There are no standard methods for surface sampling of fiberglass fibers, so we adapted published methods for other substances for our evaluation. Tape sampling was selected because gelatin tape sampling has been used to sample manmade fibers in an office building⁽³⁾ and to study mold on surfaces.^(4,5) Tape has been shown to be an easy and effective method to sample surfaces.⁽⁴⁾ One study reported that vacuum sampling was comparable to wipe sampling for asbestos; however, vacuum sampling was more efficient on the rough surfaces tested.⁽⁶⁾ Therefore, we also used vacuum sampling.

We collected fiber samples using either tape or vacuum sampling from surfaces suspected of potentially having fibers from the cut-resistant sleeves. This included work surfaces, clothing, and skin. Samples were collected throughout the steel mill to determine if fibers were being shed in different areas. A new pair of nitrile gloves was worn by the investigator when collecting each sample to avoid cross contamination. Two field blank samples for each method were collected by exposing the media briefly to ambient air.

Stick-to-it lift tape (part number 225-9809, SKC Inc.) was used following manufacturer instructions.⁽⁷⁾ Figure 1 shows a tape being used to sample on an employee’s uniform sleeve after he had removed the cut-resistant sleeves.

Polycarbonate filters (37 mm, 0.8 μm, SKC Inc.) were used with cellulose back-up pads inside conductive three-piece cassettes. Air was drawn through the cassette at 15 liters per minute by an SKC QuickTake 30 pump. Sampling was performed using a modified version of the NIOSH Method 7400 for asbestos and other fibers.⁽⁸⁾ However, an open cassette instead of a closed cassette with a nozzle was used to avoid any losses due to static electricity generated during collection. Volumes collected during surface sampling (~15 liters of air) were also lower than for air samples using NIOSH Method 7400 (>400 liters of air).

Once a surface was selected we used a 100-square-centimeter disposable cardboard template (when possible) for consistency. When a guide was not possible because of an irregular shape, we vacuumed the surface for 1 minute (for an approximate volume of 15 liters of air). During vacuuming we kept the cassette at an angle to the surface to avoid a pump fault. We used an overlapping “S” pattern to vacuum the entire surface with horizontal strokes. We then vacuumed the same area using vertical S-strokes, followed by diagonal S-strokes. Figure 2 shows vacuum sampling of a clothing surface.

Three new and two laundered cut-resistant sleeves were collected and sent to a contract laboratory for analysis. Fibers from the sleeves were compared with other fibers found in the steel mill. Other sources of fibers identified in the facility included yellow insulation used in roofs and pipelines (glass wool) and white fibrous material (Kaowool) used for thermal insulation around ovens and other industrial equipment, expansion relief, or packing behind brickwork in furnaces.

Samples were analyzed by stereomicroscope and polarized light microscopy for identification of any fiberglass, Kevlar, and cellulose fibers, as well as for fiber morphology and size.^(9-12) More than 100 standards on file were used as reference materials. Because no sample contained large amounts of particulate less than 5 micrometers (μm) in width, scanning electron microscope-energy dispersive spectroscopy was not done. Laboratory data provided percentages of each of the fiber types. We could not calculate surface loading (number of fibers per area sampled) because the analysis was qualitative.

We collected 4 surface, 33 clothing, and 6 skin surface samples. The surface samples were mostly from containers where sleeves were stored. Clothing surfaces included cut-resistant sleeves or shirts worn under the sleeves. Of the 15 employees we sampled, 60% reported using new (unlaundered) sleeves, and 40% reported using laundered sleeves. Employees who used new sleeves reported using them an average of 6 days, while those who used laundered sleeves reported using them an average of 7 days before replacing them. Skin surface samples included hands and forearms of sleeve users.

Table I presents the fiber results from all surface samples collected using the tape method (cut-resistant sleeves, clothing, and skin) and Table II presents results from all surface samples collected using the vacuum method (cut-resistant sleeves, clothing, and skin). Tape samples contained fiberglass, Kevlar, and/or cellulose fibers. Vacuum samples contained fiberglass and/or cellulose fibers. Where present, the Kevlar fibers averaged 20 μm in width, and the fiberglass fibers averaged 10 μm in width. The Kevlar and fiberglass fibers had variable lengths. None of the Kevlar, fiberglass, or cellulose fibers seen in these surface samples had sharp edges. No fibers were observed on field blanks for the vacuum samples. The field blanks for the tape samples contained fiberglass and cellulose fibers, but no Kevlar fibers. No fibers were present on the media blanks for either the tape or vacuum samples.

A photograph of a new sleeve is presented in Figure 3. Regardless of whether the cut-resistant sleeve was new or laundered, some of the Kevlar and fiberglass bundles in the sleeves were frayed and broken (Figure 4). Fibers found in the sleeves are shown in Figures 5 and Figures 6. The fiber composition, size, and morphology were similar for new and laundered sleeve samples. Other fiber sources in the steel mill were yellow insulation (100% glass wool) and white fibrous material (99% Kaowool and 1% cellulose fiber).