occupational exposure scenarios - nanex...

Development of Exposure Scenarios for Manufactured Nanomaterials

Work Package 3

Occupational Exposure Scenarios

(Including Deliverables D3.1 and D3.2)

Derk Brouwer1, Rianda Gerritsen-Ebben1, Birgit van Duuren-Stuurman1, Iris Puijk1, Gaelle Uzu2, Luana Golanski2, Celina Vaquero3, Vasilis Gkanis; Panos4 Neofytou4,

Martie van Tongeren5

December 2010

1 Netherlands Organisation for Applied Scientific Research (TNO), Utrecht Area, Netherlands 2 French Alternative Energies and Atomic Energy Commission (CEA), Grenoble, France 3 Fundaciòn CDT – (LEIA), Miñano (Álava), Spain 4 National Center for Scientific Research 'Demokritos', Athens, Greece 5 Institute for Occupational Medicine (IOM), Edinburgh, United Kingdom

iii

Project website: www.nanex-project.eu

Funded under the seventh framework programme: NMP-2009-1.3-2: Exposure scenarios

to nanoparticles

Grant agreement no.: 247794

Funding scheme: Coordination and Support Actions (supporting action)

v

Executive Summary

In order to build occupational exposure scenarios (ESs) for three types of manufactured

nanomaterials, i.e. single-walled carbon nanotubes (SWCNT) and multi-walled carbon

nanotubes (MWCNT), nano-silver (nano-Ag) and nano-titanium dioxide (nano-TiO2), the

following activities were performed.

• Identification and review of the open literature with respect to use for ES building requirements;

• Compilation of data generated by two major measurement campaigns, i.e. the NANOSH project (FP6) and the NanoINNOV project (CEA); and

• A performance check of existing exposure estimating models by comparison of modelled ES exposure estimates with actual data from the data campaigns mentioned above.

Based on 33 literature references, 22 ESs were entered into the NANEX Exposure

Scenario Database. A total of 14 ESs for CNT were developed, generating 35 contributing

exposure scenarios describing some facet of occupational exposure. Most of them were

related to ESs in the production/synthesis of carbon based nanomaterials or handling such

materials (weighing, removing, sonication, etc.) and two ESs addressed tasks related to

the machining of composites containing CNTs. A total of 5 ESs for nano-TiO2 were

developed, generating 12 contributing exposure scenarios. Of these, only two contained

sufficient information to fill in the ES; for the remaining ESs, only some elements of the

exposure scenario template could be completed. Three of these papers related to the

production of nano-TiO2 and two related to the production of materials containing nano-

TiO2.

Only two occupational ESs could be developed for nano-Ag, one describing its

manufacture in a wet chemistry process while the second was related to handling nano-Ag

in a fume hood.

In total, 35 ESs with 48 contributing scenarios were derived from the data sets of two large

measurement campaigns (NANOSH project (FP6) and the NanoINNOV project (CEA).

Most ESs were for CNTs (n=14), which were predominantly related to research-scale

activities. Most nano-TiO2 scenarios (n=8) were on commercial-scale manufacturing and

formulation. Only 2 ESs could be build for nano-Ag, whereas 11 ESs were build other

substances, including other metal oxides.

Based on the process of developing these ESs, several main conclusions could be drawn.

Most studies either reported in the literature or as part of the measurement campaigns had

vi

an explorative character and were focused on concentration/ emission analysis. Therefore,

the reports from these studies did not include most of the information necessary to build

ESs, e.g. amount used, frequency of activities. Basic characterization of the products used

was often not available and operational conditions were often not described. Most

concentration/emission-related measurement results were task-based and subsequently it

is difficult to assign a process category (PROC), as the same tasks can cover multiple

PROCs. The most important observation was the lack of harmonization either of the

measurement strategy or of the analysis and reporting of measurement data.

At this stage it was not possible to build exposure scenarios combining different

information sources (references). This was mainly due to the heterogeneity in the level and

quality of the description of the context (differences related to material characteristics,

processes, quantities handled, control systems, etc.) and in the exposure evaluation (the

absence of standards addressing different measurement strategies, equipments and data

treatment).

Information required to address the environmental release from the occupational exposure

scenarios is lacking, both from the open literature sources and from the measurement

campaigns. Information such as total quantities produced/year and some issues related to

air emissions (presence of vent hoods) was only rarely reported. No data were reported

related to water treatment.

ECETOC TRA and Stoffenmanager have been evaluated here with respect to their

applicability for estimating exposure to nanoparticles. Both models are based on a source-

receptor approach, distinguishing emission, transport, immission and personal exposure. It

was concluded that both models should, in principle, be able to predict exposure to

nanoparticles. However, the different categories within each model variable are not

particularly suitable for activities of nanomaterials and lack the required level of resolution,

which means that, in practice, many situations may fall into the same category, resulting in

the same or similar exposure estimates. Refinement of these categories in view of typical

activities for nanomaterial handling, amount of handling categories etc, is needed.

No correlation was observed between the model estimates (for mass concentration) and

the measured (particle-number) concentrations. In addition, no differences in the estimates

were observed between Stoffenmanager (activity-based) and ECETOC TRA (extrapolated

for full day exposure). Most probably the lack of correlation is largely dominated by the

variability of the (relatively few) data. The variability of the measured particle number

vii

concentrations was much larger than the variability of the Stoffenmanager and ECETOC

TRA exposure (mass concentration) predictions. This again suggests that there is a need

to refine the models to increase the resolution in the exposure estimates. As has been

demonstrated, due to lack of data or contextual information in the data sets, the entire

range within the model parameters categories could not be used, resulting in a loss of

power of discernment between exposure scenarios.

Both first tier models only provide mass concentration as proxy for exposure, where typical

devices used for nanoparticle exposure assessment use particle concentration as an

exposure metric. Since (nano) devices usually have size-windows up to 1000 nm, the

contribution of particles below 1000 nm to mass concentration will be low and might only

affect variations in the lower mass concentration ranges of the current model estimates.

This also indicates the need for recalibration of the models for nanomaterials exposure.

In summary it can be concluded that, in their current form:

• Both ECETOC TRA and Stoffenmanager are not suitable for providing estimates for nanoparticles. Neither of the models is tuned to and calibrated for nanomaterial exposure situations, and hence the actual model estimate will be inaccurate and possibly overestimate the (mass) concentration levels.

• Both models provide exposure estimates in mass concentrations, which may not be appropriate for expressing exposure to nanomaterials. Both will need to be modified to include more refined categories within several of the model variables. However, the main modification required is to address the exposure metrics.

ix

Table of Contents

Executive Summary ............................................................................................................. v Table of Contents ............................................................................................................... ix 1. Introduction ...................................................................................................................... 1

1.1 Objectives ................................................................................................................. 2 2. Methods ........................................................................................................................... 3

2.1 Introduction ............................................................................................................... 3 2.2 Review of literature .................................................................................................... 3 2.3 Data from NanoINNOV and NANOSH projects ......................................................... 3 2.4 Review of occupational exposure models and performance check ........................... 5 2.5 Building occupational exposure scenarios ................................................................ 5

3. Results ............................................................................................................................. 7 3.1. Use of information/data from open literature ............................................................. 7 3.2. Use of information/data from measurement campaigns ............................................ 8 3.3. Review of occupational exposure models and performance check ........................... 9

4. Discussion and Conclusions .......................................................................................... 11 4.1. Discussion and conclusions for ES from literature .................................................. 11 4.2. Discussion and conclusions for ES from the dataset generated by measurement

campaigns ............................................................................................................... 12 4.3. Discussion and conclusions from model review and performance check ................ 15

5. References .................................................................................................................... 19 6. Annexes ......................................................................................................................... 21

6.1. Overview of exposure scenarios in the NANEX Exposure Scenario Database ....... 22 6.2. Exposure models and exploration of their applicability for nanomaterials ............... 26

6.3.1 ECETOC TRA model ................................................................................... 26 Evaluation of model for suitability of nanomaterials under REACH ................................ 27

6.3.2 STOFFENMANAGER 4.0 ............................................................................. 28 6.3.3 Discussion and conclusion ........................................................................... 29

6.3. Performance check ECETOC TRA and STOFFENMANAGER with actual data ..... 31 6.4.1 Introduction ................................................................................................... 31 6.4.2 Assessments with ECETOC TRAv2 ............................................................. 31 6.4.2 Assessments with Stoffenmanager 4.0 ........................................................ 36 6.4.3 Performance check results ........................................................................... 37 6.4.4 Analysis of the performance check ............................................................... 47 6.4.5 Discussion and conclusion ........................................................................... 56 6.4.6 References ................................................................................................... 57

6.4. PDF extracts from the NANEX Exposure Scenario Database ................................. 58

1

1. Introduction

NANEX aimed to develop a catalogue of generic and specific exposure scenarios (ESs)

for Manufactured NanoMaterials (MNMs), taking account of the entire lifecycle of these

materials. In WP3 (Occupational exposure scenarios) measurement and contextual

information was collected and reviewed to describe and characterize occupational

exposure and available tools and models to predict occupation exposure to MNMs

reviewed.

Within WP2 (Development of generic exposure scenario descr iptions) a database for

the collection of exposure scenarios according to REACH Guidance was developed

(NANEX Exposure Scenario Database). Annex 1 provides an overview of the 57 exposure

scenarios entered in this exposure scenario (ES) database, which were developed based

on information/data from open literature or existing databases.

The deliverables from WP3 were:

D3.1: Characterization of relevant occupational exposure scenarios. D3.2: Needs/ knowledge gaps to comply with REACH.

2

1.1 Objectives

Information for describing and characterizing exposure were collected, collated and

reviewed for occupational exposure scenarios identified in WP2 relevant for the three

types of nanomaterials (High Aspect Ratio Nanomaterials (HARNs, mass produced

nanomaterials and speciality nanomaterials) using the format developed in WP2.

The Description of Work outlined the following objectives:

1. Collection, collation and summarizing of information on operational conditions and risk management measures for the scenarios identified from literature and amongst partners;

2. Collection, collation and review existing exposure data and (exposure) metrics; 3. Identification and review of available tools and models to predict occupation exposure

with respect to manufactured nanomaterials and identified scenarios; and 4. Characterization of relevant occupation exposure scenarios, if possible in terms of

operational conditions, risk management measures, estimated levels of exposure, numbers of workers involved etc, using the format developed in WP2.

The generic exposure scenarios identified in WP3 relevant for workers are:

• Carbon based nanomaterials, including SWCNT, MWCNT, carbon nanofibres and fullerenes, and their application in composite materials;

• Nano-silver (nano-Ag) and its use in textiles; and • Nano-TiO2 and its use in cosmetic products.

The development of these occupational ESs has been the core of the work of WP3.

3

2. Methods

2.1 Introduction

In order to meet the objectives of the project, the following activities were performed:

• Identification of open literature related to occupational exposure; • Literature review; • Occupational exposure model review and performance check; • Compilation of data from the projects NanoINNOV and Nanosh; and • Building occupational exposure scenarios for the three nanomaterials.

2.2 Review of literature

Information on current and past FP6/7 and other projects, as well as results from a

literature search was made available by WP2 for review. The literature was reviewed

source by source and, for those papers that included relevant information for building

occupational exposure scenarios, an entry in the NANEX Exposure References Database

(developed by WP7 and WP2) was made. In addition the database has a section to import

occupational exposure scenarios (NANEX Exposure Scenario Database) which includes

the fields required for a REACHES as these fields were taken from the latest REACH

guidance on the web-site of the European Chemicals Agency (see also WP2 reporting). A

total of 40 references were added to the NANEX Exposure References Database of which

22 were eventually used to build occupational exposure scenarios.

2.3 Data from NanoINNOV and NANOSH projects

For the building of ESs, data was used from measurement campaigns from the

NanoINNOV/NanoSafe project and the NANOSH project.

The aim of both monitoring programmes was to assess the exposure of workers to

engineered nanoparticles. In this context, background nanoparticles that may be present in

the air from other sources (e.g. welding, grinding, and transportation) and from

outdoors/ambient aerosols can interfere with this assessment. The differentiation of

engineered nanoparticles from the background is one of the most important and difficult

challenges for the sampling strategy. Since there are no direct reading instruments

available that differentiate between engineered nanoparticles and ambient aerosols,

attention was paid to the issue of background levels and other aerosol-emitting sources.

4

Determination of the air stream/ventilation patterns is a useful tool for the identification of

secondary sources. Other nano-sized aerosol emitting sources in the workplace were

identified and background concentrations were measured prior to and after the activity with

engineered nanoparticles.

As there is no firm agreement on the most appropriate exposure metric, it was decided to

employ a range of instruments for measuring mass, number and surface area

concentrations during the NANOSH project. In addition, samplers were used to collect

airborne nanoparticles for characterization purposes. Furthermore, contextual information

(e.g. description of process/situation, PPE), was collected by the completion of

questionnaires.

Direct measurements of number concentrations were obtained using Condensation

Particle Counter (CPCs) and Portacount. Number concentrations and number-based

size distributions were obtained using a Scanning Mobility Particle Sizer (SMPS) and

Electrical Low Pressure Impactor (ELPI). The SMPS characterizes particles by their

electrical mobility diameter, defined as the diameter of a spherical, singly-charged particle

having the same velocity as the particle in question in an electrostatic field. The 12-stage

ELPI monitors provides real-time aerodynamic particle size distributions ranging from

approximately 20 nm to 10000 nm. Direct measurements of surface area concentrations

of particles were obtained using Diffusion Charger (DC)-type instruments.

More specifically for the NANOSH measurement campaign, samples of airborne

nanoparticles were characterized, mainly by TEM analysis (parameters such as; shape,

size, state of agglomeration, composition). In the NANOSH study, a nanometer aerosol

sampler (NAS) was used to collect nanoparticles on Holey TEM grids (3 mm Holey Carbon

film 400 mesh Ni). Personal air samples (PAS) are taken in preference to

static/background samples (such as measurements performed using the NAS) given that,

for PAS measurements, the sample material is collected in the breathing zone of the

worker, which gives much more accurate information about the inhalation exposure. For

this purpose, a sampling device was developed consisting of an open-faced filter holder

including a 25 mm gold-coated polycarbonate filter on which a TEM grid was placed.

Data from the NanoINNOV and NANSOH projects resulted in a total of 35 entries in the

NANEX Exposure Scenario Database.

5

2.4 Review of occupational exposure models and perf ormance check

Two of the main occupational exposure models, ECETOC TRA and Stoffenmanager,

which are both mentioned in the REACH Guidance R.14 from ECHA (2008), are often

used for estimating exposure to chemicals within a REACH chemical safety assessment

for occupational exposures.

The different modules of ECETOC-TRA and Stoffenmanager were reviewed for their

applicability for nanomaterials. This review is presented in Annex 6.3. Due to the

availability of data sets from measurement campaigns and literature references, no

exposure scenarios were built using these models. Instead a performance check of the

models in comparison with the collected data set was conducted and is presented in

Annex 6.4. Summaries of the reviews are provided in Chapter 3.

2.5 Building occupational exposure scenarios

This element of the study was the main integrating activity of NANEX WP3 and constitutes

the main chapters of this report. In total, 57 exposure scenarios were built from the

literature (22) and from data obtained from two large measurement campaigns

(NanoINNOV and NANSOH, 35). Chapter 3 addresses the processes of building ES using

data from these sources.

Overall findings, conclusions and recommendations from this work are given in Chapter 4.

7

3. Results

3.1. Use of information/data from open literature

From the 57 occupational exposure scenarios in the NANEX Exposure Scenario

Database, 22 have been built from the literature, most of them related to carbon-based

nanomaterials (14), followed by nano-TiO2 (5) and nano-Ag (1). Annex 5.2 gives an

overview of the 22 literature-based exposure scenarios entered in the database and, in

Table 1, a summary of all the ES developed in WP3 is presented.

Carbon based nanomaterials

A total of 14 ESs for carbon-based nanomaterials were developed, generating 35

contributing exposure scenarios, describing some facet of occupational exposure. Most of

them were ESs from the production/synthesis of carbon-based nanomaterials or from

handling materials (weighting, removing, sonication, etc.); two scenarios addressed tasks

related to the machining of composites containing CNT (Bello, 2009; Methner, 2010).

Nano-TiO2

A total of 5 ESs for Nano-TiO2 were developed, generating 12 contributing exposure

scenarios. However, only two of the ESs describe full Operational Conditions and Risk

Management Measures relevant for occupational exposure. Three of the other ESs were

related to the production of nano-TiO2 and two references related to the production of

materials containing nano-TiO2.

Nano-Ag

Only two occupational ESs were developed for nanosilver (J. Park, 2009; S.J. Tsai, 2008),

the first one describing ES in a wet chemistry process while the second one related to

handling nanosilver in a fume hood.

8

3.2. Use of information/data from measurement camp aigns

From the 57 occupational exposure scenarios in the NANEX Exposure Scenario

Database, 35 have been built from data campaigns, most of them related to carbon-based

nanomaterials (14), followed by Others (11) (e.g ZnO (2)), nano-TiO2 (8) and nano-Ag (2).

In total, 44 contributing exposure scenarios were identified. Both surveys were considered

as one reference and included in the NANEX Exposure References Database.

Annex 5.2 gives an overview of the 35 exposure scenarios entered in the NANEX

Exposure Scenario Database and, in Table 1, a summary of all the ES developed in WP3

is presented.

Table 1 Overview of the 57 occupational ESs developed in WP3 Sources ES IDs Substances Public Literature (research/ explorative studies)

22 TiO2 (5); CNT(14); Nano-Ag (2); others (1)

Data sets campaigns NanoINNOV & NANOSH

35 CNT(14); Nano-Ag (2); TiO2 (8). Others (11) e.g (ZnO) (2)

TOTAL 57

9

3.3. Review of occupational exposure models and pe rformance check

The ECETOC TRA model estimates are based on the following input parameters:

• Vapour pressure or dustiness • Selection of a generic use scenario (25 PROCs) • Selection of limited exposure determinants (duration, LEV, RPE) • Selection of mixture or pure substance • Exposed skin surface (default for each PROC)

The inhalation exposure estimates are presented in ppm and mg/m³ and dermal values in

µgcm-2day-1 and, based on body weight, in mgkg-1day-1 and total exposure in mgkg-1day-1.

For Stoffenmanager, the following parameters and input data are needed:

• Vapour pressure or dustiness; • Percentage of substance in the product; • Handling category (7 classes for liquid, and 6 for solid); • Local controls; • Distance of worker to source; • Presence of other sources of same substance further than 1 metre from the worker; • Room volume (4 classes); • General ventilation (3 classes); • Immission control measures; • Personal Protective Equipment (5 classes); • Possibility of background exposure;

For both the ECETOC TRA and the Stoffenmanager it can be concluded that many of the

parameters used in the model will be similar for MNMs in comparison with conventional

compounds. However, the actual exposure values may not necessarily be accurate. This

is due to the possible specific characteristics of MNMs. Chemical composition, surface

structure, solubility, shape and aggregation behaviour might be completely different

compared with conventional compounds. Since the model is neither built nor calibrated for

MNMs caution is required when using either model to estimate the exposure to MNMs.

To check the performance of the models, the model predictions were compared with the

results from the particle number concentration measurements obtained during the two

surveys. Comparisons were carried out using the absolute difference in activity and

background concentrations and using the ratio of activity and background concentrations.

The results of the comparison of the model outcomes with actual workplace particle

number concentration data obtained by Condensation Particle Counters (CPC) and

Scanning Mobility Particle Sizers (SMPS) are presented in Table 2.

10

The results show that there was no correlation between the model estimates and the

measurements. In order to find out possible explanations for the lack of any correlation a

more in-depth analysis was performed, based on the NANOSH dataset. (For more details

about this analysis see Annex 6.4). The main observations were that the category

variables within the model parameters were not evenly distributed over the various

scenarios, so that the potential resolution between the scenario exposure estimates was

not fully exploited.

Table 2 Pearson correlation coefficients for the ECETOC TRA model and Stoffenmanager for the Nanosh and NanoINNOV/NanoSafe datasets Nanosh dataset Pearson correlations

CPC results (N=13) SMPS results (N=23) Absolute difference activity minus non-activity

Ratio task vs. background

Absolute difference activity minus non-activity


Stoffenmanager Inhalable dust (mg/m³) -0.15 -0.17 <-0.001 0.05

ECETOC exposure level (mg/m³) 0.063 -0.02 <-0.09 -0.15

NanoINNOV/NanoSafe dataset

CPC results Absolute difference activity minus non-activity Ratio task vs. background

Stoffenmanager Inhalable dust (mg/m³) (N=15)

0.79$ 0.866$

ECETOC exposure level (mg/m³) (N=5) 0.35 0.35

$ The high correlation is biased by one single data point; in case that data point is omitted, the correlation coefficient is close to 0.3

11

4. Discussion and Conclusions

4.1. Discussion and conclusions for ES from litera ture

In total, 22 ESs were developed from open literature and charged in the NANEX Exposure

Scenario Database (developed in WP2). Table 3 gives an overview of the main issues that

researchers have found when fulfilling the different fields in the database.

Table 3 Overview of the evaluation of the ES building process

Field in the database Evaluation of Exposure Scenarios building

Exposure Scenarios general details Title Ambiguous; standard phrases may be helpful

PROCs Difficult to assign PROCs from tasks/ processes Harmonization task description needed

Exposure Scenarios details

Product characteristics Most often incomplete (usually only name nanomaterial)

Amounts used Most often not reported

Frequency duration of use Frequency not reported; duration of activity reported

Human factors Not reported Other operational conditions Not reported Technical condition s source level Source enclosure reported

Technical conditions to control dispersion Most often reported

Organisational measures Not reported PPEs Frequently reported

Exposure Estimation

References Data sets (between) Lack of standardization Measurement strategy Measurement equipment Data analysis and report

Measurement strategy Data analysis and report

From the process of developing this ES the following main conclusions can be underlined:

• There are only a few references in the open literature that include data of workers’ exposure to nanomaterials which could be used to develop exposure scenarios. An example of this is nano-TiO2. Despite the numerous applications of TiO2 nanoparticles, it was not possible to find information on occupational exposure to TiO2 nanoparticles (no information on the size of the TiO2 particles or the concentration).

• Papers in the peer-reviewed literature often do not include most of the information needed to build comprehensive ESs.

• The majority of the ESs described are based on tasks/processes performed in research labs with few publications addressing exposure at manufacturing facilities.

• The peer-reviewed and other literature describing worker exposure do not generally include any information that could be used to estimate environmental release. On a few occasions the total quantities produced/year and information related to air

12

emissions (presence of vent hoods) is reported. No data relating to water treatment were reported.

• At this stage it has not been possible to build exposure scenarios combining different information sources (references), even if different sources appear to address similar scenarios. The main reason for this was the heterogeneity in reporting of the context (differences related to material characteristics, processes, quantities handled, control systems, etc.) and in the exposure evaluation (the absence of standard approaches to measurement strategies, equipment and data treatment).

• Information on concentration and particle size is often lacking and therefore prevented the researchers from adding more incomplete ES in the NANEX Exposure Scenario Database.

4.2. Discussion and conclusions for ES from the da taset generated by

measurement campaigns

In total, 35 ESs were developed from measurement campaigns and entered into the

NANEX Exposure Scenario Database (developed in WP2). Table 4 gives an overview of

the main issues that researches reported when trying to complete the different fields.

13

Table 4 Overview of the evaluation of the ES building process

Field in the database Evaluation of Exposure scenar ios building

Exposure Scenarios general details

Title Need of homogeneity and precision for further search in the database

List of all use descriptors related to the life cycle stage and the relevant uses under it; include market sector (by PC)

Mainly R&D workplaces/labs in dataset, few industrial cases.

PROCs Data from NANOSH, NanoINNOV and NanoSafe projects are workplace orientated so the scenario was the most often assigned with only one PROC.

Exposure Scenarios details

Product characteristics Few data are usually reported (normally only the name of nanomaterial; sometimes size/length).

Amounts used Not detailed but generally broadly reported. For some of the exposure scenarios derived from the NANOSH project detailed information for amount used is available.

Frequency duration of use Always reported. Periods of handling of nanomaterials are generally very short (<15 min), Operation of maintenance after handling the longest (~ 2h)

Human factors Generally not reported but they were assumed. (Not reported within the NANOSH project.)

Other operational conditions

Generally processes involving the handling of articles are dry whereas maintenance activities can be wet-based (washing) or dry (sanding). Information about type of workplace available (e.g. laboratory, size of company etc.) for exposure scenarios from the NANOSH project.

Tech. conditions source level It is reported when closed reactors an embedded process or fume cupboards are used, but technical measures to prevent release at source level are not usually applied.

Tech. conditions to control dispersion

Reported when tasks are performed in fume hood/vent hood, general extraction or natural ventilation.

Org. measures Not reported

PPEs Always reported. The most used PPEs are PPF3 mask, nitrile gloves, latex gloves and blouse for handling of MNMs. Maintenance operations can also be in clean room with air extraction mask.

Exposure Estimation

Measurement Techniques

Data from nanoINNOV, NanoSafe and NANOSH programs were obtained using: o Elpi (inertial impactor) o CPC (condensation particles, counter) o SMPS (Scanning Mobility Particle Sizer)

Calibration is always done before each measurement and levels of background are always measured before the activity.

Sample points o Two locations: background and near the source

Most often, during the tasks of handling, no increase of background was noticed. However, the most significant increase occurred during maintenance PROCs.

14

From the process of developing this ES the following main conclusions can be drawn:

• The majority of the ESs described is based on tasks/processes performed in research labs.

• Because datasets from NanoINNOV and Nanosafe are workplace-based and related to worker exposure only limited information is reported related to environmental exposure/release. Conditions and measures related to on site/municipal sewage treatment plant were never reported. Finally, conditions and measures at level of article production process to prevent release during service life were not reported because R&D workplaces do handle NMM but no articles.

Exposure scenarios regarding handling of nano-TiO2, CNTs and Zinc oxide were obtained

from the NANSOH project database, with 46 measured situations. From the process of

developing these ESs the following main conclusions can be drawn:

• Data from the dataset campaigns are task-based and subsequently it is difficult to assign PROCs (tasks can be part of different PROCs)

• For ‘contributing exposure scenario 1’ (controlling environmental exposure for <name substance>) no information was collected during the field work. Consequently, these fields are empty.

• It would have been easier to complete a scenario if standard phrases were used for the different types of information that should be included in the exposure scenarios, since, in the REACH guidance, it is not always clear what information should be included, and vague terms are used. Also it is difficult to give the correct use descriptors.

• Currently, it is difficult to include measurement data into the NANEX Exposure Scenario Database since there is neither an agreed or harmonized appropriate method nor approach. Several options are available to report:

• Individual results for each instrument (for each measurement separately). But this will result in a long list of measurement data for one exposure scenario.

• Background levels and levels during the activity. It is difficult to report ranges during background measurements and during task measurements, as the two results are related (in other words: from the report the user should know which background belongs to which activity).

• Ratios between activity and background levels. Drawback is that the level of the background might affect the ratio, however this has not been observed in the NANOSH dataset. Alternatively, levels during activity and ratio should be reported.

• Currently, there are no exposure limits available for nanoparticles. Within REACH the exposure scenarios should demonstrate safe use scenarios. However, for exposure nanoparticles we are not able to do so.

• Not all the entry fields of the exposure scenarios could be completed since not all the information was collected during the NANOSH report.

15

4.3. Discussion and conclusions from model review and performance

check

The following main conclusions can be drawn, based on the model review and the

performance check:

ECETOC TRA and Stoffenmanager have been evaluated here with respect to their

applicability for exposure to nanoparticles. Basically, both models use the same concepts

of the process of exposure, i.e. a source- receptor approach distinguishing emission,

transport, immission and personal exposure. Recently, Schneider et al.(2010, accepted)

proposed a conceptual model for exposure to nanoparticles based on the same

mechanism. The source strength or emission potential is determined by the activity or

process and the fugacity of the substance. In cases where the emission source is not

contained, the emission potential is further modified during transport e.g. by ventilation and

room volume, and will result in exposure. Both ECETOC TRA and Stoffenmanager follow

the same basic principles, however, both models are less detailed than the proposed

conceptual model or more elaborated models for ‘conventional’ inhalation exposure, e.g.

the Advanced Reach Tool (ART; Fransman et al., 2009). Schneider et al. (2010)

distinguished some mechanisms that will cause changes in size distributions, e.g.

coagulation; however, the effect on the mass concentration should be minimal. Therefore,

it is concluded that both models should, in principle, be able to predict exposure to

nanoparticles. However, the different categories within each model variable are not

particularly suitable for activities of nanomaterials, resulting in a lack of resolution with

many situations falling into the same category. Refinement of these categories in view of

typical activities for nanomaterial handling, the number of handling categories etc, is

needed. Calibration for nanomaterials/nanoparticle emission and exposure concentration

is another issue that should be addressed. Currently, the Stoffenmanager is being

adjusted for nanomaterials exposure (nano module Stoffenmanager) where, in contrast to

version 4.0, the model output will be only a category of exposure (exposure band) rather

than a quantitative estimate.

Both first tier models only provide mass concentration as a proxy for exposure. Currently,

in the field of risk assessment, there is no consensus on the most relevant exposure

metric. However, for insoluble particles, particle number concentration and (active) surface

area concentration are also candidates. In addition, the size frame issue is important in

16

view of exposure assessment. Nanoparticles are usually defined as particles with (mobility,

aerodynamic, optical) sizes below 100 nm, however, agglomerates or aggregates of

primary nanoparticles may also be relevant to health if deagglomeration occurs in the

body. For typical ambient and workplace air situations, the numbers of particles with sizes

above about 300 nm will be relatively low compared to the nano-size particles. Therefore,

typical devices for particle concentration have size-windows up to 300 nm. However, in

cases where the size distribution is unknown, it is not possible to calculate the mass

concentration from the particle concentration without numerous assumptions. Based on

the physical relationship between size and mass the contribution of 100 nm or 300 nm

particles to mass compared to those of 1 or 3 µm is very low. So, if nanoparticle exposure

is focused on particle size ranges up to 300 nm, the resulting mass concentration would

always be in the lower ranges of the current model estimates. This indicates a further need

for recalibration of the models for nano-exposure.

From the results of the model performance check, it can be observed that there is no

correlation between the model outputs and the actual concentration data. In addition, no

difference can be observed between Stoffenmanager (activity-based) and ECETOC TRA

(extrapolated for full day exposure).

The particle number concentration data used to compare with the model scenario outputs

was obtained from devices that measured in the size ranges 10 -1000 nm (CPC), or 6- 700

nm (SMPS). As stated before, particles with these size ranges will contribute less to mass

compared to those particles with larger size ranges. It can therefore be hypothesized that

the resolution of the models for low mass concentration might be insufficient to show any

correlation.

In addition, the lack of any correlation is largely dominated by the large variability in the

(relatively few) data. The variability of the particle number concentration within the model

exposure scenarios is large compared to the between scenario model (mass

concentration) outputs, which would be expected to have a substantial effect on any

correlation. As has been demonstrated, due to lack of data or contextual information in the

data sets, the entire range of the model parameters categories could not be used, resulting

in a loss of power of discernment between exposure scenarios.

As been discussed for the NANOSH data set, the main observations were that the

category variables within the model parameters were not evenly distributed over the

17

various scenarios, so that the potential resolution between the scenario exposure

estimates was not fully exploited. Therefore it can be concluded that the data sets used to

check the performance of the exposure models for nanoparticle exposure scenarios were

not optimal for testing.

In summary, it can be concluded that, in their current form, both ECETOC TRA and

Stoffenmanager should not be used to estimate exposure to nanoparticles. Since the

models are not tuned to or calibrated for nanomaterial exposure situations, the actual

model estimate will be inaccurate and might overestimate the mass concentration levels.

19

5. References

Bello D, Hart AJ, Ahn K, Hallock M, Yamamoto N, Garcia EJ, Ellenbecker MJ, Wardle BL. Particle exposure levels during CVD growth and subsequent handling of vertically-aligned carbon nanotube films. Carbon 2008; 46(6): 974-977.

Bello D, Wardle BL, Yamamoto N, deVilloria RG, Garcia EJ, Hart AJ, Ahn K, Ellenbecker MJ, Hallock M. Exposure to nanoscale particles and fibers during machining of hybrid advanced composites containing carbon nanotubes. Journal of Nanoparticle Research, 2009; 11(1): 231-249.

Berges M. Workplace exposure characterization at TiO2 nanoparticle production. 3rd International Symposium on Nanotechnology, 2007.

Bullock WH, Laird LT. A Pilot Study of the Particle Size Distribution of Dust in the Paper and Wood Products Industry. American Industrial Hygiene Association Journal, 1994; 836-840.

Demou A, Peter P, Hellweg S. Exposure to Manufactured Nanostructured Particles in an Industrial Pilot Plant. Ann Occup Hyg, 2008; 52(8): 695–706.

Ellis ED, Watkins J, Tankersley W, Phillips J, Girardi D. Mortality Among Titanium Dioxide Workers at Three DuPont Plants. Journal of Occupational and Environmental Medicine, 2010: 303-309.

Fransman W, Cherrie J, van Tongeren M, Schneider T, Tischler M, Schinkel J, et al. Development of a Mechanistic Model for the Advanced REACH Tool (ART), Beta release. TNO report V8667, (2009) Zeist, The Netherlands, available from www.advancedreachtool.com

Fujitani Y, Kobayashi T, Arashidani K, Kunugita N, Suemura K. Measurement of the physical properties of aerosols in a fullerene factory for inhalation exposure assessment. Journal of Occupational and Environmental Hygiene, 2008; 5(6): 380-389

Fujitani Y, Kobayashi T. Measurement of aerosols in engineered nanomaterials factories for risk assessment. Nano, 2008; 3(4): 245-249.

Garabrant DH, Fine LJ, Oliver C, Bernstein L, Peters JM. Abnormalities of pulmonary function and pleural disease among titanium metal. Scand J Work Environ Health, 1987: 47-51

Han JH, Lee EJ, Lee JH, So KP, Lee YH, Bae GN, Lee SB, Ji JH, Cho MH, Yu IJ. Monitoring multiwalled carbon nanotube exposure in carbon nanotube research facility. Inhalation Toxicology, 2008; 20(8): 741-749.

Johnson DR, Methner MM, Kennedy AJ, Steevens JA. Potential for Occupational Exposure to Engineered Carbon-Based Nanomaterials in Environmental Laboratory Studies. Environmental Health Perspectives, 2010; 118(1): 49-54.

Korhonen K, Liukkonen T, Ahrens W, Astrakianakis G, Boffetta P, Burdorf A, Heederik D, Kauppinen T, Kogevinas M, Osvoll P, Rix BA, Saalo A, Sunyer J, Szadkowska-Stanczyk I, Teschke K, Westberg H, Widerkiewicz K. Occupational exposure to chemical agents in the paper industry. Int Arch Occup Environ Health, 2004: 451-460.

Lee JH, Lee SB, Bae GN, Jeon KS, Yoon JU, Ji JH, Sung JH, Lee BG, Lee JH, Yang JS, Kim HY, Kang CS, Yu IJ. Exposure assessment of carbon nanotube manufacturing workplaces. Inhalation Toxicology, 2010; 22(5): 369–381.

Maynard AD, Baron PA, Foley M, Shvedova AA, Kisin ER, Castranova V. Exposure to carbon nanotube material: aerosol release during the handling of unrefined single-walled carbon nanotube material. Journal of Toxicology and Environmental Health (Part A), 2004; 67(1): 87-107.

Mazzuckelli LF, Methner MM, Birch ME, Evans DE, Ku BK, Crouch K, Hoover MD.

20

Identification and characterization of potential sources of worker exposure to carbon nanofibers during polymer composite laboratory operations. Journal of Occupational and Environmental Hygiene, 2007; 4(12): D125-30.

Methner M, Hodson L, Dames A, Geraci C. 2010, "Nanoparticle Emission Assessment Technique (NEAT) for the identification and measurement of potential inhalation exposure to engineered nanomaterials. Part B: Results from 12 field studies. Journal of Occupational and Environmental Hygiene, 2010; 7(3): 163-176.

Park J, Kyu Kwak B, Bae E, Lee J, Kim Y, Choi K, Yi J. Characterization of exposure to silver nanoparticles in a manufacturing facility. J Nanopart Res, 2009; 11:1705–1712.

Conceptual model for assessment of inhalation exposure to manufactured nanoparticles Schneider T, Brouwer DH, Koponen IK, Jensen KA, Fransman W, van Duuren-Stuurman

B, van Tongeren M, Tielemans E. 2011. J. Expos. Sci. Environ. Epidemiol. Nature America. (accepted for publication).

Tsai SJ, Hofmann M, Hallock M, Ada E, Kong J, Ellenbecker M. Characterization and evaluation of nanoparticle release during the synthesis of single-walled and multiwalled carbon nanotubes by chemical vapor deposition. Environmental Science and Technology, 2009; 43(15): 6017-6023.

Tsai SJ, Ada E, Isaacs JA, Ellenbecker MJ. Airborne nanoparticle exposures associated with the manual handling of nanoalumina and nanosilver in fume hoods. J Nanopart Res, 2008: DOI 10.1007/s11051-008-9459-z

Yeganeh B, Kull CM, Hull MS, Marr LC. Characterization of airborne particles during production of carbonaceous nanomaterials. Environmental Science and Technology, 2008; 42(12): 4600-4606.

21

6. Annexes

6.1: Overview of exposure scenarios build in the NANEX Exposure Scenario Database 6.2: Exposure models and exploration of their applicability for nanomaterials 6.3: Performance checks of ECETOC TRA and STOFFENMANAGER with actual data 6.4: PDF extracts from the NANEX Exposure Scenario Database

22

6.1. Overview of exposure scenarios in the NANEX E xposure Scenario

Database

In table 1 an overview of the exposure scenarios presented in the NANEX Exposure Scenario Database is presented.

Table 1: Overview of exposure scenarios in the NANEX Exposure Scenario Database

Exposure type Title exposure scenario Substance name

Literature/ Database Reference

Use of substance by workers (Including Productions)

Production of MWCNT CNT Literature Mether et al, 2010


Production of Carbon Nanofibres

CNT Literature Mether et al, 2010


Handling of Carbon Nanofibres

CNT Literature Mether et al, 2010; Mazzuckelli et al, 2007


Handling of fullerenes CNT Literature Johnson et al, 2010


Handling of MWCNT CNT Literature Mether et al, 2010; Johnson et al, 2010


Handling of CNT CNT Literature Maynard et al, 2010


Manufacturing and handling of carbon nanotubes and handling (MWCNT)

CNT Literature Lee et al, 2010


Manufacturing of MWCNT

CNT Literature Han et al, 2008


Production of carbonaceous nanomaterials (fullerenes and other carbonaceous nanomaterials)

CNT Literature Yeganeh et al, 2008


Production of carbon nanomaterial

CNT Literature Fujitani et al, 2008b


Synthesis of fullerene CNT Literature Fujitani et al, 2008a


Synthesis and handling of CNT by CVD

CNT Literature Bello et al, 2008


Production and handling of metal-based nanoparticles in a gas-phase production process

Other Literature Demou et al, 2008


Synthesis of SWCNT and MWCNT by CVD

CNT Literature Tsai, S.J., 2009

23




Machining of hybrid advanced composites containing CNT.

CNT Literature Bello et al, 2009


Manipulation of Nano-Ag into three different type of fume hoods (Conventional hood; By-pass hood; Constant velocity hood)

Nano-Ag Literature Tsai et al, 2008


Production of Nano-Ag during wet-chemistry process

Nano-Ag Literature Park, J. 2009


Autocombustion of lanthane, strontium, cobalt, iron

Other Database NanoINNOV


Cleaning of growth furnace producing nanothread Si



Maintenance of SiO2 PECVD equipment (Plasma-enhanced chemical vapour deposition)



preparation of inks with nanoZnO



Agitation of a solution of carbon black with N-methyl-pyrolidone



Opening of an epitaxy frame of an apparatus with a molecular beam with a Ti target



Opening of deposition equipment containing adsorption bed for chemical vapour deposition, used with diverse metal oxides.



Packaging of carbon black


Handling of articles by workers

Cutting of substrates coated by carbon black particles



Grinding of Nano-TiO2 TiO2 Database NanoINNOV


Maintenance of device polluted with NP with glassbead cabinets

Nano Ag Database NanoINNOV & NanoSafe


maintenance of physical vapour deposition (PVD)

Nano Ag Database NanoINNOV


Production of TiO2 by laser pyrolysis

TiO2 Database NanoINNOV


weighing of CNT CNT Database NanoINNOV

24




CNT in solution CNT Database NanoINNOV


Pouring of CNT CNT Database NanoINNOV


Handling small quantities of CNT

CNT Database NanoINNOV


Preparation of CNT pellets from CNT powder

CNT Database NanoINNOV


Production of MWCNT using gas-phase reactor

CNT Database NANOSH


production of paint TiO2 Database NANOSH


Production of pavement stones TiO2 Database NANOSH


Production of MWCNT at laboratory scale

CNT Database NANOSH


Production of TiO2 TiO2 Database NANOSH


Production of MWCNT using a tube furnace

CNT Database NANOSH


Production of TiO2 by laser ablation

TiO2 Database NANOSH


Laboratory activities on CNTs

CNT Database NANOSH


Production of filaments of CNTs

CNT Database NANOSH


Dry mounting of CNTs on to EM grids CNT Database NANOSH


Working at a research reactor to produce CNT

CNT Database NANOSH


Working at an extruder to produce polymer containing CNT

CNT Database NANOSH


CNT production using Chemical Vapour Deposition (CVD)

CNT Database NANOSH


Production of printing inks



Production of cosmetics in a laboratory


25




Test 2. Transfer of substances or preparations (charging/discharging) from/to vessels/large containers at non-dedicated facilities (activities including sieving, bin filling of powder and bagging of powder)

Other Database NANOSH


Test 1. Production in laboratory of cosmetics containing zinc oxide powder

Other Database NANOSH


Occupational exposure scenario during the production of TiO2 at Umicore SA, Belgium

TiO2 Literature Berges, 2007


Exposure scenario in paper production

TiO2 Literature Bullock et al, 1994; Korhonen et al, 2004


Exposure to TiO2 in a lab and at a manufacturer

TiO2 Literature Methner et al, 2010


Exposure to TiO2 at three DuPont plants

TiO2 Literature Ellis et al, 2010


Exposure scenario in the production of titanium metal

TiO2 Literature Garabrant et al., 1987

26

6.2. Exposure models and exploration of their appl icability for

nanomaterials

In this annex, the identification and review of available tools and models to predict occupational exposure with respect to nanomaterials and identified scenarios is addressed. The review is limited to the exposure models ECETOC TRA as a first tier tool and STOFFENMANAGER as a higher tier tool, both of which are mentioned in the REACH Guidance R.14 from ECHA (2008) dealing with the requirements and chemical safety assessment for the occupational exposure estimation.

6.3.1 ECETOC TRA model

Introduction ECETOC TRA (ECETOC, 2009) is a first tier tool developed by ECETOC (European Centre for Ecotoxicology and Toxicology Of Chemicals) to assess health and environmental risks from the supply and use of chemicals. The ECETOC TRA uses established exposure-prediction models, introduced in a more precise, structured and simplified approach as a web-based tool. The calculated basis for the approach is a modified version of the EASE (Estimation and Assessment of Substance Exposure) exposure model version 2.0, developed by the UK Health and Safety Executive (HSE) (Cherrie et al., 2003). The model is based on a relation between PROCs (product categories described in the REACH guidance) and basic exposure levels. The basic exposure levels can be modified by a limited number of operational conditions and risk management measures.

Inhalation and dermal model The EASE version 2.0 inhalation model, which forms the basis for the ECETOC TRA tool,is based on an exposure database from the HSE which contains about 100,000 exposure measurements. These measurements were grouped together in exposure ranges, based on processes with similar potential for exposure. The potential for exposure is based on the tendency of the substance to become airborne; the means of controlling exposure or of preventing the substance from entering the workroom atmosphere; and the way in which the substance is used. For liquids, the tendency to become airborne is indicated by the vapour pressure and thus the volatility of the substance and, for solids, the dustiness is considered an important parameter. There was little or no data available for dermal exposure and therefore a simple EASE dermal model was developed using the same parameters as for the inhalation model and adding the parameter ‘level of contact’ to give an indication of frequency and duration of direct exposure with the substance.

Estimating exposure using the model In the ECETOC TRA model the following input parameters are necessary to run the model and to make an occupational exposure estimate:

• Vapour pressure or dustiness

For vapour pressure the categories low (Vp >=0.00001- <0.5), medium (Vp 0.5 to 10) and high (> 10) are possible.

For dustiness the categories low (relative dustiness potential is 1 or 10-100 times dustier), medium (100 – 1,000 times dustier) and high (more than 1,000 times dustier) are also

27

possible. For these categories examples of typical materials are given.

• Selection of a generic use scenario

A choice can be made out of 25 process categories (PROCs).

• Selection of limited exposure determinants

For duration (modifying factor for > 4 hours (1), 1-4 hours (0.6), 15 min. – 1 hour (0.2) and <15 min. (0.1); use of local exhaust ventilation (yes or no); and use of RPE (no, 90 or 95% reduction).

• Selection of mixture or pure substance

No mixture (exposure modifying factor is 1); concentration >25 % w/w (modifying factor 1); concentration 5-25% w/w (modifying factor 0.6); concentration 1-5% w/w (modifying factor 0.2); and concentration is <1% w/w (modifying factor 0.1).

• Exposed skin surface

For each PROC a default for the exposed skin surface is given in order to reverse the predicted dermal exposure in µgcm-2day-1 to mgkg-1day-1.

The inhalation exposure estimates are presented in ppm and mg/m³ and dermal values in ugcm-2day-1 and, related to body weight, in mgkg-1day-1 and total exposure in mgkg-1day-1.

Validity and reliability of model The ECETOC TRA model is based on a modified version of the EASE 2.0 model. This EASE model has been validated with a number of studies in which the predictions of the model are compared with actual measurements and estimates of exposure. In addition an assessment of the degree of variation between different users of the model was undertaken, to determine whether it was capable of being used in a consistent fashion. In the ECETOC TRA model the EASE inhalation predictions have been modified for the use of ventilation and the dermal EASE predictions have been reviewed in light of available measured data (such as the RiskOfDerm and related projects) in order to ensure reasonable exposure estimates. Since the ECETOC TRA tool tends to underestimate dermal exposure in situations with LEV, the RiskOfDerm model should preferably be used for these cases (for an evaluation of the RiskofDerm model see annex 1 of the WP4 consumer report). Next to this the ECETOC TRA approach has also been validated and results showed that the approach offers the basis for a suitable cautionary scheme for the assessment of worker exposure.

The ECETOC scheme appears unsuitable for the assessment of workplace risks for:

• Mists (liquid aerosols); • Fumes arising from the use of a material within a process; • Working situations not described within the suite of generic scenarios, e.g. confined

spaces, abnormal exposure situations (e.g. spills).

Evaluation of model for suitability of nanomaterials under REACH A lot of the parameters that are used in the model will be similar for nanomaterials to those for conventional compounds, although the actual exposure values may not necessarily be equal. This is due to the possible specific characteristics of nanomaterials. Chemical

28

composition, surface structure, solubility, shape and aggregation behaviour might be completely different compared with conventional compounds. For example the same particle number for nanoparticles will result in completely different (i.e. lower) mass concentration compared to the same particle concentration for larger particles. This means that the model output as mass concentration exposure levels will probably overestimate the exposure to nanomaterials. However, since this model is a first tier model, this can be considered as a worst case estimate.

Since the model is neither built nor validated for nanomaterials, caution is necessary in using this model to estimate the exposure to nanomaterials.

Conclusion In principle ECETOC TRA will be suitable for estimate exposure to nanomaterials. However, since it is not calibrated for nanomaterials, the model estimate will most probably overestimate the exposure levels.

6.3.2 STOFFENMANAGER 4.0

Introduction The Stoffenmanager (Dutch for “substance manager”) tool originally is a web-based risk prioritizing tool for small and medium-sized enterprises. The first version of the Stoffenmanager was developed in 2003 by TNO, Arbo Unie and BECO on behalf of the Dutch Ministry of Social Affairs and Employment and the latest version is version 4.0. The exposure model developed in the latest version is a quantitative model for estimating inhalation exposure to vapours, aerosols of low volatility liquids and dusts. The Stoffenmanager would currently be regarded as a tool between a first Tier and higher Tier models. The model is based on exposure levels of tasks.

Inhalation model The model is based on the source-receptor approach (Cherrie et al., 1999, Marquart et al., 2008) but adapted in several ways. This approach indicates that exposure depends on the emission, transmission and imission factors. Determinants of exposure include task, local controls, and general ventilation and product characteristics and each are scored on a logarithmic scale. The exposure algorithm was quantified for dust and vapour scenarios for occupational exposure using 700 exposure measurements. The inhalation exposure is expressed in exposure concentration in mgm-³, based on the 90th percentiles of measured exposures.

The model uses process information, physiological characteristics and mass balance to assess exposure situations. The following parameters are needed as input data in the model:

• Vapour pressure or dustiness • Percentage of substance in the product • Handling category (7 classes for liquid, and 6 for solid) • Local controls • Distance of worker to source • Presence of other sources of same substance further than 1 metre from the worker • Room volume (4 classes) • General ventilation (3 classes) • Imission control measures

29

• Personal Protective Equipment (5 classes) • Possibility of background exposure

Estimating exposure using the model

To make an exposure assessment, values for the parameters above are entered into the model, and the resultant exposure estimates can be compared with REACH Derived No Effect Levels (DNELs) to derive safe uses.

Validity and reliability of model The Stoffenmanager is calibrated and validated with approximately 700 measurement data originating from STEAMBASE (Stoffenmanager Exposure and Modelling database). As well as the measured concentrations (mg/m³), this database contains all the contextual information necessary to make a Stoffenmanager exposure estimation.

Evaluation of model for suitability of nanomaterials under REACH As with ECETOC TRA, many of the parameters used for nanomaterials in the model will be similar to those for conventional compounds, although the actual exposure values may not necessarily be equal. This is again due to the possible specific characteristics of nanomaterials; meaning that the model output as mass concentration exposure levels will again probably overestimate the exposure to nanomaterials but can be considered as a worst case estimate.

Since the model is not built nor calibrated for nanomaterials caution is necessary in using this model to estimate the exposure to nanomaterials.

Conclusion: In principle, Stoffenmanager will be suitable for estimate exposure to nanomaterials. However, since it is not calibrated for nanomaterials, the model estimate will most probably overestimate the exposure levels.

Stoffenmanager Nano 1.0 is currently being developed. This model will contain emission, transmission and imission factors specifically for nanomaterials, greatly increasing the power of discernment for the identified scenarios. However this model has yet to be calibrated and validated with measurement data. In addition, the challenge to translate the assessment into REACH Use Descriptors remains.

6.3.3 Discussion and conclusion

The first-tier models for estimating occupational exposure under REACH, i.e. ECETOC TRA and Stoffenmanager have been evaluated here with respect to their applicability for exposure to nanoparticles. Both models use essentially the same concepts of the process of exposure, i.e. a source-receptor approach distinguishing emission, transport, imission and personal exposure. Recently, Schneider et al. (2010, accepted) proposed a conceptual model for exposure to nanoparticles, based on the same mechanism. The source strength or emission potential is determined by the activity or process and the fugacity of the substance. In cases where the emission source is not contained, the emission potential is further modified during transport e.g. by ventilation and room volume, and will result in exposure. Both ECETOC TRA and Stoffenmanager follow the same basics. However, both models are less detailed than the proposed conceptual model or more elaborated models for ‘conventional’ inhalation exposure, e.g. the Advanced Reach

30

Tool (ART; Fransman et al., 2009).

Schneider et al. (2010) distinguished some mechanisms that will affect shift in size distributions, e.g. coagulation; however, the effect on the mass concentration should be minimal. Therefore, it is concluded that both models should, in principle, be able to predict exposure to nanoparticles. However, the different categories within each model variable are not particularly well-fitted for activities involving nanomaterials. As a result, many situations may, in practice, fall into the same category. Refinement of these categories in view of typical activities for nanomaterial handling, the number of handling categories etc, is needed. Calibration for nanomaterials/nanoparticle emission and exposure concentration is another issue that should be addressed. Currently, the Stoffenmanager is in the process of adjustment for nanomaterial exposure (nano module Stoffenmanager), where in contrast to version 4.0, the model output will only be a category of exposure (exposure band) rather than a quantitative estimate.

An issue that has only implicitly been addressed during the evaluation of the models is the exposure metric. Both first-tier models only provide mass concentration as a proxy for exposure. Currently, in the field of risk assessment, there is no consensus on the most relevant exposure metric, however, for insoluble particles, particle number concentration and (active) surface area concentration are candidates. In addition, the size-frame issue is important in view of exposure assessment. Nanoparticles are usually defined as particles with (mobility, aerodynamic, optical) sizes below 100 nm. However, agglomerates or aggregates of primary nanoparticles might also be relevant for health if deagglomeration occurs in the body. For typical ambient and workplace air situations the numbers of particles with sizes above about 300 nm will be relatively low compared to the nano-sized particles. Therefore, typical devices for particle concentration have size-windows up to 300 nm. However, in cases where the size distribution is unknown, it is not possible to calculate mass concentration from particle concentration without numerous assumptions. Based on the physical relationship between size and mass the contribution of 100 nm or 300 nm particles to mass, compared to particles of 1 or 3 µm, is very low. So, if nanoparticle exposure is to be focused on particle size ranges up to 300 nm, the resulting mass concentration will always be in the lower ranges of the current model estimates. This also indicates the need for recalibration of the models for exposure to nanomaterials.

31

6.3. Performance check ECETOC TRA and STOFFENMANAG ER with

actual data

6.4.1 Introduction

To build exposure scenarios it was decided to estimate exposure for the scenarios obtained from the datasets of the NANOSH project (TNO and the NanoINNOV project (CEA). ECETOC TRA and Stoffenmanager were selected, as these models are recommended within the REACH guidance.

The applicability of these models for assessing exposure to nanoparticles is described in previous sections. The NANOSH dataset provided 46 scenarios to be assessed with the two models, with a further 15 scenarios obtained from nanoINNOV. A performance check of the models for nanomaterial exposure estimates was initiated by comparing the measurement values of the situations from the dataset with the results of the modelled exposure.

The first sections discuss how the exposure assessment has been performed with two different models and the required assumptions regarding the input. These are followed by the results, after which the consequences of the model assumptions for the results are discussed.

6.4.2 Assessments with ECETOC TRAv2

As ECETOC TRA is based on Process Category (PROC), a relationship needs to be established between the activity during the measurements from the Nanosh project and the process category. This will result in a basic exposure level which can be modified by a limited number of operational conditions and risk management measures. An overview of the different PROCs according to REACH is presented in Table 1.

32

Table 1 Overview PROCs Appendix R.12 -3: Descriptor -list for process categories [PROC]

Process category

Examples and explanations

PROC1 Use in closed process, no likelihood of exposure

Use of the substances in high integrity contained system where little potential exists for exposures, e.g. any sampling via closed loop systems.

PROC2

Use in closed, continuous process with occasional controlled exposure

Continuous process but where the design philosophy is not specifically aimed at minimizing emissions. Not high integrity and occasional exposure will arise e.g. through maintenance, sampling and equipment breakages

PROC3 Use in closed batch process (synthesis or formulation)

Batch manufacture of a chemical or formulation where the predominant handling is in a contained manner, e.g. through enclosed transfers, but where some opportunity for contact with chemicals occurs, e.g. through sampling.

PROC4

Use in batch and other process (synthesis) where opportunity for exposure arises

Use in batch manufacture of a chemical where significant opportunity for exposure arises, e.g. during charging, sampling or discharge of material, and when the nature of the design is likely to result in exposure.

PROC5

Mixing or blending in batch processes for formulation of preparations and articles (multistage and/or significant contact)

Manufacture or formulation of chemical products or articles using technologies related to mixing and blending of solid or liquid materials, and where the process is in stages and provides the opportunity for significant contact at any stage.

PROC6 Calendering operations Processing of product matrix. Calendering at elevated temperature and a large exposed surface area.

PROC7 Industrial spraying

Air dispersive techniques. Spraying for surface coating, adhesives, polishes/cleaners, air care products, sandblasting; Substances can be inhaled as aerosols. The energy of the aerosol particles may require advanced exposure controls; in case of coating, overspray may lead to waste water and waste.

PROC8a

Transfer of substance or preparation (charging/discharging) from/to vessels/large containers at non-dedicated facilities

Sampling, loading, filling, transfer, dumping, bagging in non- dedicated facilities. Exposure related to dust, vapour, aerosols or spillage, and cleaning of equipment to be expected.

PROC8b

Transfer of substance or preparation (charging/discharging) from/to vessels/large containers at dedicated facilities.

Sampling, loading, filling, transfer, dumping, bagging in dedicated facilities. Exposure related to dust, vapour, aerosols or spillage, and cleaning of equipment to be expected.

PROC9

Transfer of substance or preparation into small containers (dedicated filling line, including weighing)

Filling lines specifically designed to capture both vapour and aerosol emissions and minimise spillage

PROC10 Roller application or brushing

Low energy spreading of e.g. coatings. Including cleaning of surfaces. Substance can be inhaled as vapours, skin contact can occur through droplets, splashes, working with wipes and handling of treated surfaces.

PROC11 Non-industrial spraying

Air dispersive techniques, spraying for surface coating, adhesives, polishes/cleaners, air care products, sandblasting. Substances can be inhaled as aerosols. The energy of the aero-sol particles may require advanced exposure controls.

PROC12 Use of blow agents for foam production

33

Appendix R.12 -3: Descriptor -list for process categories [PROC]

Process category


PROC13 Treatment of articles by dipping and pouring

Immersion operations. The treatment of articles by dipping, pouring, immersing, soaking, washing out or washing in substances; including cold formation or resin type matrix. Includes handling of treated objects (e.g. after dying, plating,). Substance is applied to a surface by low energy techniques such as dipping the article into a bath or pouring a preparation onto a surface.

PROC14

Production of preparations or articles by tabletting, compression, extrusion, pelletisation

Processing of preparations and/or substances (liquid and solid) into preparations or articles. Substances in the chemical matrix may be exposed to elevated mechanical and/or thermal energy conditions. Exposure is predominantly related to volatiles and/or generated fumes, dust may be formed as well.

PROC15 Use as laboratory reagent. Use of substances at small scale laboratory (< 1 l or 1 kg pre-sent at workplace). Larger laboratories and R+D installations should be treated as industrial processes.

PROC16 Using material as fuel sources, limited exposure to unburned product to be expected.

Covers the use of material as fuel sources (including additives) where limited exposure to the product in its unburned form is expected. Does not cover exposure as a consequence of spill-age or combustion.

PROC17 Lubrication at high energy conditions and in partly open process.

Lubrication at high energy conditions (temperature, friction) between moving parts and substance; significant part of process is open to workers. The metal working fluid may form aerosols or fumes due to rapidly moving metal parts.

PROC18 Greasing in high energy conditions.

Use as lubricant where significant energy or temperature is applied between the substance and the moving parts.

PROC19 Hand-mixing with intimate contact and only PPE available.

Addresses occupations where intimate and intentional contact with substances occurs without any specific exposure controls other than PPE.

PROC20

Heat and pressure transfer fluids in dispersive, professional use but closed systems.

Motor and engine oils, brake fluids. Also in these applications, the lubricant may be exposed to high energy conditions and chemical reactions may take place during use. Exhausted fluids need to be disposed of as waste. Repair and maintenance may lead to skin contact.

PROC21 Low energy manipulation of substances bound in materials and/or articles.

Manual cutting, cold rolling or assembly/disassembly of material/article (including metals in massive form), possibly resulting in the release of fibres, metal fumes or dust;

PROC22

Potentially closed processing operations with minerals/metals at elevated temperatures in an industrial setting.

Activities at smelters, furnaces, refineries, coke ovens. Exposure related to dust and fumes to be expected. Emission from direct cooling may be relevant.

PROC23

Open processing and transfer operations with minerals/metals at elevated temperature.

Sand and die-casting, tapping and casting melted solids, drossing of melted solids, hot dip galvanising, raking of melted solids in paving. Exposure related to dust and fumes to be expected.

PROC24 High (mechanical) energy work-up of sub-stances bound in materials and/or articles

Substantial thermal or kinetic energy applied to substance (including metals in massive form) by hot rolling/forming, grinding, mechanical cutting, drilling or sanding. Exposure is pre-dominantly expected to be to dust. Dust or aerosol emission as result of direct cooling may be expected.

PROC25 Other hot work operations with metals

Welding, soldering, gouging, brazing, flame cutting Exposure is predominantly expected to fumes and gases.

34

Appendix R.12 -3: Descriptor -list for process categories [PROC]

Process category


PROC26 Handling of solid inorganic substances at ambient temperature

Transfer and handling of ores, concentrates, raw metal oxides and scrap; packaging, un-packaging, mixing/blending and weighing of metal powders or other minerals.23

PROC27a Production of metal powders (hot processes)

Production of metal powders by hot metallurgical processes (atomisation, dry dispersion).24

PROC27b Production of metal powders (wet processes)

Production of metal powders by wet metallurgical processes (electrolysis, wet dispersion).25

23 no corresponding TRA entry 24 no corresponding TRA entry 25 no corresponding TRA entry

The input parameters for the models are; • Vapour pressure or dustiness, industrial or professional setting • Selection of a generic use scenario/ process category (PROC) • Selection of limited exposure determinants: Percentage of the nanomaterials in the

product, duration of exposure , indoor/outdoor setting, use of local exhaust ventilation (LEV) and use of personal protection equipment.

Parameters to obtain the basic exposure level: 1) Vapour pressure and dustiness

These data are captured in the tool on an input data screen. They were used to categorize the material according to its fugacity (tendency of a substance to become airborne from a heterogeneous system) as defined in an availability banding for an initial assessment.

Since actual data are lacking, all nanomaterials were considered to be solids and assigned with the highest category of dustiness. Only nanomaterials in liquid suspension or solid articles were regarded as being in the lowest dustiness category.

2) Industrial or Professional setting.

Only scenarios in an industrial setting were considered.

3) Selection of a generic use scenario/ process category (PROC)

The generic use scenario/ process categories (PROC) for the measured situations were assigned as follows;

• When the measurement was performed at a completely closed system the situation was described by PROC1.

• When the measurement was performed at a closed system, but with some opportunity for exposure, the situation was described by PROC3.

• When the measurement was performed at a dedicated facility for transfer of chemicals (e.g. bagging at a dedicated facility, handling contaminated bags at a dedicated facility) the situation was described by PROC8b/9.

• When the measurement was performed at a non-dedicated facility for transfer of chemicals or included potential higher exposure due to other related activities (e.g transfer including sieving) the situation was described by PROC8a.

35

• When the measurement was performed during activities with small quantities (at least <1 l or <1 kg) the situation was described by PROC15.

• When the measurement was performed during mixing activities the situation was described by PROC5.

• When the measurement involved pelletisation or extrusion the situation was described by PROC14.

• When the measurement was performed at a site where exposure to nanoparticles was expected due to manipulation of articles or substances containing nanoparticles, the situation was described by PROC21.

When more than one of the PROC seemed applicable both were assessed, and the reasonable worst case (expert judgement) was considered valid.

Further modifications of the basic exposure level: a) Percentage of the nanomaterial in product.

Neither data set contained exact information on the concentration of nanomaterials in products. From the descriptions however it could be deduced that, in general, pure nanomaterials (concentration =100%), were handled during the activities

b) Duration of exposure.

The measurements within the datasets represent the total exposure during the specific activity that day. The assigned generic use scenario/process categories of processes and tasks contain periods of both activity and inactivity. Consequently, it was neither possible nor desirable to correct for the exposure duration of one activity and therefore the exposure measurement was considered be comparable to the full shift exposure estimate of the assigned PROC.

c) Indoor/outdoor setting

ECETOC TRA provides the option of an exposure reduction of 30% when the assessed situation is outside. Both datasets only contain scenarios in indoor settings.

This facility did not therefore offer any further modification and did not increase the power of discernment by the exposure estimates between the different measurements.

d) Risk management measures.

Within ECETOC TRA, two risk management measures could be applied: Local Exhaust ventilation (LEV); and respiratory protection.

The use of LEV is well documented within the datasets. Within ECETOC TRA, specific efficiencies are applicable for LEV for each basic exposure level. These efficiencies were applied to the exposure estimation when LEV was used during the measurements. Respiration protection is not applicable since the data in the datasets was derived from static measurements at fixed positions and not from personal measurements. However, for the calculations using the nanoINNOV dataset, respiratory protection was included.

In summary, due mainly to the lack of contextual information, the variables within ECETOC TRA potentially differentiating between the various measurement situations were limited

36

to: two categories of dustiness, i.e. high for all solids and low for liquids; the selection of a generic use scenario/ process category (PROC); and the possible application of LEV (PROC1 excepted).

6.4.2 Assessments with Stoffenmanager 4.0

This model uses process information, physicochemical characteristics, and mass balance to assess exposure situations. In contrast to ECETOC TRA, Stoffenmanager is, as with the measurement data in the dataset, activity based. This means that allocation of the handling category of Stoffenmanager to the activity in the dataset was more straightforward.

The input parameters of Stoffenmanager and considerations during the exposure estimation of the measured situations are:

1. Vapour pressure or dustiness

Since actual data were lacking, all nanomaterials were considered to be solids and assigned the highest category of dustiness. Only those nanomaterials in liquid suspension or solid articles were regarded as being in the lowest dustiness category.

2. Percentage of the substance in the product

Neither data set contains exact information on the concentration of nanomaterials in products. From the descriptions given however, it could be deduced that, in general, pure nanomaterials (concentration =100%), were handled during the activities.

3. Handling category

Solid substances can be assigned to one of six classes. These were assigned by comparing the description of the activity and quantity with the handling categories in Stoffenmanager.

4. Local controls and general ventilation (3 classes)

The use of local exhaust ventilation, general ventilation separation and segregation are well documented in the datasets and the categories were easily assigned in the Stoffenmanager tool.

5. Distance of the worker to the source (within 1 metre/further than one metre)

This type of information is not given in the datasets. It was assumed that all measurements were performed in close proximity to the source and that the near field exposure contributed to the exposure level.

6. Presence of other sources of the same substance further than one metre from the

worker (yes/no)

This information is given for most scenarios and, for the remaining ones, it was

37

assumed not to be the case.

7. Room volume (4 classes)

This information is well documented in the datasets and the actual room volume was easily assigned to the corresponding room volume class.

8. Imission control measures (worker in separate control room with clean air supply,

worker in cabin without specific ventilation system, no cabin).

Separation was not considered to be applicable for all measurement situations in the datasets.

9. Personal protective equipment used (5 classes)

Respiratory protection was not applicable since the data in the datasets was derived from static measurements at fixed positions and not from personal measurements.

10. Possibility of background exposure

For all measurement situations it was assumed that the working area was not cleaned on a daily basis and that equipment was not regularly maintained.

In contrast to ECETOC TRA, exposure duration and frequency are not required for Stoffenmanager as the event- or activity-based exposure will be comparable to the exposure measurement results.

In summary, due mainly to the lack of contextual information, the remaining variables within Stoffenmanager were limited to only four variables; i.e. two categories of dustiness (high for all solids and low for liquids); handling category; local controls/general ventilation; and room volume.

6.4.3 Performance check results

For both datasets, the particle number concentration data were compared with the estimates calculated for the same scenarios using the Stoffenmanager and ECETOC TRA models. A simple correlation was calculated, either for the absolute difference in concentrations or for the ratio of concentrations with and without activity (background).

6.4.3.1 Comparison of model outputs with data from the NANOSH data set

Table 2 shows the scenarios included in the NANEX Exposure Scenario Database which were used for comparison with exposure model outputs. Table 3 shows the type of data available, i.e. particle number concentration data obtained by Condensation Particle Counters (CPC) (usually in a size-range from 10 to 1000 nm) and data obtained by a Scanning Mobility Particle Sizer (SMPS) (usually in the size range from 6 to about 700 nm).

Table 2 Overview of exposure scenarios based on the NANOSH dataset in the NANEX Exposure Scenario Database for which a comparison with exposure models was made

38

Exposure type Title exposure scenario Substance name Contributing exposure scenario


Production of paint TiO2

14.1: Dumping large amount of powder into vessel 14.2: Dumping medium amount of powder into vessel


Production of MWCNT at laboratory scale CNT 12.1: Operating the oven


Production of TiO2 TiO2 11.2: Bag/bin filling


Production of MWCNT using a tube furnace

CNT 10.1: Production of MWCNT on a silicon substrate


Production of TiO2 by laser ablation

TiO2 9.1: Laser ablation (PROC 15, 26)


Laboratory activities on CNTs

CNT

8.1: Transfer of liquid containing CNTs (PROC 15) 8.2: Weighing of powder (PROC 15) 8.3: Manipulation of nanomaterial powder (PROC 15)


Production of filaments of CNTs

CNT 7.1: Production of filaments of CNTs


Dry mounting of CNTs on to EM grids CNT

6.1: Dry mounting of CNTs on to EM grids (PROC 15)


CNT production using Chemical Vapour Deposition (CVD)

CNT 3.1: Sampling from reactor (PROC 1) 3.2: Bagging from reactor (PROC 9)


Production of printing inks TiO2 2.1: Emptying bags in filling station (PROC 9 or 26)


Production of cosmetics in a laboratory

TiO2 1.1: Weighing powder, PROC 15 or 26

39

Table 3 Summary of available types of data for the various Contributing Exposure Scenarios (CES) used for the model performance check

CES 1.1 2.1 3.1 3.2 6.1 7.1 8.1 8.2 8.3 9.1 10.1 11.2 12.1 14.1 14.2

CPC* - + + + - - - - - - - - + + +

SMPS** + - - - + + + + + + + + + + +

* Additionally six handlings (bagging, box filling, pouring, sieving) involving ZnO2 were used (no contributing

exposure scenarios were available for ZnO2).

** Additionally 10 handlings (bagging, box filling, pouring, sieving, weighing, transferring) involving ZnO2 (no

contributing exposure scenarios were available for ZnO2).

The results of the comparisons between the model outcomes and actual workplace particle number concentration data obtained by CPC and SMPS are presented in Table 4 and plotted in figures 1-8.

Table 4 Pearson correlation coefficients for the ECETOC TRA model and Stoffenmanager

Pearson correlation coefficients CPC results N=13 SMPS results N=23 Absolute difference activity minus non-activity

Ratio of task vs. background

Absolute difference activity minus non-activity

Ratio of task vs. background

Stoffenmanager Inhalable dust (mg/m³) -0.15 -0.17 -0.0072 0.052

ECETOC exposure level (mg/m³) 0.063 -0.022 -0.097 -0.15

.

40

Figure 1 Stoffenmanager exposure estimate (mg/m³) versus ratio (GM concentration task versus GM concentration background) measured by CPC.

R2 = 0,0297

0

0,5

1

1,5

2

2,5

0,0 2,0 4,0 6,0 8,0 10,0 12,0 14,0 16,0 18,0 20,0

Stoffenmanager exposure level (mg/m3)

ratio

task

ver

sus

back

grou

nd

GM

AM

Linear (GM)

Figure 2 Stoffenmanager exposure estimate (mg/m³) versus absolute difference (GM concentration task minus GM concentration background) measured by CPC.

R2 = 0,0217

-20000

-15000

-10000

-5000

0

5000

10000

15000

20000

25000

0,0 5,0 10,0 15,0 20,0


Abs

olut

e di

ffere

nce

activ

ity m

inus

non

-ac

tivity

GM

AM

Linear (GM)

41

Figure 3 ECETOC TRA exposure estimate (mg/m³) versus ratio (GM concentration task versus GM concentration background) measured by CPC.

R2 = 0,0005

0

0,5

1

1,5

2

2,5

0,0 5,0 10,0 15,0 20,0 25,0

ECETOC TRA level (mg/m3)

ratio

task

ver

sus

back

grou

nd

GM

AM

Linear (GM)

Figure 4 ECETOC TRA exposure estimate (mg/m³) versus absolute difference (GM concentration task minus GM concentration background) measured by CPC.

R2 = 0,004

-20000

-15000

-10000

-5000

0

5000

10000

15000

20000

25000

0,0 5,0 10,0 15,0 20,0 25,0

ECETOC TRA level (mg/m3)

Abs

olut

e di

ffere

nce

activ

ity m

inus

non

-ac

tivity

GM

AM

Linear (GM)

42

Figure 5 Stoffenmanager exposure estimate (mg/m³) versus ratio (GM concentration

task versus GM concentration background) measured by SMPS.

R2 = 0,0028

0

0,5

1

1,5

2

2,5

3

3,5

0,0 2,0 4,0 6,0 8,0 10,0


ratio

task

ver

sus

back

grou

nd

GM

AM

Linear (GM)

Figure 6 Stoffenmanager exposure estimate (mg/m³) versus absolute difference (GM concentration task minus GM concentration background) measured by SMPS.

R2 = 5E-05

-15000

-10000

-5000

0

5000

10000

15000

20000

25000

30000

0,0 2,0 4,0 6,0 8,0 10,0


Abs

olut

e di

ffere

nce

activ

ity m

inus

non

-ac

tivity

GM

AM

Linear (GM)

43

Figure 7 ECETOC TRA exposure estimate (mg/m³) versus ratio (GM concentration

task versus GM concentration background) measured by SMPS.

R2 = 0,0222

0

0,5

1

1,5

2

2,5

3

3,5

0,0 10,0 20,0 30,0 40,0 50,0 60,0

ECETOC TRA exposure level (mg/m3)

ratio

task

ver

sus

back

grou

nd

GM

AM

Linear (GM)

Figure 8 ECETOC TRA exposure estimate (mg/m³) versus absolute difference (GM concentration task minus GM concentration background) measured by SMPS.

R2 = 0,0093

-15000

-10000

-5000

0

5000

10000

15000

20000

25000

30000

0,0 10,0 20,0 30,0 40,0 50,0 60,0

ECETOC TRA exposure level (mg/m3)

Abs

olut

e di

ffere

nce

activ

ity m

inus

non

-ac

tivity

GM

AM

Linear (GM)

44

6.4.3.2 Comparison of model outputs with data NanoI NNOV data set

Fourteen Stoffenmanager scenarios could be calculated from the nanoINNOV project, with model outcomes ranging from 0.03 to 8.12 mg/m³. Only five scenarios for ECETOC TRA calculations could be extracted, with model outcomes ranging from 0.05 to 5 mg/m³.

Table 5 Overview exposure scenarios calculated with model outcomes

Exposure Type Exposure Scenario Title Substance name

Stoffen -manager

ECETOC TRA


Preparation of inks with nanoZnO ZnO + -


Agitation of a solution of carbon black with N-methyl-pyrolidone

Carbon black + -


Opening of deposition equipment containing adsorption bed for chemical vapour deposition, used with diverse metal oxides.

TiO2 + -


Packaging of carbon black carbon black + -


Cutting of substrates coated by carbon black particles carbon black + -


Grinding of NanoTiO2 TiO2 + -


Maintenance of device polluted with NP with glassbead cabinets

Nano Ag + -


Maintenance of physical vapour deposition (PVD)

Nano Ag + -


Production of TiO2 by laser pyrolysis

TiO2 + -


weighing of CNT CNT + PROC 3


CNT in solution CNT + PROC 5


Pouring of CNT CNT + PROC 3


Handling small quantities of CNT CNT + PROC 8a


Preparation of CNT pellets from CNT powder

CNT + PROC 3

The results from the comparison of the model outcomes with actual workplace particle number concentration data (obtained by CPC) are presented in Table 6 and plotted in figures 9 -12.

Table 6 Pearson correlation coefficients for the ECETOC TRA model and Stoffenmanager

CPC results Absolute difference activity minus non-activity


Stoffenmanager Inhalable dust (mg/m³) N=15 0.787 0.866 ECETOC exposure level ( mg/m³ ) N=5 0.346 0.346

.

45

Figure 9 Stoffenmanager exposure estimate (mg/m³) versus absolute difference (GM concentration task minus GM concentration background) measured by SMPS.

y = 0,9184x + 0,4534

R2 = 0,7516

0

2

4

6

8

10

12

0 2 4 6 8 10

Stoffenmanager

Inhalable dust (mg/m3)

CP

C r

esu

lts

Ra

tio

ta

sk v

s b

ack

gro

un

d

Figure 10 Absolute difference activity minus non-activity for Stoffenmanager

y = 6834,6x - 3180,4

R2 = 0,6205

-10000

0

10000

20000

30000

40000

50000

60000

70000

80000

0 2 4 6 8 10

Stoffenmanager

Inhalable dust (mg/m3)

CP

C r

esu

lts

Ab

solu

te d

iffe

ren

ce a

ctiv

ity

min

us

no

n-

act

ivit

y

46

Figure 11 Ratio task versus background for ECETOC

y = -0,0194x + 1,0831

R2 = 0,1171

0

1

2

3

0 1 2 3 4 5 6

ECETOC exposure level (mg/m3)

CP

C r

esu

lts

Rati

o t

ask

vs

back

gro

un

d

Figure 12 ECETOC TRA exposure estimate (mg/m³) versus absolute difference (GM concentration task minus GM concentration background) measured by SMPS.

y = -1548,1x + 6647,3

R2 = 0,1171

0

5000

10000

15000

20000

25000

30000

0 1 2 3 4 5 6

ECETOC exposure level (mg/m3)

CP

C r

esu

lts

Ab

solu

te d

iffe

ren

ce a

ctiv

ity

min

us

no

n-a

ctiv

ity

47

6.4.4 Analysis of the performance check

The results of the performance check showed that the correlations between the model estimates and the dataset from the NanoINNOV/ NanoSafe and NANOSH were poor. In order to find possible explanations for these poor correlations a more in-depth analysis was performed analyzing the data and the parameters underlying the exposure models. This analysis was limited to the NANOSH dataset only since it was expected that possible explanations would also be valid for the NanoINNOV/ NanoSafe data.

ECETOC TRA Theoretically, 30 different combinations were possible for the ECETOC TRA assessments, based on the available scenarios. However, some of these combinations resulted in the same exposure estimate. As a result, only 13 exposure values were obtained, as shown in figure 13, ranging from 0.01 - 50 mg/m³.

Figure 13 Overview possible estimated exposure values

Only eight different PROC could be assigned to the measurement data and the data used was not evenly distributed over these various PROCs (see figure 14). Moreover, low dustiness was applicable for only eight of the measured situations and these were again not evenly distributed over the different PROCs (see figure 15).

Assignment of low dustiness resulted both from solid articles containing nanoparticles and from solid particles dispersed in a liquid, because ECETOC TRA does not have a specific activity class for the latter.

48

Figure 14 Distribution of PROCs assigned to the measurement data

PROC1

PROC14

PROC5

PROC8a

PROC8b

PROC9

PROC15

PROC21

Figure 15 Distribution of dustiness category over the PROCs

dustiness distribution

0

2

4

6

8

10

12

14

PROC15 PROC14 PROC5 PROC21 PROC1 PROC8a PROC8b PROC9

num

ber of

mes

uram

ents

low dustiness

high dustiness

Local Exhaust ventilation (LEV) however was evenly distributed over the 46 assessed situations with 43% of the total measurements having LEV. As can be seen in figure 16, many PROCs contained both LEV and no LEV.

49

Figure 16 Distribution of LEV over the 46 measurements

LEV distribution

0

2

4

6

8

10

12

14

PROC1 PROC14 PROC15 PROC21 proc5 PROC8a PROC8b PROC9

num

ber

of m

easu

rmen

ts

With LEV

Without LEV

As shown previously, CPC and SMPS results were not available for all scenarios. The results from only 13 CPC and 23 SMPS measurements could be used for the performance check.

As can be seen in figures 17 and 18, the CPC results were slightly better distributed over the PROCS than the SMPS results. However, the measurements with the CPC only contained the high dustiness category while, for the SMPS results, the low dustiness category was present (see figure 19). The distribution of the LEV category was not distributed evenly over the PROCs for SMPS or CPC results (figures 20 and 21).

The performance check using the 23 SMPS resulted in six different outcomes in ECETOC TRA, ranging from 0.01 - 25 mg/m³. The performance check using the 13 CPC results in ECETOC TRA resulted in eight different outcomes ranging from 0.01 - 50 mg/m³. So, even though the CPC had a slightly better distribution, the number of data points was very limited. The number of measurement points for relative ranking was thus somewhat better with the SMPS where the distribution over the tool variables was much more limited, hence there were only eight different outcomes for 23 assessments.

Figure 17 Distribution of SMPS results over the PROCs

PROC1

PROC14

PROC5

PROC8a

PROC8b

PROC9

PROC15

PROC21

50

Figure 18 Distribution of CPC results over the PROCs

PROC1

PROC14

PROC5

PROC8a

PROC8b

PROC9

PROC15

PROC21

Figure 19 Distribution of dustiness category in SMPS

Dustiness distribution SMPS

0

2

4

6

8

10

12

PROC1 PROC14 PROC8a PROC9 PROC15

num

ber o

f mea

rem

ents

low dustiness

high dustiness

Figure 20 Distribution of LEV category in SMPS

LEV distribution SMPS

0

2

4

6

8

10

12

14

16

PROC1 PROC14 PROC8a PROC9 PROC15

num

ber

of m

easu

rem

ents

with LEV

Without LEV

51

Figure 21 Distribution LEV category CPC

LEV distribution CPC

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

PROC1 PROC14 PROC9 PROC15 PROC8a

num

ber

of m

easu

rem

ents

with LEBV

without LEV

STOFFENMANAGER Theoretically, 31 different exposure scenarios were assessed with Stoffenmanager, based on the 46 sets of measurement data and the information on operation conditions and risk management measures. However, some of the possible combinations resulted in the same exposure estimate. Eventually 15 different model outcomes were obtained, ranging from 0.17 - 42.6 mg/m³. Unfortunately, no useful data were available for the high end model outcomes, so the actual range for comparison was 0.17 - 9.3 mg/m³.

The distribution over the six handling categories of the solids with high dustiness was rather more even, as can be seen in figure 22. The liquids were only distributed over two handling categories.

Figure 22 Distribution of solids with high dustiness over the handling categories

34%

3%

7%

23%

30%

3%Handling of prodcuts with a relatively highspeed/force which may lead to some dispersion ofdust

Handling of products, where due to high pressure,speed or force large quantities of dust aregenerated and dispersed.

handeling products in closed containers

Handling of products in small amounts or insituations where only low quantities of productsare likely to be released.

Handling of products in very small amounts or insituations where release is highly unlikely.

Handling of products with low speed or with littleforce in medium quantities.

General ventilation was widely documented in the different activities. As a result the distribution was biased towards the more common activities meaning that the data were

52

less good for those activities which occurred less frequently. One effect of this was to limit the total variance in outcome, resulting in relatively poor variance data. The distribution is shown in figure 23.

Figure 23 Distribution of general ventilation over the activities

Overview general ventilation

0

2

4

6

8

10

12

14

16

1 2 3 4 5 6 7 8 9

number of this actvity class

total number of this category

general ventilation

Risk management measures (RMM) were widely distributed across the activities. However, this meant that there was only very limited data from activities without RMM, thus limiting the total variation in outcome (see figure 24).

Figure 24 Distribution on RMM over the activities

53

Performance Check Ideally, for the performance check of Stoffenmanager using actual data, the data available should be sufficiently dispersed over the various combinations. However CPC and SMPS results are not available for all scenarios. In addition, where measurement results are available, the background and activity concentration is not reported in all cases. Consequently only 13 measurement results could be used for the CPC performance check and 23 for the SMPS performance check.

As can be seen in figures 25 and 26, the CPS results are very poorly distributed over the different activity classes; only 3 classes are presented and almost all of the measurements are located in one activity class. The SMPS results are slightly better; they are distributed over five different classes and are more equally distributed over the various classes. The measurements with the CPC also only contain the high dustiness category, while the SMPS measurements also describe some low dustiness scenarios. The performance check of the 23 SMPS results yielded six different outcomes using Stoffenmanager, ranging from 0.17 - 9.32 mg/m³. The performance check of the 13 CPC results using ECETOC TRA again yielded six different outcomes, ranging from 0.18 - 8.73 mg/m³. Although the SMPS has a better distribution the number of data points or model estimates for comparison is still very limited.

Figure 25 Distribution of CPC over the solid activity classes

72%

7%

0%

21%

0%







54

Figure 26 Distribution of SMPS over the activity classes

32%

5%

0%

18%

40%







As expected, the distribution of general ventilation over the scenarios measured by the two devices is not homogeneous. The distributions are shown in figures 27 and 28.

Figure 27 Distribution of general ventilation over activity classes for SMPS

0

1

2

3

4

5

6

7

8

9

10

1 2 3 4 5

# actvity class

# m

easu

rem

ents

total number of measurements

number of measurements with general ventilation

Figure 28 Distribution of general ventilation over activity classes for CPC

0

2

4

6

8

10

12

1 2 3

# activity class

# m

easu

rem

ents

total number of measurements

number of measurements with ventilation

55

As for General Ventilation, the distribution of the type RMMs over the scenarios measured by the two devices is not homogeneous (figures 29 and 30)

Figure 29 Distribution of RMMs over activity classes for SMPS

0

1

2

3

4

5

6

7

8

9

10

1 2 3 4 5

# activity class

# m

easu

rem

ents

total number measurements

number single RMM

number multiple RMM

Figure 30 Distribution of RMMs over activity classes for CPC

0

2

4

6

8

10

12

1 2 3

# activity class

# m

easu

rem

ents

total number of measurments

single RMM

multiple RMM

56

6.4.5 Discussion and conclusion

From the results given above, it can be seen that there is hardly any correlation between the model outputs and the actual concentration data. For both models, comparisons with either the NANOSH or the nanoINNOV data sets showed no substantial correlations. The relatively high correlation coefficients observed for the nanoINNOV data set is highly biased by only one data point and therefore it should be considered not to be meaningful. In addition, no difference can be observed between Stoffenmanager (activity-based) and ECETOC TRA (extrapolated for full day exposure).

The lack of any significant correlation is, most probably, largely attributable to the variability of the (relatively few) data. The variability of the particle number concentration within the Stoffenmanager and ECETOC TRA exposure scenarios is large compared to the between scenario model outputs. This would be expected to have a substantial effect on any correlation. As was shown earlier, in sections on the assessments using both models, due to lack of data or contextual information in the data sets not all variability within the model parameters could be used. This resulted in a loss of power of discernment between exposure scenarios. In addition, as been discussed for the NANOSH data set, the distribution of the variables within the model parameters was not even over the various scenarios so that, once more, the potential variability between the scenario exposure estimates was not fully exploited. Therefore it can be concluded that the data sets used to check the performance of the exposure models for nanoparticle exposure scenarios were not optimal for testing.

The particle number concentration data used to compare with the model scenario outputs was obtained from devices that measured in the size ranges 10 -1000 nm (CPC), or 6- 700 nm (SMPS). As stated before, particles within these size ranges will contribute less to mass compared to particles within larger size ranges, and it can be hypothesized that the resolution for low mass concentration of the models might be insufficient to show any correlation.

In summary it can be concluded that, theoretically, with respect to nanoparticle exposures, both ECETOC TRA and Stoffenmanager would be able to give an indication of exposure levels. Since the models are not attuned to, or calibrated for, nanomaterial exposure situations, the actual model estimates will be inaccurate and possibly overestimate the concentration levels.

Comparisons of the model estimates with actual data on the average particle concentrations measured during the nanomaterial handling activities revealed no significant correlation. This is most probably due to the large variability of the particle number concentration within the Stoffenmanager and ECETOC TRA exposure scenarios compared to the small between scenario model outputs. The latter is related to a combination of the typical scenarios that could be derived form the data sets; information that was lacking, and the inherent power of contrast of the models for nanomaterial exposure situations.

57

6.4.6 References

ECETOC, Addendum to ECETOC Targeted Risk Assessment Report No. 93, December 2009, Technical Report No. 107.

Cherrie, J.W., Tickner, J., Friar, J., Creely, K.S., Soutar, A.J., Hughson, G, Warren N.D., 2003. Evaluation and further development of the EASE model 2.0. Health and Safety Executive, 2003.

ECHA, Guidance on information requirements and chemical safety assessment; Chapter R.14: Occupational Exposure Estimation, May 2008.

REACH Guidance on information requirements and chemical safety assessment. Chapter R.12: Use descriptor system (DRAFT Version 2.0 07/11/2009)

Marquart H, Heussen H, Le Feber M, Noy D, Tielemans E, Schinkel J, West J, Van der Schaar D. (2008) ‘Stoffenmanager’, a web-based control banding tool using an exposure process model. Ann. Occup. Hyg.; 52 (6), 429, doi:10.1093/annhyg/men032

Schinkel J, Fransman W, Heussen H, Kromhout H, Marquart H, and Tielemans E. (2010) Cross-validation and refinement of the Stoffenmanager as a first tier exposure assessment tool for REACH. Occup. Environ. Med. 2010 (67), 125

Tielemans E, Noy D, Schinkel J, Heussen H, van der Schaaf D, West J, Fransman W. (2008) Stoffenmanager exposure model: development of a quantitative algorithm. Ann. Occup. Hyg., 52(6) 443, doi:10.1093/annhyg/men033

Tielemans E, Schneider T, Goede H, Tischer M, Warren N, Kromhout H, van Tongeren M, van Hemmen J and Cherrie JW. (2008) Conceptual model for assessment of inhalation, exposure: defining modifying factors. Ann. Occup. Hyg., 52(7), 577, doi:10.1093/annhyg/men059

ECETOC (2009). Worker exposure tool (version 2.0) from the targeted risk assessment of ECETOC. (http://www.ecetoc.org/tra; access date: August 28, 2009)

ECETOC (2010). Integrated tool exposure tool (version 2.0) from the targeted risk assessment of ECETOC. (http://www.ecetoc.org/tra; access date: April 30, 2010)

ECETOC, Technical Report No. 86. Derivation of assessment factors for human health risk assessment. February 2003. ISSN-0773-6347-86.

58

6.4. PDF extracts from the NANEX Exposure Scenari o Database

occupational exposure scenarios - nanex...

Documents