Predicting Dermal Exposure to Gas-Phase Semivolatile Organic

Subscriber access provided by TUFTS UNIV

Environmental Modeling

Predicting dermal exposure to gas-phase semi-volatile organic compounds (SVOCs): an improved description of SVOC mass transfer between clothing and skin surface lipids Jianping Cao, Xu Zhang, and Yinping Zhang Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b06485 • Publication Date (Web): 15 Mar 2018 Downloaded from http://pubs.acs.org on March 16, 2018

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 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 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.

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 26

Environmental Science & Technology

1

Predicting dermal exposure to gas-phase semi-volatile organic

2

compounds (SVOCs): an improved description of SVOC mass transfer

3

between clothing and skin surface lipids

4 5

Jianping Cao1,2, Xu Zhang1,2, Yinping Zhang1,2,*

6 7

1

Department of Building Science, Tsinghua University, Beijing 100084, China

8

2

Beijing Key Laboratory of Indoor Air Quality Evaluation and Control, Beijing 100084,

9

China

10

*

Corresponding e-mail: [email protected]

11 12

Abstract

13

Dermal exposure to indoor gas-phase semi-volatile organic compounds (SVOCs)

14

has recently received a great deal of attention, and this has included evaluating the role of

15

clothing in this process. Several models have been developed to assess dermal exposure

16

to SVOCs, based on the transient mass transfer of SVOCs from air to dermal capillaries.

17

However, these models assume either that clothing completely retards SVOC transport,

18

or that there is an air gap of constant thickness between the clothing and the surface of

19

the skin, leading to significant errors in the model calculations. To solve this problem, we

20

have improved the description of SVOC transport between clothing and epidermis by

21

considering two parallel processes: partitioning of SVOCs by direct contact (ignored in

22

existing models), and Fickian diffusion through the air gap. Predictions from this

23

improved model agree well with the experimental data found in the literature (dermal

24

uptake of diethyl phthalate (DEP) and di-n-butyl phthalate (DnBP) of a clothed

25

participant). This study provides a useful tool to accurately assess dermal exposure to

26

indoor SVOCs, especially for evaluating the effects of clothing on dermal exposure.

27

1

ACS Paragon Plus Environment


28

1. Introduction

29

Dermal exposure to gas-phase semi-volatile organic compounds (SVOCs) in the

30

indoor environment has been mostly overlooked, and in some cases ignored, in existing

31

exposure assessments of indoor SVOCs.1, 2 More recently, dermal exposure has gradually

32

been getting more attention.1-11 Several studies have demonstrated that human uptake via

33

dermal exposure is comparable to, or may even exceed, intake via inhalation for some

34

SVOCs.2-4, 8, 12 This indicates that dermal exposure to gas-phase SVOCs is an important

35

exposure pathway, that should be more thoroughly understood and more accurately

36

described.

37

Mathematical models are generally helpful, and more cost effective than

38

experimental measurements, when assessing human exposure to gas-phase SVOCs and

39

understanding the underlying mechanisms.13, 14 In the simplest approach, many studies

40

use a steady-state mass transfer model to estimate dermal exposure to gas-phase

41

SVOCs.1,

42

significant errors into the estimated exposure levels. This is because realistic exposure

43

conditions vary, and the time required to reach steady state (in common exposure

44

scenarios) is often longer than the time spent in a steady-state environment.4, 13, 17 For this

45

reason, Gong et al.13 developed a transient-state mass transfer model, which takes into

46

consideration the convective mass transfer resistance from air to the epidermis, and the

47

Fickian diffusion through the stratum corneum and viable epidermis. Gong et al.’s model

48

can predict the uptake of gas-phase SVOCs by bare skin for varying exposure conditions.

49

This model agrees much better with the experimental data (forearms and hands of

50

volunteers were exposed to m-xylene vapor for 20, 45, 120, or 180 min, and dermal

51

absorption was monitored for 8 hours

52

predictions of steady-state model differed from measured data by factors as large as 1.9.13

53

When the skin surface is covered by clothing, however, the estimates of Gong et al.’s

54

model significantly differ from the experimental data because the model assumes that

55

clothing completely retards the transport of gas-phase SVOCs.19 This assumption has

56

been experimentally proved to be invalid, e.g., the clothing can either increase (if

57

clothing has sorbed sufficient SVOCs prior to being worn) or decrease (if clothing is free

58

of SVOCs) dermal exposure to gas-phase SVOCs.6, 8, 9, 20, 21 Additionally, Gong et al.’s

14-16

However, the basic assumptions used in this model may introduce

18

) than does the steady-state model, i.e.,

2


Page 2 of 26

Page 3 of 26


59

model ignores the mass-transfer resistance provided by the skin surface lipids (SSL),

60

leading to an overestimation of dermal exposure.19

61

To more accurately estimate dermal exposure to gas-phase SVOCs, especially

62

considering the effect of clothing on dermal exposure, Morrison et al. modified Gong et

63

al.’s model by taking into account a layer of SSL, a layer of clothing, and the air gap

64

between clothing and the SSL (as shown in Figure 1 (a)).19,

65

modified model were consistent with the measured results for bare skin exposure 19 (six

66

unclothed participants were exposure to gas-phase diethyl phthalate (DEP) and di-n-butyl

67

phthalate (DnBP) for 6 hours, and the dermal uptake amount was monitored for 54 hours

68

from the start of the exposure 4). For clothed skin, the assumptions of uniform

69

distribution of SVOCs within clothing, and instantaneous equilibrium between clothing

70

and gas-phase SVOCs, led to significant overestimates for dermal exposure (for the

71

participant wore SVOC-contaminated clothes, the measured and modeled total uptake

72

values for DEP were 4.7 mg and 7.4 mg, respectively; for DnBP, the measured and

73

modelled values were 3.3 mg and 19 mg, respectively).19

22

Estimates from the

74 Figure 1. Schematic illustrating layers and parameters in: (a) dermal exposure model introduced by Morrison et al.22 (ignoring direct contact between clothing and skin surface lipids), and (b) dermal exposure model modified in this study.

75 76

Morrison et al.’s model assumes clothing to be a flat plate, and also that there is an

77

air gap with constant thickness between the clothing and the SSL. This is of course not

78

realistic since clothing does directly contact the skin due to human body movement,

79

gravity, the irregular shape of the human body, and random creases in clothes.23-27 The

80

total area of direct contact is significant compared to the total area of skin covered (e.g.,

81

up to 60% at upper chest and upper back 26). The SVOC transport rate from clothing to

82

SSL through direct contact is much faster than that through an air gap (see details in

83

Figure 2 of Morrison et al.22, where Gap = 0 represents direct contact). In addition, the

84

thickness of the air gap and the area of direct contact vary across the body.26, 27 As a

85

result, simply ignoring direct contact between clothing and the SSL, and assuming that

86

the thickness of the air gap is constant introduce significant errors to the estimated SVOC

87

dermal exposure. Furthermore, it should be noted that the size of the air gap used in 3



88

Morrison et al.22 is, in fact, an “equivalent” thickness, which they determined by varying

89

the thickness used in the model until the predicted results fit best with the experimental

90

results.22 In other words, the value of the “equivalent” air gap thickness was inversely

91

obtained based on measured results, which may be different for other exposure scenarios

92

(e.g., the exposure duration, clothing properties, or the target SVOCs).

93

The objective of this study is therefore to modify Morrison et al.’s model 22 to more

94

accurately predict dermal exposure to gas-phase SVOCs, by considering the parallel

95

transport of SVOCs from clothing to the SSL, partitioning of SVOCs by direct contact,

96

and the Fickian diffusion to the SSL through the air gap. The experimental results

97

collected by Morrison et al.20 are used to evaluate the reliability and accuracy of the

98

refined model.

99 100

2. Methods

101

2.1 Brief introduction of Morrison et al.’s model

102

Apart from the problem of the “equivalent” thickness of the air gap in Morrison et 22

103

al.’s model,

104

models. As shown in Figure 1 (a), the transport route for clothed skin of SVOCs from the

105

air to dermal capillaries is divided into 5 layers (left part of the figure): 1) clothing; 2) air

106

gap between clothing and the SSL; 3) SSL; 4) stratum corneum (SC); and 5) viable

107

epidermis (VE). For bare skin, only the last three layers (i.e., SSL, SC, and VE) are

108

considered. Morrison et al.’s model, as well as the other existing transient models, make

109

the following assumptions:22

their model provides the best accuracy when compared to other existing

110

1) There is a dermis layer below the VE layer. The SC and VE together constitute the

111

epidermis.28, 29 Given that there are abundant capillaries at the VE-dermis interface 28, the

112

SVOC concentration in the blood at the VE-dermis interface is considered to be zero.

113

This means that there is no need to consider SVOC transport below the VE-dermis

114

interface when modeling. This assumption has been shown to be valid for many

115

compounds.30

116

2) All five layers are homogeneous and isotropic media, and Fickian diffusion

117

occurs in each layer; and the SVOC transport from air to dermal capillaries is one-

118

dimensional. 4


Page 4 of 26

Page 5 of 26


119

3) The effect of SVOC transport through parts of the skin, such as hair follicles and

120

sweat glands is ignored, as is the transfer of skin surface lipids to clothing,

121

metabolization, desquamation, and ionization of SVOCs within the skin.

122 123 124 125 126 127

128

4) Convective mass-transfer coefficients above the bare skin and above the clothing surface are identical. 5) At the interfaces between each two contiguous layers, instantaneous equilibrium partitioning exists and the flux of SVOCs is conserved. With these assumptions in mind, when the skin is covered by clothing, the governing equations describing the SVOC transport process are:

 ∂Cve ∂ 2Cve = Dve , 0 < x < Lve  ∂x 2  ∂t  ∂Csc ∂ 2Csc D = , Lve < x < Lve + Lsc sc  ∂t 2 ∂ x  ∂ 2Cssl  ∂Cssl = Dssl , Lve + Lsc < x < Lve + Lsc + Lssl  ∂x 2  ∂t  ∂Cag ∂ 2Cag = D , Lve + Lsc + Lssl < x < Lve + Lsc + Lssl + Lag  ag 2 ∂ t ∂ x   ∂Ccl ∂ 2Ccl = Dcl , Lve + Lsc + Lssl + Lag < x < Lve + Lsc + Lssl + Lag + Lcl  ∂x 2  ∂t

(1)

129

where the subscripts ve, sc, ssl, ag, and cl represent the layers of viable epidermis (VE),

130

stratum corneum (SC), skin surface lipids (SSL), air gap, and clothing, respectively; C is

131

the SVOC concentration in the corresponding layer, µg/m3; D is the SVOC diffusion

132

coefficient in the corresponding layer, m2/s; L is the thickness of the corresponding layer,

133

m; t is time, s; and x is the distance to the VE-dermis interface, m (i.e., x = 0 at the VE-

134

dermis interface).

135 136 137

Given that SVOCs may have arbitrary distributions in each layer, the initial conditions are,

Cve = Cve,0 ( x ) , Csc = Csc ,0 ( x ) , Cssl = Cssl ,0 ( x ) , Cag = Cag ,0 ( x ) , Ccl = Ccl ,0 ( x ) , t = 0

(2)

138

where “0” in the subscripts represents the initial state, e.g., Cve,0(x) is the initial

139

concentration in the viable epidermis (µg/m3).

140

According to assumption 1), the boundary condition at the VE-dermis interface is, 5



C ve = 0, x = 0

141

Page 6 of 26

(3)

142

Gas-phase SVOCs transport to clothing through the SVOC concentration boundary

143

layer adjacent to the upper surface of clothing, indicating the following boundary

144

condition,

Dcl

145

 C  ∂Ccl = hm  Ca − cl  , x = Lve + Lsc + Lssl + Lag + Lcl ∂x K cl  

(4)

146

where Ca is the gas-phase SVOC concentration in the indoor air, µg/m3; and hm is the

147

convective mass-transfer coefficient, m/s, which is associated with the SVOC

148

concentration boundary layer above the clothing surface and can be determined using the

149

methods described in Gong et al.13

150 151

152

According to assumption 5), the boundary conditions at the interfaces of every two contiguous layers are as follows,

∂Cve ∂Csc  Cve Csc  K = K , Dve ∂x = Dsc ∂x , x = Lve sc  ve  Csc Cssl ∂Csc ∂Cssl  K = K , Dsc ∂x = Dssl ∂x , x = Lve + Lsc ssl  sc   Cssl = C , D ∂Cssl = D ∂Cag , x = L + L + L ag ssl ag ve sc ssl  K ssl ∂x ∂x  C = Ccl , D ∂Cag = D ∂Ccl , x = L + L + L + L ag cl ve sc ssl ag  ag K ∂x ∂x cl 

(5)

153

where K is the partition coefficient between the corresponding layer and air,

154

dimensionless.

155 156

The mass of SVOCs that transfer through the VE-dermis interface equals the exposure dose of SVOCs, which means that:13, 19 t

∂Cve,clothed

0

∂x

DEclothed = SAclothed ∫ Dve

157

dt

(6)

x =0

158

where DE is the exposure dose of SVOCs in the duration of 0 ~ t, µg; SA is the skin

159

surface area, m2; the subscript clothed represents the clothing covered part of the body;

160

and Cve,clothed or

161

∂Cve,clothed ∂x

can be obtained by solving equations (1)-(5). x =0

For bare skin, the clothing and air gap are non-existent (see the right part of Figure 1 6


Page 7 of 26


162

(a)). That means that equations related to Cag and Ccl in equations (1), (2) and (5) should

163

be eliminated. In addition, the gas-phase SVOCs will directly transport to the SSL

164

through the boundary layer adjacent to the upper surface of the SSL, i.e.,

Dssl

165 166 167

(7)

Similar to equation (6), the exposure dose of SVOCs through the bare skin can be calculated by: t

∂Cve,bare

0

∂x

DEbare = SAbare ∫ Dve

168

169

 C  ∂Cssl = hm  Ca − ssl  , x = Lve + Lsc + Lssl ∂x K ssl  

dt

(8)

x =0

where the subscript bare represents the bare part of the body; and Cve,bare or

∂Cve,bare ∂x

x =0

170

can be obtained by combining the eliminated governing equations, initial conditions, and

171

boundary conditions (i.e., equations (1), (2) and (5) eliminated as described above) with

172

equations (3) and (7), detailed equations are provided in Section S1 of the Supporting

173

Information (SI).

174 175

The total exposure dose of SVOCs (DEtotal) is: DE total = DEclothed + DE bare

(9)

176

With the above model, Morrison et al.19 found that the model predictions (i.e.,

177

DEtotal) were consistent with the measured dermal uptake of DEP and DnBP for bare-skin

178

participants (measured by Weschler et al.4), indicating that the above five assumptions are

179

reasonable, at least for the exposure scenarios they considered. However, for the clothed

180

participants, their predictions covered a wide range (see the details in Figure 2 of

181

Morrison et al.22), which was due to the incorrect description of SVOC transportation

182

from clothing to the SSL. This is illustrated by our improved model, described below.

183

Morrison et al.22 assumed that the SVOC concentration is linearly distributed within

184

the air gap and the SSL. Linearity is not assumed in equation (1). However, given that the

185

diffusivities of SVOCs in the air and the SSL are much greater than in other layers (i.e.,

186

clothing, SC, and VE), predictions with or without formal assumption of linearity are

187

identical according to our calculations.

188 7



189

Page 8 of 26

2.2 Model improvement

190

To more accurately predict the heat transfer or mass transfer processes between the

191

clothed human skin and the ambient environment, many studies have measured the

192

thickness of the air gap and the contact area between clothing and skin based on a three-

193

dimensional (3D) scan.23-27, 31, 32 In general, the methods used to quantify the air gap

194

thickness and the contact area include three steps:23, 24 1) scan the nude manikin (or in

195

some cases, the human subject); 2) scan the manikin dressed with the target clothing; 3)

196

determine the clothing-skin air gap thickness by superimposing the images of these two

197

scans (i.e., dividing the 3D image of step 2) by that of step 1)). Direct contact is assumed

198

where the obtained air gap thickness is smaller than the uncertainty of the 3D scanner

199

(e.g., less than 1 mm).24 To guarantee accuracy, the body is always sliced into several

200

parts, and quantification of air gap thickness and contact area is separately conducted for

201

each part.

202

Given that the area of direct contact is significant compared to the total area of

203

human skin, transport of SVOCs from clothing to skin surface in the direct contact parts

204

(which is by means of partitioning between clothing and the SSL) should be considered

205

when modeling dermal exposure of SVOCs, as is shown in Figure 1 (b). Where clothing

206

directly contacts the skin, the terms related to Cag in equations (1) and (2) should be

207

eliminated. In addition, the boundary conditions related to Cag in equation (5) need to be

208

replaced by: ∂C ssl ∂Ccl C ssl Ccl = , Dssl = Dcl , x = Lve + Lsc + Lssl ∂x ∂x K ssl K cl

209

(10)

210

Under this condition, the SVOC exposure dose of clothed skin (DEclothed) is the sum

211

of the exposure dose of both scenarios: clothing-skin direct contact and clothing-skin air

212

gap: DEclothed = DEcont + DEgap

213

 t ∂C ,cont , = ∑ SAclothed, p  f p ∫ Dve ve p 0 ∂x p 

x =0

t

∂Cve ,gap, p

0

∂x

dt + (1 − f p ) ∫ Dve

 (11) dt  x=0 

214

where the subscript p represents different parts of the body; the subscripts cont and gap

215

represent direct contact and air gap, respectively; f is the ratio of direct contact area to the

8


Page 9 of 26

216

217


total skin area (for each part of the body);

al.’s model 22; and

∂Cve,cont ∂x

∂Cve,gap ∂x

can be obtained from Morrison et x=0

can be obtained by combining the eliminated equations x =0

218

(1), (2) and (5) (by the means described above) with equations (3) and (10), detailed

219

equations are provided in SI Section S2.

220

The total exposure dose of SVOCs (DEtotal) can be obtained by combining equations

221

(8), (9) and (11). Because the values of f and air gap thickness (i.e., Lag) have significant

222

variations for different parts of the body, the whole body is sliced into several parts

223

(identical to the method used for determining f and Lag) and the calculation is performed

224

separately for each part. The skin area for each body part (i.e., SAp) can be obtained from

225

the literature

226

scanner method measurements 27. Other parameters required for the calculations, such as

227

Ca, initial concentrations, diffusivities, partition coefficients, thickness, and hm, can either

228

be estimated by formulas or obtained by measurement. Detailed determination of these

229

parameters has been introduced by Gong et al.13 and Morrison and colleagues 19, 22.

33, 34

, the f and Lag values of each body part can be obtained from the 3D

230 231

3. Model evaluation

232

3.1 Experimental conditions

233

Comparing model predictions and experimental data under the same exposure

234

scenarios is essential to evaluate the performance of the improved model. Recently,

235

several studies have measured the dermal uptake of certain SVOCs by participants

236

wearing clothes, under controlled conditions. These studies include dermal uptake of

237

DEP and DnBP

238

available information for the characteristic parameters describing the sorption of nicotine

239

and benzophenone-3 in clothing materials (i.e., Dcl and Kcl), only the experimental data

240

for DEP and DnBP (measured by Morrison et al.20) are used to evaluate our model.

20

, nicotine

8, 9

, and benzophenone-3

21

. However, due to the lack of

241

Using the experimental setup of Weschler et al.4, which was designed to measure

242

dermal uptake of gas-phase SVOCs by unclothed participants (they wore only shorts),

243

Morrison et al.20 measured the dermal uptake of gas-phase SVOCs by a participant

244

wearing SVOC-free clothing or SVOC-contaminated clothing (exposed to gas-phase 9



245

SVOCs prior to experiments). DEP and DnBP were the target SVOCs. The whole

246

experiment included three phases:

247

1) Before exposure: 12 hours prior to the experiment, diet and personal care products

248

were restricted, to reduce the background levels of DEP and DnBP on, or ingested by, the

249

participant; 24 hours prior to the experiment, the participant showered without soap or

250

other detergents. Two sets of clothing were prepared: clean clothes (which were put on

251

directly from the package) and exposed clothes (which had been exposed to gas-phase

252

DEP and DnBP for 9 days, the average concentrations (Ca,loading) during this period were

253

244 µg/m3 and 119 µg/m3 for DEP and DnBP, respectively). The clothes in both cases

254

were identical: an undershirt, a pair of jeans, a long-sleeved T-shirt, underwear, socks. All

255

clothes were 100% cotton (except for the socks, 85% cotton).

256

2) Chamber exposure: the participant stayed in a 55 m3 chamber for 6 hours. Gas-

257

phase concentrations of DEP and DnBP within the chamber were kept relatively stable.

258

During these 6 hours, the participant wore a breathing hood (clean air from outside the

259

chamber was supplied to the hood) to eliminate any inhalation exposure. Immediately

260

before entering the chamber, the participant changed into the clean clothes or into the

261

exposed clothes, i.e., two scenarios were included in this stage. The participant sat in

262

front of a desk for most of the 6 hours, with diet and personal care products restricted.

263

Note: the participant’s hands were bare.

264

3) Post-exposure: after 6 hours, the participant left the chamber and changed into his

265

normal clothing. This stage lasted for 48 hours, with diet and personal care products

266

restricted.

267

During the entire experiment, the urine of the participant was sampled to quantify

268

the metabolites of DEP and DnBP. The concentrations of the metabolites were then

269

converted to the dermal uptake amount (diet ingestion and inhalation uptake were either

270

ignored or corrected). Specifically, urine samples were taken immediately before entering

271

the chamber, once during the 6-hour exposure period, and for every micturition during

272

the 48-hour post-exposure period.

273 274 275

3.2 Parameterization According to Weschler et al.4 and Morrison et al.’s 20 measurements, the 10


Page 10 of 26

Page 11 of 26


276

concentrations of DEP and DnBP metabolites in the urine samples prior to exposure are

277

far lower than those after entering the chamber. They therefore assumed that the initial

278

concentrations of DEP and DnBP within the skin layers were zero, i.e., Cve,0, Csc,0, Cssl,0,

279

and Cag,0 in equation (2) equal zero:

280

Cve = 0, Csc = 0, Cssl = 0, Cag = 0, t = 0

(12)

281

There are two scenarios for the initial concentrations of DEP and DnBP in the

282

clothing: 1) for the clean clothes, the initial concentrations approximate to zero, i.e., Ccl =

283

0 (t = 0); 2) for the exposed clothes, Ccl,0 in equation (2) is the solution of a model that

284

considers the dynamic sorption process of SVOCs onto clothing. The detailed model is

285

provided in SI Section S3.22

286

Table 1 lists the parameters used in the calculations, including the parameters within

287

the clothing, the air gap, and the skin layers, as well as Lag and f for each body part.

288

Values of D and K in the skin layers (SSL, SC, and VE) and Dag are identical to those

289

used by Morrison et al.22 Kcl for DEP and DnBP are the measured results obtained by

290

Morrison et al.35 for cotton clothing. Dcl of DnBP was measured by Cao et al.36; given

291

that there is no available data for Dcl of DEP, it was estimated by Dcl,DEP =

292

Kcl,DnBP/Kcl,DEP·Dcl,DnBP (based on the correlation Dcl ~ 1/Kcl, which has been proved by

293

Cao et al.36). Mert et al.27 measured the distributions of Lag and f for different body parts

294

for various body postures; the Lag and f values listed in Table 1 are values appropriate for

295

the scenario where a manikin sits at a desk wearing a regular T-shirt, and regular trousers.

296

The division of individual body parts is shown in Figure 2 (identical to the divisions used

297

by Mert et al.27). The total area of the participant’s skin is 2.06 m2, 20, 33 the SAp for each

298

body part was obtained from the human body surface area database measured by Yu et

299

al.34 (the values for a male, as listed in Table 7 of their paper). Particularly, the

300

participant’s hands are considered to be bare skin, the head and neck are free of exposure

301

(isolated by the breathing hood), and the feet are assumed to be in direct contact with

302

clothing (wearing socks). Finally, it should be noted that the value of Ca in equations (4)

303

and (7) varies over time, i.e., Ca equals the value listed in Table 1 during the chamber

304

exposure period (from t = 0 to t = 6 hours) and is zero during the post-exposure stage

305

(from t = 6 to t = 54 hours).

306 11



Page 12 of 26

Table 1. Parameters used in model evaluation. 307

Figure 2. Schematic illustrating the division of individual parts of the body.27, 37 308 309

To numerically solve the model, the clothing, air gap (if it exists), SSL, SC, and VE

310

were divided into 50, 10, 10, 20, and 50 slices, respectively (found to be sufficient to

311

obtain a stable solution). The improved model was then discretized into a system of

312

ordinary differential equations (based on the implicit difference method

313

solved using MATLAB R2012b. Three exposure scenarios: bare skin, air gap, and direct

314

contact, were separately modeled. It should be noted that similar predictions can be

315

obtained if the air gap and the SSL are not divided into several slices, i.e., the simplified

316

method used by Morrison et al.22 (as noted just above Section 2.2).

38

), which were

317 318

3.3 Comparison between model predictions and experimental results

319

Table 2. Comparison between model predictions and experimental data. 320 321

Because the dermal uptake amount was monitored for 54 hours (6 hours of chamber

322

exposure and 48 hours of post-exposure), the total simulation time is 54 hours. The

323

predictions from the improved model when t equals 54 hours are listed in Table 2, as well

324

as the comparison between them and the experimental data from Morrison et al.20 (as

325

listed in Table 1 of their paper). It can be seen that the model predictions are consistent

326

with the measured results: when the participant wore clean clothes, the relative deviations

327

between the predicted and measured results (which are obtained by ǀPredicted DEtotal -

328

Measured DEtotalǀ/Measured DEtotal ×100%) are 62% and 43% for DEP and DnBP,

329

respectively; when the participant wore exposed clothes, the relative deviations between

330

them are 7% and 15%, respectively. Generally, the improved model slightly

331

underestimates the total exposure dose for all scenarios. The reason for this may be that

332

our calculations ignore the direct contact between the participant’s hands and the surfaces

333

inside the exposure chamber (e.g., surfaces of the desks and chairs, which may have 12


Page 13 of 26


334

sorbed significant amount of SVOCs before the participant entered the chamber).

335

Nevertheless, the good agreement between the model predictions and experimental data

336

imbues confidence in the improved model. Certainly, given that the model calculation

337

requires a number of input parameters, and most of these parameters are estimated with

338

empirical formulas, a comprehensive evaluation of the model needs further comparisons

339

between model predictions and measurements for more extensive exposure scenarios and

340

for other classes of SVOCs.

341

Additionally, the results listed in Table 2 indicate that the condition of the clothing

342

(or the SVOC concentration in the clothing) significantly influences the degree of dermal

343

exposure to SVOCs. It can be seen that in the case of clean clothes, the exposure dose of

344

the clothed body parts (DEcont + DEgap) approaches zero. This is because no SVOCs (or at

345

least, negligible SVOCs) in the chamber air have penetrated through the clothing during

346

the chamber exposure period (i.e., 6 hours). According to existing studies, the time

347

required for SVOCs to penetrate through clothes is approximately 0.2Lcl2/Dcl, i.e., 14

348

hours for DEP and 198 hours for DnBP, both significantly longer than 6 hours.36 For

349

exposed clothes, the exposure dose of the clothed body parts contributed to over 95% of

350

the total exposure (i.e., (DEcont + DEgap)/DEtotal > 95%). Furthermore, direct contact

351

between clothing and the skin surface is the primary pathway for SVOCs from clothing to

352

the dermal layer, i.e., direct contact contributes 62% and 87% of the total exposure

353

(DEcont/DEtotal) for DEP and DnBP, respectively. However, existing models ignore this

354

pathway.13, 19, 22 Thus, for a clothed human, an accurate measurement of the ratio of direct

355

contact area to the total skin area (i.e., f) is essential for accurate assessment of dermal

356

exposure to SVOCs.

357 358

4. Discussion

359

4.1 Model simplification

360

Figure 3. Variation of exposure dose through the air gap to the air gap thickness. DE*gap is the normalized exposure dose through the air gap (DEgap), i.e., DE*gap = DEgap/Ca, Ca is the gas-phase SVOC concentration. 13



Page 14 of 26

361 362

Given that the air gap thickness varies around the body, DEgap is predicted separately

363

for each part of the body in the above calculations. This may ensure the accuracy of the

364

model predictions while complicating the model implementation processes. One potential

365

way to simplify the calculation is to predict DEgap once for the whole body, based on the

366

body-average air gap thickness (designated as Lag ). If Lag is the weighted average of Lag

367

for each part of the body (weight coefficient is the ratio of the skin surface area of each

368

part of the body, i.e., Lag = ∑ SAp Lag , p SAtotal ), Lag is calculated to be 10.4 mm for the p

369

data listed in Table 1, and the relative deviation between the predicted DEgap based on

370

Lag and the predicted DEgap based on separate body parts is less than 10% for both DEP

371

and DnBP (the relative deviation for DEtotal is less than 5%). This is because the values of

372

Lag for different body parts fall in the range between 5 mm and 15 mm, and a Lag within

373

this range has no significant effect on the values of DEgap (for each Lag, the error

374

introduced to DEgap is less than 50% if Lag is replaced by Lag ), as illustrated in Figure 3.

375

Consequently, using the body-average air gap thickness ( Lag ) to predict DEgap is a

376

reliable simplification, implying that equation (11) can be simplified to:

377

 t t ∂Cve,gap ∂C DEtotal = SAclothed  f cont ∫ Dve ve ,cont dt + (1 − f cont ) ∫ Dve 0 0 ∂x x =0 ∂x  t ∂C + SAbare ∫ Dve ve ,bare dt 0 ∂x x =0

 dt   x =0

(13)

378

where fcont is the ratio of direct contact area to the total skin area for the whole body (for

379

the data listed in Table 1, fcont = 0.28);

∂ Cve,gap ∂x

is obtained based on Lag . x =0

380 381 382 383 384

4.2 Limitations and further study Given that several assumptions are made, caution should be exercised when using the present model. The limitations of this model require further study: 1) Metabolization of SVOCs within skin layers is ignored in this model, which may 14


Page 15 of 26


385

be invalid for some SVOCs. For instance, Hopf et al.39 found that 100% of di(2-

386

ethylhexyl) phthalate (DEHP) was metabolized in vitro in human viable skin. In existing

387

studies, the esterases required for the metabolization of SVOCs are always assumed to

388

not be in the SC but in the VE and dermis.27, 39, 40 Therefore, when considering SVOC

389

metabolization, the first formula in equation (1) (corresponding to the VE layer) should

390

be modified to: 40, 41

∂Cve ∂ 2Cve & D = − S , 0 < x < Lve ve  ∂t ∂x 2  ∂ 2Cve ,M &  ∂Cve, M = D + S , 0 < x < Lve ve , M ∂x 2  ∂t

391

(14)

392

where S& is the metabolization rate of SVOCs within the VE, µg/(m3·s); Cve,M is the

393

concentration of the SVOC metabolite within the VE, µg/m3; and Dve,M is the diffusion

394

coefficient of the SVOC metabolite within the VE, m2/s. In existing studies, the

395

metabolization rate is always estimated by S& = Vmax Cve

396

maximum metabolization rate of SVOCs within the VE, µg/(m3·s); and Km is the

397

Michaelis constant, µg/m3.40,

398

Additionally, if the SVOC metabolites can penetrate through the SC and SSL layers,

399

modification of the corresponding formulas in equation (1), in a similar manner to

400

equation (14), is required. The diffusion coefficient and the partition coefficient of the

401

metabolites and those of the parent SVOCs are usually not the same,41 greatly

402

complicating the whole model. Given that the predictions of the model assuming no

403

metabolization agreed well with the experimental data for DEP and DnBP uptake in bare

404

skin

405

dermal exposure model taking into consideration SVOC metabolization within the skin

406

requires further study.

19

41

( K m + Cve ) ,

where Vmax is the

Both Vmax and Km are determined experimentally.

, this assumption is also considered in the above evaluation of the model. The

407

2) The current model only considers a single-layer of clothing. When wearing multi-

408

layers, the description of SVOC transport within two contiguous layers of clothes

409

requires further study. However, given that the rate of SVOC penetration through the

410

clothing is extremely slow (indicated by the negligible dermal uptake of DEP and DnBP

411

when wearing clean clothes), SVOCs in the upper layers of clothes are hardly able to

412

transfer to the skin surface during the period of being clothed (e.g., a dozen hours). This 15



413

implies that when modeling dermal exposure to SVOCs for clothed body parts, it is only

414

necessary to consider the mass transfer of SVOCs between the bottom layer of clothes

415

and the SSL. Certainly, for the cases of multi-layer clothes, the values of Lag and fcont

416

should be different from those for single-layer clothes ( Lag and fcont tend to decrease and

417

increase, respectively), which can be quantified by a 3D body scanner.

418

3) Effects of skin surface lipid transfer to clothing, desquamation, sweating, and

419

SVOC transport through skin organs (e.g., hair follicles and sweat glands) are not taken

420

into account in the current model. Potential errors introduced by these simplifications or

421

the reasonability of ignoring these effects are not in the scope of this study, but have been

422

discussed in other studies.13, 19, 22

423

4) The improved model was preliminarily evaluated based on the measured data

424

from one experiment with one participant. Consequently, further evaluation of the

425

improved model based on more experiments is requisite, with more extensive exposure

426

scenarios, more participants (significant subject-to-subject variability in the measured

427

data has been reported, which appears to be related with the age of the participants 4), and

428

other classes of SVOCs (specifically, less volatile SVOCs, e.g., DEHP) considered in the

429

experiments.

430

Nevertheless, the primary improvement of our model, which is to take into

431

consideration the parallel SVOC transports between clothing and skin surface: direct

432

contact (which is ignored in existing models) and an air gap, increases the accuracy of

433

assessing dermal exposure to gas-phase SVOCs, especially for predicting the effects of

434

clothing on dermal exposure.

435 436

Associated content

437

Supporting Information

438 439

Supporting information is noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

440

The Supporting information includes additional detail of the model for predicting

441

SVOC exposure of bare skin (Section S1) and skin that is in direct contact with clothing

442

(Section S2), and the model describing the dynamic sorption process of SVOCs in the 16


Page 16 of 26

Page 17 of 26

443


clothing (Section S3).

444 445

Author information

446

Corresponding Author

447 448

*

E-mail: [email protected]; Phone: +86 10 62772518; Fax: +86 10

62773461.

449 450 451

Note The authors declare no competing financial interest.

452 453

Acknowledgements

454

This work was supported by the National Key Research and Development Program

455

of China (Grant No. 2017YFC0702700) and the Natural Science Foundation of China

456

(Grant Nos. 51420105010 and 51521005).

457 458

References

459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479

1. Weschler, C. J.; Nazaroff, W. W. SVOC exposure indoors: fresh look at dermal pathways. Indoor Air 2012, 22 (5), 356-377. 2. Weschler, C. J.; Nazaroff, W. W. Dermal uptake of organic vapors commonly found in indoor air. Environ. Sci. Technol. 2014, 48 (2), 1230-1237. 3. Gong, M.; Zhang, Y.; Weschler, C. J. Measurement of phthalates in skin wipes: estimating exposure from dermal absorption. Environ. Sci. Technol. 2014, 48 (13), 74287435. 4. Weschler, C. J.; Bekö, G.; Koch, H. M.; Salthammer, T.; Schripp, T.; Toftum, J.; Clausen, G. Transdermal uptake of diethyl phthalate and di(n-butyl) phthalate directly from air: Experimental verification. Environ. Health Perspect. 2015, 123 (10), 928-934. 5. Abou-Elwafa Abdallah, M.; Pawar, G.; Harrad, S. Human dermal absorption of chlorinated organophosphate flame retardants; implications for human exposure. Toxicol. Appl. Pharmacol. 2016, 291, 28-37. 6. Gong, M.; Weschler, C. J.; Zhang, Y. Impact of clothing on dermal exposure to phthalates: Observations and insights from sampling both skin and clothing. Environ. Sci. Technol. 2016, 50 (8), 4350-4357. 7. Wu, C. C.; Bao, L. J.; Tao, S.; Zeng, E. Y. Dermal uptake from airborne organics as an important route of human exposure to E-waste combustion fumes. Environ. Sci. Technol. 2016, 50 (13), 6599-6605. 8. Bekö, G.; Morrison, G.; Weschler, C. J.; Koch, H. M.; Palmke, C.; Salthammer, T.; Schripp, T.; Toftum, J.; Clausen, G. Measurements of dermal uptake of nicotine 17



480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525

directly from air and clothing. Indoor Air 2017, 27 (2), 427-433. 9. Bekö, G.; Morrison, G.; Weschler, C. J.; Koch, H. M.; Palmke, C.; Salthammer, T.; Schripp, T.; Eftekhari, A.; Toftum, J.; Clausen, G. Dermal uptake of nicotine from air and clothing: experimental verification. Indoor Air 2018, 28 (2), 247–257. 10. Lorber, M.; Weschler, C. J.; Morrison, G.; Bekö, G.; Gong, M.; Koch, H. M.; Salthammer, T.; Schripp, T.; Toftum, J.; Clausen, G. Linking a dermal permeation and an inhalation model to a simple pharmacokinetic model to study airborne exposure to di(nbutyl) phthalate. J. Expo. Sci. Environ. Epidemiol. 2017, 27, 601–609 11. Pelletier, M.; Bonvallot, N.; Ramalho, O.; Blanchard, O.; Mercier, F.; Mandin, C.; Le Bot, B.; Glorennec, P. Dermal absorption of semivolatile organic compounds from the gas phase: Sensitivity of exposure assessment by steady state modeling to key parameters. Environ. Int. 2017, 102, 106-113. 12. Gong, M.; Weschler, C. J.; Liu, L.; Shen, H.; Huang, L.; Sundell, J.; Zhang, Y. Phthalate metabolites in urine samples from Beijing children and correlations with phthalate levels in their handwipes. Indoor Air 2015, 25 (6), 572-581. 13. Gong, M.; Zhang, Y.; Weschler, C. J. Predicting dermal absorption of gas-phase chemicals: transient model development, evaluation, and application. Indoor Air 2014, 24 (3), 292-306. 14. Shi, S.; Zhao, B. Modeled exposure assessment via inhalation and dermal pathways to airborne semivolatile organic compounds (SVOCs) in residences. Environ. Sci. Technol. 2014, 48 (10), 5691-5699. 15. Xu, Y.; Cohen Hubal, E. A.; Clausen, P. A.; Little, J. C. Predicting residential exposure to phthalate plasticizer emitted from vinyl flooring: a mechanistic analysis. Environ. Sci. Technol. 2009, 43 (7), 2374-2380. 16. Little, J. C.; Weschler, C. J.; Nazaroff, W. W.; Liu, Z.; Cohen Hubal, E. A. Rapid methods to estimate potential exposure to semivolatile organic compounds in the indoor environment. Environ. Sci. Technol. 2012, 46 (20), 11171-11178. 17. Bunge, A. L.; Cleek, R. L.; Vecchia, B. E. A new method for estimating dermal absorption from chemical exposure. 3. Compared with steady-state methods for prediction and data analysis. Pharm. Res. 1995, 12 (7), 972-982. 18. Kezic, S.; Astrid, J.; Jacob, K.;, Aart, C. M.; Maarten, M. V. Percutaneous absorption of m-xylene vapour in volunteers during pre-steady and steady state. Toxicol. Lett. 2004, 153 (2): 273-282. 19. Morrison, G. C.; Weschler, C. J.; Bekö, G. Dermal uptake directly from air under transient conditions: advances in modeling and comparisons with experimental results for human subjects. Indoor Air 2016, 26 (6), 913-924. 20. Morrison, G. C.; Weschler, C. J.; Bekö, G.; Koch, H. M.; Salthammer, T.; Schripp, T.; Toftum, J.; Clausen, G. Role of clothing in both accelerating and impeding dermal absorption of airborne SVOCs. J. Expo. Sci. Environ. Epidemiol. 2016, 26 (1), 113-118. 21. Morrison, G. C.; Bekö, G.; Weschler, C. J.; Schripp, T.; Salthammer, T.; Hill, J.; Andersson, A. M.; Toftum, J.; Clausen, G.; Frederiksen, H. Dermal uptake of benzophenone-3 from clothing. Environ. Sci. Technol. 2017, 51 (19), 11371–11379. 22. Morrison, G. C.; Weschler, C. J.; Bekö, G. Dermal uptake of phthalates from clothing: Comparison of model to human participant results. Indoor Air 2017, 27 (3), 642-649. 23. Mah, T.; Guowen, S. Investigation of the contribution of garment design to 18


Page 18 of 26

Page 19 of 26

526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571


thermal protection. Part 1: characterizing air gaps using three-dimensional body scanning for women's protective clothing. Text. Res. J. 2010, 80 (13), 1317-1329. 24. Psikuta, A.; Frackiewicz-Kaczmarek, J.; Frydrych, I.; Rossi, R. Quantitative evaluation of air gap thickness and contact area between body and garment. Text. Res. J. 2012, 82 (14), 1405-1413. 25. Frackiewicz-Kaczmarek, J.; Psikuta, A.; Bueno, M.-A.; Rossi, R. M. Effect of garment properties on air gap thickness and the contact area distribution. Text. Res. J. 2015, 85 (18), 1907-1918. 26. Frackiewicz-Kaczmarek, J.; Psikuta, A.; Bueno, M.-A.; Rossi, R. M. Air gap thickness and contact area in undershirts with various moisture contents: influence of garment fit, fabric structure and fiber composition. Text. Res. J. 2015, 85 (20), 21962207. 27. Mert, E.; Psikuta, A.; Bueno, M.-A.; Rossi, R. M. The effect of body postures on the distribution of air gap thickness and contact area. Int. J. Biometeorol. 2016, 61, 363375. 28. WHO Dermal absorption: Environmental Health Criteria 235. Published under the joint sponsorship of the United Nations Environment Programme, the International Labour Organization and World Health Organization, 2006. 29. Brown, T. N.; Armitage, J. M.; Egeghy, P.; Kircanski, I.; Arnot, J. A. Dermal permeation data and models for the prioritization and screening-level exposure assessment of organic chemicals. Environ. Int. 2016, 94, 424-435. 30. Kretsos, K.; Miller, M. A.; Zamora-Estrada, G.; Kasting, G. B. Partitioning, diffusivity and clearance of skin permeants in mammalian dermis. Int. J. Pharm. 2008, 346 (1-2), 64-79. 31. Zhang, Z. H.; Wang, Y.; Li, J. Model for predicting the effect of an air gap on the heat transfer of a clothed human body. Fibres Text. East. Eur. 2011, 87 (4), 105-110. 32. Mert, E.; Psikuta, A.; Bueno, M. A.; Rossi, R. M. Effect of heterogenous and homogenous air gaps on dry heat loss through the garment. Int. J. Biometeorol. 2015, 59 (11), 1701-1710. 33. USEPA Exposure Factors Handbook. United States Environmental Protection Agency: Office of Research and Development, Washington, 2011. 34. Yu, C.; Lin, C.; Yang, Y. Human body surface area database and estimation formula. Burns 2010, 36 (5), 616-629. 35. Morrison, G.; Li, H.; Mishra, S.; Buechlein, M. Airborne phthalate partitioning to cotton clothing. Atmos. Environ. 2015, 115, 149-152. 36. Cao, J.; Liu, N.; Zhang, Y. SPME-based Ca-history method for measuring SVOC diffusion coefficients in clothing material. Environ. Sci. Technol. 2017, 51 (16), 9137– 9145. 37. Human Body Diagram Picture; http://unmasadalha.blogspot.com/2016/01/bodydiagram.html. (Accessed on December 16, 2017) 38. Bergman, T. L.; Incropera, F. P.; DeWitt, D. P.; Lavine, A. S. Fundamentals of Heat and Mass Transfer. John Wiley & Sons: Hoboken, New Jersey, United States, 2011. 39. Hopf, N. B.; Berthet, A.; Vernez, D.; Langard, E.; Spring, P.; Gaudin, R. Skin permeation and metabolism of di(2-ethylhexyl) phthalate (DEHP). Toxicol. Lett. 2014, 224 (1), 47-53. 40. Sugibayashi, K.; Hayashi, T.; Morimoto, Y. Simultaneous transport and 19



572 573 574

metabolism of ethyl nicotinate in hairless rat skin after its topical application: the effect of enzyme distribution in skin. J. Control. Release 1999, 62 (1), 201-208. 41. Boderke, P.; Schittkowski, K.; Wolf, M.; Merkle, H. P. Modeling of diffusion and

575

concurrent metabolism in cutaneous tissue. J. Theor. Biol. 2000, 204 (3), 393-407.

576

20


Page 20 of 26

Page 21 of 26

577


TOC Art

578 579

21



580

Tables

581

Table 1. Parameters used in model evaluation Parameters Ca,loading (µg/m3)a Ca (µg/m3)a hm (m/h)a Dcl (m2/s)b Kcl (-)b Lcl (mm)a Da (m2/s)a Dssl (m2/s)a Body parts SA (m2)c Lag (mm)d f (%)d Body parts SA (m2)c Lag (mm)d f (%)d

DEP 244 295 3.4 4.1 × 10-12 2.5 × 105 1.0 5.5 × 10-6 8.6 × 10-7 Head 0.15 -e Abdomen 0.10 7.7 20

DnBP 119 142 3.1 2.8 × 10-13 3.7 × 106 1.0 4.7 × 10-6 1.5 × 10-6

Parameters Kssl (-)a Lssl (µm)a Dsc (m2/s)a Ksc (-)a Lsc (µm)a Dve (m2/s)a Kve (-)a Lve (µm)a

Page 22 of 26

DEP 2.2 × 107 1.2 2.6 × 10-15 7.1 × 106 23 6.3 × 10-11 7.6 × 105 100

DnBP 2.2 × 108 1.2 4.2 × 10-15 2.7 × 107 23 6.4 × 10-12 3.3 × 106 100

Neck Upper arm Lower arm Hand Upper chest Upper back Lower chest Lower back 0.06 0.17 0.13 0.10 0.09 0.07 0.06 0.05 11 10 4.9 4.9 7.6 6.8 30 20 61 64 48 38 Lumbus Anterior pelvis Posterior pelvis Anterior thigh Posterior thigh Shin Calf Foot 0.12 0.05 0.10 0.20 0.18 0.14 0.14 0.14 7.6 14 9.9 13 21 30 19 16 9.6 38 15 38 13 10 100

582

a

Obtained from Morrison et al.22

583

b

Kcl of DEP and DnBP are obtained from Morrison et al.35; Dcl of DnBP was measured by Cao et al.36; given that there is no available data for Dcl of DEP, it is

584

obtained by Dcl,DEP = Kcl,DnBP/Kcl,DEP·Dcl,DnBP 36.

585

c

Total area of the participant’s skin is 2.06 m2,20, 33 SAp of each body part is the product of the total area and the corresponding f.34

586

d

Mert et al.27 measured the distributions of Lag and f of different body parts for various body postures; the Lag and f values listed here are subject to the

587

scenario that the manikin sat at a desk and wore a regular T-shirt and regular trousers. The division of individual body parts is shown in Figure 2.

588

e

“-” represents the bare skin (hands), the scenarios that there is no dermal exposure (head and neck) or skin surface is in direct contact with clothing (foot). 22


Page 23 of 26


589

Table 2. Comparison between model predictions and experimental data Clothing condition Clean SVOC exposed

SVOC

DEcont a

DEgap

DEhand b

Predicted DEtotal c

Measured DEtotal d

RD (%) e

DEP

0.01

0.00

0.13

0.14

0.35

62

DnBP

0.00

0.00

0.04

0.04

0.07

43

DEP

2.73

1.53

0.13

4.39

4.71

6.8

DnBP

2.45

0.33

0.04

2.82

3.33

15

590

a

The unit of DE is mg.

591

b

SVOC exposure dose through the bare hand.

592

c

Predicted by the improved model, i.e., equation (11).

593

d

Experimentally measured by Morrison et al.20

594

e

RD represents the relative deviation between model predictions and experimental data, i.e.,

595 596

RD = ǀPredicted DEtotal - Measured DEtotalǀ/Measured DEtotal ×100%.

23



597

Figures

598

Figure 1. Schematic illustrating layers and parameters in: (a) dermal exposure model

599

introduced by Morrison et al.22 (ignoring direct contact between clothing and skin

600

surface lipids), and (b) dermal exposure model modified in this study.

601 602

(a)

603

604 605

(b) 24


Page 24 of 26

Page 25 of 26


606

Figure 2. Schematic illustrating the division of individual parts of the body.27, 37

607 608

25



609

Figure 3. Variation of exposure dose through the air gap to the air gap thickness.

610

DE*gap is the normalized exposure dose through the air gap (DEgap), i.e., DE*gap =

611

DEgap/Ca, Ca is the gas-phase SVOC concentration.

612 613

26


Page 26 of 26

Predicting Dermal Exposure to Gas-Phase Semivolatile Organic

Recommend Documents