fumigation. The 40 ECD/RT units positioned around the Danville facility allowed continuous monitoring of phosphine concentrations in air with less manpower than colorimetric tubes and less instrumental expertise required than GC or openpath FTIR. However, some configurations of GCs and other instrumentation can give much lower limits of detection than the ECDs used in this study. The authors sought a
compromise between low detection limits and automation. The automation of the ECD system allowed for a single person to respond to alarms and monitor data collection. Such automation with GCs is possible for monitoring phosphine concentrations inside fumigated warehouses but not practical when attempting to detect phosphine over many acres of property surrounding the warehouses. One of the more important benefits of this computerized monitoring system is the capability to obtain data more rapidly and with fewer manpower resources with less subjectivity and greater accuracy. Data collected using the ECD/RT units were automated real-time results, which were much more comprehensive in characterizing phosphine concentrations in and around structures being fumigated than colorimetric tubes or gas chromatographs. This novel approach to comprehensive monitoring has many potential applications. The automation provided by the system limits the need for active monitoring by personnel in the field, as is the case with colorimetric tubes and other types of manual monitoring. This system can be used over large facilities since transmission distances up to 3 km from the ECD/RT units are possible. Also, this system can be configured with many different sensors types for monitoring various compounds. The system can be integrated with a weather station so that simultaneous on-site weather data can be collected and related to the effectiveness of this fumigation.
Acknowledgments The authors thank Dr. Rob Stevens for his work on this project. The authors thank Dr. Chris Coggins, Dr. Jack Reid, Dr. Deb Mereand, and Stacy Stinson all of Lorillard Tobacco for
their critical review of this article. The authors also thank Bud Dungan (Gastronics Inc.) and Ted Roper (Roper Associates).
Literature Cited (1) Childs, D. P.; Overby, J. E.; Watkins, B. J. Tobacco Sci. 1969, 13, 64-69. (2) Ruffieux, L. New Global Fumigation Standard for Tobacco. Presented at the Coresta Subgroup on Pest and Sanitation Management in Stored Tobacco Meeting, Geneva, Switzerland, November, 2000. (3) Schaefer, G. J.; Newton, P. E. Inhal. Tox. 1998, 10, 293-320. (4) Product label, Fumi-Cel and Fumi-Strip, EPA Registry Number 40285-8. (5) Ducom, P.; Bourges, C. Proc.- 4th Int. Work. Conf. Stored Prod. Prot. 1986, 630-635. (6) Leesch, J. G. J. Econ. Entomol. 1982, 75, 899-905. (7) Thorn, T. G.; Marshall, T. L.; Chaffin, C. T. Field Anal. Chem. Technol. 2001, 5 (3), 116-120. (8) Thorn, T. G.; Stevens, R. D.; Segers, D. P.; Macon, S. T. J. Proc. Anal. Chem., in press. (9) Leesch, J. G. J. Econ. Entomol. 1982, 75, 899-905. (10) Product label, Fumi-Cel and Fumi-Strip, EPA Registry Number 40285-8. (11) Robinson, E. A.; Stevens, R. D.; Mereand, D. L.; Long, G. A.; Thorn, T. G. Environ. Sci. Technol., to be published, 2001. (12) Bowman, J. T. Bowman Enterprises, private communication, 2001. (13) Robinson, E. A.; Stevens, R. D.; Mereand, D. L.; Long, G. A.; Thorn, T. G. Environ. Sci. Technol., to be published, 2001.
Received for review November 2, 2001. Revised manuscript received February 1, 2002. Accepted February 13, 2002. ES015774+
VOL. 36, NO. 9, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2053
