Identification of Alternative Vapor Intrusion Pathways Using Controlled

Subscriber access provided by Karolinska Institutet, University Library

Article

Identification of Alternative Vapor Intrusion Pathways Using Controlled Pressure Testing, Soil Gas Monitoring, and Screening Model Calculations Yuanming Guo, Chase Weston Holton, Hong Luo, Paul Dahlen, Kyle Gorder, Erik M. Dettenmaier, and Paul Carr Johnson Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b03564 • Publication Date (Web): 12 Oct 2015 Downloaded from http://pubs.acs.org on October 17, 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 34

Environmental Science & Technology

1

Identification of Alternative Vapor Intrusion Pathways Using Controlled Pressure

2

Testing, Soil Gas Monitoring, and Screening Model Calculations

3

YUANMING GUO†, CHASE HOLTON†§, HONG LUO† , PAUL DAHLEN†, KYLE GORDER‡, ERIK DETTENMAIER‡, AND PAUL C. JOHNSON*, †║

4 5 6 7 8 9 10

⊥

†School of Sustainable Engineering and the Built Environment, Ira A Fulton Schools of Engineering, Arizona State University, Tempe, AZ 85287, §CH2M, 9193 South Jamaica Street, Englewood, CO 80112, ⊥Chevron Energy Technology Company, 1200 Smith St., Houston, TX 77002, ‡ Restoration Installation Support Team, Hill Air Force Base, 7290 Weiner St., Building 383, Hill AFB, UT 84056, and ║Department of Civil and Environmental Engineering, Colorado School of Mines, Golden, CO 80401

11 12

ABSTRACT

13

Vapor intrusion (VI) pathway assessment and data interpretation have been guided by an

14

historical conceptual model in which vapors originating from contaminated soil and/or

15

groundwater diffuse upward through soil and are swept into a building by soil gas flow induced

16

by building under-pressurization. Recent studies reveal that alternative VI pathways involving

17

neighborhood sewers, land drains, and other major underground piping can also be significant VI

18

contributors, even to buildings beyond the delineated footprint of soil and groundwater

19

contamination. This work illustrates how controlled pressure method testing (CPM), soil gas

20

sampling, and screening level emissions calculations can be used to identify significant

21

alternative VI pathways that might go undetected by conventional sampling under natural

22

conditions at some sites. The combined utility of these tools is shown through data collected at a

23

long-term study house where a significant alternative VI pathway was discovered and altered so

24

that it could be manipulated to be on or off. Data collected during periods of natural and CPM

25

conditions show that the alternative pathway was significant but its presence was not identifiable

26

under natural conditions; it was identified under CPM conditions when measured emission rates

ACS Paragon Plus Environment


27

were two orders of magnitude greater than screening model estimates and sub-foundation

28

vertical soil gas profiles changed and were no longer consistent with the conventional VI

29

conceptual model.


Page 2 of 34

Page 3 of 34

30

31


INTRODUCTION Guidance for assessing the vapor intrusion (VI) to indoor air pathway varies 1-3, but most

32

emphasize multiple-lines-of-evidence (MLE) approaches involving combinations of point-in-

33

time indoor air, sub-slab soil gas, deeper soil gas, groundwater, and soil sampling, along with

34

screening-level or more complex transport modeling. The VI pathway assessment strategy and

35

data interpretation are guided by a conceptual site model (CSM). A generic conventional VI

36

pathway CSM for a site over contaminated groundwater is shown in Figure 1a: vapors diffuse

37

upward through soil and away from impacted groundwater and are swept into the building

38

through foundation cracks and perforations by advective flow induced by building under-

39

pressurization. This route to indoor air is referred to as the “soil VI” pathway in this paper, and is

40

the route focused on in most modeling and data interpretation paradigms 1, 4-7.

41

In addition to contaminated soils and aquifers, subsurface pipe networks (e.g., sewer

42

mains and land drains) may also contain contaminants of concern either from chemical discharge

43

to those systems or from inflow of contaminated groundwater or vapors originating from

44

subsurface contamination. These neighborhood sewers, land drains, and other major

45

underground piping can distribute chemical-containing water and vapor beyond delineated

46

footprints of regional dissolved groundwater plumes. Vapors in them can be drawn into indoor

47

air through two routes as shown in Figure 1b: a) flow through piping or conduits to the sub-

48

foundation region and subsequent migration to indoor air via foundation cracks and permeations,

49

and b) through direct connection of plumbing fixtures to indoor air. These alternative VI

50

pathways are referred to as the “pipe flow VI pathway” and “sewer VI pathway” here. The

51

significance of alternative VI pathways has recently begun to be reported; for example, Riis et

52

al.8 confirmed that VI impacts to homes outside a chlorinated hydrocarbon-impacted

1



53

groundwater plume were due to vapors emanating from contaminated groundwater flowing into

54

the sewer system. Similarly, Pennell et al.9 concluded that tetrachloroethylene (PCE) in indoor

55

air at their study site was the result of sewer VI.

56

Identifying VI pathways and understanding their significance is critical when VI

57

mitigation system selection and design are needed. Sub-slab depressurization (SSD), which is the

58

presumptive remedy for VI impacts10, is known to be effective where soil VI is the dominant

59

pathway, but it might not be protective for homes where pipe flow and sewer VI pathways are

60

significant. While this has not yet been demonstrated in a well-controlled study, passive sub-slab

61

ventilation was ineffective for the buildings reported by Riis et al.8, and we are aware of another

62

site where SSD has been ineffective at mitigating VI impacts in a building screened for indoor

63

sources.

64

Most buildings have plumbing connections and subsurface infrastructure and therefore

65

the potential for alternative VI pathways; however, the presence of these VI pathways and their

66

significance are not easily discerned via simple observation, building drawings, or traditional site

67

characterization. For example, alternative VI pathways were discovered by Riis et al.8 and

68

Pennell et al.9 because they had more temporally and spatially extensive data sets than is typical.

69

Riis et al.8 suspected alternative sewer VI pathways because VI-impacts were detected in homes

70

outside of a plume footprint. They determined through indoor air, sub-slab and sewer sampling

71

that sewers were serving as alternate VI pathways. Pennell et al.9 reached a similar conclusion at

72

a home where indoor air contaminant concentrations were higher on an upper level than the

73

lowest level.

2


Page 4 of 34

Page 5 of 34


74

Below we present our experiences at a well-studied and documented house11, 12 where a

75

significant alternative pipe-flow VI pathway went undetected during multi-year high-frequency

76

sampling under natural conditions, and was only discovered through the combined use of indoor

77

air and soil gas sampling during manipulation of the building pressure and screening-level

78

modeling. After detection, the alternative pathway was modified to allow on/off control of it

79

during testing. This provided a unique opportunity to collect VI pathway assessment data under

80

natural and controlled under-pressurization conditions, with and without the connection of the

81

alternative VI pathway.

82

83

STUDY SITE DESCRIPTION

84

The study site is described in Holton et al.11, 12. It includes a two-story, split-level house

85

built into a slope with a 2.5 m elevation drop from the back to front yard. There is a living space

86

and attached garage on the lower level. Multi-level soil gas and groundwater sampling points

87

were installed inside through the foundation and outside of the building, with the soil gas points

88

installed to the following depths relative to the slab elevation: sub-slab (SS), 0.9 m below slab

89

(BS) and 1.8 m BS. The building footprint and sampling locations are shown in Figure 2. The

90

house was equipped with attic blower fans to control building under-pressurization as described

91

in Holton et al.12.

92

The study house overlies a dilute dissolved chlorinated solvent groundwater plume

93

containing 1,1-dichlorethylene (1,1-DCE), 1,1,1-trichloroethane (1,1,1-TCA), and

94

trichloroethylene (TCE). Groundwater is at about 2.5 m BS. TCE concentrations in water

95

samples collected below the building foundation ranged from approximately 10 - 50 µg/L over

3



96

the four years of this study with an average concentration of 24 ± 9 µg/L and no clear long-term

97

temporal trend; the groundwater concentration history is provided in Supplemental Information

98

Figure S1.

99

The sub-foundation gravel zone is connected to a neighborhood land drain system

100

running across the southern property boundary through a lateral pipe having one end open in the

101

sub-foundation gravel near locations 5 and 6. Unknown at the beginning of this study, its

102

presence was suspected from the data presented below. A series of diagnostic tests, (land drain

103

and lateral pipe vapor sampling, land drain manhole water and vapor sampling, SF6 and Helium

104

tracer release study, videography) confirmed the active lateral pipe connection between the sub-

105

foundation region and the neighborhood land drain system. The lateral pipe was modified at

106

t=1071 d with the installation of a manual butterfly valve to control the connection between the

107

sub-foundation area and the land drain system. Tracer gases (SF6 and Helium) were released up-

108

and down-stream of the valve with it open and closed to verify its ability to seal the connection

109

between the sub-foundation area and the land drain system. Figure 2 presents a schematic of the

110

lateral pipe and valve positions; photos of the lateral pipe and valve can be found in

111

Supplemental Information Figure S2.

112

4


Page 6 of 34

Page 7 of 34

113

114


DIAGNOSTIC TOOLS OVERVIEW The diagnostic toolset employed at the study site and discussed below includes controlled

115

pressure method (CPM) testing, soil gas sampling, and screening level calculations using typical

116

site characterization data. CPM use was proposed by McHugh et al.13 for VI pathway

117

assessment and indoor source identification, and Holton et al.12 recently validated its use for

118

quickly and confidently identifying maximum VI impacts without false negative results at their

119

study home overlying a dilute chlorinated solvent plume. CPM test results can be reported as an

120

indoor air concentration and as a mass emission rate into a building. While the former is of

121

interest for human health risk assessment, the latter is of interest here. It can be compared with a

122

screening-level mass emission calculation as one line of evidence to discern if significant

123

alternative VI pathways are present as outlined below. For example, if the CPM test emission

124

rate greatly exceeds the emission rate predicted with a screening level calculation, then that

125

suggests an inconsistency between actual site conditions and the soil VI conceptual site model,

126

and this could be an indicator of a significant alternative VI pathway. CPM testing will also

127

influence soil gas profiles and the responses could be indicative of significant alternative VI

128

pathways. Specifics of these two data analyses approaches are outlined below.

129

Comparison of screening-level mass emission rate estimate with emission rate measured

130

during CPM testing: Screening-level mass emission estimates can be calculated from vertical

131

soil gas profiles or source zone vapor concentrations using a one-dimensional diffusion-

132

dominated screening model. For example, when vertical soil gas profiles Cg, i (z) [mg/m3] are

133

available for n sub-areas of the building foundation, a soil VI pathway emission estimate

134

Eestimate, i [mg/d] can be calculated for each sub-area i using the soil gas data and measured or

135

estimated overall effective diffusion coefficients Dieff [m2/d] obtained from multi-depth sampling

5



136

locations representative of each sub-area. The total emission then can be obtained by the

137

summation of all sub-area emission rates:

138

i =n  D eff E estimate =∑  i i =1  Li

  ∆C g ,i ∆AF ,i 

Page 8 of 34

(1)

139

where ∆AF,i [m2] is the area of sub-area i, Σ∆AF,i = AF is the total building foundation area, and

140

within sub-area i ∆Cg,i is the soil gas concentration difference over the vertical distance Li, , and

141

Dieff [m2/d] is the overall in situ effective diffusion coefficient for the vertical interval Li. The

142

effective diffusion coefficient can be measured using the Johnson et al.14 tracer method or

143

estimated using the empirical Millington-Quirk expressions as described in Johnson and

144

Ettinger4. When the interval Li has multiple estimated or measured values Di,jeff over m sub-

145

layers of thickness ∆Li,j, where Σ∆Li,j = Li, and j denotes the sub-layer, then:

146

 Dieff   Li

 1 = j = m ∆Li , j  ∑ eff j =1 Di , j

147

When only vapor source concentrations are available, the USEPA spreadsheet

148

implementation of the Johnson and Ettinger model15 can be used to generate a screening-level

149

emission estimate. In that case EJ&E-estimate is calculated from the user-specified building

150

exchange rate EB [1/d] and building volume VB [m3], and the indoor air concentration estimate

151

CJ&E-indoor [mg/m3] output in the spreadsheet:

(2)

152

EJ&E-estimate = EB x VB x CJ&E-indoor

153

While the building volume VB and building exchange rate EB are inputs to the Johnson

154

(3)

and Ettinger model and the concentration output C is dependent on them, the emission rate EJ&E-

6


Page 9 of 34


is not sensitive to their choice for reasonable values. Once Eestimate or EJ&E-estimate is

155

estimate

156

obtained, it is compared with the measured emission rate Emeasured from CPM testing12. When

157

Emeasured is more than an order or magnitude greater than Eestimate (or EJ&E-estimate), then this might

158

indicate a significant alternative pathway, or other discrepancies between actual site conditions

159

and a simplistic VI site conceptual model. The discrepancies could also include

160

mischaracterization of soil properties and soil gas concentrations, or the presence of constituent

161

production mechanisms (e.g., daughter product production from parent decay) not accounted for

162

in the screening-level modeling.

163

Response of soil gas profiles during CPM testing: Soil gas profile response to CPM

164

testing could be different in the presence and absence of alternative VI pathways, as is illustrated

165

below for the study site. For example, for a site with only the soil VI pathway present, shallow

166

soil gas concentrations might increase or decrease during CPM testing, but they should always

167

remain lower than vapor source concentrations (these are referred to as “conforming” soil gas

168

profiles here). With the pipe flow VI pathway present and connected to a relatively high

169

concentration vapor source, it is conceivable that shallow sub-slab soil gas concentrations could

170

become greater than intermediate depth soil gas concentrations. Thus, observation of “non-

171

conforming” soil gas profiles (those that do not match the conventional conceptual model) could

172

indicate significant pipe flow VI at a site. The absence of non-conforming soil gas profiles,

173

however, does not necessarily prove the absence of a significant alternative VI pathway. For

174

example, it is unlikely that non-conforming soil gas profiles would be observed for sewer VI

175

pathways as their contaminant vapor sources are directly connected to indoor air.

7



176

Below we illustrate use of the diagnostic tools and analyses discussed above for a home

177

where a significant pipe flow VI pathway was discovered and then modified to be manipulated

178

on and off.

179

180

181

EXPERIMENTAL METHODS Data presented below were obtained over four years involving natural and controlled

182

building under-pressurization conditions, and both with the lateral pipe valve open and closed.

183

The time sequence of experimental conditions is summarized in Table 1.

184

Indoor air concentrations of chlorinated chemicals and the SF6 tracer, indoor - outdoor

185

and sub-slab soil gas - indoor pressure differentials, and external environmental conditions were

186

monitored continuously at frequencies of minutes to hours as described in Holton et al. 11, 12.

187

TCE concentrations in soil gas beneath and around the building foundation were

188

measured 25 times over four years. Soil gas samples were collected in Tedlar bags using a

189

vacuum box. TCE concentrations were quantified on-site using an SRI 8610C gas

190

chromatograph equipped with a dry electrolytic conductivity detector (DELCD). Both direct

191

injection and sorbent-concentration methods were used. The method detection limit (MDL) is 4.9

192

ppbv (26 µg/m3) for the former and 0.019 ppbv (0.1 µg/m3) for the latter.

193 194

Effective diffusion coefficients were measured at the sampling points during five of the soil gas sampling events using the method presented by Johnson et al.14 with helium as the tracer.

195

8


Page 10 of 34

Page 11 of 34

196

197


DATA REDUCTION Measured TCE emission rates to indoor air (Emeasured) were determined for CPM test

198

conditions using the Holton et al.12 approach: Emeasured = Ci x (Cotracer/Ctracer) x Qtracer, with known

199

indoor SF6 tracer release rate (Qtracer) and concentration (Cotracer) and measured indoor air tracer

200

and TCE concentrations (Ctracer and Ci). Building air exchange flow rates (QB) can also be

201

calculated from SF6 tracer release rate (Qtracer) and tracer concentration data (Ctracer) QB =

202

(Cotracer/Ctracer) x Qtracer.

203

For comparison, screening-level TCE emission estimates (Eestimate and EJ&E-estimate) were

204

generated using equations (1) – (3) and two different data reduction approaches. In both cases a

205

single “high resolution” estimate was generated using all data collected beneath the foundation

206

footprint and multiple “low resolution” estimates were generated using the data from each

207

individual 1.8 m BS sampling location exterior to the foundation. This was done to assess if

208

reliance on low resolution exterior sampling yields emission estimates similar to high resolution

209

through-the-foundation sampling, as the former is more likely to be implemented in practice than

210

the latter.

211

More specifically, Eestimate, values were generated using TCE soil gas data sets and

212

equations (1) and (2). For the high-resolution estimates, soil gas concentrations from locations 1

213

to 6 were assigned to 6 foundation footprint sub-regions with 14.1 m2 areas as shown in Figure 2.

214

The sub-slab and 1.8 m BS concentrations were used to calculate ∆Cg,i . Equation (2) was used to

215

calculate (Di,jeff /Li) by conceptualizing a three-layer soil system beneath the house, with layers of

216

uniform effective diffusion coefficients from 0-30 cm, 30-90 cm and 90-180 cm BS. Effective

217

diffusion coefficients for each sub-region (Di,jeff ) were obtained by averaging results from the

9



218

five field surveys for each depth interval. Using equation (1), low resolution estimates were

219

generated using each individual 1.8 m BS TCE exterior sampling point concentration from the

220

25 soil gas surveys collected across the 4 year study. Each low resolution estimate utilized the

221

same average effective diffusion coefficient (Di,jeff) calculated for the five Di,jeff measurements at

222

that 1.8 m BS sampling location.

223

EJ&E-estimate values were generated using equation (3) and the USEPA spreadsheet

224

implementation15 of the Johnson and Ettinger model4. As above, high resolution calculations

225

made use of the sub-foundation data and low resolution calculations employed the 1.8 m BS

226

exterior data values. For high resolution estimates, the 1.8 m BS TCE concentrations averaged

227

within the building footprint from each soil gas data set were used as the vapor source

228

concentration input. A three-layer soil system was also modeled as above with soil properties

229

selected to obtain layer-specific effective diffusion coefficients that are the same as those used

230

above for high resolution Eestimate calculations. Low resolution EJ&E-estimate values were computed

231

in a similar fashion, except with use of individual exterior sampling location data consistent with

232

the low resolution Eestimate calculations above. With respect to soil permeability, each layer was

233

assigned a generic value from the USEPA spreadsheet15 based on qualitative soil descriptions;

234

these included sand for the sub-slab layer and sandy clay for the next two layers. All model

235

inputs are summarized in Table 2.

236 237 238

RESULTS AND DISCUSSION The raw data used for calculation of measured and estimated emission rates are presented

239

in Supplemental Information Tables S1 and S2 and Figures S3 and S4. These include measured

240

TCE effective diffusion coefficients Di,jeff (Table S1), soil gas TCE concentrations (Table S2),

10


Page 12 of 34

Page 13 of 34


241

indoor air exchange flow rates QB (Figure S2) and indoor air TCE concentrations (Figure S3).

242

Figure S2 also presents daily (24-h) average pressure differentials. In brief, Di,jeff values range

243

from 0.001 to 0.02 cm2/s for tests conducted at 0.9 m BS and 1.8 m BS and for sub-slab depth

244

locations outside the foundation footprint. Beneath the foundation at the sub-slab depth the Di,jeff

245

values are consistently the largest of all locations, ranging from 0.01 to 0.03 cm2/s. The (Di,jeff /Li)

246

values calculated using equation (2) are also presented in Table S1. With respect to pressure

247

differential and QB values, it was noted that under natural and controlled pressure conditions,

248

both were not significantly influenced by the state (open/closed) of the land drain lateral control

249

valve.

250

Calculation and Comparison of Measured and Estimated TCE Emission Rates:

251

Figure 3 presents the TCE emission rates measured during CPM testing, first with the land drain

252

lateral connection open to the sub-foundation region (780 – 1045 d) and then later with it closed

253

(1071 – 1157 d). Summary statistics are presented in Table 3 for both conditions. The results for

254

780 – 1045 d were presented previously in Holton et al.12, while the latter are being published

255

here for the first time. Under both experimental conditions, the emissions were relatively

256

consistent with time, having minimal variations from day-to-day and across long time periods.

257

TCE emission rate estimates are also presented in Table 3 for comparison. There are four

258

columns capturing the different combinations possible with the two calculation approaches

259

(equations 1 and 3) and the high- and low-resolution data analysis approaches discussed above.

260

The ranges of the emission rate estimates are also presented in Figure 3 with the measured values.

261

The following observations come from a review of Table 3 and Figure 3:

11



262

•

The contribution of the alternative vapor intrusion pathway (land drain lateral valve open)

263

is clearly evident as the mean measured emission rate with the land drain lateral valve

264

open (0.18 g/d) is about two orders of magnitude greater than the mean emission rate

265

measured with the land drain valve closed (0.0013 g/d).

266

•

The emission rate measured with the land drain valve open is about two orders of

267

magnitude or more greater than any of the emission rate estimates. This supports the

268

hypothesis that an inconsistency between estimated and measured emission rates can be a

269

line of evidence for identifying alternative VI pathways, especially when the emission

270

rate measured during CPM testing is much greater than estimated values.

271

•

All mean emission rate estimates are within about 2X to 4X of the mean emission rate

272

measured with the land drain valve closed; this provides confidence in the use of simple

273

screening equations to estimate the maximum impact from the soil VI pathway.

274

•

The high-resolution method emission rate estimates span less than an order of magnitude

275

and 14 of the 21 values (67%) are within about 50% of their mean value, independent of

276

the screening calculation approach used. This suggests that only a few sampling events

277

would be required to generate a reliable emission estimate and it provides some

278

confidence in the use of the high resolution approach, even though its practicability is

279

questionable.

280

•

The less data intensive and arguably more practicable low resolution method leads to

281

emission rate estimates spanning about three orders of magnitude at this site, independent

282

of the screening calculation approach used. This variation reflects both spatial and

283

temporal variability in the soil gas concentration data. While they span a wide range, all

284

estimates are significantly less than the emission rate measured with the land drain lateral

12


Page 14 of 34

Page 15 of 34


285

valve open, so comparison of any with the measured CPM test emission rates would lead

286

to suspicion of the presence of an alternative VI pathway. These results suggest,

287

however, that practitioners should be cautious about relying on a single exterior sampling

288

location and a single sampling event when estimating soil VI pathway emission rates.

289

•

While not shown in Table 3, the mean of the estimates for each of the four exterior

290

location data sets is generally within about 50% of the mean measured emission rate

291

during CPM testing after lateral valve was closed, independent of the screening

292

calculation approach used. This suggests that reliable emission rate estimates might be

293

obtained at other sites with a small number of exterior sampling locations and a few

294

sampling events.

295

Soil gas distribution response to CPM testing: Figures 4 and 5 present representative

296

soil gas distributions across the four years of this study. These contour plots were prepared using

297

the soil gas concentrations and locations, and Surfer 12 (Golden Software, Inc.) with its Kriging

298

gridding algorithm. Each plot presents TCE concentration distributions at sub-slab (SS), 0.9 m

299

BS and 1.8 m BS depths. For location C, the ground surface is below the sub-slab elevation, so a

300

0 ppbv TCE concentration was assigned to this point when creating contours. The building

301

footprint is shown as a dashed outline on the sub-slab depth plot with the back of the house being

302

the north side of the plot.

303

Holton et al.11 characterized the indoor air concentration vs. time behavior at this house

304

under natural conditions and with an open land drain lateral valve as having “VI-active” and

305

“VI-inactive” periods. The VI-active behavior was prevalent in fall, winter and early spring,

306

while the VI-inactive behavior was prevalent in late spring and summer. Causes for VI-active

307

and –inactive periods were not identified in that work, although increasing VI activity appeared

13



308

to be related to increasing indoor-outdoor temperature difference more than any other factor.

309

Figures 4 and 5 present representative TCE soil gas distributions from VI-active and VI-inactive

310

periods, respectively. There are similarities between them at the 0.9 m BS and 1.8 m BS depths,

311

with TCE concentrations generally decreasing when moving from the north to the south (back to

312

front of the house). This reflects the influence of the sloping ground surface, which decreases in

313

elevation by about 2 m from back to front of the house, so sampling points at equivalent

314

elevations are closer to ground surface in the front of the house.

315

The soil gas profiles under natural conditions do not provide any indication that a

316

significant alternative VI pathway is present at this site. The distributions in Figures 4 and 5 are

317

both consistent with the conventional diffusion-driven soil gas VI pathway conceptualization

318

prevalent in the vapor intrusion literature4, 16. Soil gas concentrations decrease from the source to

319

the building and ground surface as expected. For example, the concentration attenuation from 1.8

320

m BS to the sub-slab depth ranges from 10-2 to 10-3 and this is comparable to the modeling

321

results for “soil VI” only conceptual models5-7.

322

Four soil gas sampling events separated by one to three months occurred during the long-

323

term CPM test. Representative results are presented in Figure 6. TCE concentration distributions

324

at 0.9 m BS and 1.8 m BS depths remain similar to those in Figures 4 and 5 under natural

325

conditions, with concentration differences at each location and depth being within 3X of values

326

in Figures 4 and 5. A significant change, however, can be seen in the sub-slab depth TCE

327

concentrations beneath the house living area (right-hand portion of the footprint). The increases

328

are 100X or greater in comparison to concentrations measured under natural conditions.

14


Page 16 of 34

Page 17 of 34

329


Under CPM conditions, the vertical distribution of soil gas concentrations is no longer

330

consistent with the conventional diffusion-driven soil gas VI pathway conceptualization.

331

Concentrations decrease from 1.8 m BS to 0.9 m BS but then increase to the sub-slab depth.

332

Sub-slab depth concentrations at some locations are now greater than 1.8 m BS near-source

333

concentrations. For example, at one central sampling location the sub-slab TCE concentration

334

was 91.1 ppbv while it was 6.6 ppbv and 43.3 ppbv for the 0.9 m BS and 1.8 m BS samples. Thus,

335

the data support the hypothesis that soil gas profiles under CPM test conditions can at some sites

336

provide an indication of a significant alternative VI pathway.

337

Once the “pipe VI” pathway was closed, CPM testing did not significantly alter the soil

338

gas distribution at this site from that observed under natural conditions. Figure 7 presents a

339

representative TCE soil gas distribution for CPM test conditions with the lateral drain valve

340

closed (no alternative VI pathway). In comparison to Figure 6, the previously elevated TCE

341

concentrations at the sub-slab depth beneath the house living area decreased after the valve was

342

closed and the deep soil gas (0.9 m and 1.8 m BS) concentrations remained relatively unchanged.

343

The soil gas profile resembles that anticipated for a soil VI-dominated setting. It is also very

344

similar to that shown in Figure 8, which presents data measured under natural conditions and

345

with the land drain valve closed.

346

Reflection on key lessons-learned and future research: The experiences and results

347

from this study illustrate that the presence of a significant alternative VI pathway is not easily

348

detected by visual observation or routine VI pathway assessment measurements under natural

349

conditions. In particular, the soil gas profiles under natural conditions conformed to typical soil

350

VI-dominated conceptual models at this site, with and without the presence of the significant

351

alternative VI pathway. The presence of the significant pipe flow VI pathway was only revealed

15



352

by data collected during CPM testing; more specifically the observations that measured emission

353

rates greatly exceeded emission rate screening estimates and soil gas profiles that changed and

354

no longer conformed to traditional soil VI conceptual models of the vapor intrusion pathway.

355

In summary, this work in addition to the work of Riis et al.8 and Pennell et al.9 suggest

356

that the following conditions might be indicative of the presence of significant pipe flow and

357

sewer VI pathways: a) VI impacts under natural or CPM testing conditions in buildings outside

358

the delineated boundaries of the vapor source(s) indicate one or more alternative VI pathways, b)

359

CPM test emission rates that greatly exceed screening-level estimates in combination with

360

conforming soil gas profiles might indicate a significant sewer VI pathway, and c) CPM test

361

emission rates that greatly exceed screening-level estimates in combination with non-conforming

362

soil gas profiles might indicate a significant pipe flow VI pathway.

363

There are a number of reasons why there should be interest in being able to quickly

364

identify significant alternative VI pathways. One is that conventional pathway characterization

365

paradigms, data analyses, and decisions have been built on a soil VI-only conceptualization of

366

the VI pathway, and these might lead to erroneous decisions when significant alternative VI

367

pathways are present. The second is that VI mitigation system design and monitoring is also

368

based on the soil VI-only conceptualization and it is not known if presumptive remedies are

369

effective when significant alternative VI pathways are present. This should be examined in

370

future research studies.

371

The proposed method was tested at a chlorinated hydrocarbon impacted site with a

372

known “pipe flow VI” pathway, and its effectiveness was well demonstrated. However, when

373

assessing petroleum hydrocarbon-impacted sites or other site conceptual models, such as sites

16


Page 18 of 34

Page 19 of 34


374

with a “sewer VI” pathway present, its effectiveness is unknown. Further research is necessary to

375

evaluate this method under different scenarios.

376

377

ASSOCIATED CONTENT

378

Supporting Information

379

Additional information as noted in the text. Supporting information is available free of charge

380

via the Internet at http://pubs.acs.org.

381

382

ACKNOWLEDGMENTS

383

This research was funded by the U.S. Department of Defense, through the Strategic

384

Environmental Research and Development Program (SERDP). The findings and conclusions in

385

this article are those of the authors and do not necessarily represent the views of the U.S. Air

386

Force.

387

388

389

AUTHOR INFORMATION

390

Corresponding Author

391

*Email: [email protected]

17



392

REFERENCES

393

1. U.S. Environmental Protection Agency. OSWER Draft Guidance for Evaluating the Vapor

394

Intrusion to Indoor Air Pathway from Groundwater and Soils (Subsurface Vapor Intrusion

395

Guidance); EPA: Washington, DC, 2002.

396 397 398 399 400 401 402

2. Interstate Technology & Regulatory Council. Vapor Intrusion Pathway: A Practical Guideline; ITRC: Washington, DC, 2007. 3. New Jersey Department of Environmental Projection. Vapor Intrusion Technical Guidance; NJDEP: Trenton, NJ, 2013. 4. Johnson, P.C.; Ettinger, R.A. Heuristic Model for the intrusion rate of contaminant vapors into buildings. Environ. Sci. Technol. 1991, 25 (8), 1445-1452. 5. Abreu, L.D.; Johnson, P.C. Effect of vapor source-building separation and building

403

construction on soil vapor intrusion as studied with a three-dimensional numerical model.

404

Environ. Sci. Technol. 2005, 39 (12), 4550-4561.

405

6. Bozkurt, O.; Pennell, K.G.; Suuberg, E.M. Simulation of the vapor intrusion process for non-

406

homogeneous soils using a three dimensional numerical model. Ground Water Monit. Rem.

407

2009, 29 (1), 92-104.

408 409 410

7. U.S. Environmental Protection Agency. Conceptual Model Scenarios for the Vapor Intrusion

Pathway. EPA: Washington, DC, 2012. 8. Riis, C. E.; Christensen, A. G.; Hansen, M. H.; Husum, H.; Terkelsen, M. Vapor Intrusion

411

through Sewer Systems: Migration Pathways of Chlorinated Solvents from Groundwater to

412

Indoor Air. Presentation at the 7th Battelle International Conference on Remediation of

413

Chlorinated and Recalcitrant Compounds, Monterey. 2010.

18


Page 20 of 34

Page 21 of 34


414

9. Pennell, K. G.; Scammell, M. K.; McClean, M. D.; Ames, J.; Weldon, B.; Friguglietti, L.;

415

Suuberg, E. M.; Shen, R.; Indeglia, P. A.; Heiger-Bernays, W. J. Sewer gas: An indoor air

416

source of PCE to consider during vapor intrusion investigations. Ground Water Monit. Rem.

417

2013, 33 (3), 119-126.

418 419

10. U.S. Environmental Protection Agency. Engineering Issue: Indoor Air Vapor Intrusion

Mitigation Approaches. EPA: Washington, DC, 2008.

420

11. Holton, C.; Luo, H.; Dahlen, P.; Gorder, K. A.; Dettenmaier, E. M.; Johnson, P. C. Temporal

421

variability of indoor air concentrations under natural conditions in a house overlying a dilute

422

chlorinated solvent groundwater plume. Environ. Sci. Technol, 2013, 47, 13347-13354.

423

12. Holton, C.; Guo, Y.; Luo, H.; Dahlen, P.; Gorder, K. A.; Dettenmaier, E. M.; Johnson, P. C.

424

Long-Term evaluation of the controlled pressure method for assessment of the vapor

425

intrusion pathway. Environ. Sci. Technol, 2015, 49, 2091-2098.

426

13. McHugh, T. E.; Beckley, L.; Bailey, D.; Gorder, K.; Dettenmaier, E.; Rivera-Duarte, I.;

427

Brock, S.; MacGregor, I. C. Evaluation of vapor intrusion using controlled building pressure.

428

Environ. Sci. Technol, 2012, 46, 4792-4799.

429

14. Johnson, P. C.; Bruce, C.; Johnson, R. L.; Kemblowski, M. W. In situ measurement of

430

effective vapor-phase porous medium diffusion coefficient. Environ. Sci. Technol, 1998, 32,

431

3405-3409.

432

15. U.S. Environmental Protection Agency. Johnson and Ettinger (1991) Model for Subsurface

433

Vapor Intrusion into Buildings (3-Phase System Models and Soil Gas Models); EPA:

434

Washington, DC, 2000; http:// www.epa.gov/oswer/riskassessment/airmodel/

435

johnson_ettinger.htm.

19



436 437

16. Atteia, O.; Hohener, P. Semianalytical model predicting transfer of volatile pollutants from groundwater to the soil surface. Environ. Sci. Technol, 2010, 44, 6228-6232.

438

20


Page 22 of 34

Page 23 of 34

439

440


FIGURES AND TABLES Table 1. Building operation conditions and indoor air sampling methods. Period Building pressure condition Lateral pipe valve Mean of the 24-h averaged pressure differentials (outdoor - indoor)

120 d to 740 da

1071 d to 1157 d Controlled underpressurization Closed

1157 d +

Open (NI)

780 d to 1045 db Controlled underpressurization Open (NI)

0.02 ± 0.9 Pa

11 ± 4 Pa

12 ± 1 Pa

0.7 ± 2 Pa

Natural

Natural Closed

Lower level: Lower level: Lower level: TDTD-GC/MS TD-GC/MS GC/MS Attic: TD-GC/MS Attic: TD-GC/MS Attic: GC/ECD Attic: GC/ECD a Note: Between 740 d to 780 day, blower system was installed and tested. b Note: Blower speed changed from “High” to “Low” at 1046 d, and switched back to “High” at 1071 d. NI – butterfly not installed on the land drain lateral during this phase of the study. Indoor air sample location: analysis method

Lower level: TD-GC/MS

441 442

21



443 444

Table 2. Johnson and Ettinger Model USEPA Spreadsheet15 Inputs Depth below grade to bottom of enclosed floor [cm] 30 Soil gas sampling depth below grade [cm] 210 Average soil temperature, C 25 Thickness of soil stratum A [cm] 60 Thickness of soil stratum B [cm] 60 Thickness of soil stratum C [cm] 90 Enclosed floor thickness [cm] 10 Building under-pressurization [Pa] 5 Enclosed space floor length [cm] 1140 Enclosed space floor width [cm] 740 Enclosed space floor height [cm] 210 Floor-wall crack with [cm] 0.1 Air exchange rate [1/h] 0.5 2 Stratum A soil type permeability (sand) [cm ] 1.02 ×10-7 Stratum B soil type permeability (sandy clay) [cm2] 1.79 ×10-9 Stratum C soil type permeability (sandy clay) [cm2] 1.79 ×10-9 High Resolution Approach Effective Diffusion Coefficient [cm2/s] Stratum A 1.42 ×10-2 Stratum B 4.52×10-3 Stratum C 3.80×10-3 Low Resolution Approach Effective Diffusion Coefficients [cm2/s] are provided in Supplemental Information Table S1

445

446

447

22


Page 24 of 34

Page 25 of 34

448 449


Table 3. Summary statistics for measured and estimated TCE emissions rates.

Pressure condition Land Drain Lateral Valve Condition Mean Maximum Minimum 90th Percentile 10th Percentile

Measured TCE Emission Rates [g/d] Controlled Pressure Controlled Pressure Method Test* Method Test* Open Closed [780 – 1040 d] [1071 – 1157 d] 0.18 1.3× 10-3 0.29 6.3× 10-3 0.09 1.2× 10-4 0.26 4.5× 10-3 0.12 2.8× 10-4

TCE Emission Rate Estimates [g/d] High Resolution Data Low Resolution Data Reduction Approach** Reduction Approach*** Eestimate

EJ&E-estimate

Eestimate

EJ&E-estimate

8.0× 10-4 1.9× 10-3 1.3× 10-4 1.3× 10-3 3.2× 10-4

2.7× 10-4 6.2× 10-4 4.6× 10-5 4.2× 10-4 1.0× 10-4

7.9× 10-4 4.6× 10-3 1.3× 10-6 2.5× 10-3 5.6× 10-6

3.3× 10-4 1.7× 10-3 5.4× 10-7 9.5× 10-4 2.4× 10-6

450

451 452 453 454 455 456 457 458

* ** ***

- summary statistics of all daily 24-h average emission rate values determined during the measurement period, - summary statistics of high resolution estimates; one high resolution estimate calculated for each of 20 soil gas snapshot sampling events - summary statistics of low resolution estimates; four low resolution estimates calculated for each soil gas snapshot sampling events (one for each of four exterior sampling locations)

23



459 460 461

Figure 1a. Conventional vapor intrusion pathway conceptualization showing only the “soil VI” pathway.

462

463 464 465

Figure 1b. Vapor intrusion pathway conceptualization showing the “pipe flow VI” and “sewer VI” alternative VI pathways.

466

24


Page 26 of 34

Page 27 of 34


467

468 469

Figure 2. Schematic of building footprint, sample locations and lateral land drain pipe with

470

valve installed for this study. Red dashed lines delineate sub areas used for high-resolution

471

screening-level emission estimates.

472

25



473 474

Figure 3. Measured 24-h average TCE emission rates for the four building conditions tested with

475

ranges of screening level model estimates, including: A) Eestimate using the high-resolution

476

approach, B) Eestimate using the low-resolution approach, C) EJ&E-estimate using the high-resolution

477

approach, and D) EJ&E-estimate using the low-resolution approach. Horizontal bars on the estimated

478

emission rate ranges indicate the maximum, mean and minimum modeling results (ordered from

479

top to bottom).

480

26


Page 28 of 34

Page 29 of 34


N

ppbv

ND: None Detected. N/A: No data available.

481 482

Figure 4. Representative TCE soil gas concentrations collected from t=368 d to 370 d during a

483

VI-active period under natural conditions with the land drain lateral valve open. SS, 0.9 m BS

484

and 1.8 m BS contours are shown from top to bottom. The bold dashed line in the SS surface

485

delineates the building perimeter.

27



Page 30 of 34

N

ppbv


486 487

Figure 5. Representative TCE soil gas concentrations collected from t=514 d to 516 d during a

488

VI-inactive period under natural conditions with open land drain lateral valve. SS, 0.9 m BS and

489

1.8 m BS contours are shown from top to bottom. The bold dashed line in the SS surface

490

delineates the building perimeter.

28


Page 31 of 34


N

ppbv


491 492

Figure 6. Representative TCE soil gas concentrations collected from t=910 d to 911 d during

493

CPM conditions with open land drain lateral valve. SS, 0.9 m BS and 1.8 m BS contours are

494

shown from top to bottom. The bold dashed line in the SS surface delineates the building

495

perimeter.

29



Page 32 of 34

N

ppbv


496 497


498

CPM conditions with closed land drain lateral valve. SS, 0.9 m BS and 1.8 m BS contours are

499


500

perimeter

30


Page 33 of 34


N

ppbv


501 502


503

natural conditions with closed land drain lateral valve. SS, 0.9 m BS and 1.8 m BS contours are

504


505

perimeter.

31



Page 34 of 34

Air flow due to pressure differential

“Sewer VI”

“Pipe flow VI”

Impacted Groundwater ACS Paragon Plus Environment

This picture is adopted from Figure 1b in the manuscript with minor changes

Identification of Alternative Vapor Intrusion Pathways Using Controlled

Recommend Documents