3M FORMALDEHYDE MONITOR (MODEL 3721)

Introduction

A field and laboratory evaluation of the 3M model 3721 formaldehyde monitor was conducted by the OSHA Health Response Team (OSHA-HRT) and the OSHA Analytical Laboratory. The model 3721 monitor had not, to our knowledge, been thoroughly tested at the time of our evaluations (Field Study - 1986, Lab Study - 1987-88). A preliminary study (9.1.) had been performed using a model 3721 prototype; however, a different sampling rate than what is now recommended was used for calculations.

Previous studies (9.1.-9.3.) of formaldehyde passive sampling and analysis were conducted with the model 3751 monitor. The OSHA permissible exposure limit (PEL) was a 3-ppm time weighted average (TWA) at the time of these studies. The model 3751 monitor was replaced by 3M with the model 3721 monitor. Design changes introduced in the new monitor are Section 1. The OSHA PELs for formaldehyde were changed in are summarized:

These changes signaled a need to re-examine the performance of this monitor. This re-examination was accomplished using two studies:

Field Study:   Samples were collected by the OSHA-HRT at various wood products industries. Side-by-side samples were taken at field sites to compare the 3M sampling and analytical method (9.5., 9.6.) with OSHA method no. 52 (9.7.) and to determine occupational exposure to formaldehyde in wood products manufacturing. Personal and area samples were taken. The manufacturing processes sampled included assembly of particleboard, medium density fiberboard, hardwood and softwood plywood, and paneling.

Lab Study:   This portion of the product evaluation was conducted at the OSHA Laboratory to determine if the re-designed 3M monitor is capable of sampling accurately and precisely at different humidity and concentration levels.

1. History

The simplicity and freedom of the 3M formaldehyde passive monitor (model 3751) showed promise when first offered in 1981 as an industrial hygiene sampling alternative for formaldehyde (9.8.); however, subsequent independent studies indicated analyte loss when sampling at low humidities (9.1., 9.3.). Consequently, the model 3751 monitor was removed from the market by 3M in April, 1984. The model 3721 3M monitor, capable of sample humidification, was introduced in 1985 as a replacement. The changes instituted by 3M and incorporated into the model 3721 are:

1. A water saturated pad in the bottom section of the monitor has been added for sample humidification.
2. Each monitor is now packaged in a sealed metal container. Previously, the model 3751 monitor was enclosed in a resealable plastic bag.
3. The calculated sampling rate has been changed from 0.0659 liters per minute (L/min) to 0.0614 L/min*.

With the exception of the moisturizing pad, the model 3721 appears physically identical to the model 3751 monitor.

The model 3751 monitor has been extensively evaluated by independent laboratories (9.1.-9.3.). Results from these studies did not indicate serious problems with desorption efficiency, face velocity, reverse diffusion, or post-collection sample storage stability. The recent modifications instituted by 3M do not suggest sampling performance would be significantly affected in these areas. As long as the face velocity of the sampled environment is above 3.1 to 4.6 m/min (10 to 15 ft/min), the sampling rate of the monitor does not appear to be significantly altered (9.1.-9.3. , 9.9.).

Sampling and analytical procedures are identical for either model monitor; however, result calculations are different since slightly different sampling rates are used.

2. Principle of the 3M Monitor

The 3M formaldehyde monitor is a diffusion-type air monitoring assembly. The monitor is worn in the breathing zone of personnel in order to evaluate potential exposure to formaldehyde vapors contained in the atmosphere. The formaldehyde vapor passes through a diffusion barrier and is adsorbed on bisulfite-impregnated paper contained within the monitor.

The formaldehyde collected by the monitor is laboratory analyzed by desorbing the formaldehyde-bisulfite adduct from the treated paper with formaldehyde-free deionized water. Chromotropic acid and sulfuric acid are added to an aliquot of the sample to form a purple mono-cationic chromogen. The absorbance of this colored solution is read in a spectrophotometer at 580 nm and is proportional to the amount of formaldehyde collected.

3. Evaluation Criteria

Field Study:   The monitor lot numbers used during the field study were unavailable. All monitors were tested before their expiration dates. The field experiments involved taking side-by-side personal and area samples with 3M monitors and treated XAD-2 solid sorbent contained in glass tubes. These sampling tubes are further described in reference 9.7.

Personnel and area samples were taken in various occupational environments where exposures to formaldehyde were near 1 ppm as a TWA. Exposure levels were estimated using short-term detector tubes. A portable, gas-phase infrared analyzer was used during the beginning of the study to estimate exposures but was soon discontinued due to difficulties with positive interferences. The wavelength recommended for analyzing formaldehyde with the infrared analyzer measures C-H band stretching. Other organic compounds in the sampled air tended to provide a positive interference at this wavelength.

Results for the 3M monitor were compared to the treated XAD-2 sampling device using linear regression statistics.

Lab Study:   The laboratory experiments were mainly concerned with:

1. Comparing the performance of the monitors near the 1 ppm PEL with a reference method
2. Assessing performance in a low humidity environment (<50% RH)
3. Taking STEL measurements (approximately 2 ppm for 15 min)

The 3M sampling and analytical method (9.5., 9.6.) was compared side-by-side at the OSHA Laboratory to OSHA method no. ID-102 (9.10., 9.11.). This method uses 10% methanol in midget fritted glass bubblers (MFGB) as the sampling device.

Monitors obtained from 3M for the lab portion of the evaluation were from lots 6125 002 (exp. date 10/20/87) and 8092 001 (exp. date 10/6/89). All monitors were exposed and analyzed in the lab before their expiration dates. The former lot was tested about 1 week before the expiration date. (Note:   A previous study (9.3.) of the 3751 monitor uncovered a shelf-life problem where the bisulfite-treated pad appeared to dehydrate before the expiration date. If the monitors were not used soon after manufacture, the impregnated pad would not capture formaldehyde effectively in the dehydrated state.)

Monitor performance was determined using formaldehyde test atmosphere concentrations of approximately 0.4, 1, 1.5, 3, and 5 × PEL and calculated using the OSHA Laboratory Inorganic Method Statistical Evaluation Protocol (9.12.). Testing was mainly conducted for time-weighted average (TWA) determinations. One experiment to determine the ability of the monitor to sample STEL exposures was also performed at 50% RH. Lot 6125 002 was used for experiments at 1.5, 3, and 5 × PEL while the other lot was used for the remaining concentrations and the STEL experiment. Three different RH levels (30, 50, and 80% RH) were used for each TWA concentration tested. One exception is a test performed at 30% RH and 0.25 × PEL. This concentration was used instead of 0.4 × PEL to evaluate monitor performance under "worst case" conditions (low humidity - low concentration).

In addition, a separate experiment was conducted at the OSHA Laboratory comparing the Field Study reference method with the Lab Study reference method. This experiment was conducted by taking side-by-side treated XAD-2 sorbent and 10% methanol MFGB samples at 50% RH and again at 80% RH.

All bubbler, tube, and monitor samples collected for the Field or Lab Studies were analyzed at the OSHA Lab. All monitor samples were analyzed according to the analytical procedure provided by 3M (9.6.).

4. Formaldehyde Test Atmospheres

Field Study:   The formaldehyde concentration of samples taken in different wood products industries was dependent on the type of adhesive or resin mixture used, amount of production, and press temperature. The source of exposure was mainly from either urea-formaldehyde (UF) or phenol-formaldehyde (PF) resins with the former source predominating in the areas sampled. A variety of industrial operations were sampled and the largest exposures to formaldehyde were confined to areas near wood-pressing operations. The hot presses used to press the various woods together were heated to temperatures ranging from 250 to 450 °F. The formaldehyde concentration in any specific area near the presses varied according to the press operation. When the press was closed and the wood/resin was being heated, the concentration was low; the formaldehyde concentration would increase when the press was opened. Therefore, most formaldehyde concentrations at the field sites were not constant throughout the sampling periods and fluctuated with press operation.

Lab Study:   The apparatus used to generate dynamic test atmospheres of formaldehyde for lab tests is shown as a block diagram in Figure 1. The system consisted of five essential elements:

1. A vapor generating chamber (permeation oven)
2. A humidity control system
3. A mixing manifold
4. An aluminum passive monitor exposure channel
5. A sampling manifold for active samplers

The formaldehyde was generated using permeation tubes containing reagent grade paraformaldehyde. The permeation device used for generation of formaldehyde vapor was a model 450 Dynacalibrator (VICI/Metronics, Santa Clara, CA). The permeation device chamber was maintained at a constant temperature of 110 °C. Concentrations of formaldehyde in the generation system were preliminarily determined by the weight change of the permeation tube. Weight changes are shown below:

Purified air for contaminant dilution was prepared by flowing compressed air through a particulate filter, a silica gel bed, and a charcoal bed. Humidified air for dilution was produced by flowing the purified air into a flow, temperature, and humidity control system (Miller-Nelson Research Inc., model HCS-301).

A controlled formaldehyde concentration was produced by taking filtered room air and passing it over the permeation tube, and then mixing with the temperature-controlled (20 to 25 °C) humidified air in a glass mixing manifold.

The contaminant-air mixture then entered a Teflon sampling manifold and eventually a calibrated dry test meter (Singer Co., model no. DTM 115) for flow rate measurement. The flow rate of the formaldehyde vapor to the mixing chamber was approximately 0.17 L/min. Depending on the final concentration desired, the flow rate range for dilution air was 5 to 25 L/min.

An aluminum exposure channel containing six openings, obtained from 3M and also used in a previous OSHA study (9.2.), was connected to an arm of the sampling manifold. This channel, as shown in Figure 2, was used to test 5 monitors per experiment. The sixth port was sealed using an empty monitor.

Monitor results were compared to the reference method sample results to determine recovery ratios. The bias and overall error in a previous study of the reference method (9.10., 9.11.) were +2.9% and ±19% respectively. Theoretical results using the weight change of permeation tubes were not used in the statistical analysis due to a small leak that developed in the generation system during low concentration testing (tests done at <1 × PEL) . This leakage altered the actual generated concentration and apparently was a result of the increased dilution volume necessary for low concentration testing.

5. Sampling

Field Study:   The field sampling technique provided by 3M (9.5.) was used for taking monitor samples. Full shift samples were taken. Tubes containing XAD-2 treated with 2-(hydroxymethyl) piperidine were taken with Du Pent P125 sampling pumps at a flow rate of approximately 0.050 L/min. For personal measurements, a monitor and a treated XAD-2 sampling device were placed within 2 inches of each other on the lapel of the worker's clothing. Area samples were placed in strategic areas near the operation with each paired monitor and tube sample also being within 2 inches of each other. Samples were sealed after sampling and submitted to the OSHA Lab for analysis.

Lab Study:   The monitor and reference samples were exposed using the following step-by-step procedure:

1. Adjust diluent flow rates in the dynamic generation system to determine an approximate generated HCHO concentration. Adjust the system to deliver either 30, 50, or 80% RH.
2. Record the sample number and the exposure start time on the back of each monitor.
3. Before exposing, wrap parafilm around the edge of each monitor to prevent leakage between each monitor and the exposure channel.
4. For the reference samples, calibrate personal sampling pumps to approximately 0.5 L/min sampling rates. Add approximately 15 mL of 10% methanol (MeOH) to midget fritted glass bubblers (MFGB).
5. Place the monitors in the exposure channel and attach MFGBs to the sampling manifold. Simultaneously expose the bubblers and the monitors.
6. At the conclusion of the sampling period, immediately remove and discard the white plastic film and purple retaining ring from the top of the monitor. Also remove the bottom portion which contains the metal clip.
7. Firmly snap the closure caps onto the monitor faces.
8. Place the MFGB solutions into 25-mL volumetric flasks and dilute to the mark with 10% MeOH. Post-calibrate the sampling pumps.
9. Record the exposure end time on the back of the monitor.
10. Enclose the monitor in the original aluminum container until ready for analysis.

This procedure (a to j) was repeated for each humidity-concentration level tested.

As previously mentioned, reference MFGB samples were taken according the procedures outlined in references 9.10 and 9.11. The MFGB sampling times for the TWA determinations varied between 30 and 240 min. Six MFGB and five monitor samples were tested at each concentration/humidity level. All monitor samples were exposed for 240 min except for the STEL experiment which took 15 min.

The exposure channel inlet was connected to the manifold and the outlet was connected to a Du Pont P4000 sampling pump whose sampling rate was set at 3 to 5 L/min. This flow range produced face velocities inside the channel of 9.2 to 15.3 m/min (30 to 50 ft/min). As shown in Figure 2, the monitor samples are placed in series in the channel such that the formaldehyde mixture contacts the first and then subsequent monitors. If the flow rate through the channel is less than 6 L/min, a "starvation effect" is produced from one monitor to another. The flow rates in this 3M sampling chamber were less than 6 L/min and all results were corrected for this "starvation" using the following equation (9.9.):

Where:

6. Analysis

Field and Lab Studies - 3M Monitors: The analytical procedure used was the NIOSH P&CAM 125 procedure for analysis of formaldehyde (9.13.) which has been slightly modified by 3M. Modifications were apparently instituted to alter the analytical working range and minimize interferences. These modifications include:

1. A smaller total sample volume (3 mL) and greater sample/aliquot ratio (3/2) are used.
   
2. A change from a ratio of 4.0: 0.1: 6.0 mL (NIOSH method) to 2.0: 1.0: 5.0 mL (3M monitor method) for the analytical reagents listed below.

   sample: 1% chromotropic acid: concentrated sulfuric acid

The formaldehyde adsorbed on the bisulfite-impregnated adsorbent was analyzed according to the procedure described below (9.6.):

1. The formaldehyde-bisulfite adduct was desorbed by adding 3 mL of deionized water (DI H₂O) through the center port of the elutriation cap (outer cap) using a pipette. The port was re-sealed.
2. The sample was allowed to sit for 30 min with occasional gentle agitation. A 2-mL aliquot of the solution was then transferred into a 20-mL screw-cap glass vial.
3. A purple color was produced in this aliquot after addition of 1 mL of 1% chromotropic acid solution along with slow and careful addition of 5 mL concentrated sulfuric acid. This mixture was agitated and then allowed to cool to room temperature.
4. The absorbance was measured with a UV-VIS Model Spectronic 100 single beam spectrophotometer (Bausch & Lomb, Rochester, NY) or a double beam CARY 219 UV-VIS spectrophotometer (Varian Associates, Palo Alto, CA) at 580 nm using a 1-cm cell.

Field Study - Solid sorbent samples: The resulting derivative [formaldehyde plus 2-(hydroxymethyl) piperidine] was desorbed from the XAD-2 treated sorbent with toluene and then nitrogen selective detector as specified in reference 9.7.

Lab Study (MFGB samples): The 10% MeOH solutions were buffered, hydrazine was added, and the resulting hydrazone compound was analyzed by square wave polarography. This technique was used instead of the differential pulse polarographic technique mentioned in reference 9.10. and 9.11. because it offered a decrease in analysis time with comparable results.

7. Results

The weight of formaldehyde collected by each monitor was calculated a polynomial regression concentration-response curve. Samples were using blank corrected and the average concentration of formaldehyde in the sampled ambient air was then calculated by (9.6.):

Where:

Since RC   =   1.00 (9.2., 9.6.), equation (1) was simplified to:

Sample results for solid sorbent and bubbler samples were calculated according to their respective methods (9.7., 9.10.) The results of the experiments can be found in these Tables:

When comparing two different methods by linear regression, ideal agreement between methods is displayed if the correlation coefficient and the slope are equal to a value of 1. As shown in Figure 3, the dashed line corresponding to ideal agreement does not differ greatly from the regression line calculated from field results. When considering the slope and the standard deviation of the slope at 95% confidence (0.952 ± 2 × 0.037), the slope does not significantly differ from a value of 1. The correlation coefficient is also close to 1. Therefore, the 3M and XAD-2 sampling media gave similar results throughout the concentration range tested at the field sites. Both sampling devices also gave similar results regardless of the workplace source for formaldehyde or industrial operation.

Average blank values for monitors analyzed during the Field Study were approximately 1 µg.

Lab Study:   Table 2 contains individual monitor results for the three RH levels and five concentrations tested. A positive bias is noted when comparing monitors with MFGBs. As previously mentioned, generation system leakage at low concentrations prevented theoretical results from being used and therefore, monitor results were only compared with MFGB results. Theoretical results at generated concentrations greater than the PEL were available and a comparison with monitor results gave an acceptable overall error (9.12.) of ±21.7%. When comparing the monitor and MFGB recoveries using theoretical values in this concentration range, the MFGB average recovery was 8.7% less than the monitors.

Table 3 contains results for the five monitors taken at the STEL. The STEL measurements indicate somewhat variable results are obtained when sampling for a short duration.

Table 4 contains a summary of Table 2 data and F test results. As shown in Table 4, the monitor displayed a total recovery ratio of 1.064 and thus, a positive overall bias of 6.4%. Figure 4 graphically displays the comparison of recovery ratios for the monitor and MFGB samples. Each point on the graph is an average of five or six sample recovery ratios at a given concentration and RH. In this Figure, ideal agreement between the two methods is shown by the dashed line. Bias is displayed when the solid line (slope of actual measurements) moves towards either axis. Positive bias is shown by the solid line leaning toward the 3M axis.

An F test was used to detect any significant variability in results across the three humidity levels tested. The F test monitor results in Table 4 show a significant difference does exist at the 99% confidence level. An examination of recovery ratios at the different RH levels (Table 2) indicates an enhancement at 80% RH. Recovery ratios were approximately 10% higher at 80% RH. A shift in performance due to humidity was not noted for the reference method.

Blank values for monitors analyzed during the Lab Study ranged from 0.91 to 3.39 µg with a CV of 0.55. If the 3.39 µg measurement is omitted, the blank values are much closer with a range of 0.91 to 1.5 µg and a CV of 0.21.

The comparison of the treated XAD-2 sampling device with MFGBs showed agreement between the two methods (Table 5). A slight positive bias of the XAD-2 device over the MFGB is also shown. This bias tended to increase with increasing RH.

8. Summary

The 3M model 3721 monitor and selected reference methods gave similar formaldehyde results in field and lab experiments when samples were taken for at least 240 min. The field study indicates the monitor results were not significantly different from the treated XAD-2 sampling device. A positive bias was noted when the monitor was compared in the Lab Study to the 10% methanol sampling device; however, this bias can be accounted for since the MFGB method appears to be giving results approximately 5 to 10% lower than either the monitor or the treated XAD-2 sampling device. At high humidity (80% RH), tests in the laboratory indicated the monitor gave a slight positive bias when compared to the MFGBs.

Short-term sampling with the monitor may give highly variable results. This variability could be due to insufficient equilibration time or to the decreased amount of analyte taken. The time necessary to passivate the surfaces of the monitor and to achieve a homogeneous sampling rate may be longer than 15 min. Any analytical method has difficulty when measuring amounts of analyte near the method's detection limit. The total amount of formaldehyde generated at the STEL was approximately 2 µg. This amount is near the analytical detection limit described by 3M (9.9.). Even if this amount is detectable using the procedure described, the blank values noted in the Lab Study were approximately 1 to 3 µg. If the blank value is not consistent, the measurement of any sample near the blank value can become erratic and inaccurate.

As previously mentioned, during the Field Study the concentration of formaldehyde usually fluctuated over the sampling period. The variability in concentration did not appear to affect the ability of the monitor to sample. These results agree with a previous study (9.2) performed using fluctuating formaldehyde concentrations under laboratory conditions.

Concentrations and humidities used to test the monitor were in the range normally expected in workplace situations. Overall results indicate the monitor gave similar results to the 10% methanol MFGBs and treated XAD-2 sampling devices and is an alternative for formaldehyde TWA sampling.

9. References

9.1. National Council of the Paper Industry for Air and Stream Improvement Inc.: A Laboratory Evaluation on the Performance of Passive Diffusion Badge Monitors and Detector Tubes for Determination of Formaldehyde. (Technical Bulletin No. 451), NY: NCASI, 1985.

9.2. Occupational Safety and Health Administration Analytical Laboratory: Evaluation of 3M Formaldehyde Monitors (Model 3751) by J.C. Ku (USDOL/OSHA-SLCAL Product Evaluation no. ID-139). Salt Lake City, UT. 1982 (unpublised).

9.3. Kennedy, E.R. and R.D. Hull: Evaluation of the Du Pont Pro-Tek Formaldehyde Badge and the 3M Formaldehyde Monitor. Amer. Ind. Hyg. Assoc. J. 47:94-105 (1986).

9.4. "Formaldehyde," Code of Federal Regulations Title 29, Pt. 1910, Section 1048. 1988. pp. 992-1028.

9.5. Occupational Health and Safety Products Division/3M: 3M Formaldehyde Monitor #3721 Instructions for Use. St. Paul, MN: 3M Company. No publication date given.

9.6. Occupational Health and Safety Products Laboratory: Organic Vapor Method no. 4D. St. Paul, MN: 3M Company, May, 1985.

9.7. Occupational Safety and Health Administration Analytical Laboratory: OSHA Analytical Methods Manual (USDOL/OSHA-SLCAL Method No. 52). Cincinnati, OH: American Conference of Governmental Industrial Hygienists" (Pub. no. ISBN: 0-936712-66-X), 1985.

9.8. Rodriguez, S.T., P.B. Olson, and V.R. Lund: "Colorimetric Analysis of Formaldehyde Collected on a Diffusional Monitor." Paper presented at Amer. Ind. Hyg. Assoc. Conference, Portland, OR, May 1981.

9.9. Occupational Health and Safety Products Division/3M: 3M Brand Formaldehyde Monitor #3750/3751. St. Paul, MN: 3M Company, Internal document - No publication date given.

9.10. Occupational Safety and Health Administration Analytical Laboratory: OSHA Analytical Methods Manual (USDOL/OSHA-SLCAL Method and Backup Report no. ID-102). Cincinnati, OH: American Conference of Governmental Industrial Hygienists (Pub. no. ISBN: 0-936712-66-X), 1985.

9.11. Septon, J.C. and J.C. Ku: Workplace Air Sampling and Polarographic Determination of Formaldehyde. Amer. Ind. Hyg. Assoc. J. 43:845-852 (1982).

9.12. Occupational Safety and Health Administration Analytical Laboratory: Precision and Accuracy Data Protocol for Laboratory Validations. In The OSHA Laboratory Methods Manual. Cincinnati, OH: American Conference of Governmental Industrial Hygienists (Pub No. ISBN: 0-936712-66-X), 1985.

9.13. National Institute of Occupational Safety and Health: NIOSH Manual of Analytical Methods, 1st ed. (P&CAM 125) edited by D. Taylor (DHHS/NIOSH Pub. 77-157-A). Cincinnati, OH: NIOSH, 1977.

Ar   =   Area Sample               P   =   Personal Sample

Block Diagram of the Major components in a Dynamic Generation System


Figure 1

3M Monitor Aluminum Exposure Channel

Field Study



Figure 3

Lab Study



Figure 4

Footnote (1) Research Chemist, Branch of Inorganic Methods Development, OSHA Technical Center, Salt Lake City, UT.

Footnote (2) Senior Industrial Hygienist, Health Response Team, OSHA Technical Center.