These hazards are listed from top to bottom in decreasing order of risk. The level of risk was based upon two factors:

1. likelihood of occurrence
2. severity of consequence

Each risk ranking is of a qualitative nature based upon the professional judgment of the authors. Because of limited resources, the authors concentrated on the following health and safety hazards in the drycleaning industry:

Study Design

Ten drycleaning shops were visited during walk-through and in-depth surveys between August 1993 and January 1995. Following literature searches and reviews of previous NIOSH research reports, NIOSH investigators conducted four walk-through surveys to observe a variety of drycleaning operations. Photographs were taken during these initial surveys, as well as limited air sampling and ventilation measurements.

Results of the walk-through surveys helped determine specific businesses of interest for the six in-depth surveys. NIOSH engineers, industrial hygienists, and ergonomists surveyed each site, evaluating equipment, controls, and potential hazards (exposures to PERC and spotting chemicals, ergonomic risk factors, and fire hazards). Air sampling, video-exposure monitoring, ventilation system measurements, and process and workplace observations were performed. These data were evaluated to determine the effectiveness of each control system. Plant reports were prepared, which presented the data and results of the analysis, conclusions, and recommendations. These plant reports have served as the basis for the findings outlined in this report, as has information gathered during a World Health Organization (WHO) fellowship to study drycleaning in Europe.

Typical Drycleaning Process

The walk-through and in-depth surveys of various commercial drycleaning shops in the U.S. revealed that although practices varied from one shop to another, the overall drycleaning process was similar. The typical process begins when garments are brought to the shop by customers or arrive by van from another "satellite shop," which does no drycleaning on site. Garments are initially tagged for identification. Prior to spotting or being loaded into the drycleaning machine, garments are typically inspected and sorted according to weight, color, and finish. Modern garments are made from many different types of fabrics, and most drycleaners have the tools to clean all types.

Particular attention was paid to spotting during the surveys to determine if excessive exposures occurred during this process. The spotting process might assume greater importance if wet cleaning methods in the U.S. were more widely used. Information was gathered prior to shop visits to determine which spotting chemicals a particular shop used most frequently. Purchasing records and Material Safety Data Sheets were reviewed.

In a typical shop, garments with visible stains are routinely treated at the spotting station. Spotting involves the selective application of a wide variety of chemicals and steam to remove specific stains from the garments. The three general categories of stains are water soluble, solvent soluble, and insoluble. Stains rarely consist of a single substance. Spotting chemicals and chemical mixtures are either solvent-based liquids (dryside) or water-based detergents (wetside). Wetside chemicals, which may contain water, remove water-soluble stains; and dryside chemicals, which never contain water, remove solvent-soluble stains.

The spotting chemicals, contained in small, plastic squeeze bottles, are applied to the stain. Spotting usually occurs on a spotting board equipped with pressurized air, steam, and water guns designed to flush the chemicals and stains from the garment. Air, steam, a small brush, a spatula, and fingers are all used to help break up the stains and wash them away. A pedal-actuated vacuum is used to capture the spotting chemicals, which are either held in a local storage reservoir or transported to a vacuum canister for discard.

In addition to PERC, some of the more common chemicals and chemical families for spotting are other chlorinated solvents, amyl acetate, petroleum naphtha, oxalic acid, acetic acid, esters, ethers, ketones, dilute hydrofluoric acid, hydrogen peroxide, and aqueous ammonia. Each of these chemicals are used in limited quantities. Most spotting chemicals are purchased from a company that supplies proprietary products to the industry. However, some shops use their own concoctions, which are prepared by individuals highly skilled in the art of stain removal.⁵⁰

The spotting process has two components: pre-spotting, which involves dryside chemicals to remove or loosen solvent-soluble stains; and post-spotting, which utilizes wetside chemicals to remove water-soluble stains. Water-soluble stains, which may further set during the drycleaning process, are pre-spotted before drycleaning. Post-spotting is often used if the stain was not visible before drycleaning, or if the stain was not completely removed after pre-spotting and drycleaning. However, some shops utilize the more time-intensive post-spotting exclusively since many spots are removed during drycleaning. Some spotting processes involve little use of tools; the solvent is simply squirted liberally on all the stains before the garments are cleaned in the machine.

Drycleaning, a three-step process, involves washing, extracting, and drying. A diagram of this process can be seen in Figure 1. Before washing, a worker adds detergent to the solvent in one of two ways: 1) A charge detergent is added to the solvent in set concentrations. The charged solvent is re-used in successive loads, and fresh detergent is periodically added to maintain the proper detergent concentration. 2) Injection or no-charge detergent is added to each load, based on each load's weight. Water is added to the system before or during drycleaning and aids in removing water soluble soils from the fabric.

To begin washing, clothes are manually loaded into the machine, followed by the solvent. The contents of the machine are then agitated for a period of time, allowing the solution to remove soils. Next, the clothes are spun at a high speed to extract the solvent.

After extraction, the fabric is tumbled dry. The drying process may occur in the same machine or a different one, depending on the system. Recirculated warm air vaporizes the residual solvent. Unheated air is then passed through the system during the cool-down cycle. This step reduces wrinkles. Following cool-down in vented machines, fresh air is passed through the system to freshen and deodorize the clothing during the aeration step.²⁷ Garments are then removed from the machine prior to pressing.

Figure 1. The drycleaning process flow diagram.

Drycleaners use filtration and/or distillation to recover and purify solvent. Filtration removes insoluble soils, nonvolatile residues, and loose dyes from the solvent and, in some cases, soluble soils. Filtration is usually a continuous process in which the solvent passes through either an adsorbent powder or filter cartridge, both needing periodic replacement. Additionally, new, powderless, spin-disc filters⁵¹ significantly reduce the generation of hazardous waste because they are regenerated instead of discarded.⁵²

Distillation, used by 90% of the industry, removes soluble oils, fatty acids, and greases not removed by filtration. Drycleaning stills typically consist of a kettle, condenser, and separator. Distillation occurs by heating PERC to its boiling point. The PERC vaporizes and condenses back into a liquid form. During this process, nonvolatile impurities, which do not evaporate, remain at the bottom of the still and are discarded as hazardous waste. Both filtration and distillation produce solid wastes containing PERC.

Pressing is generally a dynamic and repetitive task that requires reaching and precision gripping. Pressers usually stand at nonadjustable workstations for much of the day and reach well overhead for hanging garments. When a garment is placed on a pressing machine, it is pressed between two surfaces, at least one of which is heated to a temperature around 149°C (300°F).

Some of the equipment used by pressers includes general utility presses; puff irons; pants toppers; finishers; electric irons; bosom, body, and yoke presses; collar, cuff, and yoke presses; and sleevers. Because this equipment is highly specialized, drycleaners do not necessarily have all of these presses. Once the garments have been completely pressed, they are returned to the overhead rack and wrapped in plastic for customer pick-up.

Drycleaning Machine Control Technologies

Shops included in the current NIOSH study were collectively selected to represent a cross-section of the controls available within the drycleaning industry. The most important factor for site selection was the effectiveness of the control system for reducing exposures to PERC. The NIOSH study evaluated the primary working hypothesis that four general categories of drycleaning machines were capable of maintaining an 8-hour TWA personal exposure to PERC of less than or equal to 25 ppm.

The four general categories of drycleaning machines evaluated were:

1. transfer machines.
   
2. vented and nonvented dry-to-dry machines.
   
3. machines with refrigerated condensers and/or carbon adsorbers.
   
4. "fifth generation," nonvented, dry-to-dry machines, having residual controls.
Two Basic Types of Drycleaning Machines

Two basic types of machines are generally used in drycleaning: transfer and dry-to-dry. Transfer machines are older, less expensive, and require manual transfer of solvent-laden clothing between the washer and dryer, a source of high worker exposure. Dry-to-dry machines eliminate clothing transfer because items are processed in one step, entering and exiting the machine dry. Transfer machines process more clothing than comparably sized dry-to-dry machines because the process time is approximately half that of a dry-to-dry machine. To compensate, some owners of dry-to-dry machines reduce the cycle time, thereby increasing productivity. Such practices can also increase exposures that result from unevaporated residuals in the drycleaned clothing.⁵³

Transfer machines are generally vented, but dry-to-dry machines can be vented or nonvented. Vented dry-to-dry machines exhaust residual solvent vapors directly into the atmosphere or through some form of vapor recovery system, usually during the aeration process. Nonvented dry-to-dry machines are essentially closed systems, which are open to the atmosphere only when the machine door is opened. These machines recirculate the heated drying air through a vapor recovery system and back to the drying drum, eliminating the aeration step. Unfortunately, significant concentrations of PERC remain in the drum at the end of the cycle.

Approximately 70% of the drycleaning machines found today in the U.S. are of the newer dry-to-dry design. Transfer machines are no longer manufactured in the United States because of the high solvent usage rate, emissions, and exposures during transfer. However, used or reconditioned transfer machines can still be purchased. Either PERC or petroleum-based solvents can be used in dry-to-dry or transfer machines, but PERC is the solvent of choice for the majority of U.S. drycleaners.

Two Primary Vapor Recovery Technologies

The two primary technologies used to recover PERC vapors from drycleaning machines are the carbon adsorber and the refrigerated condenser. Carbon adsorbers are used in approximately 35% of controlled machines and refrigerated condensers, in approximately 65%. Carbon adsorbers remove PERC molecules from the air by passing solvent-laden vapors over activated carbon with a high adsorption capacity. The carbon is then desorbed and the PERC recovered, or the carbon is discarded when saturated. Carbon desorption typically occurs with steam or hot air. Desorption can be done automatically after each load, at the end of the day, or it can be done at other intervals depending on the PERC concentration and amount of carbon. If not done regularly, the carbon bed will become saturated and ineffective.

Refrigerated condensers use a refrigerant to cool the solvent-laden air below the dew point of the vapor to recover the PERC. The number of transfer machines using refrigerated condensers is roughly equivalent to the number of transfer machines using carbon adsorbers. However, on dry-to-dry machines there are approximately twice as many refrigerated condensers in use (see Table 4).



Drycleaning Machine Types in the U.S.

In 1988, the International Fabricare Institute (IFI) conducted a national drycleaning plant and equipment survey in which 5.8% of U.S. drycleaners participated. Results of that survey indicated that 79% of the shops responding used PERC only, and another 9% of shops used PERC and another solvent. For all PERC drycleaning machines, regardless of age, approximately 34% were dry-to-dry nonvented, 32% were dry-to-dry vented, and the remaining 34% were transfer machines. For PERC drycleaning machines that were less than five years old at the time, approximately 62% were dry-to-dry nonvented, 31% were dry-to-dry vented, and 7% were transfer. The majority of PERC drycleaning machines had a capacity of 30-39 lb, followed by 40-49 lb, and 50-59 lb, respectively. The average cleaning volume was 1,926 lb of clothing per week.⁵⁴

The Evaluated Drycleaning Machines

During the six in-depth surveys conducted in various U.S. commercial drycleaning shops, NIOSH investigators evaluated sources of emissions and exposures from machines in the four general categories given above. Descriptions of the specific drycleaning equipment evaluated follow.

The Evaluated Transfer Drycleaning Machine.—The transfer unit studied consisted of two separate machines: a J & T® Model 60, 60-lb washer and a Model SF-145 Hoyt® Solvo-miser reclaimer. Both machines were over ten years old. The washer and reclaimer were connected to a Kleenrite® Vapor Condenser. The reclaimer had an internal, water-cooled condenser and separate refrigerated, vapor condenser. The reclaimer operated in three different modes: dry, cool, and aerate. During the dry mode, recirculated air was heated and then cooled in the condenser to recover PERC. During the cool mode, the air was diverted to the separate vapor condenser and cooled to recover PERC. During the final aeration cycle, fresh air was drawn into the reclaimer and exhausted outside of the building.

The Evaluated Dry-to-Dry, Vented Drycleaning Machine.—An Omega® Model CE55, 55-lb dry-to-dry machine, which was about six years old, was evaluated. This machine had a refrigerated condenser as the primary vapor recovery device. The secondary vapor recovery device consisted of a small, centrifugal fan rated at 110 cubic feet per minute (cfm) that was ducted to a carbon canister. When the machine door was opened, a microswitch energized the fan to draw PERC-laden air from the cylinder through the activated carbon.

The Evaluated Dry-to-Dry, Nonvented Machine.—Two nonvented, refrigerated, dry-to-dry machines, each less than five years old, were evaluated: a Boewe® Permac Model P540, 46-lb dry-to-dry machine and a Fluormatic® Model M242, 30-lb dry-to-dry machine. Both had a refrigerated condenser as the primary vapor recovery device. Both machines were connected to a Rite-Temp® Model RTR1003AWC Chiller used to chill water for reclamation, distillation, and maintenance of proper solvent temperature. There were no secondary vapor recovery devices.

Modern "Fifth Generation" Machine Technology.—Modern, dry-to-dry, nonvented drycleaning machines with residual controls are commonly referred to as "fifth generation" machines. The design of "fifth generation" machines began in Germany where their use is now mandated primarily by the Second German Emission Standard. "Fifth generation" drycleaning machines have engineering controls which dramatically reduce residual PERC in the machine's cylinder at the end of the dry cycle. These closed-loop, dry-to-dry machines rely on both an integrated, refrigerated condenser and a large carbon adsorber to recover PERC vapors during the dry cycle. "Fifth generation" machines are much more effective at recovering solvent vapors while drying garments than are machines equipped with only a carbon adsorber or refrigerated condenser alone. They are capable of lowering the PERC concentration in the machine's cylinder to below 290 ppm at the end of the dry cycle.

"Fifth generation" machines use a single beam, infrared photometer to monitor the PERC concentration in the machine cylinder. An interlock on the machine door prevents the operator from opening the machine door until the PERC concentration in the cylinder is below 290 ppm. The drying cycle will continue to operate until the cylinder concentration is sufficiently low. Such sophisticated machines are expensive and not widely used in the U.S.

Two Boewe Passat® dry-to-dry, nonvented drycleaning machines with residual controls were evaluated. One machine was a model P546, 46-lb machine, and the other was a model P536, 36-lb machine. Both machines had an integrated, refrigerated condenser and a large regenerable carbon adsorber to recover PERC vapors during the dry cycle. These machines were designed to lower the PERC concentration in the machine cylinder at the end of the dry cycle. This was accomplished using a large carbon adsorption system to capture PERC in the airstream and a single beam infrared photometer to continuously monitor the PERC concentration in the machine cylinder. An interlock on the machine door prevented opening until the PERC concentration in the cylinder was sufficiently low. The machines had a cleaning cycle of 35 to 40 minutes; however, the drying cycle was often automatically extended to ensure that the PERC concentration in the cylinder was below 290 ppm. The operator could determine the PERC concentration in the cylinder by reading a printout from the infrared photometer. The emission-free, still-cleaning device required still-raking only once every three weeks. A dosing unit enabled the operator to perform garment waterproofing within the machine.

METHODOLOGY

The principle measurements in this study were air concentrations of PERC measured both with conventional air-sampling pumps with charcoal tubes, and real-time instruments.

Air Sampling

Personal, area, and background air sampling was conducted using NIOSH Method 1003 for PERC and other halogenated hydrocarbons. In this method, organic vapors were drawn through 100 mg/50 mg coconut shell charcoal tubes. Carbon disulfide was used to desorb the solvents from the charcoal tube. Analysis was done using a gas chromatograph, fitted with a flame ionization detector to determine PERC concentrations and, in some instances, to determine all halogenated hydrocarbons concentrations. Samples were collected over 120-minutes at a flow rate of 0.1 liters/minute and a volume of 12 liters. The limit of detection for this method is 0.01 mg/sample.⁵⁵

Two-hour consecutive personal sampling was conducted on all of the machine operators, as well as on some of the spotters and pressers who worked in proximity to the drycleaning machines. Two-hour consecutive samples were used to determine full-shift TWA exposures.The purpose of conducting personal sampling was to compare worker perchloroethylene exposures to a TWA of 25 ppm. In addition to TWAs, 5-minute and 15-minute exposures were measured during some manual tasks. This was done at a flow rate of 0.1 liters/minute and a volume of 0.5 and 1.5 liters respectively.

Area sampling determined which areas of the shop had the highest concentrations and were collected at various distances and locations around the periphery of the drycleaning machine. Air samples were generally collected in front of and behind the drycleaning machines, in the pressing area, in the spotting area, in the hazardous waste storage area, near the customer counter, and outside of the building. These locations were similar for each shop evaluated. This data helped in the development of strategies to reduce the exposures.

Air sampling was also used to evaluate the risk of chemical inhalation during the spotting process. Air sampling was conducted for two days to evaluate worker exposures to PERC, trichloroethylene, 2-butoxyethanol, hexylene glycol, MIBK, and n-butyl acetate. Additionally, five area samples were taken to determine concentrations of PERC and trichloroethylene. The ratio of equilibrium vapor concentration to TLV was used to compute a vapor/hazard ratio. Peak and 8-hour TWA personal sampling of spotters were conducted for halogenated hydrocarbons that could be sampled using NIOSH Method 1003. When combined exposures of PERC and other spotting chemicals were involved, the combined exposure (C_E) was determined as follows.

CE = C₁/T₁ + C₂/T₂ + . . . + C_n/T_n

Where:

C_n = Exposure to an individual contaminant (ppm)
T_n = The OSHA PEL for the individual contaminant (ppm)

If the value of CE is less than 1, the combined exposure is believed to be acceptable. When this calculation was used during the in-depth spotting evaluation, the OSHA PEL for each chemical was applied if available. For hexylene glycol, no OSHA PEL was available; therefore, the NIOSH ceiling of 25 ppm was used.

Real-Time Monitoring

Real-time monitoring was used to study how specific manual tasks and maintenance operations affected worker exposure to PERC. Some drycleaning procedures occurred frequently throughout the day, such as loading/unloading the machine, while others, such as cleaning the lint and button traps, were less often. Most of these tasks took between 5 and 30 minutes. Real-time monitoring of PERC exposures were performed using a MicroTIP® IS3000® (PHOTOVAC Inc, Thornhill, Ontario) with a 10.6 EV ultraviolet lamp. This instrument uses a photoionization detector (PID) to provide an analog output response proportional to the concentration of ionizable chemicals present in the air. The MicroTIP® was spanned, using 100 ppm isobutylene span gas, and calibrated for both PERC and mineral spirits, using five standard concentrations of PERC vapor and mineral spirits vapor. Instrument readings and actual concentrations were used to construct two calibration curves and find predictive equations. The following formula was used to convert the output of the PID (volts) to concentration of contaminant (ppm):

C(t) = IR(t) * CF* MR

where:

C(t) = concentration of vapor at time t (ppm)
IR(t) = instrument response at time t (volts)
CF = conversion factor from calibration equation
MR = MicroTIP® range

Information gathered using the MicroTIP® was electronically recorded on a Rustrak® data logger (Rustrak® Ranger, Gulton, Inc., East Greenwich, RI) and downloaded to a portable computer, using Pronto® software. During the gathering of real-time data, a video camera recorded worker activities. This videotape permitted detailed analysis of tasks for quantitative determination of work activities causing the highest exposures.

Real-time monitoring was used for brief periods of time and focused on worker exposure during manual tasks and maintenance operations which might be related to high exposures. The frequency of manual tasks varied from shop to shop. Some common activities and their approximate frequencies follow:

            Maintenance procedure           Approximate frequency

            Still cleaning                  Daily
            Filter replacement              Demand related
            Filling the tank                Demand related
            Water separator                 Weekly maintenance
            Button/lint trap cleaning       Daily

Specific questions addressed from real-time monitoring results included:

p Are certain controls more effective at reducing exposures around cleaning and pressing stations?
p Do specific actions result in higher exposures?
p How does air contaminant concentration decay over time?

Real-time monitoring was also used to study off-gassing of garments and to compare vapor recovery efficiency of the machines. A standard test swatch, approximately 5 in. by 6 in. and made of 51% rayon and 49% polyester, was added to several runs of the drycleaning machine. At the completion of the dry cycle, the test swatch was placed in a small glass test chamber. As the solvent residuals in the swatch vaporized, PERC concentrations were monitored and recorded, using the MicroTIP® and Rustrak® data logger. This method detailed garment off-gassing and provided continuous concentration profiles, which were used to quantitatively compare the vapor recovery system's performance for each machine evaluated.

Evaluation of Ventilation Systems

Another aspect of this study was the evaluation of ventilation system performance. The various components of both the local and general ventilation systems were documented in order to assess the role of the entire ventilation system in controlling worker exposures. Data gathered during study of the ventilation system was compared to recommended values in the literature.

The following factors were examined when appropriate:

p exhaust hood design, dimensions, and location
p hood face velocity, capture velocity, and flow rate
p airflow patterns and velocity around the source of exposure
p system design specifications

A Kurz® Model 1440 Air Velocity Meter was used to measure air speed. Smoke tubes were used to qualitatively evaluate airflow patterns near the machine and within the building.

General dilution ventilation was evaluated at each drycleaning shop studied. General ventilation can reduce contaminant concentrations in large areas that are not controlled locally. As part of the evaluation, a layout of the facilities was obtained, showing locations of exhaust ducts and makeup air. Flow rates, face velocities, and airflow patterns in the building were documented. All of the data gathered were compared to design specifications when possible.

Ergonomic Evaluation

Hazards and risk factors present during the pressing operation were evaluated by examining work station design, anthropometry, and conducting time studies of the tasks performed. Based upon walk-through surveys, preliminary information indicated that high repetition/insufficient recovery time, and awkward postures might be risk factors present during pressing operations. Interaction of these and other risk factors have been shown to result in cumulative trauma disorders.^35,55

Repetitiveness and recovery time were evaluated by examining cycle time. This was accomplished by videotaping and analyzing tasks in their elemental forms. Awkward and sustained postures performed during the pressing operation were examined to determine whether they were a problem in light of the general guidelines which can be found in the current literature. Each of the tasks were videotaped and measurements were made to determine height, reach, anthropometric envelope, and physical layout of the workstation.

Work Practice Observations

Because work practices often have a dramatic effect on exposures, observations recorded during site visits were an important aspect of this study. Of particular importance were activities related to loading/unloading the machines, maintenance procedures, pressing operations, spotting, and use of personal protective equipment.

Back to the Table of Contents

CONTROL OF DRYCLEANING SOLVENT EXPOSURES

The following section gives results from the six in-depth drycleaning surveys, discusses exposures and emissions, and suggests control options for U.S. commercial, drycleaning shops. Information gathered during the WHO fellowship in Europe is included primarily in the sections addressing substitution and "fifth generation" drycleaning machines.

Sampling and Monitoring Results

Air Sampling

Results of personal air sampling can be seen in Table 5. Generally, air samples gathered near the drycleaning machine had higher PERC concentrations than those farther away from the machine. Similarly, the more time a worker spent near the drycleaning machine, the higher the PERC exposures tended to be. Those drycleaning machines that were designed to lower the PERC concentration, emitted from the machine cylinder, were more effective than traditional machines at reducing worker exposures. Operators received the highest exposures from loading and unloading the machine; whereas, pressers were primarily exposed from garment off-gassing.

Table 5
Time-Weighted Average Worker Exposures to PERC
(for entire survey)


Transfer Equipment.—All but one of the daily TWA personal samples taken on the transfer machine operator and two pressers were below 25 ppm. The transfer machine operator was exposed to 19.5 ppm TWA PERC for the entire survey. The operator's two-hour samples ranged from 3.9 to 42.3 ppm. The two pressers, who did not work in as close proximity to the transfer machine as the operator, were exposed to 3.3 and 3.8 ppm TWA PERC during the entire survey. Statistical analysis showed a significant difference between the operator's and pressers' TWA exposures.

NIOSH investigators gathered 5- and 15-minute personal air samples on charcoal tubes during transfer operations and filter changing (Figure 2). Neither of these activities exceeded the OSHA ceiling of 200 ppm or OSHA peak of 300 ppm. Exposure during transfer operations was 57.5 ppm for a 5-minute period and 21 to 22 ppm during a 15-minute period. The reason for the differences between 5- and 15-minute concentrations was due to time taken for tasks. Exposure during filter changing was approximately 121 ppm for 5 minutes and 118 ppm for 15 minutes.


Figure 2. Five-minute and fifteen minute personal
exposures to PERC from transfer equipment.

The highest area concentrations were above the reclaimer door. The geometric mean concentration in that area was 15.2 ppm, which was significantly different from all other areas. It seems logical that these area concentrations would have been the highest because the air was heated, and the PERC was in the vapor state during loading/unloading the reclaimer, but the solvent and air in the washer were not heated.

Dry-to-Dry, Vented Equipment.—All of the individual personal samples taken for the dry-to-dry, vented machine were below 25 ppm. The machine operator was exposed to 15.8 ppm TWA PERC for the entire survey. Two-thirds of these exposures resulted from loading and unloading the machine. The two pressers were exposed to 5.0 and 2.5 ppm TWA PERC for the entire survey. The highest geometric mean area concentrations measured, 9 and 12 ppm PERC, were near the drycleaning machine where a vapor leak was detected. The next highest geometric mean concentrations, 6 and 8 ppm PERC, were above the machine door.

Dry-to-Dry, Nonvented Equipment.—The machine operator, a presser, and a clothing inspector were sampled at the shop having two nonvented dry-to-dry machines, and all individual personal samples were below 25 ppm. The machine operator was exposed to 7.8 ppm TWA PERC for the entire survey. Nearly half of this exposure resulted from loading and unloading. The presser and inspector were exposed to less than 1 ppm TWA PERC. The highest area samples, 5.4 and 5.7 ppm PERC, were taken above the drycleaning machines' doors. The next highest concentrations were behind the machines.

Dry-to-Dry, Nonvented with Residual Control.—Figure 3 summarizes personal air samples gathered for the two dry-to-dry, nonvented machines with residual controls. All of the personal samples were well below 25 ppm and were dramatically lower than exposures on any other machines evaluated. The operator of the "fifth generation" machine had the highest exposure to PERC, ranging from 0.31 to 4.9 ppm TWA. Almost all of the operator two-hour samples were below 2 ppm PERC, and most samples were below 1 ppm. The only exception to these low exposures occurred when the operator raked the still bottoms of both machines. During the last morning of sampling, the operator was exposed to approximately 12 ppm over a two-hour period and 4.9 ppm full-shift TWA PERC primarily because of cleaning the stills. If the stills had not been cleaned, the TWA exposure probably would have been less than 2 ppm.



Figure 3. Time-weighted average worker exposure to
PERC from "fifth generation" drycleaning
machines. (Four days of air sampling.)

The pressers were protected from PERC vapors originating from the machine by ventilation and barriers between the pressing station and drycleaning machine. All of the PERC concentrations measured near the presser's breathing zone were at or below the limit of detection, 0.01 mg/sample. This minimal exposure resulted from almost no PERC retention in the clothing. Little PERC was detected on samples outside of the drycleaning room. This nondetection of PERC can be attributed to excellent machine design and ventilation near the machines. The highest area concentrations, ranging from 0.0 to 1.9 ppm, were detected above and behind the drycleaning machines.

Statistical Analysis of Air Sampling Data

Summaries of the personal air sample results and their statistical analysis are in Tables 5, 6, 7, and 8. All of the TWA and mean exposures to PERC in these tables were below 25 ppm. However, some of the individual two-hour samples exceeded 25 ppm, and so did a small number of the daily TWA exposures. Time-weighted average exposures would have been somewhat lower if sampling had occurred for a full 8-hour shift; however, air sampling generally occurred when the drycleaning machines were in operation, typically six or seven hours per day.

Statistical analysis was performed on log transformed air sampling data for PERC. A two-way analysis of variance (ANOVA) showed that job title and shop had a significant effect upon concentration (p<0.0001). A multiple comparison test with 5% significance level, least significant difference (LSD), was used to analyze concentration differences.

Table 5 gives the TWA personal air samples by job for each shop evaluated. The TWA exposures in the table are for the duration of the entire survey. At each shop evaluated, the machine operator had consistently higher exposures than did any other job title. Most of the workers performing other jobs were exposed to less than one-third of the machine operator's exposure. Much of the difference can be attributed to the fact that the other jobs did not include the peak exposures associated with loading and unloading the machine.

Table 6 provides arithmetic and geometric mean personal exposures to PERC by shop for machine operators and the results of the two-way analysis of variance (2-WAY ANOVA). It also includes the concentration range and geometric standard deviation. This statistical analysis indicates that the operator of the advanced, "fifth generation" drycleaning machines was exposed to significantly lower PERC concentrations than operators of any other type of machine. Differences between operator exposures at all other shops were not statistically significant.

*Different letters indicate a statistically significant difference using the least significant difference test. ( = .05)



Table 7 gives the results of a one-tailed ( = 0.05) t-test, comparing operator exposure to PERC at each shop evaluated to 25 ppm. All of the machines studied maintained operator exposures to PERC below 25 ppm with the exception of the shop having two dry-to-dry, vented machines. Another shop with similar equipment, but only one machine, was able to control exposures below 25 ppm. Part of this difference may be due to production volume; this shop had two machines and a higher volume of garments, which required machine loading/unloading nearly twice as often.

Statistically significant at ( = .05)



Table 8 is a summary of personal exposures to PERC by job title across shops. As expected, the machine operator had the highest mean exposure to PERC, which was 13.3 ppm for all shops evaluated. The machine operators' exposures were significantly different from exposures experienced by pressers or "other" workers. The bulk of operator exposures resulted from loading/unloading the machine, or in some cases from transferring garments. The next highest mean exposure by job title, 6.8 ppm, occurred in a group of workers labeled as "other." This group included workers who were not full-time operators or pressers. Finally, the pressers from all of the shops studied were exposed to the lowest mean concentration, 2.1 ppm.


Real-Time Monitoring

Results of real-time analysis tended to support the air sampling results. Real-time results gave further insight into the reason why exposures from one type of machine were lower than exposures from another machine. The most significant source of PERC exposure occurred during loading/unloading the machines and garment transfer. Loading the machine frequently resulted in concentrations as high as or higher than unloading the machine due to air displacement. This was because of large quantities of PERC contaminated air being forced from the machine when dirty garments were loaded. Since high peak exposures occurred during these activities, the frequency of machine loading, unloading, and transfer played an important role in affecting total exposure. Exposure sources other than those directly from the machine, such as garment off-gassing and pressing, were substantially lower than exposures that occurred directly from the machine.

Transfer Equipment.—Figure 4 shows real-time data during garment transfer, loading the washer, and hanging the garments. Exposures during loading and unloading reached instantaneous concentrations between 1,000 and 1,500 ppm. The highest average exposures, 500-600 ppm, occurred during garment transfer from the washer to the reclaimer. The next highest average exposures occurred during loading the washer, unloading the reclaimer, and hanging clothing, respectively. Based on a comparison of real-time and air sampling results, over one-half of the operator's exposure resulted from loading/unloading the machine and garment transfer. A comparison of unloading the reclaimer, transfer, and loading the washer indicated that transfer generally occupied more time. After the garments were dried, hanging the garments took even more time than other tasks, but the average exposure during handling was relatively low, 14-21 ppm. Some of the bulkier garments, which retained solvent and took longer to dry, resulted in instantaneous PERC exposures near 70 ppm.

Figure 4. Operator exposure to PERC during clothing
transfer, loading washer, and hanging clothing.


Dry-to-Dry, Vented Equipment.—The process of unloading took nearly twice as long as loading. Real-time monitoring (Figure 5) showed that average PERC exposure during loading was much higher than unloading the cleaned garments. The average exposure during loading was 846 ppm; average exposure during unloading was 271 ppm. The integrated exposure (area under the curve) was also higher during loading the machine, approximately 11,850 ppm*sec. versus 7,050 ppm*sec. Real-time measurements near the carbon canister on the top of the machine indicated that concentrations of PERC (approximately 1,500 ppm) were blown into the work environment each time the machine door was opened (Figure 6).The carbon canister was ineffective at capturing PERC in the exhausted air.

Figure 5. Operator exposure to PERC from a dry-to-dry, vented
drycleaning machine during unloading/loading.

 

Figure 6. Real-time measurements of PERC concentration
near carbon canister exhaust of a dry-to-dry, vented
drycleaning machine.


Dry-to-Dry, Nonvented Equipment.—Concentrations during loading/unloading were higher than any other activity; however, the instrument used to conduct real-time monitoring in the present NIOSH study became saturated and did not read the highest concentrations. This was because the maximum concentration the instrument was capable of reading was 156 ppm. Real-time measurements taken by NIOSH researchers at other shops using dry-to-dry, nonvented machines were approximately 1,500 to 2,000 ppm during machine loading and unloading. Figure 7 gives operator exposure during loading and unloading a machine and hanging clothing. The tops of the largest peaks are truncated because of the upper limit of the real-time detector. The larger machine generally took longer to unload than the smaller machine. Hanging the clothing took approximately 9 minutes. For both machines, the average PERC exposure while loading the machine with dirty clothing was almost the same as unloading garments cleaned in PERC. The similar exposures were caused by PERC contaminated residual air being forced from the cylinder while the door was opened to admit unclean clothing.

Figure 7. Operator exposure to PERC from a dry-to-dry,
nonvented drycleaning machine during loading,
unloading, and hanging clothing.

When the 46-lb machine was loaded/unloaded, the integrated exposure to PERC was higher during unloading (1,955 ppm-sec. vs. 1,572 ppm-sec.); however, the average exposure was actually higher during loading (98 ppm vs. 89 ppm). The difference is made more profound through an examination of the median exposure during loading (147 ppm) versus unloading (98 ppm). Average exposure from loading and unloading the 30-lb machine was slightly higher than that from the 46-lb machine. There are two factors at work here: less efficient solvent recovery from the 30-lb machine and better local ventilation near it. Average exposure while hanging the clothing was approximately 5 ppm. The operator was exposed to higher PERC concentrations when hanging bulky items, which had not completely dried.

Dry-to-Dry, Nonvented with Residual Control.—Although these full-shift, TWA exposures were lower than at other shops, the greatest source of operator exposure from "fifth generation" machines continued to be from loading and unloading the machines. Exposures during this procedure peaked at approximately 160 ppm, which is an order of magnitude lower than most peak exposures at shops using different types of machines.

Figure 8 shows operator exposure during the first cycle of the day when only loading occurred. When dirty clothing was added to the cylinder, contaminated air was forced from the cylinder into the worker's breathing zone. This was characterized by a rapid increase and an almost instantaneous peak which approached 160 ppm. The concentration in the worker's breathing zone dissipated over the next 10 to 20 seconds and then returned to zero. This peak occurred because the machine cylinder, which was not fully isolated from the other sources of PERC within the machine, was not purged immediately prior to the door's being opened. PERC vapors apparently diffused from the vapor loop, into the machine cylinder, and eventually out into the work environment because of air displacement during loading. Operator exposure for loading the 46-lb machine was higher during this first cycle than for loading a smaller machine. This result may have been caused by differences in general ventilation near the machine doors, greater air displacement from the machine cylinder, and/or differences in the PERC concentrations within the machine cylinder.

Figure 8. Operator exposure to PERC from "fifth generation"
machine during the first cycle of the day.

Figure 9 gives operator exposure during unloading and loading both machines. For most cycles, average operator exposures and integrated exposures were higher during unloading than loading. The exception to this general finding occurred when dirty garments were processed in the 36-lb machine (Figure 9). The average operator exposure for loading was 101 ppm, which was significantly higher than the average exposure of 34.6 ppm for unloading. Again, in this instance, the machine had been idle for several minutes prior to the door's being opened. Because of this inactive period, PERC was able to diffuse from the sources within the vapor loop into the machine cylinder. PERC vapors were forced from the cylinder because the centrifugal fan provided insufficient airflow when the machine door was opened.

Figure 9. Operator exposure to PERC from "fifth generation"
machine during unloading/loading.


Garment Off-gassing.—Garment off-gassing was evaluated for each machine during a typical cycle. Evaluation of garment off-gassing revealed that the "fifth generation" machines were much more effective at recovering solvent from the garments. This efficiency was directly related to machine design.

During an average cycle of the transfer machine, the total PERC off-gassing from the test swatch was 89.0 mg PERC/kg cloth. The dry-to-dry, vented machine had a total PERC off-gassing from the test swatch of 31.8 mg PERC/kg cloth. The total PERC off-gassing from the test swatch in the dry-to-dry, nonvented machines varied according to the size of the machine. The PERC concentration from the 30-lb machine was 126 mg PERC/kg cloth and 69.4 mg PERC/kg cloth for the 46-lb machine. Both "fifth generation" machines were extremely effective at recovering solvent from the garments. The total PERC off-gassing from the test swatch was 1.34 mg PERC/kg cloth. Figure 10 provides a comparison of typical off-gassing from a swatch cleaned in a "fifth generation" machine with a swatch from a refrigerated, dry-to-dry machine.

Figure 10. A comparison of typical swatch off-gassing from a
refrigerated, dry-to-dry drycleaning machine versus a
"fifth generation," dry-to-dry drycleaning machine. The
total PERC off-gassed was approximately 31.8 mg
PERC/kg cloth and 1.34 mg PERC/kg cloth, respectively.
Machine Maintenance

Real-time evaluation of drycleaning machine maintenance showed that exposures during maintenance were less than exposures during loading/unloading or transfer. Maintenance activities that were evaluated included cleaning the drycleaning machine's lint trap, button trap, and still; changing solvent filters; and disposing of hazardous waste. One reason for lower exposures from machine maintenance tasks than from loading/unloading was that the former occurred much less often than the latter.

Figure 11 provides operator's exposures during lint and button trap cleaning on a dry-to-dry, vented machine. The activity took less than 10 minutes, and the average exposure to the worker was approximately 22 ppm. Figures 12 and 13 give maintenance exposures on two modern, dry-to-dry machines with residual controls. One machine had a 46-lb capacity, and the other a 36-lb capacity. Figure 12 shows exposures for cleaning the lint traps. This task was done every day and involved cleaning the lint/button traps and disposing of hazardous waste. Normally, lint traps were cleaned in the morning. The average exposure during maintenance on the 46-lb machine was approximately 44 ppm PERC, and on the 36-lb machine, 16 ppm. For "fifth generation" machines, instantaneous exposures could be higher during maintenance tasks than during loading and unloading.


Figure 11. Operator exposure to PERC during maintenance
on lint and button traps of a dry-to-dry, vented
drycleaning machine.

 


Figure 12. Operator exposure to PERC while cleaning lint traps
on a "fifth generation" machine.

Figure 13 presents operator exposures during still rake out on each machine. These stills were fitted with a pump that enabled the residue to be pumped directly into a safety can. This device reduced still rake out from a daily or weekly task to an approximately once-every-three-week task. Operator exposure to PERC while raking the still of the 46-lb machine was in excess of 156 ppm, the limit of detection, and that of the 36-lb machine averaged 35 ppm. During the cleaning of the 46-lb machine still, the instrument became saturated, which is shown by the horizontal line. The proximity of the 36-lb machine to a supply fan, which moved 2,100 cfm of air, helped to reduce airborne PERC concentrations.

Figure 13. Operator exposure to PERC while cleaning the
stills on two "fifth generation" machines.


Figure 14 gives exposures for changing the solvent filters. There was no still on this machine; rather, clay and carbon filters removed both soluble and insoluble soils. There were several tubes which held five filters each. Each tube was changed once for every 10,000 pounds of clothing cleaned, or approximately once a month. Real-time monitoring revealed that the greatest exposures occurred when the filters were placed into the hazardous waste container and when the new filters were placed into the tubes. Higher exposures probably resulted from contaminated air being displaced by filter deposition or insertion, as Figure 14 demonstrates.

Figure 14. Worker exposure to PERC during changing
clay/carbon solvent filters on a transfer machine.


Figure 15 shows worker exposure during machine maintenance on two nonvented, dry-to-dry machines. As was the case for most machines observed in the study, the lint and button traps were located behind the machines, near a wall. Fairly significant concentrations were discovered in this area. This background concentration can be attributed to exhaust from the secondary control device, located behind the machines. Additionally, lint from the lint trap and other hazardous waste was inappropriately stored behind the machines in a cardboard cylinder. The highest maintenance exposure occurred while the lint trap of the 80-lb machine was cleaned. The average exposure during this operation was approximately 180 ppm, as compared to the average exposure for cleaning the lint trap of the 70-lb machine, which was approximately 149 ppm.

Figure 15. Exposure to PERC while cleaning lint traps
on two dry-to-dry, nonvented drycleaning machines.
Ventilation

Most of the shops studied had inadequate local and general ventilation systems to control worker exposures. Local ventilation can be used to reduce worker exposure during a variety of tasks, such as loading/unloading the machine, maintenance, spotting, or waterproofing, but loading/unloading is the most important activity to control.

None of the evaluated machines provided 100 feet per minute (fpm) air velocity at the face of the drycleaning machine doors. The door face velocity recommended by much of the literature is 100 fpm.^56,57,58,59 In most cases, the airflow was significantly less, and in many cases there was no ventilation. This reduction and elimination of ventilation can be attributed to environmental regulations restricting solvent vapor emissions. Many machine designers have eliminated process ventilation to reduce emissions. Ventilation was used on machines in several evaluated shops as part of a secondary control that exhausted into the work environment. The air passed through a vapor recovery system that did not capture all of the PERC, and workers continued to be exposed to the remaining hazardous concentrations emitted from the machine. In one evaluated shop that exhausted air outside, air currents carried the contaminated air back into the shop through a nearby doorway.

General ventilation, appeared to be primarily used for comfort rather than for dilution of air contaminants. Many shops were located in buildings that were originally constructed for purposes other than drycleaning, and the ventilation systems were ill-equipped to remove or dilute hazardous vapors. Several shops had no operable windows to provide natural ventilation. Observed ventilation cases follow.

CASE 1:

At one shop evaluated, two modern "fifth generation" machines were used. These machines reduced the instantaneous PERC concentration in the cylinder to below 290 ppm (an order of magnitude lower than concentrations in machines of a different design). Because the PERC concentrations were so low, local exhaust ventilation was less of a concern. Although there was minimal airflow at the face of the open machine door, it was not a significant problem with respect to the overall control system because the concentration in the cylinder was so low.

Several propeller fans were located in the drycleaning area to provide dilution ventilation. Most of the PERC contaminated air originating from the machines was prevented from escaping into other areas because the room was under negative pressure. Figure 16 shows airflow measurements moving to and from the drycleaning area.

Figure 16. Airflow balance in/out of drycleaning room having
effective general ventilation.

CASE 2:

A transfer unit was used at another shop consisting of a washer and reclaimer, which were connected to a vapor condenser. During the final aeration cycle, fresh air was drawn into the reclaimer and exhaust