Accepted Manuscript

Assessment of strength properties of cemented paste backfill by ultrasonic pulsevelocity test

Tekin Yılmaz, Bayram Ercikdi, Kadir Karaman, Gökhan Külekçi

PII: S0041-624X(14)00043-2DOI: http://dx.doi.org/10.1016/j.ultras.2014.02.012Reference: ULTRAS 4771

To appear in: Ultrasonics

Received Date: 19 November 2013Revised Date: 11 February 2014Accepted Date: 12 February 2014

Please cite this article as: T. Yılmaz, B. Ercikdi, K. Karaman, G. Külekçi, Assessment of strength properties ofcemented paste backfill by ultrasonic pulse velocity test, Ultrasonics (2014), doi: http://dx.doi.org/10.1016/j.ultras.2014.02.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting proof before it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Assessment of strength properties of cemented paste backfill by ultrasonic 1

pulse velocity test 2

Tekin Yılmaza, Bayram Ercikdia*, Kadir Karamana, Gökhan Külekçib 3

aDepartment of Mining Eng., Karadeniz Technical University, 61080 Trabzon, Turkey 4

bDepartment of Mining Eng., Gümüşhane University, Gümüşhane, Turkey 5

*Phone: +90-462-3773171; Fax: +90-462-3257405 6

E-mail address: bercikdi@ktu.edu.tr

2

Abstract: 7

Ultrasonic pulse velocity (UPV) test is one of the most popular non-destructive techniques 8

used in the assessment of the mechanical properties of concrete or rock materials. In this 9

study, the effects of binder type/dosage, water to cement ratio (w/c) and fines content (<20 10

µm) of the tailings on ultrasonic pulse velocity (UPV) of cemented paste backfill (CPB) 11

samples were investigated and correlated with the corresponding unconfined compressive 12

strength (UCS) data. A total of 96 CPB samples prepared at different mixture properties were 13

subjected to the UPV and UCS tests at 7, 14, 28 and 56–days of curing periods. UPV and 14

UCS of CPB samples of ordinary Portland cement (CEM I 42.5 R) and sulphate resistant 15

cement (SRC 32.5) initially increased rapidly, but, slowed down after 14 days. However, 16

UPV and UCS of CPB samples of the blast furnace slag cement (CEM III/A 42.5 N) steadily 17

increased between 7-56 days. Increasing binder dosage or reducing w/c ratio and fines content 18

(<20 µm) increased the UCS and UPV of CPB samples. UPV was found to be particularly 19

sensitive to fines content. UCS data were correlated with the corresponding UPV data. A 20

linear relation appeared to exist between the UCS and UPV of CPB samples. These findings 21

have demonstrated that the UPV test can be reliably used for the estimation of the strength of 22

CPB samples. 23

Keywords: Ultrasonic pulse velocity, unconfined compressive strength, cemented paste 24 backfill, mixture properties 25

3

1. Introduction 26

Ultrasonic pulse velocity (UPV) test, a non-destructive and easy method to apply in both field 27

and laboratory conditions, is increasingly being used to determine the geotechnical properties 28

of rock or concrete materials in mining, civil and geotechnical engineering. It employs the 29

principle of measuring the travel velocity of ultrasonic pulses through a material medium. 30

Knowing the UPV, changes of the geotechnical properties can be evaluated using known 31

relationships between velocities and mechanical properties [1]. There are many studies on the 32

use of ultrasonic pulse velocity (UPV) test. Karpuz and Paşamehmetoğlu [2] utilized UPV test 33

to determine the weathering degree of Ankara andesites. Kahraman et al. [3] reported that the 34

quality classification and estimation of slab production efficiency of the building stones can 35

easily be made by ultrasonic UPV test. Others attempted to assess grouting and 36

blasting/fragmentation efficiencies in a rock mass [4,5] and thermal conductivity of any rock 37

[6] by laboratory UPV test. In addition, UPV test was used for predicting the stress 38

distribution around the mine tunnel and estimating the thickness of damaged zones caused by 39

the tunnel excavation [7]. Many researchers have found that there is a good relationship 40

between UPV and unconfined compressive strength of rock or concrete materials [8-12]. 41

Cemented paste backfill (CPB) is primarily composed of mill tailings (75–85% solids by 42

weight), a hydraulic binder (usually 3–9% by weight) and mixing water [13-15]. The 43

unconfined compressive strength (UCS) of CPB at a given time is one of the most important 44

parameter since the CPB structure must remain stable during the extraction of adjacent stopes 45

to ensure the safety of the mine workers and to avoid ore dilution. Although UPV test as a 46

low-cost, less time consuming and practical method is known to be extensively exploited for 47

estimating the UCS of rock and concrete [8-12,16], there are no detailed studies on the 48

utilization of UPV test for predicting the mechanical performance of CPB. In this regard, the 49

UPV test method can be beneficial for the rapid estimation of the UCS of CPB instead of 50

4

conventional compressive strength test. Chou et al. [1] determined the geotechnical properties 51

such as Young, shear and bulk modulus of rockfill by measuring P and S waves over a curing 52

period of 120 days. The P–and S–waves velocity changes in CPB having 3 to 5 wt.% binder 53

content at early curing ages were monitored by Diezd’Aux [17]. He suggested that the 54

evaluation of ultrasonic properties should be performed after several days of curing due to the 55

inaccurate results obtained in the short term. Galaa et al. [18] also measured the P– and S–56

waves in CPB samples to understand the development of strength and stiffness of CPB over a 57

curing period of 7 days. However, they did not correlate UCS data with UPV. 58

In the present study, ultrasonic pulse velocity (UPV) was evaluated as a non-destructive, low-59

cost and practical method for the estimation of CPB strength. The effects of binder type, 60

binder dosage, water to cement ratio (w/c) and fines content (<20 µm) of the tailings on the 61

strength and, particularly, ultrasonic properties of CPB produced from the mill tailings was 62

investigated over 7-56 days of curing periods. The UCSs of CPB samples were correlated 63

with the UPV results in an attempt to use the UPV measurement to predict the strength of 64

CPB. Potential benefits of UPV test in CPB were discussed. 65

2. Materials and methods 66

2.1. Tailings and binders 67

In this study, a tailings sample was obtained from the tailings dam of a copper flotation plant 68

(Kastamonu Küre, Turkey). The tailings sample was collected from the point that is 40 m far 69

away from the tailings discharge point by hydraulic excavator (Fig. 1a). Annually, 70

approximately 0.55 Mt of sulphide tailings were produced as a result of milling operations 71

and disposed into the tailings dam. The tailings sample was deslimed by hydrocyclone at the 72

plant in order to investigate the effect of fines content (<20 µm) on the strength and ultrasonic 73

5

properties of CPB. Table 1 shows the physical, chemical and mineralogical properties of the 74

as-received and deslimed tailings. Compared with the as-received tailings, the deslimed 75

tailings were determined to contain less fines (e.g. 35% c.f. 58.4% finer than 20 µm) and 76

higher pyrite content (43.5% c.f. 52.2%) presumably due to the removal of silicate minerals in 77

slimes fraction as suggested by the decrease in SiO2+Al2O3 contents of tailings (Table 1). The 78

coefficient of uniformity (Cu) was determined to be 11.0 and 6.54 for the as-received and 79

deslimed tailings, respectively. X-ray diffraction (XRD) analysis indicated that the major 80

mineral phase was identified to be pyrite. Although both the as-received and deslimed tailings 81

contain quartz and chlorite, other silicates such as muscovite and albite were identified only in 82

the as-received tailings (Table 1). 83

In this study, blast furnace slag cement (CEM III/A 42.5 N) was used as the main binder in 84

the tests where the effect of binder dosage, w/c ratio and fines content (<20 µm) of the tailings 85

on the strength and ultrasonic properties of CPB were evaluated. Ordinary Portland cement 86

(CEM I 42.5 R) and sulphate resistant cement (SRC 32.5) were also used to study the effect of 87

binder type on the UCS and ultrasonic pulse velocity (UPV) of CPB. The physical, chemical 88

and mineralogical characterisations of the cements were summarized in Table 1. SRC was 89

characterized by its low C3A content (Table 1), which is an important phase that controls the 90

long term performance of CPB against acid and sulphate attack [19,20]. However, only the 91

short term (up to the 56 days) performances of binders were evaluated in this study. It is 92

relevant to note that CEM III/A 42.5 N contains granulated blast furnace slag (35%) as 93

additive to improve its binding performance and to reduce the binder costs. 94

6

2.2. Preparation of CPB samples 95

A total of 96 CPB samples in triplicate were prepared by mixing and homogenizing the 96

tailings samples (as-received and deslimed), binder and tap water in a Univex Stand model 97

blender equipped with a double spiral (Fig. 1b). The paste was then placed in the slump cone 98

in one-third length increments, each being tamped 25 times with a small rod. The CPB 99

samples were prepared at a slump consistency of 16.51-21.59 cm, which were verified using 100

the 30.48 cm high concrete cone test according to ASTM C 143 [21]. The solids contents of 101

paste mixtures were set to 73.58-80.17 wt.% The effect of w/c ratio, binder type and fines 102

content was examined at a fixed binder dosage of 7 wt.% while binder dosage was tested at 5, 103

6 and 7 wt.% (Table 2). 104

The CPB mixtures after being thoroughly mixed were poured into the plastic cylinders with a 105

diameter x height of 10x20 cm. Bottom of these cylinders were perforated (seven holes with 2 106

mm diameter) to allow the drainage of excess water. The cylinders were sealed in plastic bags 107

and allowed to cure in a curing room maintained at 20 ±1°C. 108

2.3. UPV and UCS tests 109

CPB samples were subjected to the ultrasonic pulse velocity (UPV) tests according to ASTM 110

C 597 [22] at 7, 14, 28 and 56 day curing periods. The UPV was measured on CPB samples 111

by a Portable Ultrasonic Nondestructive Digital Indicating Tester (PUNDIT) that measures 112

the time of propagation of ultrasound pulses with a precision of 0.1 µs and its transducers 113

were 42 mm in diameter with 54 kHz (Fig. 1c). Length of the measuring base was determined 114

within an accuracy of 0.1 mm. End surfaces of the CPB samples were polished to provide a 115

good coupling between the transducer face and the sample surface to maximize accuracy of 116

the transit time measurement. A thin film of vaseline was applied to the surface of the 117

7

transducers (transmitter and receiver) in order to ensure full contact and to eliminate the air 118

pocket between transducers and the test medium. The direct transmission technique, as the 119

most satisfactory and reliable method, was used in the test in which the transmitter and 120

receiver were positioned on the opposite end surfaces of the specimens tested. Repeated 121

readings at a particular location were taken and a minimum value of transit time was taken as 122

the experimental result [16]. After the measurements, the velocity of P–wave, UPV, was 123

calculated from the measured travel time and the distance between the transmitter and 124

receiver as below: 125

UPV (x,t) = x/t (Eq.1) 126

Where UPV (x,t) is the velocity of P–wave in CPB, x is the distance between the transmitter 127

and receiver and t is the travel time. 128

After UPV tests, the UCS tests [23] were performed on the same CPB samples using a 129

computer–controlled mechanical press (ELE Digital Tritest) (Fig. 1d), which had a load 130

capacity of 50 kN and a displacement speed of 0.5 mm per minute. All the experiments were 131

carried out in triplicate and the mean UPV and UCS values were presented in the results. 132

3. Results and discussion 133

3.1. Effect of binder type and dosage 134

Fig. 2 illustrates the strength and UPV of CPB samples prepared from as-received tailings 135

using CEM I 42.5 R, CEM III/A 42.5 N and SRC 32.5 at a constant binder dosage of 7 wt.%. 136

The UPV and UCSs of all the CPB samples increased over the curing time of 56 days 137

irrespective of the binder type and dosage (Figs 2a,b). CPB samples of CEM I 42.5 R 138

8

produced consistently higher UCS than those of CEM III/A 42.5 N and SRC 32.5 at all curing 139

times. CPB samples of CEM III/A 42.5 N yielded relatively low UCSs at early ages, but, with 140

the elapse of curing time, a steady increase was observed. The low early UCSs for CEM III/A 141

42.5 N could be attributed to the low heat of hydration of blast furnace slag [9]. Furthermore, 142

increasing curing time led to a reduction in the gap between the UCSs of CEM I 42.5 and 143

CEM III/A 42.5 N (Fig. 2a), which could be interrelated with the pozzolanic activity of blast 144

furnace slag [24]. In CPB practice, a 28-day UCS of 0.7-2.0 MPa is often required by mine 145

operators to be threshold for the self–supporting stopes and the mined stopes adjacent to ore 146

extraction [25]. In this regard, the CPB samples prepared from these binders at 7 wt.% dosage 147

failed to achieve a 28–day strength of 0.7 MPa, when the tailings sample was used as 148

received. 149

It is well known that as the curing time increases, CPB gains stiffness with the hydration of 150

cement phase, leading to the formation of solid products (i.e. C-S-H) that fill the pore space 151

and creates bonds between the particles of mine tailings. Self–weight consolidation and 152

evaporation of water from the CPB samples can also contribute to the development of solid 153

stiffness [18,26]. In this regard, the UPV increased as the samples cured irrespective of the 154

CPB mixture. However, the UPV of CPB samples of different binders increased by 8.3-19.4% 155

compared with 31.5-120% increase in UCSs between 7 and 56 days. These findings agree 156

well with Gesoğlu [27] who observed similar behavior for the UCS and UPV values of 157

concrete samples between 7 and 28 days. It can be inferred that compressive strength tends to 158

increase faster than UPV due to the increased stiffness and density of materials. A substantial 159

increase in the UPV of the CPB samples of CEM I 42.5 and SRC 32.5 were observed to occur 160

between 7 and 14 days. This indicates that more hydration products become connected to each 161

other at early curing periods due to the inherent chemical characteristics of CEM I 42.5 R and 162

9

SRC 32.5. Thereafter, a trend of slight increase in the UPV was observed. The UPV profiles 163

for these binders were similar in character to the corresponding UCS profiles. However, a 164

different trend of UPV was observed for CEM III/A 42.5 N in that the UPV increased linearly 165

between 7-56 days. This could be attributed to the continuing development of strength in 166

these samples i.e. the formation of hydration products is still in progress, probably linked with 167

inherently slow pozzolanic reactions. These observations are well consistent with the findings 168

of Ye et al. [28] who stated that an increase in UPV is limited after the fully connected solid 169

frame occurred in cement-based materials. It is relevant to note that UPV values in CPB 170

samples varied between 1328 and 1616 m/s at the 28–day curing period (Figs. 2-6). Various 171

researchers [9,11,29-31] measured higher (>3640 m/s) UPV values in concrete samples cured 172

for 28 days. Concrete samples have higher cement contents (≥300 kg/m3) and lower w/c ratios 173

(usually ranges between 0.3–0.6) than CPB. Additionally, the type and volume of the more 174

rigid aggregates result in high UPV of concrete samples. Therefore, the comparatively low 175

UPV values in CPB samples could be ascribed to the high void ratio due to the lower cement 176

content (82–136 kg/m3) and higher w/c ratios (3.81–6.82) (Table 2). Furthermore, CPB does 177

not contain rigid aggregate. Trtnik et al. [16] reported that binder type does not significantly 178

affect the P-wave velocity of concrete samples. In this regard, the variation of UPV and UCS 179

values in CPB samples of different binders at the same curing periods was determined to vary 180

in the range of 1–9% and 5.8-77.7%, respectively (Fig. 2). Accordingly, binder type appeared 181

to have a significant effect on the strength development of CPB while its effect on UPV is 182

insignificant. 183

As expected, the UCS and UPV values of CPB samples increased with increasing the binder 184

dosage irrespective of the curing periods (Figs. 3a,b). The CPB samples prepared at 7 wt.% 185

binder dosage had slightly greater UPV values than those of 5 wt.% binder dosage, in spite of 186

10

a marked difference observed in the compressive strength. Similar results were also found in 187

previous studies [17,31]. Mahure et al. [31] reported that an increase in the binder dosage 188

from 257 kg/m3 to 425 kg/m3 have resulted in 15% and 76% increase in the UPV and UCS of 189

concrete samples, respectively, at 28 days of curing period. Similarly, Diezd’Aux [17] 190

observed that the CPB samples prepared at 5 wt.% binder content generated approximately 191

5% higher UPV values than those prepared at 3 wt.% binder content over the same curing 192

period. The beneficial effect of increasing binder dosage on UPV and UCS can be ascribed to 193

the increase in quantity of hydration products (CH and C–S–H) with the resultant reduction in 194

void ratio and porosity [13]. Notwithstanding this, the binder dosage should be increased 195

beyond 7wt.% to achieve the required 28 day UCS of ≥0.7 MPa for CPB samples of the as-196

received tailings when the CEM III/A 42.5 N is used as the binder. 197

3.2. Effect of w/c ratio and fines content 198

It has been reported that water to cement ratio (w/c) significantly affects the strength and P-199

wave velocity (UPV) of cementitious materials [16,28,30]. Galaa et al. [18] reported that the 200

lower the w/c ratio the higher is the cementing bonds leading to a rapid strength development 201

of CPB. Lafhaj et al. [30] also demonstrated that the porosity of mortar samples increased 202

with increasing w/c ratio with the resultant reduction in the UPV values. In agreement with 203

these studies, CPB mixtures with a lower w/c ratio had higher UPV and UCS values (Figs. 204

4a,b). This could be ascribed to the higher solid volume fraction and lower porosity and void 205

ratios of the mixtures with lower w/c ratios [24,28]. The CPB samples prepared at 4.62 w/c 206

ratio were observed to develop consistently 1.06–1.10 and 1.18–1.26 times higher UPV and 207

UCSs than those at 5.13 w/c ratio, respectively. These findings suggest that the UPV is less 208

sensitive to w/c ratio changes. It is pertinent to note that those CPB samples prepared from the 209

11

as-received tailings failed to produce a 28-day UCS of ≥0.7 MPa even at the lowest w/c ratio 210

of 4.62 tested. 211

Fig. 5 indicates that decreasing the fines content (<20 µm) resulted in an increase in the UPV 212

and UCS of CPB samples at all curing periods. Like compressive strength, there was a 213

systematic increase in the UPV values of CPB samples with increasing curing time being 214

more distinct particularly at 28 and 56 days. Additionally, the CPB samples of the deslimed 215

tailings produced higher UPV and UCS values (by 1.07-1.2 and 1.22-1.28 fold, respectively) 216

than those of reference tailings at the same curing periods (Figs 5a,b). In contrast to the binder 217

type, dosage and w/c ratio, the UPV appears to be more sensitive to fines content (<20 µm). 218

The quantity of pores in a CPB sample has been reported to increase with increasing fines 219

content (<20 µm) of the tailings [32,33]. Singh and Kripamoy [34] and Karaman et al. [35] 220

have also indicated that UPV decreases with increasing silicate minerals (i.e. mica, clay, 221

quartz), which is presumably due to the water retention potential of these minerals [25]. 222

Furthermore, Ercikdi et al. [36] demonstrated that CPB samples produced from coarse tailings 223

release more water (by drainage) than those of medium or fine tailings. The loss of water by 224

drainage leads to the settling of the paste backfill (increasing of the packing density) and the 225

consequent reduction of total porosity and void ratio of the backfill material [32]. In this 226

regard, the comparatively high UPV and UCSs in CPB samples of the deslimed tailings can 227

be attributed to their lower water retention capacity (i.e. lower water–to–cement ratio) mainly 228

due to its lower silicate (SiO2+Al2O3) content (26.6% c.f. 32.5% for the as-received tailings) 229

and fines content (35.0% c.f. 58.4% for the as-received tailings) (Table 1). 230

Wichtmann and Triantafyllidis [37] investigated the effect of grain size distribution curve on 231

UPV of quartz sand and found that UPV decreases with increasing the coefficient of 232

12

uniformity (Cu) due to the improvement in the gradation of sand. Similarly, Trtnik et al. [16] 233

observed that UPV increased with increasing the aggregate content. The high UPV and UCS 234

values in CPB samples of the deslimed tailings can be also linked with their higher solids 235

content (80.17%), lower w/c ratio (3.81) and lower coefficient of uniformity (Cu=6.54) of the 236

deslimed tailings (Table 1 and 2). Consistent with these findings, the highest UPV (1821 m/s) 237

was obtained from the deslimed CPB samples over a curing period of 56 days. Moreover, 238

only the CPB samples of the deslimed tailings produced the desired 28-day UCS of ≥0.7 MPa. 239

3.3. Relationship between UCS and UPV 240

There are a number of empirical equations developed between UPV and UCS for concrete in 241

the literature [9,16,31]. In the current study, the relationship between UPV and UCS of the 242

CPB samples was sougth after by using simple regression analysis. Linear (y= ax+b), 243

logarithmic (y= a+ Inx), exponential (y= aex) and power (y= ax

b) curve fitting 244

approximations were undertaken. Fig.6 represents the relationship between UPV and UCS for 245

all CPB samples (96 specimens) regardless of the composition of the material. As shown in 246

Fig. 6, the UPV increased with increasing UCS irrespective of the mixture properties (i.e. 247

binder type, w/c ratio).The plots of UPV versus UCS in Figs. 7-10 show that there is a linear 248

relationship between the UCS and UPV for all CPB samples. A relatively high correlation 249

coefficient of 0.86 was found (Eq. 2) when all the UCS and UPV results obtained from CPB 250

samples (96 specimens). Compared with all CPB specimens, lower correlation coefficient of 251

0.83 was found (Eq. 3) when all the mean UCS and UPV results obtained from CPB samples 252

of different binders were taken into account (Fig. 7a). However, a higher correlation 253

coefficients for a particular set of data e.g. data collected from the same binder type (Fig. 7b), 254

binder dosage (Fig. 8b), w/c ratios (Fig. 9b) and fines content (<20 µm) (Fig. 10b) were 255

obtained. Taking into account all the average UCS and UPV results obtained from CPB 256

13

samples for different binder types (Eq. 3) (Fig. 7a), dosages (Eq. 4) (Fig. 8a), w/c ratios (Eq. 257

5) (Fig. 9a) and fines content (Eq. 6) (Fig. 10a), the following general equations were 258

determined, respectively: 259

UCS = 0.0011UPV - 1.0397 (r=0.86) (Eq.2) 260

UCS = 0.0015UPV - 1.5151 (r=0.83) (Eq.3) 261

UCS = 0.0013UPV - 1.3212 (r=0.90) (Eq.4) 262

UCS = 0.0014UPV - 1.3917 (r=0.94) (Eq.5) 263

UCS = 0.0009UPV - 0.7190 (r=0.92) (Eq.6) 264

Although the high correlation coefficients (r value) were found for the obtained equations, 265

they do not necessarily indicate the goodness-of-fit of these equations. Thus, t- and F tests 266

were conducted in order to check the validity of these equations. The t-test compares the 267

computed values with tabulated values using null hypothesis. According to the t-test, when 268

computed t value is greater than tabulated t-value, the null hypothesis is rejected and obtained 269

correlation coefficient (r-value) is acceptable. The significance of the regressions was 270

determined by analysis of variance (F-test). In these tests, a 95 per cent level (p<0.05) of 271

confidence was chosen. As seen in Table 3, the computed t and F values are greater than the 272

tabulated t and F values, indicating the significance of r values and the validity of the derived 273

equations. In this regard, these empirical equations (Eqs. 2-6) can provide an estimate of the 274

UCS of CPB samples using UPV data. 275

Conclusions 276

In this study, the effect of binder type/dosage, water to cement ratio and fines content (<20 277

µm) of the tailings on the mechanical and ultrasonic properties of CPB samples produced 278

14

from mill tailings was evaluated. CPB samples prepared at different mixture properties were 279

subjected to UPV and UCS tests at 7, 14, 28 and 56 days of curing periods. A change in 280

mixture properties was observed to produce a significant effect on the UCS development of 281

CPB samples, whilst, its effect on UPV was relatively low. CPB samples of CEM I 42.5 and 282

SRC 32.5 demonstrated similar UCS and UPV behavior at a given curing time while those of 283

CEM III A 42.5 N exhibited a different trend apparently due to the inherent hydration 284

characteristics of this binder. Strength and ultrasonic properties of CPB samples increased 285

with increasing the binder dosage (5 to 7 wt.%) or reducing the w/c ratio (5.13 to 4.62) and 286

fines content (58.4% to 35.0% finer than 20 µm). The strength development and ultrasonic 287

properties of CPB samples were found to be highly sensitive to the fines content (<20 µm) of 288

the tailings. Only the CPB samples of the deslimed tailings prepared at 7 wt.% binder dosage 289

were able to achieve the 28–day UCS of ≥0.7 MPa. A linear correlation between the UCS and 290

corresponding UPV values was obtained at a particular mixture property (i.e. binder dosage, 291

w/c ratio). Furthermore, the relationship between the UPV and corresponding UCS values 292

was acceptable according to the statistical analysis by t- and F-tests. These findings suggest 293

that the UPV test as a low-cost, less time consuming and practical method can be reliably 294

used to predict the UCS of CPB samples. 295

Acknowledgement 296

The authors would like to express their sincere thanks and appreciation to the EtiBakır A.S., 297

for the material and financial support and, to Prof. Dr. Hacı Deveci and Associate Prof. Dr. 298

Gülten Yaylalı Abanuz for improving paper quality. 299

15

References 300

[1] C.L.Chou, M. Chouteau, M. Benzaazoua, Laboratory characterization of mining cemented 301

rockfill by NDT methods: experimental set up and testing. In: Proceedings of the 302

International Symposium on Nondestructive Testing of Materials and Structures, Istanbul, 303

Turkey, 2011. p. 935-942. 304

[2] C. Karpuz, A.G. Pasamehmetoglu, Field characterisation of weathered Ankara andesites, 305

Eng. Geol. 46 (1997) 1–17. 306

[3] S. Kahraman, A quality classification of building stones from P-wave velocity and its 307

application to stone cutting with gang saws, J. S. Afr. I. Min. Metall. 107 (2007) 427–430. 308

[4] R.P. Young, J.J. Hill, I.R. Bryan, R. Middleton, Seismic spectroscopy in fracture 309

characterization, Q. J. Eng. Geol. Hydroge. 18 (1985) 459–479. 310

[5] N. Turk, W.R. Dearman, Assessment of grouting efficiency in a rock mass in terms of 311

seismic velocities, B. Eng. Geol. Environ. 36 (1987) 101–108. 312

[6] H.T. Özkahraman, R. Selver, E.C. Işık, Determination of the thermal conductivity of rock 313

from P-wave velocity, Int. J. Rock. Mech. Min. 41 (2004) 703–708. 314

[7] C. Wright, E.J. Walls, D.J. Carneiro, The seismic velocity distribution in the vicinity of a 315

mine tunnel at Thabazimbi, South Africa, J. Appl. Geophys. 44 (2000) 369–382. 316

[8] S. Kahraman, Evaluation of simple methods for assessing the uniaxial compressive 317

strength of rock, Int. J. Rock. Mech. Min. 38 (2001) 981–994. 318

[9] R. Demirboğa, İ. Türkmen, M.B. Karakoç, Relationship between ultrasonic velocity and 319

compressive strength for high-volume mineral-admixtured concrete, Cement Concrete Res. 320

34 (2004) 2329–2336. 321

[10] E. Yaşar, Y. Erdoğan, Correlating sound velocity with the density, compressive strength 322

and Young’s modulus of carbonate rocks, Int. J. Rock. Mech. Min. 41 (2004) 41: 871–875. 323

16

[11] Z.C. Ulucan, K. Türk, M. Karataş, Effect of mineral admixtures on the correlation 324

between ultrasonic velocity and compressive strength for self-compacting concrete, Russ. J. 325

Nondestruct. 44 (2008) 367–374. 326

[12] S. Yağız, P-wave velocity test for assessment of geotechnical properties of some rock 327

materials, B. Mater. Sci. 34 (2011) 947–953. 328

[13] B. Ercikdi, A. Kesimal, F. Cihangir, H. Deveci, İ. Alp, Cemented paste backfill of 329

sulphide–rich tailings: importance of binder type and dosage, Cement Concrete Comp. 31 330

(2009a) 268–274. 331

[14] M. Fall, M. Pokharel, Coupled effects of sulphate and temperature on the strength 332

development of cemented tailings backfill: Portland cement-paste backfill, Cement 333

Concrete Comp. 31 (2010) 819–828. 334

[15] E. Yılmaz, T. Belem, B. Bussiere, M. Benzaazoua, Relationships between 335

microstructural properties and compressive strength of consolidated and unconsolidated 336

cemented paste backfills, Cement Concrete Comp. 33 (2011) 702–715. 337

[16] G. Trtnik, F. Kavcic, G. Turk, Prediction of concrete strength using ultrasonic pulse 338

velocity and artificial neural networks, Ultrasonics 49 (2009) 53–60. 339

[17] M. Diezd’aux, Ultrasonic wave measurement through cement paste backfill [MSc], 340

Toronto, Canada: University of Toronto; 2008. 341

[18] A.M. Galaa, B. D. Thompson, M.W. Grabinsky, W.F. Bawden, Characterizing stiffness 342

development in hydrating mine backfill using ultrasonic wave measurements, Can. 343

Geotech. J. 48 (2011) 1174–1187. 344

[19] S.U.A. Dulaijan, M. Maslehuddin, M.M.A. Zahrani, A.M. Sharif, M. Shameem, M. 345

İbrahim, Sulfate resistance of plain and blended cements exposed to varying concentrations 346

of sodium sulphate, Cement Concrete Comp. 25 (2003) 429–37. 347

17

[20] M. Şahmaran, O. Kasap, K. Duru, İ.O. Yaman, Effects of mix composition and water–348

cement ratio on the sulfate resistance of blended cements, Cement Concrete Comp. 29 349

(2007) 159–67. 350

[21] ASTM C 143-90, Standard test method for slump of hydraulic cement concrete. Annual 351

Book of ASTM Standards, American Society of Testing Material; 2002. 352

[22] ASTM C 597, Standard test method for pulse velocity through concrete. Annual Book of 353

ASTM Standards, American Society of Testing Material; 2009. 354

[23] ASTM C 39, Standard test method for compressive strength of cylindrical concrete 355

specimens. Annual Book of ASTM Standards, American Society of Testing Material; 2002. 356

[24] B. Ercikdi, F. Cihangir, A. Kesimal, H. Deveci, İ. Alp, Utilization of industrial waste 357

products as pozzolanic material in cemented paste backfill of high sulphide mill tailings, J. 358

Hazard. Mater. 168 (2009b) 848–856. 359

[25] F.W. Brackebusch, Basics of paste backfill systems, Eng. Min. J. 46 (1994) 1175–1178. 360

[26] E. Yılmaz, M. Benzaazoua, T. Belem, B. Bussiere, Effect of curing under pressure on 361

compressive strength development of cemented paste backfill, Miner. Eng. 22 (2009) 772–362

785. 363

[27] M. Gesoğlu, Influence of steam curing on the properties of concretes incorporating 364

metakaolin and silica fume, Mater. Struct. 43 (2010) 1123–1134. 365

[28] G. Ye, P. Lura, K.V. Breugel, A.L. Fraaij, Study on the development of the 366

microstructure in cement based materials by means of numerical simulation and ultrasonic 367

pulse velocity measurement, Cement Concrete Comp. 26 (2004) 491–497. 368

[29] L.M. Riodel, A. Jimenez, F. Lopez, F.J. Rosa, M.M. Rufo, J.M. Paniagua, 369

Characterization and hardening of concrete with ultrasonic testing, Ultrasonics 42 (2004) 370

527–530. 371

18

[30] Z. Lafhaj, M. Goueygou, A. Djerbi, M. Kaczmarek, Correlation between porosity, 372

permeability and ultrasonic parameters of mortar with variable water/cement and water 373

content, Cement Concrete Res. 36 (2006) 625–633. 374

[31] N.V. Mahure, G.K. Vijh, P. Sharma, N. Sivakumar, M. Ratnam, Correlation between 375

pulse velocity and compressive strength of concrete, Int. J. Earth Sci. 4 (2011) 871–874. 376

[32] M. Fall, M. Benzaazoua, S. Ouellet, Experimental characterization of the influence of 377

tailings fineness and density on the quality of cemented paste backfill, Miner. Eng. 18 378

(2005) 41–44. 379

[33] B. Ercikdi, H. Baki, M. İzki, Effect of desliming of sulphide rich mill tailings on the long 380

term strength of cemented paste backfill, J. Environ. Manage. 115 (2013) 5–13. 381

[34] T.N. Singh, S. Kripamoy, Geotechnical investigation of Amiyan landslide hazard zone in 382

Himalayan region, Uttaranchal India, In: Proceedings of the First International Conference 383

on Geotechnical Engineering for Disaster Mitigation and Rehabilitation, Singapore, 2005. 384

p. 355-360. 385

[35] K. Karaman, F. Cihangir, B. Ercikdi, A. Kesimal, The effect of specimen length on 386

ultrasonic P-wave velocity in clayey-carbonate rocks, Madencilik 49 (2010) 37–45 (in 387

Turkish). 388

[36] B. Ercikdi, F. Cihangir, A. Kesimal, H. Deveci, İ. Alp, Effect of drainage conditions on 389

the strength of paste backfill, Madencilik 47 (2008); 47: 15–24 (in Turkish). 390

[37] T. Wichtmann, T. Triantafyllidis, On the influence of the grain size distribution curve on 391

P-wave velocity, constrained elastic modulus Mmax and Poisson’s ratio of quartz sands, Soil 392

Dyn. Earthq. Eng. 30 (2010) 757–766. 393

394

395

396

19

397

398

399

400

LIST OF FIGURES 401

Figure 1. Preparation and testing of CPB samples: tailings collection (a); mixing (b); UPV 402

tests (c) and UCS tests (d). 403

Figure 2. Effect of binder type on the strength (a) and ultrasonic (b) properties of CPB 404

samples prepared from the as-received tailings at a fixed binder dosage. 405

Figure 3. Effect of binder dosage on the strength (a) and ultrasonic (b) properties of CPB 406

samples prepared from the as-received tailings. 407

Figure 4. Effect of w/c ratio on the strength (a) and ultrasonic (b) properties of CPB samples 408

prepared from the as-received tailings. 409

Figure 5. Effect of fines content (<20 µm) on the strength (a) and ultrasonic (b) properties of 410

CPB samples prepared from the deslimed tailings. 411

Figure 6. Relationship between UCS and UPV for all CPB samples (96 specimens) 412

Figure 7. Relationship between UCS and UPV for CPB samples produced from all (a) and 413

each (b) binders at 7 wt.% binder dosage. 414

Figure 8. Relationship between UCS and UPV for CPB samples produced from all (a) and 415

each (b) binder dosages. 416

Figure 9. Relationship between UCS and UPV for CPB samples produced from all (a) and 417

each (b) w/c ratios. 418

Figure 10. Relationship between UCS and UPV for CPB samples produced from all (a) and 419

each (b) fines content (<20 µm). 420

20

Fig. 1 421

(a) (b)

(c) (d)

21

Fig. 2 422

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 14 28 42 56

Com

pres

sive

str

engt

h (M

Pa)

Curing time (days)

CEM I 42.5 R CEM III/A 42.5 N SRC 32.5

1200

1300

1400

1500

1600

0 14 28 42 56

P-

wav

e ve

loci

ty (

m/s

)

Curing time (days)

CEM I 42.5 R CEM III/A 42.5 N SRC 32.5

(a)

(b)

22

Fig. 3 423

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 14 28 42 56

Com

pres

sive

str

engt

h (M

Pa)

Curing time (days)

5 wt. % 6 wt. % 7 wt. %

1100

1200

1300

1400

1500

1600

0 14 28 42 56

P-

wav

e ve

loci

ty (

m/s

)

Curing time (days)

5 wt. % 6 wt. % 7 wt. %

(a)

(b)

23

Fig. 4 424

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 14 28 42 56

Com

pres

sive

str

engt

h (M

Pa)

Curing time (days)

w/c: 4.62 w/c: 4.87 w/c: 5.13

1100

1200

1300

1400

1500

1600

1700

0 14 28 42 56

P-

wav

e ve

loci

ty (

m/s

)

Curing time (days)

w/c: 4.62 w/c: 4.87 w/c: 5.13

(a)

(b)

24

Fig. 5 425

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 14 28 42 56

Com

ress

ive

stre

ngth

(MP

a)

Curing time (days)

As-received Deslimed

1100

1200

1300

1400

1500

1600

1700

1800

1900

0 14 28 42 56

P-

wav

e ve

loci

ty (

m/s

)

Curing time (days)

As-received Deslimed

(a)

(b)

25

Fig. 6 426

UCS = 0.0011UPV - 1.0397r= 0.86

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1100 1200 1300 1400 1500 1600 1700 1800 1900

UC

S (

MP

a)

UPV (m/s)

26

Fig. 7 427

UCS = 0.0015UPV - 1.5151r= 0.83

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1200 1300 1400 1500 1600

UC

S (

MP

a)

UPV (m/s)

CEM I 42.5 R

UCS = 0.0013UPV - 1.1473r = 0.98

CEM III/A 42.5 N

UCS = 0.0013UPV - 1.3370r= 0.94

SRC 32.5

UCS = 0.0011UPV - 1.0318r = 0.99

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1200 1300 1400 1500 1600

UC

S (

MP

a)

UPV (m/s)

(a)

(b)

27

Fig. 8 428

UCS = 0.0013UPV - 1.3212r= 0.90

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1100 1200 1300 1400 1500 1600

UC

S (

MP

a)

UPV (m/s)

5 wt.%

UCS = 0.0009UPV - 0.817r= 0.97

6 wt.%

UCS = 0.001UPV - 0.9699r= 0.93

7 wt.%

UCS = 0.0013UPV - 1.337r= 0.94

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1150 1250 1350 1450 1550 1650

UC

S (

MP

a)

UPV (m/s)

(a)

(b)

28

Fig. 9 429

UCS = 0.0014UPV - 1.3917r= 0.94

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1150 1250 1350 1450 1550 1650

UC

S (

MP

a)

UPV (m/s)

(a)

(b)

w/c=4.87

UCS = 0.0013UPV - 1.337r= 0.94

w/c= 4.62

UCS = 0.0015UPV - 1.5676r= 0.96

w/c= 5.13

UCS = 0.0016UPV - 1.6286r= 0.94

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1150 1250 1350 1450 1550 1650

UC

S (

MP

a)

UPV (m/s)

(b)

29

Fig. 10 430

431

UCS= 0.0009UPV- 0.7190r= 0.92

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1150 1350 1550 1750 1950

UC

S (

MP

a)

UPV (m/s)

As-received

UCS = 0.0013UPV - 1.3386r= 0.94

Deslimed

UCS = 0.0008UPV - 0.6827r= 0.93

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1150 1350 1550 1750 1950

UC

S (

MP

a)

UPV (m/s)

(a)

(b)

30

LIST OF TABLES 432

Table 1. Chemical, physical and mineralogical properties of tailings and binders. 433

Table 2. A summary of the experimental conditions used in the preparation of CPB samples. 434

Table 3. Results of t and F tests for the linear models obtained for the relationships between 435

the UCS and UPV 436

31

Table 1. 437

D30= Particle size at 30% passing 438

Characteristics

Tailings

As-received

(%)

Deslimed Tailings

(%)

CEM I 42.5 R (%)

CEM III/A 42.5 R (%)

SRC 32.5 (%)

Chemical composition

SiO2 25.80 21.21 20.57 27.58 20.88 Al2O3 6.68 5.42 4.81 7.04 3.84 Fe2O3 39.83 45.43 3.67 2.37 4.52 MgO 2.14 2.21 1.35 3.91 1.49 SO3 - - 2.97 2.91 2.84 CaO 2.79 1.64 65.27 52.75 64.56 Na2O 0.35 0.24 0.41 0.25 0.31 K2O 0.42 0.31 0.85 1.06 0.67 TiO2 0.43 0.38 0.45 0.40 0.33 P2O5 0.03 0.03 0.13 0.03 0.10 MnO 0.06 0.05 0.11 1.00 0.12 Cr2O3 0.02 0.017 0.075 0.015 0.177 Free CaO - - 1.19 - 0.43 Loss–on–ignition (LOI) 20.6 21.9 2.1 2.8 2.8 Total 99.15 98.84 99.90 99.21 99.87

Sulphide content (S-2) (%) 23.18 27.82 - - -

Pyrite content (FeS2) (%) 43.47 52.16 - - -

Physical properties Specific gravity (g/cm3) 3.66 3.81 3.14 3.08 3.27 Specific surface area (cm2/g) 4630 1810 4335 4260 3170 Coefficient of curvature (Cc=(D30)2/(D10 ×D60)

0.99 1.17 - - -

Coefficient of uniformity (Cu=(D60/D10)

11.00 6.54 - - -

Fines content (<20 µm) ( %) 58.40 35.00 - - - Mineralogical properties

Pyrite Quartz

Chlorite Calcite

Muscovite Albite

Pyrite Quartz

Chlorite Calcite

Ankerite

C3S: 58.44 C2S: 14.95 C3A: 6.54

C4AF: 11.16

- C3S: 61.96 - C2S: 13.18 - C3A: 2.54 - C4AF:13.74

32

Table 2. 439

Tailings

type

Solids content,

(SC)1, wt%

Binder dosage

(BD)2, wt%

Water to cement ratio

(w/c)3

Slump

(cm)

Binder dosage

74.58 5 6 7

6.82 19.05 As-received 5.68

4.87

Water to cement ratio 75.58

7 4.62 16.51

As-received 74.58 4.87 19.05 73.58 5.13 21.59

Binder type

As-received 74.58 7 4.87 19.05

Fines content (<20 µm) Deslimed 80.17 7 3.81 19.05

1

)(

)(100:

waterbinderdrytailingsdry

binderdrytailingsdry

MMM

MMxSC

++

+

−−

−−

; 2

)

)(100:

tailingsdrybinderdry

binderdry

MM

MxBD

−−

−

+

; 3

binderdry

water

M

Mcw

−

:/ ; (M: Weight) 440

33

Table 3. 441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

Parameter Correlation

coefficient (r) tcomputed ttabulated Fcomputed Ftabulated Equation number

Total (96 specimens) 0.86 108.649 ±1.664 11778.478 1.592 (2)

Binder type 0.83 4.586 ±1.80 21.029 2.82 (3)

Binder dosage 0.90 6.502 ±1.80 42.278 2.82 (4)

w/c ratio 0.94 8.814 ±1.80 77.684 2.82 (5)

Fines content (<20µm) 0.92 5.697 ±1.90 32.453 3.79 (6)

34

460

461

462

463

464

465

HIGHLIGHTS 466

►Desliming significantly affects the UPV and UCS properties of CPB. ►Binder type plays 467 an important role for UCS and UPV of CPB ►The UPV of CPB increases with increasing 468 binder dosage and reducing w/c ratio. ►There is a linear relation between UCS and UPV of 469 CPB. ►The UCS of CPB can be estimated by ultrasonic UPV test. 470

471 472