Journal of Exposure Analysis and Environmental Epidemiology (2003) 13, 325340 r 2003 Nature Publishing Group All rights

reserved 1053-4245/03/$25.00

www.nature.com/jea

Reduction of occupational exposure to perchloroethylene and trichloroethylene in metal degreasing over the last 30 years: inuences of technology innovation and legislation
RLIMANN, MARTIN SCHERINGER AND KONRAD HUNGERBU HLER JULIA VON GROTE, CHRISTIAN HU
Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology, Zurich, Switzerland

Occupational exposure to trichloroethylene (TRIC) and perchloroethylene (PERC) in metal degreasing is analyzed by calculating airborne concentrations for a large set of possible exposure scenarios (Scenario-Based Risk Assessment, SceBRA). Different types of degreasing machines ranging from open-top machines used until the 1980s to closed-loop nonvented machines used since the 1990s are investigated; the scope of the study is Germany. Concentrations are calculated for different kinds of releases (emissions from open baths, leakage, release of contaminated air during loading and unloading) with a dynamic two-box model for the near-eld and the far-eld. The concentration estimates are in good agreement with measured data. The airborne concentrations are compared to maximum workplace concentrations (MAK values). The full set of scenarios shows for which situations MAK values were exceeded and how the transition to newer degreasing machines reduced the occupational exposure by more than one order of magnitude. In addition, numbers of exposed workers are estimated for different years. While more than 25,000 workers in the near-eld were exposed to TRIC and PERC in 1985, the number is below 3000 since 1996, which is mainly due to technology changes, rationalization, automatization, and replacement of TRIC and PERC by nonchlorinated solvents. Journal of Exposure Analysis and Environmental Epidemiology (2003) 13, 325340. doi:10.1038/sj.jea.7500288

Keywords: mass-balance models, occupational exposure, metal degreasing, trichloroethylene, perchloroethylene, risk screening, risk assessment

Introduction
Efcient control of human exposure to the solvents trichloroethylene (TRIC) and perchloroethylene (PERC) is important because these solvents cause a variety of toxic effects, possibly including cancer (BUA, 1993, 1994; ECETOC, 1994, 1999), and because they are widely used as cleaning and degreasing agents with a potentially high number of exposed workers. Assessments of the occupational exposure to solvents such as TRIC and PERC are impeded by the complexity of the activity patterns of workers and the temporal and spatial variability of workplace conditions (Matthiessen, 1986; Jayjock and Hawkins, 1993; Fehrenbacher and Hummel, 1996). One difculty is that it is often impossible to ascertain exactly the conditions that determine the exposure in a specic case. A second problem is that the variability of a broad range of workplace conditions has to be included.

1. Address all correspondence to: Dr. Martin Scheringer, ETH Hoenggerberg, HCI G127, CH-8093 Zu rich, Switzerland. Tel.: +41-1-632-30-62. Fax: +41-1-632-11-89. E-mail: scheringer@tech.chem.ethz.ch Received 20 November 2002; accepted 18 April 2003.

With respect to the second problem, which is in the focus of the present study, sufciently exible methods for covering the wide range of possible exposure situations are required. To this end, we adapt the method Scenario-based risk assessment (SceBRA), which has been developed for characterizing exposure to solvents used in a variety of applications (Scheringer et al., 2001), to the assessment of the occupational exposure to TRIC and PERC in metal degreasing. Metal degreasing is performed in highly variable settings; for metal parts of different sizes and shapes; and with machines of different capacities and technological standard. Therefore, the variety of scenarios employed in the SceBRA method is a suitable approach to characterize this technology. The objectives of this paper are (i) to quantify reliably the exposure levels resulting from the use of TRIC and PERC in various metal degreasing machines, ranging from highly emissive machines developed in the 1950s and 1960s to closed-loop machines used since 1990, and (ii) to demonstrate the effect of technology development and stricter legislation on the two dimensions of risk considered in SceBRA, namely the risk quotient, that is, the ratio of airborne concentrations and an effect level, and the number of exposed workers. The scope of the study is Germany

von Grote et al.

Occupational exposure to PERC and TRIC in metal

because this country has one of the strictest legislations regulating the use of TRIC and PERC and because data on machine types and numbers as well as working conditions could be obtained. The metal degreasing technology is described and the SceBRA method is introduced in the Materials and methods section. The mass-balance models employed and the model parameters describing metal degreasing settings are dened in the section Model and Exposure Scenarios. The results obtained for airborne TRIC and PERC concentrations, risk quotients, and numbers of exposed workers for different machine types and years are presented in the Results section.

Materials and methods

Metal Degreasing Metal degreasing can be carried out either with aqueous systems, with hydrocarbons, or with chlorinated solvents. Chlorinated solvents are often used in vapor degreasing equipments for difcult cleaning tasks such as metal parts that need to be totally dry, are very small, are temperature sensitive, or have lots of cavities (Leisewitz and Schwarz, 1994). The fact that TRIC and PERC are both suspected to cause cancer (BUA, 1993, 1994; ECETOC, 1994, 1999) led to several revisions in the German law regulating the use of PERC and TRIC. The German 2nd Federal Immission Protection Directive of 1986 (Zweite Bundes-Immissionsschutzverordnung: 2nd BImSchV) (BImSchV, 1986)),

the 2nd BImSchV of 1990 (BImSchV, 1990), and the amendment of the 2nd BImSchV 1990 in 2001 (BImSchV, 2001) prescribe not only the chemicals to be used in different applications but also technological standards of metal degreasing machines, which have changed substantially over the last 30 years. The technological development of metal degreasing machines can be grouped into ve machine types (see Figure 1). Information on these types was taken mostly from sales brochures (Manufacturers, 19602001). Type I machines are fully emissive open-top machines with watercooling at 151C. They consist of several solvent baths and a vapor bath and have a suction device at the rim of the baths. Type I machines are followed by open-top machines with electro-cooling at 21C (type II). In type III, the baths are encased. Type IV machines are the rst one-chamber metal degreasers (there for the rst time the solvent is brought to the metal parts and not vice versa) with an integrated recirculated air dryer condensing the solvent and a recycling loop. The parts are dried with refrigeration-cooling at temperatures between 201C and 401C. Type IV machines and some upgraded type III machines fulll the 2nd BImSchV of 1986. The modern type V machines, which are in use today, are closed one-working-chamber machines, with closed-loop drying and recycling systems with refrigeration-cooling where the cooled air is additionally directed over activated carbon before re-entering the working chamber to dry the metal parts. Type V machines do not release any exhaust air to the environment. The chamber concentration is monitored continuously and the metal parts are only released if the concentration in the cleaning chamber is below 1 g/m3.

types I and II

type III

vent

bath 1

bath 2

vapor

bath 1

bath 2

vapor

type IV

vent

type V

drying cycle refrigeration tank activated carbon

drying cycle refrigeration

working chamber

working chamber

Figure 1. Technological progress of metal degreasing machines represented by ve different machine types.
326

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

Occupational exposure to PERC and TRIC in metal

von Grote et al.

This technological standard is prescribed in the 2nd BImSchV of 1990. The amendment of this regulation of 2001 requires type V machines that are operated under vacuum (type V B machines) for the use of TRIC. The fully open type I machines were rst built in the early 1950s and type II machines in the late 1950s. The encased type III machines were developed in the middle of the 1960s. These three types of machines were in use until the 2nd BImSchV was enacted in 1986. Type IV machines that fulll the standard of 1986 came up in the middle of the 1970s. After 1986 and after the transitional regulation, only type IV machines and upgraded type III machines were allowed. In the late 1980s, type V machines were developed. The regulations of 1990 (BImSchV, 1990) require closed-loop type V machines for all metal degreasing installations using PERC and, after the amendment of 2001, closed-loop type V B machines with integrated vacuum system for degreasers using TRIC.

Model and exposure scenarios

Mass-balance Model As there are very few measured data reported in the past and, in addition, the available data often lack a detailed description of the machine technology and the workplace surrounding, it is necessary to simulate the airborne concentrations in metal degreasing facilities (Nicas and Jayjock, 2002). To this end, indoor air quality models based on mass-balance equations that account for incoming and outgoing material are used. The mass-balance equation for a one-box model reads . dC t E 1 kC t V dt where C(t) is the pollutant concentration in air (g/m3), t is the . time (h), E is the emission rate (g/h), V the box volume (m3), and k is the overall loss rate constant (h1). If all production and loss mechanisms are constant with time, Eq. (1) may be integrated to give . E 1 ekt 2 C t C0 ekt Vk where C0 (g/m3) is the concentration of the pollutant at time . t 0 and E=Vk C 1 is the steady-state concentration. In many cases, however, a simple one-box approach is inadequate. For room volumes greater than 500 m3, which are common in metal degreasing facilities, complete mixing is often not given (Gmehling et al., 1989). Therefore, some studies have recommended the use of mixing factors to account for noncomplete mixing (Drivas et al., 1972; Matthiessen, 1986; Jayjock, 1988). Another approach is to use a two-compartment mass-balance model (Cherrie et al., 1996; Furtaw et al., 1996; Keil, 1998, 2000). Nicas (1996) discusses the advantages of two-box models versus the use of mixing factors. In a two-box model, the concentrations in the vicinity of the source (near-eld) and further away (far-eld) can be distinguished. This results in two coupled linear differential equations, see Eq. (3). A two-box model may still be insufcient to describe complex situations such as the concentration in the breathing zone during spray painting. Under these conditions, additional boxes may be introduced to account for the complexity or computational uid dynamic models may be applied (Flynn and Sills, 2000; Nicholson et al., 2000), but the need for adequate measurements or estimates of the transfer coefcients precludes the use of full, multibox models for many problems of practical interest (Ryan et al., 1988). In this study, a two-box mass-balance model is used to calculate the airborne concentrations for the near-eld and the far-eld of metal degreasing devices (see Figure 2). In the . simplest case, there is a constant mass ow EA (g/h) from the machine into the ctitious box A. The air exchange between box A with volume VA (m3) and concentration CA (g/m3),
327

Scenario-based Risk Assessment (SceBRA) The SceBRA method is intended to represent the potential health risk through exposure to chemicals used in many different applications during different stages of the products lifetime such as production, formulation, transport, and end-use (Scheringer et al., 2001). In SceBRA, various scenarios for possible exposure situations are investigated. The resulting risk is then characterized by two indicators: (i) the risk quotient, r, which is the ratio of concentration and effect level, and (ii) the number of exposed workers, N. Here, inhalative exposure to volatile solvents is calculated with mass-balance models for the vicinity of the machine (near-eld) where the machine operators are situated, as well as for the far-eld where other work is carried out. The long-term airborne concentration C, calculated as time weighted average (TWA) for an 8-h working day, is compared with the German MAK (maximum workplace concentration) value, corresponding to threshold limit values (TLVs), which leads to the risk quotient r C/MAK. The MAK values of 50 ppm (0.345 g/ m3) for PERC and of 50 ppm (0.270 g/m3) for TRIC have not been changed since 1982 for PERC and since 1970 for TRIC (BUA, 1994, 1993). In the SceBRA calculations, good working practice is assumed, as there is hardly any possibility to abuse the machines that are in use today. For older machines, however, working practice in some instances may have differed largely from good working practice; it has been reported that under time pressure, cleaned parts had been taken out before being entirely dried (Mannheim et al., 1979; Leisewitz and Schwarz, 1994). However, because of lack of data, it was not possible to include bad working practice for older machines.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

von Grote et al.

Occupational exposure to PERC and TRIC in metal

box B
concentration CB

fresh air

exhaust air kL kB kA

kL
box A

air exchange between box A and box B

concentration CA

machine
.

emission

EA

Figure 2. Two-box model with ctitious box A for the near-eld of the machine and box B representing the far-eld.

and box B with volume VB (m3) and concentration CB (g/ m3) is given by the exchange rates kA (h1) and kB (h1). kB and kA are related by kB (VA/VB)kA. For the air exchange of box B with the environment, denoted by kL (h1), it is assumed that the background concentration of the fresh air can be neglected. The mass-balance equations for the two boxes are . dCA EA VB CA kA CB kB dt VA VA 3 dCB VA CA kA CB kB kL dt VB Equation (3) represents a system of two coupled rst-order linear inhomogeneous differential equations that can be solved for CA(t) and CB(t). Different emission factors can be used to characterize the variability of the release of chemicals from intermittent sources (Franke and Wadden, 1987; Wadden et al., 1991; Keil, 1998). Here, the exposure models for the use of TRIC and PERC in metal degreasing include three different . emission factors: (i) a continuous factor Ec1 (g/h) for diffuse emissions from open bath surfaces (open-top machines) or from leakage of sealing (closed machines); (ii) a second . continuous factor Ec2 (g/h) for emissions from cleaned parts and metal parts with cavities, and (iii) a periodic factor Ep (g/batch) for repetitive emissions from chamber air during loading and unloading (Wadden et al., 1989; Wadden et al., 1991; Scheff et al., 1992). For the two continuous emission factors, the inhomogeneous differential equation system is given by Eq. (3) where . . . EA is replaced by Ec1 or Ec2 . For the periodic emission factor, the homogeneous differential equation system is used, . which means that Eq. (3) has no continuous source term EA .
328

Instead, periodic emissions Ep (g/batch) are added at each time when the machine is unloaded so that the new 0 background concentrations Cp ;A;i1 for box A are obtained: Ep 0 Cp 4 ;A;i1 Cp;A t itb VA where i indicates the number of unloads and tb the batch time. With these three sets of equations, the airborne concentrations are calculated for an 8-h working day with the degreasing devices continuously in use (see Figure 3). The concentrations in boxes A and B are obtained as CA t Cc1;A t Cc2;A t Cp;A t CB t Cc1;B t Cc2;B t Cp;B t 5

Metal Degreasing Scenarios To represent the different exposure situations in metal degreasing, a range of possible and plausible user scenarios are dened. The scenarios are described by machine parameters (such as sizes, technology types, and loads of machines) and emission factors. Another factor that inuences the occupational exposure is the workplace surrounding, which is described by workplace parameters such as room size and air exchange rates. Machine Parameters The machine technology has changed fundamentally since the 1950s. Therefore, ve types of machines are dened according to information from sales brochures (Manufacturers, 19602001) (see the section Metal Degreasing). To account for small differences in each type due to different manufacturers, subtypes are dened which vary in the number of baths (types I and II), in the cooling temperature (type IV), or in the existence of a vacuum drying device (type V) (see Table 1). Each machine
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

Occupational exposure to PERC and TRIC in metal

von Grote et al.

1.6 1.4

PERC concentration (g/m3)

1.2 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8

dynamic concentration box A,CA(t) dynamic concentration box B,CB(t) long-term concentration box A long-term concentration box B

MAK value: 0.345 g/m3

10

12

time (h)

. Figure 3. Dynamic concentration of PERC used in a type III, size 5 metal degreasing machine. Parameter values are: Ep: 117 g/batch, Ec1 : 17.2 g/h, . 3 1 1 Ec2 : 73.4 g/h, a: 0.75 h, tb: 0.75 h, V: 400 m , kL: 5.5 h , kA: 6 h (for parameter denition, see the section Metal Degreasing Scenarios).

Table 1. Machine parameters: characteristics of types and subtypes.

Types and subtypes IA IB II A II B III IV A IV B VA VB Characteristics Two baths Two baths and vapor degreasing Two baths Two baths and vapor degreasing No subtypes Cooling temperature: PERC 201C, TRIC 301C Cooling temperature: PERC 401C, TRIC 401C No vacuum drying Vacuum drying

degreasing, the parts are heated till they have the same temperature as the vapor, then condensing stops and the vapor degreasing process is nished (Mannheim et al., 1979)). Therefore, size 5 machines have longer batch times than size 1 machines (see Table 2) (Manufacturers, 1960 2001). . Emission Factors The diffuse emission factor Ec1 (g/h) describes a continuous discharge out of the machine into box A. For open-top machines (types I and II), this corresponds to releases from open bath surfaces, which emit as long as the baths are heated. In fully emissive machines, the solvent vapor layer at the height of the air-cooler coils separates the vapor phase of the bath and the workplace air. Additionally, a rim exhauster directs the ascending hot vapor toward the side panel where the coolers are situated; the recirculation ow draws in fresh air from the workplace. In addition to the temperature of cooling and the capacity of the rim exhauster, also the width of the baths and the freeboard rate, which is the ratio of freeboard height and bath width, inuences the emission mass ow. Owing to lack of adequate information on the volumetric ow rate and the dimensions and location of the exhauster inlet, an exact calculation of the emission mass ow is not feasible (Heinsohn, 1991). However, a simplied calculation makes it possible to determine reasonable estimates. The emissions from open bath surfaces are approximated with Ficks rst diffusion law: Jz D DC Dh 6

Table 2. Machine parameters: load and batch time.

Machine size 1 2 3 4 5 Load (kg) 4050 5060 120150 Approx. 600 Approx. 1000 Batch time tb (h) 0.080.12 0.10.17 0.140.2 0.330.5 0.51

type is manufactured in different sizes adjusted to specic purposes. For this study, the ve most common sizes are chosen. They range from small size 1 machines with a loading of 4050 kg to large size 5 machines with loadings of more than 1000 kg (Manufacturers, 19602001). Among the different sizes, size 3 machines are most commonly used. The batch time, tb, depends on the material, the geometry, and the weight of the parts to be cleaned (during vapor
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

where Jz (g/(m2 h)) is the diffusion mass ow in the z direction, D (m2/h) the molecular diffusion coefcient, DC (g/m3) the concentration gradient between the saturation
329

von Grote et al.

Occupational exposure to PERC and TRIC in metal

concentration in the cooling zone and the air in box A, and Dh (m) the diffusion length. Dh was tted so that the mass ow equals the results reported by CEFIC (1984), which state that if common freeboard rates of 0.5 for open-top degreasers are increased to 0.75, solvent emissions can be reduced by 2530%. As a cross-check, mass ows based on the tted Dh were calculated for different cooling temperatures. The results agree with the nding that emissions can be reduced by 50% if the cooling temperature in open-top machines is reduced from 51C to 301C (CEFIC, 1984). Multiplication of Jz with the bath . surface area results in the continuous emission factor Ec1 . . For encased and closed machines (types IIIV), Ec1 is taken as a leaking rate proportional to the solvent volume present in the machine (CerclAir, 1993). The cleaned parts that are taken out of the machine after washing and drying are sometimes not completely dry, especially if parts with cavities or cupped parts are cleaned. Emissions from cleaned parts depend on the drying time, the surface of the cleaned parts, and on the existence of cavities and cupped areas (Scheff et al., 1992). In degreasing, metal parts vary considerably in size, weight, and shape (e.g., from small needles to whole car bodies). Therefore, a range of standard metal parts is dened here with different amounts, sizes, and arrangements of plates and closely packed spheres. Between the closely packed spheres there are cavities where residual solvent can remain; this amount determines the emissions from cavities. Together with the amount that is dragged out by parts without cavities, this leads to the total amount of residual solvent Ec2 (g). As a worst-case scenario, it is assumed that all of the residual solvent will evaporate continuously over a certain period a (h), which lies between 0.5 and 1.5 times the batch time. The . quotient of Ec2 and a leads to the emission factor Ec2 (g/h). The residual PERC amounts on noncupped parts dragged out of fully emissive open-top machines are assumed to be 3.94.3 g PERC per m2 of metal surface after a solvent bath and 3.23.6 g/m2 after vapor degreasing. The lower boiling point of TRIC and therefore the shorter time till the metal parts have achieved the boiling temperature is the reason that metal parts that are cleaned with TRIC drag out smaller amounts of residual solvent. These gures correspond well with reported amounts of solvent that had to be relled in open-top machines (Mannheim et al., 1979). For the spheres it is assumed that there are some cavities between the spheres, which are not easily dried. It is assumed that for types IIII machines, 5% of the sphere surface can be taken as a cupped area, which drags out 32.535.7 g PERC per m2 of sphere surface (Mannheim et al., 1979). Each time when the machine is loaded or unloaded with a basket of metal parts, solvent-charged air from the chamber volume is exchanged with workplace air, which is described by the periodic emission factor Ep (g/batch). The concentration in the working chamber can be calculated for the open330

top machines as the solvents saturation concentration at the height of the cooling device. During loading, the vapor volume proportional to the volume of the metal parts is displaced; during unloading, the basket volume minus the volume of the parts is dragged out. The rim exhauster of types I and II machines captured 50% of the solvent, whereas for types III and IV 90% could be kept back from emitting into the workplace (Mannheim et al., 1979; CEFIC, 1984; Leisewitz and Schwarz, 1994). In type V machines, the cleaned parts are released only if the chamber concentration is below 1 g/m3 (BImSchV, 1990). The emission factor is calculated as intermittent emissions that occur at the very moment when the door is opened. In Table 3, the three different emission factors for each machine size are shown for PERC as an example. Note that the emissions through loading and unloading contribute most . to the overall emissions (Ep and Ec2 ). Wadden et al. (1991) report emission factors of 16.9 g TRIC/basket of parts for an open-top degreaser where . emissions from cleaned parts (Ec2 ) were not taken into consideration. The described machine is similar to a type II and size 3 machine. In the SceBRA calculations, the amount . of Ep added to the amount of Ec1 is 18.521.0 g/basket, which shows good correspondence.

Workplace Parameters Not only do the degreasing equipment and the operation of the machine inuence the occupational exposure, but also the workplace surrounding. The airborne concentration is dependent on the room volume and the air exchange rates. Production hall volumes can vary from small volumes of about 300 m3 up to several thousand cubic meters. Here, relatively small volumes of 400 and 600 m3 are assumed for the whole workplace. The volume is divided into a ctitious working box of a volume of 100 m3 and the rest of the volume, where other work not related to the degreasing is carried out. According to Gmehling et al. (1989) this is an appropriate assumption. Typical air exchange rates in industrial facilities lie between 4 and 15 times per hour (Hayes, 1991; Recknagel et al., 2000). Wadden et al. report an air exchange of 6 h1 for metal degreasing with TRIC (Wadden et al., 1989) and 5.1 h1 in an offset printing shop without local exhaust (Wadden et al., 1995). Workplaces for metal degreasing are equipped with local exhaust devices. Therefore, an efcient and complete mixing for the near-eld (box A) can be assumed. On the other hand, air exchange rates above 8 h1 are not likely, as there would be an uncomfortable draught for the workers. Here, kL values of 6 and 6.5 h1 are used for open-top types I and II machines, whereas for encased and closed machines values of 5.5 and 6 h1 are employed. The air exchange rate between box A and box B kA is assumed to be greater than the overall exchange rate kL because particularly the older machines are heat sources and create an upward ow of warm air (Heinsohn, 1991; Bach et al., 1992). Therefore, the
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

Occupational exposure to PERC and TRIC in metal

von Grote et al.

Table 3. Emission factors for PERC.

Type . Diffuse emission factor Ec1 (g/h) IA Min Max IIA Min Max III Min Max IVA Min Max VA Min Max . Emission factor from cleaned plates Ec2 IA Min Max IIA Min Max III Min Max IVA Min Max VA Min Max Size 1 1.27E+01 1.40E+01 7.05E+00 7.76E+00 2.12E+00 2.33E+00 1.50E+00 2.40E+00 1.14E+00 1.50E+00 (g/h) for period a 3.20E+02 1.06E+03 3.20E+02 1.06E+03 3.34E+01 1.10E+02 8.34E+00 2.75E+01 4.17E+00 1.38E+01 Size 2 1.82E+01 2.00E+01 1.01E+01 1.11E+01 3.03E+00 3.33E+00 2.17E+00 3.47E+00 1.65E+00 2.17E+00 Size 3 2.26E+01 2.49E+01 1.26E+01 1.38E+01 3.77E+00 4.15E+00 2.92E+00 4.67E+00 2.22E+00 2.92E+00 Size 4 4.12E+01 4.53E+01 2.29E+01 2.52E+01 6.86E+00 7.55E+00 3.33E+00 5.33E+00 2.53E+00 3.33E+00 Size 5 1.03E+02 1.13E+02 5.72E+01 6.29E+01 1.72E+01 1.89E+01 8.33E+00 1.33E+01 6.33E+00 8.33E+00

3.21E+02 1.06E+03 3.21E+02 1.06E+03 3.34E+01 1.10E+02 8.35E+00 2.76E+01 4.17E+00 1.38E+01

4.00E+02 1.32E+03 4.00E+02 1.32E+03 4.17E+01 1.38E+02 1.04E+01 3.44E+01 5.21E+00 1.72E+01

6.07E+02 2.00E+03 6.07E+02 2.00E+03 6.32E+01 2.09E+02 1.58E+01 5.22E+01 7.90E+00 2.61E+01

7.05E+02 2.33E+03 7.05E+02 2.33E+03 7.34E+01 2.42E+02 1.84E+01 6.06E+01 9.18E+00 3.03E+01

Periodic emission factor from chamber air Ep (g/batch) IA Min 1.47E+00 Max 1.79E+00 IIA Min 8.14E01 Max 1.06E+00 III Min 6.36E01 Max 8.12E01 IVA Min 1.77E01 Max 2.82E01 VA Min 3.53E03 Max 7.06E03

3.05E+00 3.73E+00 1.70E+00 2.20E+00 1.32E+00 1.69E+00 3.68E01 5.89E01 7.36E03 1.47E02

8.55E+00 1.05E+01 4.75E+00 6.18E+00 3.71E+00 4.74E+00 1.03E+00 1.65E+00 2.06E02 4.12E02

4.59E+01 5.61E+01 2.55E+01 3.32E+01 1.99E+01 2.54E+01 5.53E+00 8.85E+00 1.11E01 2.21E01

2.70E+02 3.30E+02 1.50E+02 1.95E+02 1.17E+02 1.50E+02 3.25E+01 5.21E+01 6.51E01 1.30E+00

kA values are assumed to be 7 and 7.5 h1 for open-top machines and 6 and 6.5 h1 for types IIIV machines.

Number of Exposed Workers In addition to the risk quotient, the number of exposed workers is estimated in SceBRA. For this purpose, a distinction between the two boxes with their different concentrations must be made. The number of exposed workers in the near-eld can be estimated from the number of machines in use in different years. It is assumed that between 1.5 and 3 workers (average of 2.25) were required to handle the mostly manually operated opentop types I and II machines. The encased and fully closed machines are automated and therefore only between 1 and 1.6 workers (average of 1.3) are needed (Nader, 2001). Furthermore, it is assumed that the workers operating the machines are present in the near-eld during the whole 8-h working day.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

The estimation of the number of workers in the far-eld is more difcult. As metal degreasing machines are used all over in the metal processing industries, workplaces can vary considerably from small production halls with only a few workers to whole production lines. As the open-top machines were mostly used in times when the production processes were not automated, it is assumed that for types I and II machines 310 workers work in the far-eld, whereas for types IIIV machines the number is set at 28 workers. In Table 4 the industrial estimates of the numbers of metal degreasing machines using PERC are listed for the years 1985 (only former West German states, 1 year before the 2nd BImSchV of 1986) (Adams and Jeker, 1986), 1991 (only former West German states, during transitional regulation for 2nd BImSchV 1990) (Adams, 1993), 1996 (after the 2nd BImSchV 1990) (Adams, 1997), and 1999 (Nader, 2001).
331

von Grote et al.

Occupational exposure to PERC and TRIC in metal

Table 4. Estimates of the number of metal degreasing machines using PERC in Germany for the years 1985*, 1991*, 1996, and 1999 (*only former West German states).
Machine type Sizes 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 6261 3280 4028 3127 1985* (a) 1444 149 122 50 6 1444 149 122 49 5 0 452 379 341 38 0 451 379 341 38 18 72 90 32 4 13 52 13 0 0 1 6 1 0 0 1991* (b) 648 67 55 22 3 649 67 55 22 2 9 35 44 16 2 9 35 44 16 2 88 353 442 157 20 27 106 133 47 6 27 106 133 47 6 44 177 221 79 10 24 94 118 42 5 70 282 352 125 16 23 94 117 42 5 1408 1408 75 300 375 133 17 38 150 188 67 8 1351 1351 1996 (c) 1999 (d)

IA

IB

IIA

IIB

III

IVA

IVB

VA

VB

Total Vapor degreasing only

Sources: (a) Adams and Jeker (1986); (b) Adams (1993); (c) Adams (1997); (d) Nader (2001).

332

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

Occupational exposure to PERC and TRIC in metal

von Grote et al.

Results
The two-box model allows us to specify the dynamic and the long-term concentration of each single scenario. By combining all independent input parameters, up to 2160 possible and plausible exposure scenarios can be obtained for each type of machine. From these, the maximum and minimum scenarios are selected for each size of machine. Thereby the number of scenarios is reduced to 20 for each machine type (two values for ve sizes and two subtypes). On this basis, the ve types of machines can be compared, and characteristic differences between the two solvents can be depicted. Finally, for selected years the number of exposed workers, N, is combined with the risk quotients, r, of the different types and sizes of machines. On this basis, the changes in the number of exposed workers and the magnitude of the risk quotient can be compared over a period of 15 years.

2.5

basic scenario
Ep and Ec2, . max Ep and Ec2, . av E .c1, max Ec1, av V, max kL and kA, max
spheres instead of plates

long-term TRIC concentration (g/m3)

1.5
tb, av

tb, max

Dynamic Concentration for a Single Scenario With one set of input parameters for a specic machine (machine parameters, emission factors, workplace parameters), the dynamic and the long-term concentrations over an 8-h working day can be calculated for a single scenario. In Figure 3, these concentrations are shown for PERC used in an encased type III and size 5 metal degreasing machine as an example. In this scenario, metal plates are washed with a batch time tb of 0.75 h. Every time when the basket with metal plates is unloaded, a peak concentration can be observed. The long-term concentration is calculated as the integral of the concentration curve over one batch time, divided by tb. After the second unloading (after 1.5 h), the steady state has been achieved; thereafter a regular pattern of subsequent peaks follows. After 8 h, the operation of the metal degreasing device stops and the airborne concentration decreases to the initial value. The long-term concentration lies at 0.42 g/m3 in box A and at 0.11 g/m3 in box B, which means that for this specic scenario the long-term concentration in the near-eld, the working area, is 1.2 times the MAK level of 0.345 g/m3, whereas the long-term far-eld concentration reaches only 32% of the MAK level. Parameter Combination For one machine of type II A and size 5, the contribution of the different input parameters to the range of airborne concentrations is depicted in Figure 4 by systematically combining minimum, average, and maximum values of all input parameters. The basic scenario indicated by the circle in Figure 4 represents the minimum values of the emission factors, the room volume, the air exchange rate, and the batch time and the case that plates are cleaned. The black diamonds show the concentrations obtained if only a single parameter is increased from the minimum to the maximum . value (Ep, Ec2 , tb: also average values). Changing several parameters at a time results in the combination of the
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

scenarios
Figure 4. Effects of changes in different model parameters on the longterm concentration in box A of a type IIA, size 5 metal degreasing device using TRIC. Starting from the basic scenario (circle), the different parameters are varied from their minimum to maximum values (emission factors and batch time: also average values), which leads to the concentrations indicated by black diamonds. The gray marks indicate the concentrations obtained by combinations of two or more parameter changes. For this machine, the minimum and the maximum concentrations are 2.2 and 0.9 g/m3.

individual effects; these scenarios are represented by the gray marks. Parameters that vary only sparsely between different scenarios have a relatively small inuence on the long-term concentration. Such parameters are the continuous emission factor from open baths and leakage of . closed machines (Ec1 ) and the room volume (V). This is . because Ec1 represents emissions caused by machines in standby mode with only small uctuations. The room volume has a small effect because it inuences mainly the concentration level in the far-eld and the air exchange from box B to box A (kB). Parameters that show stronger effects are primarily the batch time tb, the kind of cleaned parts, the coupled air exchange rates kL and kA as well as the coupled periodic . emissions Ep (g/batch time) and Ec2 (g/h). The length of tb inuences the long-term airborne concentration most strongly, as it determines the spacing between the emission peaks. The difference between the two kinds of metal parts is due to the different volume and surface ratio of plates and . spheres. The variations in Ep and Ec2 represent the differences in the work process; they depend on the saturation concentration in the machines, the geometry and
333

concentration (g/m3)

334
von Grote et al.

10.000

TRIC

10.000

PERC

1.000
MAK-value: 0.270 g/m3

1.000
MAK-value: 0.345 g/m3

0.100

0.100

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

0.010

0.010

0.001
5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1

0.001
5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1

II

III

IV A

IVB

VA

VB

II

III

IV A

IVB

VA

VB
Occupational exposure to PERC and TRIC in metal

types and sizes

Figure 5. Comparison of near-eld concentrations (logarithmic scale) of the ve types (and relevant subtypes) subdivided into ve sizes of metal degreasing machines for TRIC and PERC. Arrow: machine characterized in Figure 4.

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

335

Occupational exposure to PERC and TRIC in metal

Table 5. Exposure measurements of TRIC used in metal degreasing.

Year 1942 Country USA Machine type I I I Machine description n.d. n.d. n.d. Mostly open-top degreasers n.d. Vapor degreasing with exhaust Vapor degreasing without exhaust n.d. Measurement During normal operation n.d. n.d. 510 min samples n.d. n.d. n.d. n.d. Sample Number of Number of Range machines samples (ppm) n.d. 200 n.d. 84 50 n.d. n.d. 116 n.d. n.d. n.d. 5600 n.d. n.d. 2247 65.0165.0 n.d. 90%-pct. 100 75%-pct. 50 n.d. Range (mg/m3) n.d. n.d. 273240 n.d. n.d. 119254 351891 n.d. Average Average References (ppm) (mg/m3) 128.9 120.9 n.d. 49.8 59.6 n.d. n.d. 55.9 696.1 652.9 n.d. 268.9 321.8 n.d. n.d. 301.9 n.d. Morse and Goldberg (1943) Raaschou-Nielsen et al. (2002) Bardodej and Vyskocil (1956) Grandjean et al. (1955) Raaschou-Nielsen et al. (2002) Nicklin (1972) Nicklin (1972) Raaschou-Nielsen et al. (2002) CEFIC (1984)

Personal 108 Air Air Air Air n.d. n.d. n.d. n.d.

194759 Denmark 1956 1955* n.d.

Switzerland I and II I or II I or II I or II I or II I and II

196096 Denmark 1972 1972 Germany Germany

Personal 1 Personal 1 Air and personal Air n.d. n.d.

197079 Denmark 1984 Western Europe

Open-top equipment n.d.

90%-pct. 540 n.d. 75%-pct. 270 n.d. 13.0

198089 Denmark 198591 Germany 199095 Germany 199095 Germany

II and III

n.d.

n.d. TWA and range measurements TWA TWA

III and IV n.d. IV and V IV and V Manually operated Automatic without exhaust Automatic with exhaust n.d.

Air and personal n.d.

n.d. 267

371 748 14 23

70.2 83.2 41.0 22.1

90%-pct. 101.3 90%-pct. 547 15.4 90%-pct. 56.3 90%-pct. 304 7.6 95%-pct. 109.6 95%-pct. 110 90%-pct. 52.2 90%-pct. 282 4.1 95%-pct. 104.8 95%-pct. 566 90%-pct. 44.3 90%-pct. 239 10.0 95%-pct. 50%-pct. 90%-pct. 95%-pct. 90%-pct. 67.4 44.6 131.7 168.9 52.4 95%-pct. 50%-pct. 90%-pct. 95%-pct. 90%-pct. 364 241 n.d. 711 912 283 11.3

Raaschou-Nielsen et al. (2002) BK-Report (1999) Bock et al. (1999) Bock et al. (1999)

Personal 7 Personal 18

199095 Germany

IV and V

TWA

Personal 74

159

54.0

Bock et al. (1999)

199095 Germany

IV and V

o1 h

Personal 22

30

n.d.

Bock et al. (1999)

199297 Germany

IV and V

n.d.

1997* 2000

Australia Germany

IV and V V

n.d. n.d.

n.d. 97 TWA and only for 1992 range measurements TWA Personal n.d. Information from Air n.d. manufacturers

179

61.0

BK-Report (1999)

5 n.d.

n.d. 2.05.0

n.d. 10.827.0

8.9 n.d.

48.1 n.d.

Neghab et al. (1997) Manufacturers (19602001)

von Grote et al.

n.d. not dened; *no year of measurement given, year of publication used; pct. percentile.

336
von Grote et al.

Table 6. Exposure measurements of PERC used in metal degreasing.

Year 198285 1985 1988 1988 1991 1991 199095 Country Finland Germany Germany Germany Germany Germany Germany Machine type n.d. n.d. IV IV IV and V IV and V IV and V Machine description n.d. n.d. n.d. n.d. n.d. n.d. Manually operated without exhaust Automatic without exhaust Automatic without exhaust n.d. Measurement n.d. n.d. n.d. TWA n.d. TWA TWA Sample Air Air Air Personal Air Personal Personal Number of machines 3 n.d. n.d. n.d. n.d. n.d. 8 Number of samples n.d. n.d. n.d. n.d. n.d. n.d. 12 Range (ppm) 1.85.0 12.520 7.210 n.d. 1.36.2 1.42.0 90%-pct. 92.3 Range (mg/m3) 12.427.0 86138 5069 n.d. 943 1014 90%-pct. 637 Average (ppm) 3.0 n.d. n.d. 8.4 n.d. n.d. 1.0 Average (mg/m3) 20.7 n.d. n.d. 58.0 n.d. n.d. 6.9 References Rantala et al. (1992) Niemann (1995) Niemann (1995) Niemann (1995) Niemann (1995) Niemann (1995) Bock et al. (1999)

199095

Germany

IV and V

TWA

Personal

19

31

95%-pct. 116.4 90%-pct. 46.7 95%-pct. 73.2 90%-pct. 48.0 95%-pct. 50%-pct. 90%-pct. 95%-pct. 110 3.816.4 o0.7 n.d. 95%-pct. 78.8 5.1 39.6 46.2

95%-pct. 803 90%-pct. 322 95%-pct. 505 90%-pct. 331 95%-pct. 50%-pct. 90%-pct. 95%-pct. 769 26113 o5 n.d. 95%-pct. 544 35 273 319

2.6

17.9

Bock et al. (1999)

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

199095

Germany

IV and V

TWA

Personal

135

279

5.9

40.7

Bock et al. (1999)

199095

Germany

IV and V

o1 h

Personal

22

30

n.d.

n.d.

Bock et al. (1999)

1996 1996 1993 1993 1998

Finland Finland Germany Germany Germany

IV and V IV and V V V V

n.d. n.d. n.d. n.d. n.d.

n.d. n.d. n.d. TWA Control of TWA according to 2nd BImSchV 1990

Air Air Air Personal Air

1 1 n.d. n.d. 9

n.d. n.d. n.d. n.d. 52

10.1

70

5.2 10.1 n.d. 2.8 0.5

35.9 69.7 n.d. 19.3 3.5

OECD (1996) OECD (1996) Niemann (1995) Niemann (1995) Thoulass (1998)

Occupational exposure to PERC and TRIC in metal

n.d. not dened; * no year of measurement given, year of publication used: pct. percentile.

Occupational exposure to PERC and TRIC in metal

von Grote et al.

16000 14000 1985 PERC number of exposed workers 12000 10000 8000 6000 4000 1991 TRIC 2000 1996 PERC 1999 PERC 1996 TRIC 1999 TRIC` -2 -1.5 -1 -0.5 log risk quotient 0 0.5 1 1.5

1985 TRIC

1991 PERC

0 -2.5

Figure 6. Cumulative log r, N-plot of PERC and TRIC in the near-eld for the years 1985*, 1991*, 1996, and 1999 in Germany (*only former West Germany).

volume of the metal parts, and the amount of solvent that is dragged out. The air exchange rates kL and kA directly inuence the removal of the pollutant. In addition to the effects depicted in Figure 4, the major impact on the airborne concentration within one machine type is caused by the size of the machine, see concentration ranges given in Figure 5 for all types and sizes. The machine size directly inuences the bath volumes and bath surfaces, the volumes of the baskets and, therefore, the surfaces of the cleaned metal parts.

Comparison of Machine Technologies The different machine technologies can be compared by analyzing the concentration ranges for each size of machine. In Figure 5, where the airborne concentrations are given for all ve machine types (and for different subtypes if they are discernible in terms of concentrations), it can be seen that for both solvents, the concentration ranges decrease with decreasing machine size, and are reduced by more than two orders of magnitude from fully emissive open-top type I to closed-loop type V B machines. Fully emissive types I and II and encased type III machines using TRIC show a bigger variation between the different sizes of machines than machines operated with PERC. This can be attributed to the fact that the vapor pressure of TRIC is higher for the temperature range in which these machines are used. Therefore, higher concentrations are achieved by increasing the bath surface and the volume of the basket from small size 1 to size 5 machines. In Tables 5 and 6, measured concentration data are listed for TRIC and PERC. In general, the measurement reports
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

contain little information about the machines that were measured; for types IV and V machines, relatively good information with indicated number of samples, measurement ranges, and long-term concentrations is available (Thoulass, 1998; BK-Report, 1999; Bock et al., 1999). In this situation, it is not possible to assign the measured data to machine sizes or even individual scenarios. However, if the ranges of the measured concentrations are compared to the ranges spanned by the model results obtained for all machine sizes of a certain type (IV), a good agreement of these ranges is observed (data in Tables 5 and 6 and graphs in Figure 5). All measured average concentrations lie within the SceBRA ranges.

Exposed Workers Finally, the average risk quotients of the different machine types and sizes that were in use at a certain time are combined with the number of workers exposed to TRIC and PERC in metal degreasing and compared for different years. Figure 6 shows a cumulative representation of the results for the neareld (box A) over 15 years. In this graph, each point (log r; Ncum) of the curve shows the total number of exposed workers, Ncum, with a risk quotient at least as high as indicated by log r. This kind of plot combines the information from all scenarios of a solvent for a specied year in a single line and thus several years and both solvents can be compared in one plot. For each year, the total number of exposed workers and the number of exposed workers with a risk quotient above 1 (or any other particular value) can be derived from the plot. The total number of exposed workers in the near-eld in 1985 is 13,800 for PERC and 10,830 for
337

von Grote et al.

Occupational exposure to PERC and TRIC in metal

TRIC compared to 1760 for PERC and 391 for TRIC in 1999. In 1985, 1 year before the 2nd BImSchV was enacted in 1986, mainly open-top types I and II and encased type III machines were in use, whereas in 1991 all machines had to fulll the 2nd BImSchV of 1986. In 1996, after the transition period for the amended regulations, and in 1999, all machines had to conform to the 2nd BImSchV of 1990, meaning that only closed-loop type V machines were allowed. In 1985, 53,000 tonnes of PERC and 34,000 tonnes of TRIC were sold (only former West German states), but only 4500 tonnes PERC and 5100 tonnes TRIC in 1999 (Nader, 2002). This is a reduction by 91.5% for PERC and by 85% for TRIC. The reduction in the solvent amount used in metal degreasing is due to changes in environmental legislation, which led to: (i) technology changes (closed-loop machines with integrated recycling, where the solvent has a residence time in the machine of about 1 year (Nader, 1993), whereas Mannheim et al. (1979) report for open-top types I and II machines a total solvent loss of 0.9 kg/h for PERC and 0.7 kg/h for TRIC, which means that solvent had to be relled once or twice a week); (ii) solvent substitution; and (iii) more efcient processing in metalworking industry with less greasing and degreasing.

Discussion
The core part of the calculations with SceBRA presented in this paper is the airborne concentrations obtained with the two-box model. As far as a comparison with measured data is possible, these model results are in agreement with concentrations measured in metal degreasing facilities, see the values listed in Tables 5 and 6 and the model results shown in Figure 5. Owing to highly variable workplace conditions, the measured data show some scatter that is not reproduced by the model, but the model results match the average values of the measured data and there are no systematic deviations. All of the model parameters are based on empirical information and none of them was adjusted in order to increase the correspondence of calculated and measured concentrations. A local sensitivity analysis for the model input parameters indicated that the size of machine, the batch time tb, and the air exchange rate kA are the most inuential parameters for the airborne concentration in box A. The machine sizes are taken from sales brochures and are therefore well known. The batch time depends on the loads of the machine, that is, the weight, the material, and the shape of the parts. A specication of the parts and therefore more specic batch times would only make sense if the different parts cleaned in the region under consideration (Germany) and at a certain time could be described in more
338

detail. As this is not possible in the context of a screening method, it is reasonable to use a realistic range of batch times, which was also the result of discussions with machine manufacturers. The air exchange rate between boxes A and B is an assumption specically made for the two-box model. More specic air exchange rates for certain workplaces could be estimated by tracer gas measurements (Sohn and Small, 1999). However, as we want to estimate the concentrations for a broad range of industry facilities, it seems to be sufcient to take reasonable kL values for rooms with local exhaust and to assume that for older machines, which are stronger heat sources, the assumed value of kA is somewhat higher than kL. Parameters with a medium inuence are the emission . . factors Ec2 , Ep, and Ec1 . Taken together, the three emission factors correspond well with total solvent losses for the opentop machines reported by Mannheim et al. (1979). Better estimations could be achieved if emission factors are measured for different machine technologies, which, however, is not easily feasible as in Germany today only type V machines are in use. Parameters that have hardly any inuence on the near-eld concentration are V and kL. Overall, the concentrations obtained with the scenarios correspond to the range of measured concentrations. This means that the scenarios cover a relevant range of exposure situations. A strategy for selecting a reasonable set of scenarios is to dene the scenarios bottom up by setting up all reasonable parameter combinations (here a total of 2160 scenarios for each machine size) and then to analyze which scenarios lead to very similar concentrations. On this basis, redundant scenarios can be omitted and those that span the relevant concentration range can be selected for further investigation. Using a broadly dened set of scenarios makes it possible to include the high variability of metal degreasing facilities into a comprehensive exposure analysis. In the present application of the SceBRA method, this variability is described by clusters of scenarios dened by the parameter values, for example, machine size, types of parts cleaned, etc. (see Figure 4), but not by frequency distributions. (Note that the set of scenarios can be seen as a discrete frequency distribution of concentrations resulting from identical frequencies for all model input parameter values.) In a companion study on dry cleaning, which deals with a much less heterogeneous functional unit (textiles), we derived frequency distributions for the different model input parameters from a questionnaire providing information on the parameters describing typical dry cleaning facilities. In conclusion, the SceBRA method is suitable for characterizing the occupational risk through TRIC and PERC in metal degreasing. The numbers of exposed workers and the risk quotients in Figure 6 clearly indicate the combined effect of technological innovation and stricter
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

Occupational exposure to PERC and TRIC in metal

von Grote et al.

legislation, leading to strongly reduced airborne concentrations and also much lower numbers of exposed workers. The reduction in the number of workers is also due to automatization and a replacement of TRIC and PERC by nonchlorinated solvents. A similar assessment for the application of nonchlorinated solvents would also be desirable.

Acknowledgments
We thank Dow Europe S.A., the Swiss Reinsurance Company, and the Swiss Federal Ofce of Public Health (BAG) for nancial support of this study. Thanks go to for helpful H.-N. Adams, Dow Europe, and Jo rg Pastre discussions.

References
Adams, H.-N. Metallentfettung mit CKW-Lo semitteln und Ersatzstoffe und alternative Technologien in diesem Anwendungsbereich. Enquete-Kommission Schutz des Menschen und der Umwelt, Stellungnahme der Sachversta ndigen zu dem Fragenkatalog fu r die o ffentliche Anho rung am 3. und 4. Juni 1993 zu dem Thema Chlorchemie, pp. 3118. Deutscher Bundestag Enquete-Kommission Bonn, 1993. Adams, H.-N. Entwicklung CKW-Lo semittel Markt. Dow Europe S.A., Horgen, 1997, pp. 117. Adams, H.-N., and Jeker, H. The Inuence of BIMSCH on the Degreasing Equipment Used in Germany and on the Solvent Consumption F Estimation. Dow Europe S.A., Horgue 1986, pp. 112. Bach, H., Becher, K., Biegert, B., Detzer, R., Diebschlag, W., Dittes, W., Hocke, M., Jungba ck, E., Madjidi, M., and Scholer, W. Gezielte Belu ftung der Arbeitsbereiche in Produktionshallen zum Abbau der Schadstoffbelastung, pp. 1157. Bundesministerium fu r Forschung und Technologie (BMFT), Bonn, 1992. Bardodej, Z., and Vyskocil, J. Arch Ind Hyg Occup Med 1956: 13: 581. BImSchV. Zweite Verordnung zur Durchfu hrung des Bundes-Immissionsschutzgesetzes (Verordnung zur Emissionsbegrenzung von leichtu chtigen Halogenkohlenwasserstoffen F 2. BImSchV). 21 April 1986. BImSchV. Zweite Verordnung zur Durchfu hrung des Bundes-Immissionsschutzgesetzes (Verordnung zur Emissionsbegrenzung von leichtu chtigen Halogenkohlenwasserstoffen F 2. BImSchV). 10 December 1990. BImSchV. Gesetzesnovelle vom 21.8.2001 gema ss Art. 2 VO zu: Zweite Verordnung zur Durchfu hrung des Bundes-Immissionsschutzgesetzes (Verordnung zur Emissionsbegrenzung von leichtu chtigen Halogenkohlenwasserstoffen F 2. BImSchV). 21 August 2001. BK-Report. Polyneuropathie oder Enzephalopathie durch organische Lo semittel oder deren Gemische, pp. 110. BK, BIA, 1999. Bock, W., Brock, T.H., Stamm, R., and Wittneben, V. BGAA-Report (Berufsgenossenschaftlicher Arbeitskreis Altstoffe), Altstoffe F Expositionen am Arbeitsplatz, pp. 1105. Hauptverband der gewerblichen Berufsgenossenschaften (HVBG), Sankt Augustin, 1999. BUA. Trichlorethen, BUA-Stoffbericht 95 (Stand: Juni 1991). Hirzel Verlag, Stuttgart, 1993, pp. 1398. BUA. Tetrachlorethen (PER), BUA-Stoffbericht 139. Hirzel Verlag, Stuttgart, 1994, pp. 1366.

CEFIC. Trichloroethylene in metal cleaning and other industrial en de lIndustrie applications, pp. 118. CEFIC, Conseil Europe Chimique, Brussel, 1984. CerclAir. Massnahmen zur Verringerung der Umweltbelastung beim Reinigen und Entfetten von Metallen (VOC-Kommission), pp. 167. Amt fu r Umweltschutz, Glarus; Schweizerische Gesellschaft der Lufthygieniker in Zusammenarbeit der Eduard-Sta hlin-Organisation, Glarus, 1993. Cherrie, J.W., Schneider, T., Spankie, S., and Quinn, M. A new method for structured, subjective assessments of past concentrations. Occup Hyg 1996: 3: 7583. Drivas, P.J., Simmonds, P.G., and Shair, F.H. Experimental characterization of ventilation systems in buildings. Environ Sci Technol 1972: 6(7): 609614. ECETOC. Trichloroethylene: assessment of human carcinogenic hazard. Technical report no. 60, pp. 177. ECETOC, Brussels, 1994. ECETOC. Joint assessment of commodity chemicals no 39: tetrachloroethylene CAS no. 127-18-4, pp. 1271. ECETOC, Brussels, 1999. Fehrenbacher, M.C., and Hummel, A.A. Evaluation of the mass balance model used by the Environmental Protection Agency for estimating inhalation exposure to new chemical substances. Am Ind Hyg Assoc J 1996: 57: 526536. Flynn, M.R., and Sills, E.D. On the use of computational uid dynamics in the prediction and control of exposure to airborne contaminants F an illustration using spray painting. Ann Occup Hyg 2000: 44(3): 191202. Franke, J.E., and Wadden, R.A. Indoor contaminant emission rates characterized by source activity factors. Environ Sci Technol 1987: 21(1): 4551. Furtaw, E.J., Pandian, M.D., Nelson, D.R., and Behar, J.V. Modeling indoor air concentrations near emission sources in imperfectly mixed rooms. J Air Waste Manage Assoc 1996: 46: 861868. Gmehling, J., Weidlich, U., Lehmann, E., and Fro hlich, N. Verfahren zur Berechnung von Luftkonzentrationen bei Freisetzung von Stoffen aus u ssigen Produktgemischen. Staub - Reinhalt Luft 1989: 49: 227230 and 295299. Grandjean, E., Mu nchinger, R., and Turrian, V. Investigations into the effects of exposure to trichloroethylene in mechanical engineering. Br J Ind Med 1955: 12: 131142. Hayes, S.R. Use of an indoor air quality model (IAQM) to estimate indoor ozone levels. J Air Waste Manage Assoc 1991: 41: 161170. Heinsohn, R.J. Industrial Ventilation: Engineering Principles. WileyInterscience Publication, New York, 1991. Jayjock, M.A. Assessment of inhalation exposure potential from vapors in the workplace. Am Ind Hyg Assoc J 1988: 49(8): 380385. Jayjock, M.A., and Hawkins, N.C. A proposal for improving the role of exposure modeling in risk assessment. Am Ind Hyg Assoc J 1993: 54(12): 733741. Keil, C.B. The development and evaluation of an emission factor for a toluene parts-washing process. Am Ind Hyg Assoc J 1998: 59: 1419. Keil, C.B. A tiered approach to deterministic models for indoor air exposure. Appl Occup Environ Hyg 2000: 15(1): 145151. Leisewitz, A., and Schwarz, W. Metallobera chenreinigung mit CKW, KW und wa ssrigen Reinigern. Stoffstromanalyse und Umweltbelastungsvergleich, pp. 1324. Umweltbundesamt, 1994. Mannheim, R., Budelmann, C., and Winkel, P. Ersatzstoff fu r Trichlorethylen bei der Entfettung metallischer Obera chen, pp. 1151. Siemens AG Berlin, im Auftrag des Umweltbundesamtes, Berlin, 1979. Manufacturers. Sales brochures and information of metal degreasing RR, MULTImachines provided by PERO, ROLL, FISA, DU CKH, ILSA and EMO. 19602001. MATIC, HO Matthiessen, R.C. Estimating chemical exposure levels in the workplace. Chem Eng Prog 1986: 3034. 339

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)

von Grote et al.

Occupational exposure to PERC and TRIC in metal

Morse, K.M., and Goldberg, L. Chlorinated solvent exposures. Ind Med 1943: 12: 706713. Nader, F.W. Stellungnahme zu Fragenkatalog Nr. 2 Chlorkohlenwasserstoffe. In: Enquete-Kommission Schutz des Menschen und der Umwelt Stellungnahme der Sachversta ndigen zu dem Fragenkatalog fu ffentliche Anho rung am 3. und 4. Juni 1993 zu dem r die o Thema Chlorchemie, pp. 193234. Deutscher Bundestag Enquete-Kommission Bonn, 1993. Nader, F.W. CKW-Lo semittel-Markt. VCI, 2000. Nader, F.W. VCI survey of German metal degreasers using TRIC. Frankfurt, 2001. Nader, F.W. CKW-Lo semittel-Markt (Frischware) Bundesrepublik Deutschland, pp. 13. VCI, Frankfurt, 2002. Neghab, M., Qu, S., and Bai, C.L. Raised concentrations of serum bile acids following occupational exposure to halogenated solvents, 1,1,2trichloro-1,2,2-triuoroethane and trichloroethylene. Int Arch Occup Environ Health 1997: 70: 187194. Nicas, M. Estimating exposure intensity in an imperfectly mixed room. Am Ind Hyg Assoc J 1996: 57: 542550. Nicas, M., and Jayjock, M.A. Uncertainty in exposure estimates made by modeling versus monitoring. Am Ind Hyg Assoc J 2002: 63: 275283. Nicholson, G.P., Clark, R.P., and De Calcina-Goff, M.L. Computational uid dynamics as a method for assessing fume cupboard performance. Ann Occup Hyg 2000: 44(3): 203217. Nicklin, H.G. Arbeitshygienische Gesichtspunkte beim Einsatz von chlorierten Lo semitteln in Reinigungsverfahren, Vol. 11, pp. 18. Die Berufsgenossenschaft, 1972. Niemann, S. Efzienz von Arbeitsschutzmassnahmen bei einem Perchlorethylen-belasteten Kollektiv. Universita t Hamburg, Hamburg, 1995, pp. 118. OECD. SIDS initial assessment report for the 5th SIAM, tetrachloroethylene. OECD (Organisation for Economic Co-operation and Development), Paris, 1996. Raaschou-Nielsen, O., Hansen, J., Thomsen, B.L., Johansen, I., Lipworth, L., McLaughlin, J.K., and Olsen, J.H. Exposure of

Danish workers to trichloroethylene, 19471989. Appl Occup Environ Hyg 2002: 17(10): 693703. Rantala, K., Riiginen, H., and Anttila, A. Altisteetyo ssa , 22 halogeenihiitivedyt [exposure at work, 22 halogenated hydrocarbons]. Finnish Institute of Occupational Health, Helsinki, 1992. Recknagel, H., Sprenger, E., and Schramek, E.R. Taschenbuch fu r Heizung und Klimatechnik. R. Oldenbourg Verlag, Mu nchen, 2000. Ryan, P.B., Spengler, J.D., and Halpfenny, P.F. Sequential box models for indoor air quality: application to airliner cabin air quality. Atmos Environ 1988: 22(6): 10311038. Scheff, P.A., Friedman, R.L., Franke, J.E., Conroy, L.M., and Wadden, R.A. Source activity modeling of freons emissions from open-top vapor degreasers. Appl Occup Environ Hyg 1992: 7(2): 127134. Scheringer, M., Vo gl, T., von Grote, J., Capaul, B., Schubert, R., and Hungerbu hler, K. Scenario-based risk assessment of multi-use chemicals F application to solvents. Risk Anal 2001: 21(3): 481497. Sohn, M.D., and Small, M.J. Parameter estimation of unknown air exchange rates and effective mixing volumes from tracer gas measurements for complex multi-zone indoor air models. Building Environ 1999: 34: 293303. Thoulass, R. Occupational Hygiene Report: Workplace personal exposure to tetrachloroethene (perchloroethylene) during the use of modern metal cleaning equipment, pp. 110. 1998. Wadden, R.A., Hawkins, J.L., Scheff, P.A., and Franke, J.E. Characterization of emission factors related to source activity for trichloroethylene degreasing and chrome plating processes. Am Ind Hyg Assoc J 1991: 52(9): 349356. Wadden, R.A., Scheff, P.A., and Franke, J.E. Emission factors for trichloroethylene vapor degreasers. Am Ind Hyg Assoc J 1989: 50(9): 496500. Wadden, R.A., Scheff, P.A., Franke, J.E., Conroy, L.M., Javor, M., Keil, C.B., and Milz, S.A. VOC emission rates and emission factors for a sheetfed offset printing shop. Am Ind Hyg Assoc J 1995: 56: 368376.

340

Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(5)