Detection of Carbon Nanotubes in Indoor Workplaces Using

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN

Article

Detection of Carbon Nanotubes in Indoor Workplaces Using Elemental Impurities Pat E Rasmussen, Mary-Luyza Avramescu, Innocent Jayawardene, and H. David Gardner Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b02578 • Publication Date (Web): 09 Oct 2015 Downloaded from http://pubs.acs.org on October 10, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

Environmental Science & Technology

ACS Paragon Plus Environment


1

Detection of Carbon Nanotubes in Indoor Workplaces Using Elemental

2

Impurities

3 4 5

Pat E. Rasmussen1,2*, Mary-Luyza Avramescu1, Innocent Jayawardene1, and H.

6

David Gardner1,2

7 8 9 10 11 12 13 14 15

1. 2.

Environmental Health Science and Research Bureau, HECSB, Health Canada, 50 Colombine Driveway, Tunney’s Pasture 0803C, Ottawa, Ontario, Canada, K1A 0K9 University of Ottawa, Earth Sciences Department, Ottawa, ON, Canada K1N 6N5

16 17 18

*corresponding author: Pat E. Rasmussen; Environmental Health Science and Research

19

Bureau, HECSB, Health Canada, 50 Colombine Driveway, Tunney’s Pasture 0803C,

20

Ottawa, Ontario, Canada, K1A 0K9; phone: (613) 868-8609; fax: (613) 952-8133; email:

21

[email protected]

22 23 24 25

1


Page 2 of 37

Page 3 of 37


26

ABSTRACT

27 28 29

This study investigated three area sampling approaches for using metal impurities in

30

carbon nanotubes (CNTs) to identify CNT releases in workplace environments: air

31

concentrations (µg/m3), surface loadings (µg/cm2), and passive deposition rates

32

(µg/m2/hr). Correlations between metal impurities and CNTs were evaluated by

33

collecting simultaneous co-located area samples for thermal-optical analysis (for CNTs)

34

and ICP-MS analysis (for metals) in a CNT manufacturing facility. CNTs correlated

35

strongly with Co (residual catalyst) and Ni (impurity) in floor surface loadings, and with

36

Co in passive deposition samples. Interpretation of elemental ratios (Co/Fe) assisted in

37

distinguishing amongst CNT and non-CNT sources of contamination. Stable isotopes of

38

Pb impurities were useful for identifying aerosolized CNTs in the workplace environment

39

of a downstream user, as CNTs from different manufacturers each had distinctive Pb

40

isotope signatures. Pb isotopes were not useful for identifying CNT releases within a

41

CNT manufacturing environment, however, because the CNT signature reflected the

42

indoor background signature. CNT manufacturing companies and downstream users of

43

CNTs will benefit from the availability of alternative and complementary strategies for

44

identifying the presence/absence of CNTs in the workplace and for monitoring the

45

effectiveness of control measures.

46 47

2



48

INTRODUCTION

49

Distinguishing releases of engineered nanomaterials from background aerosols (arising

50

from natural sources or incidental sources such as welding fumes or vehicle exhaust) is a

51

major challenge that must be addressed to improve assessments of risk associated with

52

emerging nanotechnologies.1,2 Direct reading instruments used for monitoring airborne

53

nanoparticles are non-specific and do not have the capacity to distinguish engineered

54

nanomaterials from background aerosols.3-5 In the case of carbon nanotubes (CNTs),

55

determination of elemental carbon is recommended6 for quantification of occupational

56

exposures to CNTs (using NIOSH Method 5040 or an equivalent method). Difficulties

57

arise in distinguishing CNTs where the background aerosol is also largely composed of

58

elemental carbon (e.g., diesel soot), requiring additional off-line analytical techniques

59

such as transmission electron microscopy (TEM) and scanning electron microscopy with

60

energy dispersive X-Ray spectroscopy (SEM-EDX) to be used in combination with direct

61

reading and elemental carbon approaches to verify the presence of CNTs in workplace

62

environments.6-8

63 64

Several studies have employed metal catalyst impurities (imbedded in CNTs during the

65

manufacturing process) as an index or proxy for CNTs using inductively coupled plasma

66

atomic emission spectroscopy (ICP-AES),9,10 inductively coupled plasma mass

67

spectroscopy (ICP-MS),11-13 or emerging instrumental approaches.14-16 A survey of CNTs

68

produced by different manufacturers indicated that most catalyst residues are transition

69

metals, such as Fe, Ni, Mo, Y, Co and Cr, and that other unexpected impurity elements

70

(including As, Gd, W, Yb, and Sm) also occur in CNTs due to large-scale production 3


Page 4 of 37

Page 5 of 37


71

procedures, post fabrication and post-purification treatments.17 That survey, which used

72

neutron activation analysis, found that metal impurities in CNTs contribute between 0.44

73

and 3 wt%, even after purification.17 This important observation opens up the possibility

74

of using metal impurities to detect CNT releases to the environment. For example,

75

Schierz et al.13 used the catalyst elements Mo and Co to fingerprint CNT occurrence in

76

sediment cores after a simulated spill into an outdoor wetland mesocosm. Olson et al.12

77

used filter-based air sampling followed by ICP-MS analysis to quantify a variety of water

78

soluble metals (Y, Ti, Fe, Cu, As, Zn, Cd, Pb, and Ag) in CNT materials that were

79

aerosolized in a controlled environmental chamber, and reported concentrations similar to

80

those observed in urban ambient PM2.5 samples on a mass per mass basis, with the

81

exception of Y which was higher in the studied CNTs than in ambient air samples.

82 83

Technologies to identify CNT releases are especially important for monitoring the

84

effectiveness of control measures to reduce worker exposure across common process

85

tasks.18 Maynard et al.9 used Fe and Ni catalyst impurities as a CNT index to investigate

86

airborne and dermal exposures to SWCNT in production environments. Birch et al.5

87

investigated spatial correlations of carbon nanofibers (CNFs) and Fe catalyst impurities in

88

manufacturing facilities using air filter samples, and reported that the Fe catalyst was not

89

a useful CNF proxy due to interferences from non-CNF sources of Fe in the facility.

90 91

The present study investigates the potential usefulness of other CNT impurities (such as

92

Co, Mo, Ni and Y) which are less prevalent than Fe in the background environment, for

93

identifying the presence/absence of CNTs in the workplace and for monitoring the 4



94

effectiveness of control measures. Several alternative sampling strategies are presented,

95

as some approaches are more suited to certain workplaces, such as CNT production

96

facilities versus research laboratory environments. The concept of using metal impurities

97

to monitor CNTs in a production facility is evaluated by examining correlations between

98

metals and CNTs in co-located surface wipe samples, passive deposition samples and air

99

samples. As an alternative to filter-based air sampling methods, a wet electrostatic

100

precipitation method is investigated in a laboratory environment for its capacity to

101

capture transient airborne Pb isotopic signatures of aerosolized CNTs. As metal

102

impurities in themselves present a potential health hazard, and some studies indicate that

103

elemental carbon and metal particles together produce oxidative effects greater than

104

either particle type alone,5,19 the alternative strategies presented in this study add breadth

105

to existing sampling strategies used for CNT exposure and risk assessments.

106 107

EXPERIMENTAL

108 109

Reference Materials and Sampling Methods

110

Two standard reference materials (SRMs) for CNTs were available with certified metal

111

concentrations in the relevant range for evaluating metal impurity recoveries: NIST 2483

112

Single Wall CNT (SWCNT; from National Institute of Standards and Technology,

113

Gaithersburg, MD, USA), and NRC Certified Reference Material SWCNT-1 (from

114

National Research Council, Canada). Additionally two non-SRM CNT test materials

115

were obtained for laboratory experiments: SWCNT from Sigma-Aldrich Co. (Gillingham,

116

UK no. 698695 MKBB3788) which has been characterized previously,11 and SWCNT 5


Page 6 of 37

Page 7 of 37


117

research material (NRC Test) provided by National Research Council, Canada. Further

118

characterization of the NRC SRM, Aldrich CNT, and NRC Test material using SEM and

119

TEM is provided as Supporting Information (Table S-1 and Figures S-1 to S-4).

120 121

In a research laboratory setting, aerosolized CNTs were collected using wet electrostatic

122

precipitation under controlled environmental conditions detailed previously.20 An

123

Aerosol-to-Liquid Particle Extraction System (ALPXS; Meinhard Glass Products,

124

Golden, CO) was used to capture aerosolized particles of Aldrich CNT, NIST 2483 SRM,

125

and NIST 1648a (Urban Particulate Matter) for stable Pb isotope determination using

126

quadropole ICP-MS. The particulate matter was aerosolized using a mechanical shaker

127

with an ALPXS sampling duration of 30 min (300 L/min flow rate).20 Background air

128

was sampled using the ALPXS (2 hr duration) in two laboratory locations before

129

aerosolization experiments were initiated.

130 131

Area sampling was conducted in five areas of a CNT manufacturing environment: the

132

vicinity of the reactor, the catalyst preparation area, an administrative office, the control

133

room, and the packaging area. In the preliminary mapping survey, five surface wipe

134

samples were collected from a variety of floor and unused shelf surfaces in each of the

135

first three locations. GhostWipes (Environmental Express, Charleston, South Carolina,

136

SC 4250; pre-moistened with deionized water in individually sealed packets) were used

137

for wipe sampling in the preliminary mapping. GhostWipes, which are composed of

138

polyvinyl alcohol copolymer, meet the ASTM E-1792 and OSHA Method ID-125G wipe

139

sampling protocols, and have been used for monitoring lead and other metals.21 To 6



140

determine surface loading (in µg/cm2), wipe samples were collected using a mechanical

141

device designed by TNO (Utrecht, Netherlands) that systematically samples a surface

142

area of 22 cm2 by moving a 9.6 cm2 disc of wipe material in a back-and-forth motion

143

using consistent pressure.

144 145

In the follow-up survey of the CNT manufacturing facility, duplicate samples were

146

collected in each location: one for ICP-MS and the other for thermo-optical carbon

147

determination. Filter-based area air samples were collected using cassette samplers

148

loaded with quartz fiber filters (Whatman 25 mm, Cat No 1851-025) at all five locations

149

within the facility (2 L/min flow rate; 4-6 hr sampling duration in work process sites; 8 hr

150

in the administrative area used as background ). Co-located wipe samples (using 25 mm

151

quartz fiber filters and 9.6 cm2 Ghost Wipe discs) were collected from the floor and from

152

shelf surfaces not used by workers, at four out of the five air sampling locations (i.e. all

153

but the control room). To determine deposition rates (in µg/cm2/hour), passive

154

accumulation samples were collected by laying out a GhostWipe (15cm x 15cm) in a

155

large pre-cleaned petri dish (for each sample, 2 consecutive wipes were exposed for a

156

total of 40 h 15 min and combined for digestion and analysis) at each of the four locations

157

within the facility where surface wipe samples were collected. The preliminary mapping

158

was conducted during downtime (for metals only) and the follow-up sampling (for metals

159

and CNTs) was conducted one month later while the facility was in production phase.

160

Samples were shipped to Health Canada (Ottawa, Canada) for metal determination and to

161

TNO (Utrecht, Netherlands) for carbon determination.

162 7


Page 8 of 37

Page 9 of 37


163

Sample Preparation and Elemental Analysis

164

A combined hot-block/microwave (HBMW) digestion method with nitric acid and

165

hydrogen peroxide22 was used to extract metals from the air filter samples, wipe samples

166

and deposition samples collected from the manufacturing site. Recoveries of 87% and

167

110% respectively were obtained for Co and Mo in NIST 2483 combined with

168

GhostWipes, as described previously.22 For extraction of the quartz filter samples, an

169

additional washing step (with de-ionized water) was added at the end of the HBMW

170

method, as advised by USEPA.23 A NexION 300s ICP-MS with Dual-channel Universal

171

Cell (Perkin Elmer, Canada) equipped with a SC-Fast autosampler (Elemental Scientific,

172

Omaha, NE), a high temperature apex-ST PFA MicroFlow nebulizer, cyclonic spray

173

chamber and a PC3x chiller operated at 2oC, and a triple cone interface (nickel-platinum

174

skimmer and sampler cones, and aluminium hyper cone) was operated in the standard

175

mode for all elements except Fe for which reaction mode was used. A preliminary semi-

176

quantitative scan of the digests identified Ge, In and Re as appropriate internal standards.

177

A short ultrasonic digestion method24 with a strong acid mixture (HNO3-HF) was used

178

for the preliminary mapping samples.

179 180

ALPXS samples were preconcentrated using hot block evaporation, then digested using a

181

nitric acid/ultrasonic bath digestion procedure described previously.20 Pb isotopic ratios

182

were determined using an Elan DRC II-6100 Inductively Coupled Plasma Mass

183

Spectrometer (ICP-MS; Perkin Elmer, Woodbridge, ON, Canada). For each run, the

184

NIST 981 Common Lead Isotopic Standard (National Institute of Standards and

185

Technology, Gaithersburg, MD, USA) was determined after every 5 samples for quality 8



186

assurance (observed values for 3 reported isotopes in NIST 981 were within 0.3% of

187

certified values).

188 189

Sample pre-treatment for thermal-optical carbon analysis consisted of dissolving the

190

GhostWipe sample in a 1:1 nitric acid/water mixture and filtering the dispersion onto a

191

pre-fired (900 °C) quartz fiber filter with a vacuum filtration system, followed by rinsing

192

with de-ionized water. A 1 cm2 sample of each quartz filter was analyzed for elemental

193

carbon (EC) and organic carbon (OC) using a thermal/optical carbon monitor (Sunset

194

Laboratory Inc., USA) according to NIOSH 5040.6 A modified IMPROVE protocol was

195

used for the temperature and atmospheric gas settings.25 OC was removed from the filter

196

in the temperature range of 120-550 °C in a non-oxidizing carrier gas (helium). EC was

197

then removed in the temperature range of 550-920 °C in a mixture of helium and 2%

198

oxygen (2% O2/He). The resulting CO2 was then converted to methane and detected by

199

flame ionisation detection (FID). Correction for pyrolysis of OC to EC was carried out by

200

measurement of light transmission. EC was categorized into EC1 (550 °C), EC2 (650 °C)

201

and EC3 (920 °C) according to the oxidized temperature. The sum of EC2 and EC3 was

202

used for a quantitative estimate of the CNT concentration. Doudrick et al26 showed that

203

thermal analytical methods for measuring CNTs require customized optimization for

204

different environmental matrices; therefore, the use of a single protocol for different

205

matrices (filters and wipes) may affect CNT recoveries in one matrix relative to the other.

206

Although this factor does not impact the trends reported in this study (due to consistent

207

recoveries within each matrix type), future research is needed to optimize thermal carbon

208

analyses for the various sampling substrates that may be used for CNT collection. 9


Page 10 of 37

Page 11 of 37


209 210

Quality control

211

High purity acids (SEASTAR Chemicals Inc., Sidney BC, UN2031, CAS 7697-37-2) and

212

ultrapure Milli-Q water (18.2 MΩ cm) were used for preparation of samples and

213

standards. High purity standard stock solutions (Delta Scientific Laboratory Products

214

Ltd., Mississauga, ON) were used to prepare the calibration and internal standard

215

solutions. To guard against sample contamination, powder-free nitrile gloves were worn

216

and changed frequently (at least in every new sampling location, if not after every

217

sample) and samples were double-bagged before shipment. The sampling matrices used

218

in this study (GhostWipes and quartz fiber filters) contained variable concentrations of

219

contaminant elements introduced during their manufacturing and packaging processes;

220

therefore, 3-8 matrix blanks (filters or wipes as appropriate) were included in every

221

analytical batch to enable the calculation of matrix blank corrections. See Supporting

222

Information (Tables S-2 to S-4) for ICP-MS quality control details including LODs,

223

reagent blanks and matrix blanks). Field blanks were collected by subjecting the sampling

224

medium (wipe or filter) to all the same steps as a sample, minus the actual sample

225

collection. Matrix blanks were subtracted from all samples and field blanks, but field

226

blanks were examined separately and were not subtracted from samples.

227 228

RESULTS AND DISCUSSION

229 230

Selection of CNT impurities as candidate elemental tracers

10



Page 12 of 37

231

Determining whether elemental impurities in CNTs are likely to be useful as

232

environmental tracers requires preliminary knowledge of concentrations of those

233

elements in the surrounding environment arising from background sources (i.e. any

234

natural or anthropogenic sources other than the CNTs being investigated).

235

illustrated in Figure 1 which compares background elemental signatures (represented by

236

global soil averages27) against those of two CNT standard reference materials (SRMs;

237

Figure 1a) and two test CNT materials supplied without elemental concentration

238

information (unknowns; Figure 1b), as is commonly experienced by downstream users of

239

CNTs. Elemental concentrations in Figure 1 were determined using ICP-MS

240

determinations of digested subsamples of all four CNT materials, three of which were

241

also characterized using SEM and TEM (SI Table S-1 and Figures S-1 to S-4).

This is

242 243

Identification of potential elemental tracers is enabled by the horizontal lines in Figure 1,

244

each representing an order-of-magnitude increase in concentration (log scale). Co, Mo

245

and Ni in the NRC SRM and Co and Mo in the NIST SRM (Figure 1a) emerge as

246

potentially useful tracers, as their concentrations in the CNTs exceed global background

247

by 2 to 3 orders of magnitude. With respect to the two unknowns (Figure 1b) Ni, Y and

248

Co in the NRC Test material, and Ni and Y in the Aldrich material, emerge as likely

249

candidates as they exceed global background by more than 2 orders of magnitude.

250 251

Note that the reverse is true for Fe and Al: despite having elevated concentrations in all

252

four CNTs (100-1000 µg/g; Figure 1), the global background averages for these elements

253

are orders of magnitude higher (about 3.5% for Fe and 8% for Al27), making them less 11


Page 13 of 37


254

useful as potential CNT tracers. A recent study19 indicates that the above trends hold for

255

CNTs that are significant in commercial applications: impurity concentrations (µg/g) tend

256

to be higher for Ni (GM 816; GSD 15; max 11,757) than for Fe (GM 217; GSD 3.1; max

257

2008) in 21 commercially relevant multi-wall CNTs (MWCNTs). Current usage of other

258

metal catalysts is indicated by their maximum values (µg/g) for this set of MWCNTs (Co

259

5656; Cu 5830; Mn 2236; Mo 1863) and for seven commercially relevant SWCNTs (Co

260

3225; Cr 1931; Fe 1408; Mo 1442).19

261 262

Preliminary mapping of a CNT manufacturing environment

263

Ramachandran et al.28 recommended concentration mapping as a quantitative tool to

264

investigate spatial and temporal variability and identify contaminant sources in

265

nanotechnology workplaces, or as a pre-survey tool to determine optimal sampling

266

locations for subsequent measurements. Although indoor environments may be relatively

267

free of outdoor soil and dirt, contributions of metals from indoor sources can significantly

268

interfere with the use of CNT metal impurities as tracers. Thus the first goal of the

269

present study was to investigate background levels of alternative elemental tracers such

270

as Co, Ni and Mo in a CNT manufacturing facility, by comparing surface loadings of the

271

candidate elements in an administrative room remote from the CNT operations (as the

272

reference indoor background site) and two work process sites (around the catalyst

273

preparation area and around the reactor where the CNTs are produced). The catalyst used

274

for CNT production in this facility was Co, with CNT concentrations of Co within the

275

range reported for typical SWCNT and MWCNT products currently in commerce

12



276

(

Detection of Carbon Nanotubes in Indoor Workplaces Using

Recommend Documents