Accepted Manuscript

Effects of surface washing on the mitigation of concrete corrosion under sewerconditions

Xiaoyan Sun, Guangming Jiang, Tsz Ho Chiu, Mi Zhou, Jurg Keller, Philip L. Bond

PII: S0958-9465(16)30025-7

DOI: 10.1016/j.cemconcomp.2016.02.013

Reference: CECO 2607

To appear in: Cement and Concrete Composites

Received Date: 31 July 2015

Revised Date: 25 November 2015

Accepted Date: 10 February 2016

Please cite this article as: X. Sun, G. Jiang, T.H. Chiu, M. Zhou, J. Keller, P.L. Bond Effects of surfacewashing on the mitigation of concrete corrosion under sewer conditions, Cement and ConcreteComposites (2016), doi: 10.1016/j.cemconcomp.2016.02.013.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Effects of surface washing on the mitigation of concrete corrosion under 1

sewer conditions 2

Xiaoyan Sun, Guangming Jiang*, Tsz Ho Chiu, Mi Zhou, Jurg Keller, Philip L. Bond 3

Advanced Water Management Centre, Gehrmann Building, Research Road, The University 4

of Queensland, St. Lucia, Queensland 4072, Australia 5

* Corresponding author. Tel.: +61 (0) 7 334 67205; fax: +61 (0) 7 336 54726. 6

Email: x.sun@awmc.uq.edu.au; g.jiang@awmc.uq.edu.au; tsz.chiu@uqconnect.edu.au; 7

mi.zhou2@uq.net.au; phil.bond@awmc.uq.edu.aul; j.keller@awmc.uq.edu.au 8

9

10

11

12

13

14

15

16

17

18

19

20

21

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

Abstract 22

This study systematically investigated the potential of mitigating sulfide induced sewer 23

concrete corrosion by surface washing. Washing interrupted the corrosion activity of concrete 24

coupons by increasing the surface pH and decreasing the H2S uptake rates (SUR). The SUR 25

recovered to the level prior to washing within 60-140 days. The slowest recovery rate was 26

from the most severely corroded coupon. However, no significant difference was observed 27

for concrete mass loss of the washed and unwashed coupons after 54 months. The results 28

suggest that frequent washing at short intervals of a few months might be needed to control 29

corrosion over a long term. 30

Key words: Sewer; Concrete; Corrosion control; Sulfide; Washing 31

Abbreviations 32

BSE Backscattered electron

MLA Mineral Liberation Analyzer

SEM Scanning electron microscopy

SOB Sulfide oxidizing bacteria

SUR Sulfide uptake rate

33

34

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

1. Introduction 35

Urban sewer networks collect and transport domestic and industrial wastewater (sewage) to 36

centralized facilities for treatment prior to discharge of the treated effluent into receiving 37

waters. In industrialised countries, the establishment of sewer networks has been achieved 38

through continuous public investment for more than a century. However, the deterioration of 39

sewer concrete pipes, caused by sulfide induced concrete corrosion, is a major economic and 40

infrastructure burden in many countries [1]. It shortens the service life of sewer pipes and 41

results in expensive replacement of prematurely failed structures. For example, the estimated 42

corrosion caused cost in wastewater catchment infrastructure per year is around $20 billion in 43

the USA [1]. As such, the development of effective technologies to mitigate corrosion is 44

imperative for extending the service life of sewer pipes and reducing the huge annual 45

maintenance expense. 46

Sulfide is produced within the anaerobic regions of the sewer, mainly in the fully filled 47

pressure pipes [2, 3]. During the pumping of sewage from a pressure pipe to a gravity pipe, 48

the sulfide is partially released into the sewer atmosphere of the gravity pipe [4]. In the 49

presence of oxygen, gaseous H2S is taken up by the moist concrete surface exposed to gas 50

phase and there it is oxidized to sulfate and other sulfur species by sulfur oxidizing bacteria 51

(SOB) [5, 6]. The acid formed during sulfide oxidation will react with alkaline compounds in 52

the concrete and form corrosion products [7]. This process causes decreased concrete surface 53

pH (from 12-13 to below 1-2), loss of concrete mass, cracking and weakened sewer structure 54

[6, 8]. A corrosion layer develops on the concrete surface that is largely gypsum, this has a 55

soft texture and can be several mm in thickness. Severe corrosion will eventually result in the 56

structural collapse of the sewer network. 57

To achieve effective corrosion control, various technologies have been developed. These 58

technologies mitigate the corrosion through either preventing the build up of sulfide in the 59

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

liquid phase or by preventing the concrete surface from H2S attack [9-14]. In particular, one 60

treatment is to directly remove the corrosion layer by washing the concrete surface. This is 61

considered a cost efficient approach and hence is of great interest to water utilities [15]. 62

However, surprisingly there has been no comprehensive evaluation on the effectiveness of 63

surface washing for the purpose of mitigating concrete corrosion. 64

Currently, there is disagreement on whether the formed concrete corrosion layer accelerates 65

or slows the corrosion process [16-18]. It is proposed that the thick corrosion layer acts as a 66

barrier that slows the sulfuric acid attack onto the intact concrete surface, and consequently it 67

is suggested that the removal of the corrosion layer would accelerate the corrosion activity 68

[16, 18]. In contrast it is argued that the flushing of the concrete pipe with sewage removes 69

the corrosion layer and this disturbs and removes the low pH environment of the sulfur 70

oxidizing bacteria [15, 17, 19]. Therefore, there are contradictory theories of the effect that 71

surface washing may have on the corrosion processes and it is not clear whether washing 72

would mitigate the corrosion activity. 73

It is indicated that gentle flushing is not a successful control measure. Two early studies 74

flushed concrete samples with wastewater for a few seconds and repeated it 1-3 times daily or 75

weekly [15, 19]. This immediately increased the concrete surface pH, although, this returned 76

to its low level found prior to flushing, in a few hours. Likely the gentle flushing only 77

removed soluble components (e.g. acid) but not the corrosion layer containing the acid-78

generating microbes. They measured the surface pH and indirectly measured the corrosion 79

rates from the polarization resistance. It was postulated that heavy and frequent washing is 80

necessary to effectively reduce corrosion rates (in terms of producing high surface pH) [15, 81

19]. A more recent study washed concrete pipes using a hose and a brush, but this only 82

temporarily reduced the corrosion activity (measured as sulfide oxidation rate). In just 10 83

days the sulfide oxidizing rate was seen to increase and by 30-40 days this had reached the 84

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

pre-washing level [20]. In this study the bioactive layer would have been removed from 85

concrete surface, although the effect on corrosion activity was only measured through the 86

sulfide oxidation rates [20]. A comprehensive and systematic study evaluating the effect of 87

high pressure washing on the concrete surface pH, sulfide oxidation rate and concrete loss in 88

the long term is required to delineate the overall impacts of surface washing for mitigating 89

sewer corrosion. 90

The aim of this study was to assess the effectivity of high pressure washing on controlling 91

corrosion of concrete in corrosive sewer environment. To simulate the concrete exposure 92

condition in real sewers, concrete coupons were exposed to the gas phase of laboratory 93

chambers with controlled levels of relative humidity, temperature and H2S concentration. 94

After being exposed in the chamber for specified periods, the coupons with various corrosion 95

levels were washed using a high pressure washer with water. The sulfide oxidation activities 96

of the coupons were measured and concrete losses of washed and unwashed concrete 97

coupons were compared in a long term study of over 4.5 years. The experiment enabled a 98

comprehensive understanding and evaluation of the effect of washing on corrosion control. 99

2. Materials and methods 100

2.1. Concrete coupons 101

Two types of concrete coupons, i.e. fresh and pre-corroded concrete coupons, were prepared 102

and established into the laboratory corrosion chambers as described [6]. To briefly explain, 103

the fresh concrete coupons were cut from a newly manufactured spun cast standard 104

reinforced concrete sewer pipe, which has 1.2 m internal diameter and a standard strength of 105

load class 2 (HUMES, Australia). The concrete had a minimum of 400 kg m-3 cementitious 106

content and 45-50% aggregate content by volume. The standard sewer pipes are suitable for 107

most sewer systems. The pre-corroded concrete coupons were cut from reinforced concrete 108

that previously served as a sewer wall for 70 years from Sydney Water Corporation, Australia. 109

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

The dimensions of each coupon were approximately 100 mm (length) × 70 mm (width) × 70 110

mm (thickness). After cutting, the coupons were washed in fresh water and then dried at 111

60 °C for 3 days to achieve similar and stable initial water content. 112

The internal surface of the sewer pipe, was designed as the coupon surface to be exposed to 113

H2S. After cutting, the coupons were embedded in epoxy (FGI R180 epoxy & H180 hardener) 114

in specially designed stainless steel frames. Each frame containing one fresh and one pre-115

corroded coupon and forms a concrete coupon pair (Figure 2). The stainless steel frame 116

provided a reference point for determining the change in thickness due to corrosion. The 117

coupons were exposed to the gas phase of corrosion chambers as described in section 2.2. 118

2.2. Corrosion chambers and exposure conditions 119

Four identical corrosion chambers were constructed of glass panels of 4 mm thickness to 120

achieve a controlled environment simulating that of real sewers and the detailed description is 121

made elsewhere [6, 21]. Briefly, the dimensions of the chambers were 550 mm (length) × 450 122

mm (width) × 250 mm (height). The conditions in the chambers were controlled at four H2S 123

levels (i.e. 5 ppm, 10 ppm, 25 ppm and 50 ppm) with the gas-phase temperatures at about 124

25 °C and the relative humidity level at 100%. Each chamber contained 2.5 L domestic 125

sewage (characteristics of sewage was reported in Table S1 in Supplementary Information) 126

collected from a local sewer pumping station which was replaced every fortnight. Coupons 127

were exposed to the gas phase of the chambers with the exposure surface facing downwards 128

about 110 mm above the sewage. This arrangement simulates the crown area of sewer pipe, a 129

region that is highly susceptible to corrosion [16]. 130

To control the H2S gaseous concentrations inside the chamber at the specified level, Na2S 131

solution was pumped (Bio-chem fluidics, model: 120SP2440-4 TV) into a plastic container 132

partially filled with hydrochloric acid (16%). The H2S concentrations were monitored using a 133

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

H2S gas detector (OdaLog Type 2, App-Tek International Pty Ltd, Brendale, Australia) with a 134

range between 0 and 200 ppm. A programmable logic controller was applied to monitor the 135

H2S concentration and to trigger the pump for Na2S addition to maintain the H2S 136

concentration at the specified level. The chambers were stored in a cabinet and the 137

temperature of the sewage was controlled by recirculating temperature controlled water 138

through glass tubes submerged in the sewage. Through this way, the relative humidity was 139

controlled at 100% for all the chambers. 140

2.3. A test to monitor the corrosion recovery after high pressure washing 141

This test investigated the recovery of corrosion activity of corroding concrete coupons after 142

the high pressure washing. The coupons used were from the 5 ppm (F5 and P5, namely the 143

fresh and pre-corroded coupon respectively) and 50 ppm (F50 and P50, namely the fresh and 144

pre-corroded coupon respectively) H2S level chambers exposed to 100% relative humidity 145

and 25 °C for 45 months. In this test the H2S uptake rate (SUR, a previously proven indicator 146

of corrosion activity [22]), sulfur species in the corrosion layers and surface pH were 147

measured (see details in section 2.5). 148

The washing was carried out on the coupon exposure surface using a high pressure washer 149

(Karcher K 5.20 M, 12 MPa). Each coupon surface was washed by 4-8 L of deionized water. 150

H2S uptake tests and surface pH measurements were carried out immediately before and after 151

the high pressure washing, as described in section 2.5.1 and 2.5.2, respectively. The washed 152

coupons were returned to the original exposure chamber for re-exposure. This was followed 153

by weekly measurements of SUR for 1-4 weeks, and monthly measurements of SUR for 1-6 154

months untill the H2S uptake rate recovered to the pre-washing levels. Both elemental sulfur 155

and sulfate on the fresh coupon surface were determined whereas only sulfate on the pre-156

corroded coupon was determined as the level of elemental sulfur was insignificant according 157

to previous study by Jiang et al. [6] and preliminary analysis (see section 2.5.3). 158

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

2.4. A test to determine the concrete loss of washed and unwashed coupons 159

This test measured the actual concrete loss of 28 coupons (Figure S1 & S2 in Supplementary 160

Information) from two corrosion chambers at different gaseous H2S concentrations, i.e. 10 161

ppm and 25 ppm after 54 months of exposure. There were 7 sets of coupons in each chamber 162

(i.e. 7 fresh and 7 pre-corroded coupons). Each coupon set, No. 1-6 were subjected to one 163

high-pressure washing event during the exposure period in the chambers, this occurring after 164

6, 12, 18, 24, 34, and 44 months exposure respectively. After the wash the coupon sets were 165

returned to their respective corrosion chambers for the remainder of the exposure period. 166

Then after 54 months exposure in the chambers all coupon sets were removed and subjected 167

to the high pressure washing and the coupon concrete loss was determined as described in 168

section 2.5.4. Coupon surface pH before and immediately after washing was determined as 169

described in section 2.5.2. 170

2.5. Analytical methods 171

2.5.1. H2S uptake rate (SUR) 172

Coupon pairs were retrieved from the corrosion chambers for the SUR measurements. Each 173

fresh and pre-corroded coupon of the coupon pair was sealed in the two separated 174

compartments of the H2S uptake reactor, with the gas phase environment controlled at 100% 175

relative humidity and temperature of 25 °C as previously described [22]. To measure the 176

SUR of a coupon at its historical exposure H2S concentration in the corrosion chamber, 177

gaseous H2S was injected into each compartment to about 10-12 ppm higher than historical 178

exposure H2S level. The changes of H2S level inside the compartment were monitored. The 179

obtained H2S uptake profiles were used to calculate the SUR of coupon at the historical 180

exposure H2S level as described previously [22]. The SUR of each coupon was determined as 181

the average of 3-5 repeated measurements. 182

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

2.5.2. Surface pH 183

The coupon surface pH was measured using a flat surface pH electrode (Extech PH150-C 184

concrete pH kit, Extech Instruments, USA). The pH meter was allowed to obtain a steady 185

reading of the coupon surface after being wetted by 1 ml of milliQ water. Four independent 186

measurements were carried out on four randomly selected spots to calculate the average value. 187

2.5.3. Sulfur species 188

For sulfate measurement of fresh coupons, a known surface area of the fresh coupon was 189

scraped using a clean scalpel blade. The sample was dispersed into sulfide anti-oxidant buffer 190

solution [23]. For pre-corroded coupons, the wash-off water was homogenized using a 191

magnetic mixer for 2 h and then subsamples were taken into the sulfide antioxidant buffer 192

solution. The prepared solutions were transferred into air-tight vials for the measurement of 193

sulfate using a Dionex ICS-2000 IC with an AD25 absorbance (230 nm) and a DS6 heated 194

conductivity detector (35 °C). 195

The analysis of elemental sulfur was based on converting elemental sulfur to thiosulfate at 196

high pH [24]. A known surface area was scraped from the fresh coupon using a clean scalpel 197

blade. Soluble sulfur species were removed through mixing the sample with Milli-Q water in 198

a 10 ml vial, centrifuation for 10 min at 10, 000 rpm, and removal of the supernatant. The 199

obtained pellet was washed 3 times with Milli-Q water. The pellet was resuspended in 10 ml 200

of water mixed with 1.5 ml of Na2SO3 and 0.15 ml of 1 M NaOH, and then incubated in an 201

orbital shaker (120 rpm) overnight at 60 °C for the conversion of elemental sulfur to 202

thiosulfate. After incubation, samples were cooled to room temperature and centrifuged at 10, 203

000 rpm for 10 min. The thiosulfate concentration of the supernatant was measured by ion 204

chromatography (the same method as shown in the previous paragraph) and this value was 205

used to calculate the concentration of elemental sulfur according to the reaction stoichiometry. 206

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

2.5.4. Concrete loss 207

Five photos of each coupon surface prior to exposure in the corrosion chamber and after high 208

pressure washing were taken to determine the concrete loss using photogrammetry [25]. The 209

photogrammetry generates a 3D image of the exposed coupon surface and from that the 210

average surface height of the coupon relative to the stainless steel frame was determined. The 211

decrease of coupon thickness due to corrosion was then calculated by subtracting the surface 212

height after washing from the surface height prior to exposure. This technique provides an 213

accurate change in coupon thickness be determined irrespective of the surface roughness. 214

Additionally, it also provides a detailed record of the spatial distribution of the losses that 215

occurred. 216

2.5.5. Mineral analysis 217

To determine the bulk composition of concrete, advanced mineral analysis was performed on 218

the thin sections prepared from both fresh and pre-corroded concrete coupons. 219

For each concrete coupon, a thin section was prepared by cutting through the coupon at 45° at 220

a sample preparation lab operated by Petrographic International Pty Ltd (Australia) [26]. 221

Cutting was performed at the exposure surface to obtain a section of 2 cm × 2 cm. Then, the 222

sections were ground using a fixed diamond lap to the thickness of about 2 mm, and further 223

ground using 3 μm diamond slurry (with an aliphatic hydrocarbon base as a suspension media) 224

on ceramic lap, and 1 μm diamond on a textile cloth. The prepared sections were then coated 225

with a layer of carbon approximately 25 nm thick, using a JEOL JEE-420 vacuum evaporator. 226

The thickness of carbon was determined by a copper standard. 227

The advanced mineral analysis was performed on the prepared concrete sections using 228

Mineral Liberation Analyzer (MLA) [26]. MLA is an automated scanning electron 229

microscopy (SEM)-based mineralogical characterization tool. It uses a combination of 230

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

backscattered electron (BSE) intensity and X-ray analysis to identify minerals or phases 231

present in particles of the prepared concrete section. During measurement the BSE image is 232

typically used to identify individual particles and define the boundaries of mineral phases, an 233

X-ray is then obtained from each of the segmented phases which is then used during offline 234

processing to identify the mineral or phase based on its chemical composition. In cases where 235

phases cannot be separated by BSE gray levels, X-ray mapping at an accelerating voltage of 236

25 kv can be used. The output from a measurement includes the sample BSE image, a false-237

tool classified particle map and a database which contains particle and grain based data 238

(Figure S3). 239

3. Results and discussion 240

3.1. Concrete composition 241

The MLA mapping results from one section of fresh coupon and one section of pre-corroded 242

coupon are presented in Figure 1A and 1B, respectively. The distribution of major mineral 243

compounds on the surface of the concrete sections was determined. Then, the weight 244

percentages of the mineral compounds on the surface of the section were determined. 245

[Insert Figure 1] 246

Quartz, representing the area of aggregates and sand particles, accounts for around 57.3% and 247

47.8% of the fresh and pre-corroded concrete, respectively. The second most abundant 248

mineral on both the fresh and pre-corroded concrete was calcium silicate (i.e. the primary 249

hydrated cement), at about 13.1 % and 22.2 %, respectively. The rest major compounds are 250

other calcium/iron/potassium aluminium silicate (e.g. amphibole, orthoclase, and plagioclase), 251

which are also cement hydration products. The results suggest that the major mineral 252

compounds of the pre-corroded and fresh coupon were similar, with comparable abundance 253

of these components. In addition, there is clear difference between the parts near the exposure 254

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12

surface and the intact parts of the pre-corroded concrete section (Figure 1B) whereas there is 255

no such difference on the fresh concrete section (Figure 1A). The pre-corroded coupon was 256

found to be corroded with the presence of corrosion products, such as gypsum (at the weight 257

percentage of ~1.8%) and ettringite (at the weight percentage of ~3.3%), around the exposure 258

surface area (Figure 1B). This implies that significant sulfide induced corrosion occurred in 259

the pre-corroded coupon but no obvious corrosion occurred in the fresh coupon. 260

3.2. Recovery of corrosion activity after high pressure washing 261

3.2.1. Visual inspection 262

Visual comparison of the concrete coupons before and after the high pressure washing was 263

performed (Figure 2). It was evident there was limited corrosion on the coupons exposed at 5 264

ppm H2S. In contrast, a porous and whitish layer of corrosion product was evident on the 265

coupons exposed at 50 ppm H2S. This layer was soft and loosely bound material with little 266

mechanical strength, which is most likely the corrosion products, mainly gypsum and 267

ettringite, formed through the reactions between sulfuric acid and alkaline compounds in the 268

concrete [7]. For both the 5 ppm exposed coupons (F5 and P5), washing caused no obvious 269

change to the surface by visual inspection. However, washing removed a significant 270

corrosion layer from the 50 ppm coupons, especially from the pre-corroded coupon. In 271

addition, after washing away of the surface corrosion layer, there were still remaining 272

solidified corrosion layer on the surface of coupon P5, F50 and P50 (Figure 2C & 2D). This 273

is similar to a previous study showing that washing only removed the loosely bound 274

corrosion products from the concrete pipe [20]. The yellowish and brownish color of this 275

remaining layer after washing is likely due to the formation of elemental sulfur or iron rust 276

[26]. 277

[Insert Figure 2] 278

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13

3.2.2. Surface pH and sulfur species 279

Before exposure in the corrosion chambers, the fresh and pre-corroded coupon surface pH 280

was about 10.6 and 8.0 respectively. The surface pH of the 5 ppm and 50 ppm H2S exposed, 281

fresh and pre-corroded coupons, clearly indicates that the four coupons were at different 282

stages of corrosion (Figure 3). F5 was still at a very early stage of corrosion, with a slight 283

surface pH neutralisation, which is likely caused through acidification by CO2 and H2S [8]. 284

Low levels of sulfur species were detected on the F5 coupon surface (Table 1). Coupons P5 285

and F50 were likely at an intermediate stage of corrosion, where biological acid production 286

(e.g. sulfate at the levels of 20-120 g m-2 shown in Table 1) occurred to lower the surface pH 287

to 4-5 [27, 28]. Subsequently, coupon P50 was at an advanced stage of corrosion with pH at 288

around 2, suggesting that acidophilic microorganisms were well established and prevalent on 289

the coupon [29]. On this coupon the biogenic sulfuric acid produced here (the measured 290

sulfate concentration was 282.9 g m-2, as shown in Table 1) would react actively with 291

cementitious materials of the concrete causing formation of the observed corrosion layer 292

(Figure 2B). 293

[Insert Figure 3] 294

[Insert Table 1] 295

The surface pH of the coupons increased by various extents after washing (Figure 3). The rise 296

of surface pH caused by washing was more obvious on the 50 ppm exposed coupons than on 297

the 5 ppm coupons. For the 50 ppm coupons, the surface pH increased by 2.9 and 2.0 units on 298

the fresh and pre-corroded concrete, respectively. The change for 5 ppm exposed coupons 299

was only around 0.2 for both the fresh and pre-corroded concrete. This difference was due to 300

the higher level of corrosion occurring on the 50 ppm coupons, and this acidic corrosion layer 301

having a greater difference in pH to that of the concrete surface measured after washing. It is 302

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14

shown in corrosion layers that the pH increases abruptly from the surface of the corrosion 303

layer to the concrete surface after washing. The concrete surface pH of these four coupons 304

after washing varied between 5 and 10 (Figure 3). Obviously, in our study the high pressure 305

washing did not completely strip off the corrosion layer as the measured pH was still below 306

that of intact concrete, and this agrees with the observation of remained corrosion layer on 307

concrete surface after washing (Figure 2C & 2D). 308

The increase of pH after washing is consistent with the findings reported previously [15, 19, 309

26]. In the study reported by Mansfeld, Shih [15], concrete samples were set up in a corrosion 310

chamber of a local sewer to monitor the effect of periodic flushing on the surface pH. 311

Flushing the severely corroded concrete for 8 s using sewage increased the surface pH from 312

below 1 to levels below 3. Similar results were obtained by Islander, Devinny [19]. However, 313

a rapid (i.e. within a few hours) re-establishment of low pH environment was found in both 314

cases probably due to the continuous acid production by sulfide oxidizing bacteria. 315

3.2.3. Sulfide uptake activity 316

The sulfide uptake rates were determined before (SURb) and after (SURr) washing the 317

coupons (Figure 4A). For all concrete coupons, the SUR decreased immediately after 318

washing probably due to the loss of SOB and thus decrease of the microbial sulfide oxidizing 319

activity. The remained SUR after washing is likely driven by chemical sulfide oxidation and 320

the residual microbial catalysed sulfide oxidation. The largest decrease of SUR, measured 321

immediately after the high pressure washing, was about 60% on the P50 coupon. The 322

smallest such decrease of about 35% occurred for the F50 coupon. The larger decrease of 323

SUR, measured on the more severely corroded P50 coupon, was probably due to the higher 324

degree of SOB activity on that corrosion layer prior to washing. The decrease of SUR for the 325

F50 coupon, albeit smaller, indicates a poorer removal of the corrosion layer, as we detected 326

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15

visually (Figure 2). The decrease immediately after washing was positively correlated to the 327

pre-washing corrosion levels. 328

[Insert Figure 4] 329

For the 5 ppm exposed coupons, the SUR of F5 recovered to prewashing levels (recovery 330

ratio ≥1) after 59 days, which is 25 days shorter than the recovery of P5 (Figure 4). The 331

difference between the full recovery time for F50 and P50 was 65 days, with F50 reaching 332

full recovery within 70 days. The recovery of SUR was most likely due to the re-growth of 333

sulfide oxidizing bacteria. The quick recovery for F50 again possibly indicates the 334

incomplete removal of the corrosion layer. Overall, it is clear that the recovery period was 335

positively correlated to the pre-washing corrosion levels, i.e. the higher the pre-washing 336

corrosion extent, the longer the full recovery period after the washing treatment. 337

The SUR recovery of the corroded coupons of our study were slower than those in a study 338

which observed full recovery in 30-40 days [20]. The previous study intermittently supplied 339

gaseous H2S up to 1000 ppmv to the corroded concrete pipe (surface pH below 1-2), which is 340

much higher than typical H2S levels in real sewers. Our study used gaseous H2S 341

concentrations of up to 50 ppm, which is much more representative of sewer conditions. Also, 342

we used deionized water to avoid unwanted interference from the washing liquid. In contrast, 343

the previous study using sewage would likely add nutrients and trace minerals to residual 344

bacteria on the concrete surface. Consequently, the lower recovery rates observed in this 345

study could be due to both the lower H2S levels used for exposure and the different washing 346

methods. 347

3.3. Long term concrete loss of washed coupons after re-exposure 348

The effect of washing on the concrete loss during the long term exposure to corrosive sewer 349

conditions was determined (Figure 5A and B). After a total exposure period of 54 months, the 350

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16

corrosion layer was clearly observed on the surface of the coupons (Figure S1&S2 in 351

Supplementary Information). For the fresh coupons exposed to both 10 ppm and 25 ppm H2S, 352

the average concrete losses from the washed coupons (coupons numbered 1 to 6) after 54 353

months were limited, i.e. 0.86 ± 0.16 mm and 0.64 ± 0.18 mm, respectively (Figure 5A). Also, 354

there was no significant difference of concrete loss between the washed (No. 1-6) and the 355

non-washed (No. 7) fresh coupons (Figure 5A). The concrete loss of the pre-corroded 356

coupons after the 54 months of H2S exposure was determined (Figure 5B). The average 357

concrete loss for the high pressure washed pre-corroded coupons (numbered 1-6) was 3.33 ± 358

0.65 mm and 5.92 ± 0.67 mm for the 10 ppm and 25 ppm H2S exposure concentrations 359

respectively. The unwashed pre-corroded coupons had concrete losses of 2.85 mm and 6.67 360

mm for the 10 ppm and 25 ppm H2S exposure conditions respectively. These losses were 361

almost within the 95% confidence ranges, i.e. 2.81 – 3.84 mm and 5.39 – 6.46 mm, for the 362

washed pre-corroded coupons exposed to 10 ppm and 25 ppm gaseous H2S concentrations, 363

respectively. It is thus concluded that the high pressure washing did not cause significant 364

differences to the concrete loss from the fresh coupons at both levels of H2S exposure and 365

from the pre-corroded sewer concrete exposed to 10 ppm H2S (P<0.05). However, a slight 366

decrease in concrete loss caused by the high pressure washing was observed on the 25 ppm 367

H2S exposed, pre-corroded coupons. These results were also supported by the similar surface 368

pH detected on the washed and non-washed coupons for both the fresh and pre-corroded 369

concrete (Figure 5C & D). 370

[Insert Figure 5] 371

3.4. Potential mechanisms of corrosion mitigation by surface washing 372

Our results demonstrated that high pressure washing reduced the H2S uptake activity of the 373

coupons that had a range of corrosion levels. The full recovery of sulfide uptake was fairly 374

quick on the fresh concrete, around 60-80 days, whereas the full recovery took over 4 months 375

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17

for the severely corroded pre-corroded concrete (i.e. coupon P50). Therefore, in the short 376

term (i.e. at the magnitude of month), the high pressure washing was possibly more effective 377

to control the corrosion on the more heavily corroded concrete. However, the washing did not 378

result in significant changes of concrete loss over the long-term exposure period of 4.5 years. 379

Frequent washing at short intervals every few months might be needed to mitigate the 380

corrosion over a long term. 381

Undoubtedly, the high pressure washing removes the corrosion layer containing corrosion 382

products (e.g. gypsum and ettringite), SOB and acid formed through chemical and biological 383

sulfide oxidation. The loss of SOB would cause decrease of the microbially induced H2S 384

uptake by the concrete and hence decrease of the concrete SUR, as was detected (Figure 4). 385

Meanwhile, the removal of the corrosion layer would eliminate the barrier for acid attack and 386

hence facilitate the penetration of acid, that could be produced by chemical sulfide oxidation 387

and the residual SOB catalysed sulfide oxidation, towards the inner intact concrete. 388

Consequently, washing may pose positive and negative effects on the SUR and acid 389

penetration of corroding sewer concrete. However, these effects were negated in this study, 390

thus, having no overall effect on the long term concrete loss. 391

During the re-exposure of coupons in the corrosion chambers after washing, the SUR 392

gradually recovered and the thickness of the corrosion layer gradually increased. However, 393

increasing corrosion development is not expected to accompany with an infinite increase of 394

SUR. On a thick corrosion layer sulfide oxidation would be limited to certain regions within 395

the layer depending on the diffusion gradients of dissolved oxygen and H2S and the 396

distribution of SOB. Additionally, extremely low pH resulting from the acid production 397

within the corrosion layer would in turn inhibit the activity of SOB [20, 27, 30]. Likely, an 398

increasing corrosion thickness would increase the barrier for acid diffusion to the corrosion 399

front. Then within the layer, the diffusion rate would limit the overall corrosion rate, as this is 400

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18

generally much lower than the neutralizing reactions at the corrosion front [31]. Thus, a semi-401

stable corrosion rate will be achieved eventually [32]. Overall, the washing only affects the 402

redevelopment of the corrosion layer, which may only take some months. 403

4. Conclusions 404

This study investigated the effect of high pressure washing on the corrosion activity of sewer 405

concrete and the main conclusions are: 406

• High pressure washing increased the coupon surface pH and decreased the SUR. The 407

highest increase of pH and largest decrease of SUR occurred on the coupons with the 408

most severe levels of corrosion. 409

• The SUR recovered to the level prior to washing within 60-140 days. The slowest 410

recovery was observed on the coupons with the most severe corrosion. 411

• Washing did not cause significant difference of the concrete loss over the long-term 412

exposure in sewer conditions of 4.5 years. 413

Acknowledgements 414

We acknowledge the Australia Research Council and many partners from the Australian 415

water industry for funding the Sewer Corrosion and Odour Research (SCORe) Project 416

(LP0882016). Ms Xiaoyan Sun acknowledges the University of Queensland and the Chinese 417

Research Council for funding the Tuition Fee International Scholarship and the Living 418

Allowance Scholarship. Dr Guangming Jiang is the recipient of a Queensland State 419

Government’s Early Career Accelerate Fellowship. 420

References 421

[1] US Environmental Protection Agency. State of technology for rehabilitation of 422

wastewater collection systems: US Environmental Protection Agency, Office of Research and 423

Development. EPA/600/R-10/078, July 2010; 2010. 424

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

19

[2] Jiang G, Sun J, Sharma KR, Yuan Z. Corrosion and odor management in sewer systems. 425

Current Opinion in Biotechnology. 2015;33:192-7. 426

[3] Gutierrez O, Jiang G, Sharma K, Yuan Z. Biofilm development in sewer networks In: 427

Romaní AM, Guasch H, Balaguer MD, editors. Aquatic biofilms: Ecology, Water Quality 428

and Wastewater Treatment: Caister Academic Press; 2015. 429

[4] Hvitved-Jacobsen T, Vollertsen J, Nielson ARH. Sewer processes: microbial and 430

chemical process engineering of sewer networks: CRC Press, Boca Raton London New York 431

Washington, D.C.; 2013. 432

[5] Dopson M, Johnson DB. Biodiversity, metabolism and applications of acidophilic sulfur-433

metabolizing microorganisms. Environmental Microbiology. 2012;14(10):2620-31. 434

[6] Jiang G, Keller J, Bond PL. Determining the long-term effects of H2S concentration, 435

relative humidity and air temperature on concrete sewer corrosion. Water Research. 436

2014;65:157-69. 437

[7] Zivica Vr, Bajza A. Acidic attack of cement based materials - a review. Part 1. Principle 438

of acidic attack. Construction and Building Materials. 2001;15(8):331-40. 439

[8] Joseph AP, Keller J, Bustamante H, Bond PL. Surface neutralization and H2S oxidation at 440

early stages of sewer corrosion: Influence of temperature, relative humidity and H2S 441

concentration. Water Research. 2012;46(13):4235-45. 442

[9] Ganigue R, Gutierrez O, Rootsey R, Yuan Z. Chemical dosing for sulfide control in 443

Australia: An industry survey. Water Research. 2011;45(19):6564-74. 444

[10] Sydney R, Esfandi E, Surapaneni S. Control concrete sewer corrosion via the crown 445

spray process. Water Environ Res. 1996;68(3):338-47. 446

[11] Pikaar I, Sharma KR, Hu S, Gernjak W, Keller J, Yuan Z. Reducing sewer corrosion 447

through integrated urban water management. Science. 2014;345(6198):812-4. 448

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

20

[12] Sun X, Jiang G, Bond PL, Keller J, Yuan Z. A novel and simple treatment for control of 449

sulfide induced sewer concrete corrosion using free nitrous acid. Water Research. 450

2015;70:179-87. 451

[13] Haile T, Nakhla G, Allouche E. Evaluation of the resistance of mortars coated with 452

silver bearing zeolite to bacterial-induced corrosion. Corrosion Science. 2008;50(3):713-20. 453

[14] Glass GK, Reddy B, Buenfeld NR. Corrosion inhibition in concrete arising from its acid 454

neutralisation capacity. Corrosion Science. 2000;42(9):1587-98. 455

[15] Mansfeld F, Shih H, Postyn A, Devinny J, Islander R, Chen C. Corrosion monitoring 456

and control in concrete sewer pipes. Corrosion. 1991;47(5):369-76. 457

[16] Satoh H, Odagiri M, Ito T, Okabe S. Microbial community structures and in situ sulfate-458

reducing and sulfur-oxidizing activities in biofilms developed on mortar specimens in a 459

corroded sewer system. Water Research. 2009;43(18):4729-39. 460

[17] Monteny J, Vincke E, Beeldens A, De Belie N, Taerwe L, Van Gemert D, et al. 461

Chemical, microbiological, and in situ test methods for biogenic sulfuric acid corrosion of 462

concrete. Cement and Concrete Research. 2000;30(4):623-34. 463

[18] Mori T, Nonaka T, Tazaki K, Koga M, Hikosaka Y, Noda S. Interactions of nutrients, 464

moisture and pH on microbial corrosion of concrete sewer pipes. Water Research. 465

1992;26(1):29-37. 466

[19] Islander RL, Devinny JS, Mansfeld F, Postyn A, Shih H. Microbial ecology of crown 467

corrosion in sewers. Journal of Environmental Engineering. 1991;117(6):751-70. 468

[20] Nielsen AH, Vollertsen J, Jensen HS, Wium-Andersen T, Hvitved-Jacobsen T. Influence 469

of pipe material and surfaces on sulfide related odor and corrosion in sewers. Water Research. 470

2008;42(15):4206-14. 471

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

21

[21] Joseph AP, Keller J, Bond PL. Examination of concrete corrosion using a laboratory 472

experimental set up simulating sewer conditions. 6th International Conference on Sewer 473

Processes and Networks. Surfers Paradise, Gold Coast, Australia2010. 474

[22] Sun X, Jiang G, Bond PL, Wells T, Keller J. A rapid, non-destructive methodology to 475

monitor activity of sulfide-induced corrosion of concrete based on H2S uptake rate. Water 476

Research. 2014;59:229-38. 477

[23] Keller-Lehmann B, Corrie S, Ravn R, Yuan Z, Keller J. Preservation and simultaneous 478

analysis of relevant soluble sulfur species in sewage samples. 2nd International IWA 479

Conference on Sewer Operation and Maintenance, Vienna, Austria2006. 480

[24] Jiang G, Sharma KR, Guisasola A, Keller J, Yuan Z. Sulfur transformation in rising 481

main sewers receiving nitrate dosage. Water Research. 2009;43(17):4430-40. 482

[25] Wells T, Melchers RE, Bond P. Factors involved in the long term corrosion of concrete 483

sewers. Australia Corrosion Association Proceedings of Corrosion and Prevention. 2009. 484

[26] Jiang G, Wightman E, Donose BC, Yuan Z, Bond PL, Keller J. The role of iron in 485

sulfide induced corrosion of sewer concrete. Water Research. 2014;49:166-74. 486

[27] Gutiérrez-Padilla MGD, Bielefeldt A, Ovtchinnikov S, Hernandez M, Silverstein J. 487

Biogenic sulfuric acid attack on different types of commercially produced concrete sewer 488

pipes. Cement and Concrete Research. 2010;40(2):293-301. 489

[28] Jiang G, Sun X, Keller J, Bond PL. Identification of controlling factors for the initiation 490

of corrosion of fresh concrete sewers. Water Research. 2015;80:30-40. 491

[29] Okabe S, Odagiri M, Ito T, Satoh H. Succession of sulfur-oxidizing bacteria in the 492

microbial community on corroding concrete in sewer systems. Appl Environ Microbiol. 493

2007;73(3):971-80. 494

[30] Mahmood Q, Zheng P, Hayat Y, Islam E, Wu D, Ren-cun J. Effect of pH on anoxic 495

sulfide oxidizing reactor performance. Bioresource Technology. 2008;99(8):3291-6. 496

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

22

[31] Houst YF, Wittmann FH. Depth profiles of carbonates formed during natural 497

carbonation. Cement and Concrete Research. 2002;32(12):1923-30. 498

[32] Wells T, Melchers R. An observation-based model for corrosion of concrete sewers 499

under aggressive conditions. Cement and Concrete Research. 2014;61:1-10. 500

501

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Table 1. Sulfur species measured on surface of coupons exposed to 5 and 50 ppm of H2S

H2S exposure level (ppm) Fresh coupons Pre-corroded coupons

S0 (g m-2) SO42- (g m-2) S0 (g m-2) SO4

2- (g m-2)

5 7.3 4.9 NA 119.4

50 27.6 22.8 NA 282.9

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure 1. The MLA mapping of major (i.e. percentage of weight higher than 0.3%) mineral

compounds identified on the fresh (A) and pre-corroded (B) concrete sections.

A B

0µm 2000µm

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure 2. Photos of a 5ppm (A&C) and a 50 ppm (B&D) H2S exposed concrete coupon set before and after the high pressure washing. Each coupon set consists of a fresh (the left side) and a pre-corroded (the right side) concrete coupon.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure 3. Surface pH of coupon F5, P5, F50 and P50 (F and P indicate fresh and pre-corroded coupon, respectively, 5 and 50 indicate the exposure H2S concentration (ppm)) before the high-pressure washing and immediately after that. Each error bar represents the standard deviation of four measurements on each concrete coupon surface.

0

2

4

6

8

10

12

F5 P5 F50 P50

Sur

face

pH

Before washing After washing

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure 4. Recovery ratio (SURb and SURr indicate the SUR of coupon prior to and after

washing, respectively) of the coupon F5, P5, F50 and P50 (F and P indicate fresh and pre-

corroded coupon, respectively, 5 and 50 indicate the H2S concentration (ppm) in exposure

chamber) after the high pressure washing is shown in Figure A and the full SUR recovery

time of the four concrete coupons after washing is shown in Figure B.

0.0

0.5

1.0

1.5

0 30 60 90 120 150 180

Rec

over

y ra

tio (

SU

Rr/

SU

Rb)

Time (day)

F5 P5 F50 P50 SURb

A

0

30

60

90

120

150

F5 P5 F50 P50

Ful

l rec

over

y tim

e (d

ay)

B

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure 5. Concrete loss from the fresh (A) and pre-corroded coupons (B) exposed to H2S at

10 ppm and 25 ppm for 54 months and the corresponding surface pH of the fresh (C) and pre-

corroded (D) coupons after 54 months of exposure. Coupons numbered 1 to 6 were subjected

to one high-pressure washing event during the exposure period, this occurring after 6, 12, 18,

24, 34, and 44 months exposure respectively. Coupons numbered 7 received no washing

during the exposure period. Average concrete losses were calculated from those measured on

coupons 1 to 6. Each error bar represents the standard deviation of four pH measurements on

each coupon surface.

0.0

0.3

0.6

0.9

1.2

1.5

1 2 3 4 5 6 7

Con

cret

elo

ss (

mm

)

Coupon No.

10 ppm25 ppmAverage 10 ppmAverage 25 ppm

A

0

2

4

6

8

10

1 2 3 4 5 6 7

Con

cret

e lo

ss (

mm

)

Coupon No.

B

0

1

2

3

4

5

6

7

1 2 3 4 5 6 7

Sur

face

pH

Coupon No.

C

0

1

2

3

4

5

6

7

1 2 3 4 5 6 7

Sur

face

pH

Coupon No.

D