January 6, 1992

MEMORANDUM FOR:          ELLEN I. ROZNOWSKI, ACTING DIRECTOR
                        OFFICE OF COMPLIANCE ASSISTANCE

FROM:                    DANIEL T. CRANE, SUPERVISORY PHYSICAL SCIENTIST
                        SALT LAKE TECHNICAL CENTER

SUBJECT:                 ENVIRON REPORT

I have read and reviewed both the executive summary and the supporting data volumes submitted by the Resilient Floor Covering Institute prepared by Environ. Two of the jobs described cause me some concern. The first is the Potomac Presbyterian Church. The perimeter sample is in excess of 0.1 fiber/cc. There is some mention that the custodian vacuumed during the sampling period in another room. This value is quite a bit higher than those reported for the workers or the observer. There is no TEM data to explain this sample. It is possible that fibers on the filter were from this source. Without TEM, the identity of the fibers is not known. Also, while the actual values reported for the partial shift samples are below the action level, the upper 90% confidence level indicates that there is greater than 5% chance of exceeding the action level. By the way, it is usual to discuss 95% confidence levels. This is a wider band, specifying a higher possible measurement. Calculation of that number would require more detailed information. It appears from quick calculation that values from the Leidel-Busch curve were used. As an example of this use the value for sample P-W1P-09 (an excursion level sample). Using the equation:

Upper 90% confidence limit = mean - 1.645(CV)mean

CV = .31 for 18 fibers counted (22.9 fibers/mm²) (sample #P-W1P-09)

Using instead, the confidence interval of 1.965CV to get the 95% interval, the value will be 0.2219 fibers/cc rather than 0.2080 fibers/cc gotten from using 1.645CV.

One other location was the Newtown samples N-W1F-03, N-W1P-06 and N-W1P-09. N-W1F-03 was a full-shift sample for worker #1, while the other two were partial shift samples covering the floor removal time for worker #1. The full-shift sample showed a value of 0.0108f/cc 8hr TWA. The other two showed values of 0.2160 and 0.3597 respectively. The full-shift sample is below the action level. If the two partial shift samples were treated as the only exposure time, a TWA calculated on a basis of 480 minutes would give a value of 0.0475f/cc. This might be considered to be below the level. The concern is that the square footage of this floor was 286 square feet (13' x 22'). According to the record, the partial samples covered only the tile removal portion. The removal time is estimated from the times given (8:17 to 9:48 or 91 minutes). This translates to 3.14 square feet of removal per minute. A larger floor could be removed in a longer time. Using this figure (3.14 sq. ft./min) a floor measuring over 1500 feet could be done in an 8 hour day. This is not an unreasonable size floor. In the letter from ENVIRON to me dated 2 August 1991, Table A-3 indicates that the floor of the training room at the Armstrong facility is 1260 square feet.

The letter states that the levels were below the Action level for that facility. I did not go back and check that out. I make the point that the floor at the facility was relatively large and would require almost 7 hours to remove flooring at the rate described at the Newtown location. This translates for that employee that the level would result in a TWA in excess not only of the action level but, also the PEL. I note that the floor cover institute observer was present for the work done at Newtown.

An effort to minimize the impact of these numbers was attempted by the use of TEM. Unfortunately, the TEM sheds no light on the samples. Although the same samples were analyzed by both TEM and PCM, no "Federal Fibers" (PCM-visible fibers) were detected in either of the first partial shift sample or the full-shift sample. The analysis of N-W1P-03 only had a 50% chance of seeing 1 fiber at the loading and analysis rates used. It is not surprising that no federal fiber was seen. However, because none were seen, the TEM is nearly valueless in assessing level. It does show that of the total fibers, 116 of 125 fibers analyzed were indeed asbestos. This does not indicate that fibers counted in PCM should be discounted. The other sample, N-W1P-06 showed no federal fibers while 10 of 14 fibers (fiber #2 in grid opening 10 was less than 3:1 aspect ratio) were asbestos. This sample should have had the power to see 3 federal fibers and saw none. Again, there is no information upon which to discount any PCM visible fibers.

For the most part, the results show that the methods when used strictly as defined by the guidelines, result in fiber values lower than the action level. It is also apparent, that the PEL could be exceeded even when company representatives are present. It is recommended that when large jobs are contemplated or situations not strictly in accord with those outlined in this report are encountered, sampling should be done to assess the levels in the job.