Experimental Study on the Effects of Stress-Induced Damage on the Microstructure and Mechanical Properties of Soft Rock

Wang, Hao; Wang, Yong-Yan; Yu, Zhuo-Qun; Li, Jian-Guang

doi:https://doi.org/10.1155/2021/6696614

Advances in Civil Engineering

On this page

Abstract Introduction Materials and Methods Results and Analysis Conclusions Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Research Article | Open Access

Volume 2021 | Article ID 6696614 | https://doi.org/10.1155/2021/6696614

Experimental Study on the Effects of Stress-Induced Damage on the Microstructure and Mechanical Properties of Soft Rock

Hao Wang,^1,2Yong-Yan Wang,¹Zhuo-Qun Yu,¹and Jian-Guang Li¹

Academic Editor: Guoyang Fu

Received28 Nov 2020

Revised13 Jan 2021

Accepted18 Jan 2021

Published31 Jan 2021

Abstract

After rocks are damaged under stress loading, the changes of their microstructural and mechanical properties are major factors that affect construction safety in geotechnical engineering projects. Studying the microstructures and mechanical behaviors of stress-damaged rocks can help better guide construction and reduce construction risks for geotechnical engineering projects. In this study, a sandstone was first artificially predamaged and then subsequently subjected to scanning electron microscopy (SEM) analysis, computed tomography (CT) scanning, and uniaxial compression testing. Afterwards, the rock microstructures were three-dimensionally (3D) reconstructed, and the pores were classified and characterized based on their diameters. Moreover, the microstructural and mechanical parameters of the rock were subjected to significance analysis. The results showed that as the stress-induced damage () increased, the uniaxial compressive strength () of the soft rock decreased by 13.7–31.8%; as increased from 11.2 to 19.6 MPa, the elastic modulus (E) of the soft rock increased by up to 28.8%; and as increased beyond 19.6 MPa, there was a significant (22.3%) decrease in E. Stress-induced damage significantly affected the spatial distribution of the pores’ structure of the soft rock. Changes in the spatial structure of the pores led to the formation of cracks. The microstructural parameters of the stress-damaged soft rock were correlated with its mechanical parameters.

1. Introduction

At the microscopic level, rocks are discontinuous and inhomogeneous and contain natural defects, such as pores, cracks, and joints [1]. After rocks sustain damage (e.g., stress-induced damage, chemical erosion, and fatigue-induced damage) from external causes, their original defects can develop and subsequently change their mechanical properties [2, 3]. Stress-induced damage is the principal type of damage that occurs during geotechnical engineering construction (e.g., tunneling and mining for underground resources). The mechanical properties of stress-damaged rocks affect construction safety in geotechnical engineering projects. Meanwhile soft rock is a type of porous material of which the internal microstructure dictates its mechanical properties. Therefore, it is critically important to accurately characterize and evaluate damage-induced microstructural changes in rocks.

With advancements in detection technology, scanning electron microscopy (SEM), acoustic emission (AE) testing [4, 5], and computed tomography (CT) scanning [6, 7] have been gradually applied to microstructural characterizations and evaluations of rocks. Chen et al. [8] examined the AE characteristics of sandstone with a uniaxial compression failure test and found that AEs could be used to characterize various stages of the uniaxial compression failure process of the sandstone. Shuang et al. [9] characterized the internal microstructural damage of rock at various loading rates using the AE technique and found that this technique could characterize the specific locations where internal damage occurred. To further explain the microscopic mechanisms by which rocks sustain damage, some researchers have employed SEM to characterize rocks at the microscopic level. Qin et al. [10] investigated the effects of water invasion-water loss cycles on the microstructure of rock by SEM and found that microstructural changes in the rock under water invasion-water loss cycles exhibited four stages. Yang et al. [11] studied microstructural changes of a rock under wetting-drying cycles and determined the changes in pore parameters (e.g., porosity, pore size, and pore shape) through SEM by increasing the number of wetting-drying cycles. Tao et al. [12] examined the evolution of cracks in elliptical-hole-bored granitic rocks under dynamic loading and used SEM to characterize the responses of the constituent minerals at different stages of the failure process. SEM can be used to determine the microstructural forms and morphologies of damaged rocks. Compared to the AE technique, SEM is capable of characterizing damage-induced microstructural changes in rocks in a more accurate and comprehensive fashion. However, both the SEM and AE methods are only able to qualitatively characterize rock microstructures. Inspired by medical CT scanning, some researchers have applied CT to obtain quantitative characterizations of rock microstructures. As evidenced by its working principle [13, 14], CT scanning is superior to AE and SEM techniques for characterizing microstructural damage in rocks. Some researchers have characterized and evaluated damaged rocks by the use of CT scanning [15–18]. All of these research results have demonstrated the tremendous advantages of CT scanning for characterizing rock microstructures.

Currently, CT scanning is the most advanced technical method for characterizing and evaluating rock microstructure. Two-dimensional (2D) CT scan slices can accurately characterize rock microstructures. However, 2D CT scan slices are unable to characterize the spatial distributions of rock microstructures and their changes. So three-dimensional (3D) spatial characterization of rock microstructures has become a research hotspot and has been pursued by numerous researchers. Nevertheless, the use of CT and 3D reconstruction techniques to examine the effects of stress-induced damage on rock microstructures has yet to be studied adequately in depth. In this study, SEM, CT, and 3D reconstructions were combined to qualitatively and quantitatively characterize microstructural changes in rock after sustaining damage under stress loading. In addition, the mechanism of microstructural changes was elucidated. Moreover, the mechanical and microstructural parameters of the rock were analyzed using statistical analysis methods.

2. Materials and Methods

Figure 1 shows the experimental procedures and equipment. There were four steps in the experiments. Firstly, rocks were machined into standard samples, which were maintained for 28 days. Secondly, a Hitachi SU8000 SEM system was used to observe the slices extracted from the samples. Then, a nanoVoxel 3000 high-resolution micro-CT scanner was employed to scan the samples. Finally, TWA-200 electronic material mechanical testing machine was used to perform uniaxial compression tests on the samples which were scanned by nanoVoxel 3000 scanner.

2.1. Sample Preparation

A kind of yellow fine-grained sandstone, which was collected from the Yanzhou coal mine in Jining, Shandong, China, was investigated in this study. According to international rock mechanics standards [19, 20] and GB/T 50123-1999, these rocks were machined into standard samples as shown in Figure 1(a). Each sample was 100 ± 2 mm in height and 50 ± 1 mm in diameter and had a mass of 385 ± 5 g. The end-surface flatness tolerance was set to ±0.05 mm. The uniaxial compressive strength () and elastic modulus (E) of the samples were measured to be 28 ± 0.5 MPa and 4.5 ± 0.3 GPa, respectively. According to the study of Bieniawski and Bernede [20], when the load () is 40–80% of its , the internal pores and cracks develop in an unstable fashion, which can damage the rock. On this basis, in this study, the samples were artificially damaged. Specifically, that were 40%, 50%, 60%, 70%, and 80% of were applied separately to five samples. Thus, samples that were damaged to various extents were obtained. These samples are denoted as B-S-2, C-S-1, D-S-3, E-S-3, and F-S-2, respectively.

2.2. SEM Observations

The aim of this section was to determine the minimum diameter of pores in the samples by SEM observations and therefore provide a reliable parameter for selecting a suitable resolution for CT scanning. A Hitachi SU8000 SEM system which was shown in Figure 1(e) was used. Three cross sections of a rock sample that had not been artificially damaged were selected from top to bottom at height intervals of 25 mm for observations. Three rectangular slices with identical dimensions were then selected from each of the three cross-sectional surfaces. Figure 1(c) shows the locations of the selected slices. These slices were subsequently observed at two primary magnifications (1000x and 2000x) and one complementary magnification (5000x). Figure 2 shows the SEM images obtained. Based on the SEM images, the minimum diameter of the internal pores of the sample was determined to be 18–25 μm.

2.3. CT Scanning

nanoVoxel 3000 high-resolution micro-CT scanner (Tianjin San Ying Precision Instrument Co., Ltd., Tianjin, Hebei, China), as shown in Figure 1(g), was used to scan the samples. This scanner provided a true spatial resolution of 500 nm and minimum voxel size of 70 nm and thus met the accuracy required for tests. The minimum pore diameter (25 μm), as determined by SEM observations, was taken as the minimum resolution of the CT scans. In the tests, 1440-frame helical CT scanning was performed at a voltage of 180 kV and current of 40 mA. A total of 4,000 2D CT slices with dimensions of 1,800 pixels × 1,800 pixels were obtained for each sample. The rock substrate, cement components, and pores and cracks are shown as white, gray, and black areas in Figure 3, respectively.

2.4. Uniaxial Compression Tests

The stress-damaged soft-rock samples were subjected to uniaxial compression failure testing using a TAW-200 electronic material mechanical testing machine. First, each sample was placed on the test platform of the testing machine, while the center of the sample, center of the base, and center of the indenter were located on the same axis. Then, a computer was used to control the downward movement of the testing machine indenter until its lower end contacted the upper surface of the sample. In addition, a certain prestress (150 N) was applied to prevent slipping of the indenter, sample, and test platform during the test. Finally, the load control mode was used during the test. According to the test standards [19, 20], each sample was loaded at a rate of 0.1 MPa/s up to 60 MPa or until failure. During testing, the computer was used to record the test data.

2.5. 3D Reconstruction

Damage-induced microstructural changes in a rock can be qualitatively characterized by 3D reconstruction. Microstructural parameters extracted from 3D reconstructions can quantitatively characterize microstructural changes in rocks. In this study, 3D reconstruction was achieved in four steps. First, an image from each obtained CT slice was cropped. Second, the cropped images were subjected to median filtering, threshold segmentation, and binarization. Third, the processed CT slices were imported in sequence into Avizo software (version 3.0) for 3D spatial modeling. Finally, the obtained 3D pore models were classified and characterized. Figure 4 shows the 3D reconstruction procedures for the rock.

2.6. Image Processing of CT Slices

In step 1 of the image processing, the CT slices were cropped. When preparing the samples, the pores at the edges of each sample were damaged during the cutting or polishing process, which could affect the observations and analyses of the true pore conditions [21]. An image of 459 pixels × 459 pixels was cropped from each CT slice, as shown in Figure 4(b). However, 4,000 CT slices greatly exceed the available computational capacity for 3D reconstructions. Thus, 1,000 consecutive CT slices were used for 3D reconstructions. In step 2, images that were cropped from the CT slices were subjected to filtering and threshold segmentation. A median filter was used in this study to eliminate noise points in the CT slices. Figure 4(c) shows images processed with a median filter using 3 × 3, 10 × 10, and 20 × 20 neighborhood windows. The principle is that the grayscale values of adjacent pixels are weighted and averaged to eliminate noise points. Subsequently, a scan line passing through both the pores and substrate in each median-filtered image was selected, as shown in Figure 4(d) [21]. Figure 4(e) shows the variations in grayscale values along the scan line. The grayscale values above and below the scan line are those of the substrate and pores, respectively. This line was used to determine the segmentation threshold T (T = 177). Finally, the grayscale values of the image were converted to 0 or 1 by using equation (1) to convert to black/white contrast images that reflect microscale characteristics, as shown in Figure 4(d).where and are the initial and postbinarization grayscale values of pixel , respectively.

2.7. Classified Characterization of Microstructure

The processed CT slices were imported in sequence into Avizo software (version 3.0). The pore structure of each sample was extracted and rendered to produce a 3D graph, as shown in Figure 5. All pores and fractures were reconstructed and rendered in the same colors as in the 3D reconstruction process. However, in this way, the effects of stress-induced damage on the microstructure of sample could not be easily represented. The SEM observations indicated that as increased, significant changes in the pore diameters of the sandstone occurred. Thus, the pores were classified and characterized based on their diameters. Yu used a pore classification that is applied in oil and gas reservoir geology area as a reference [22, 23], and the pores were classified into four types based on their diameters, namely, types I (25–75 μm), II (75–130 μm), III (130–250 μm), and IV (>250 μm). Figures 6–11 show the pore classifications and characterization graphs for the samples (types I, II, III, and IV pores are shown in blue, yellow, red, and green, respectively).

(a)

(b)

(c)

(d)

(e)

(f)

3. Results and Analysis

As shown in Figures 6–11, stress-induced damage significantly affected the spatial distributions of the pores. When = 0, all pore types were distributed uniformly in the sample, most pores had small diameters (25–130 μm), and the pores were poorly interlinked. In terms of the pore numbers (N), type I pores made up the largest proportion and were followed by type II, III, and IV pores. As increased, both the connectivity and N values for type I and II pores increased, and both the connectivity and N values for type III and IV pores decreased. The patterns for type III and IV pores were opposite those of type I and II pores. The connectivity and N values for type I and II pores peaked when = 19.6 MPa. As increased beyond 19.6 MPa, cracks formed in the samples, but the distribution of type I pores was not notably related to that of the cracks and was similar to that of the type I pores in the undamaged sample. Moreover, under this condition, type II, III, and IV pores were concentrated around the cracks.

Six parameters, namely, N, porosity (φ), and the proportions of type I, II, III, and IV pores (P_I, P_II, P_III, and P_IV, respectively), were calculated for the rock samples using Avizo software (version 3.0). As shown in Figure 12, N decreased as increased. The smallest N occurred for a value of 22.4 MPa. As increased, φ first decreased and then increased, with a minimum value of 5.7% and maximum value of 14.9%. The statistical results showed that as increased, P_I and P_II first increased and then decreased, with maximum values of 81.3% and 16.35% and minimum values of 65.5% and 9.4%, respectively. Notably, as increased, P_III and P_IV first decreased and then increased. This pattern for P_III and P_IV was opposite that of P_I and P_II.

(a)

(b)

(c)

To further investigate the mechanism of these microstructural changes, all samples were observed by SEM. Figure 13 shows the SEM images. Evidently, samples in their native state had intact crystal faces, crystals in close contact, and regular pores. Compared with the samples in their native state, some pores were closed in samples B-S-2 and C-S-1 under stress-induced damage, and their pore diameters were smaller. However, the microstructures for both samples B-S-2 and C-S-1 differed nonsignificantly from those of the samples in their native state. As a result of stress-induced damage, fine and long microcracks appeared on the internal crystal faces in sample D-S-3, and there were no accompanying cracks around the microcracks. In this sample, the crystal faces remained intact, but the pores were conspicuously closed. Due to the stress-induced damage, cracks began to develop on the crystal faces in sample E-S-3, and accompanying “wing-shaped” cracks appeared around the main cracks. In addition, intergranular cracks began to develop, and the cracks became wide. However, in this sample, the crystal faces remained intact. Moreover, notable pore closure occurred, and free detritus emerged near the pores in sample E-S-3. Because of the stress-induced damage, significant microstructural deterioration occurred for sample F-S-2. Its crystal faces showed breakage, shedding, and dislocation, and its pores fractured.

(a)

(b)

(c)

(d)

(e)

(f)

The uniaxial compression test results showed that stress-induced damage reduced the values of rock samples. However, as increased from 11.2 to 19.6 MPa, the E values of the rock samples increased, as is shown in Figure 14. The correlations of the microstructures with and E were analyzed by a significance test [24]. Microstructural and mechanical parameters were subjected to significance tests using SPSS software. Table 1 summarizes the results. The values of the soft rock were positively correlated with N. The E values of the soft rock were uncorrelated with N and φ. The E values of the soft rock were positively correlated with P_I but were negatively correlated with P_III and P_IV. The microstructural and mechanical parameters were fitted. Figure 15 shows the results.

(a)

(b)

(c)

(d)

4. Conclusions

In this study, samples of a soft rock which were damaged to various degrees under different stresses were subjected to SEM analysis, CT scanning, 3D reconstruction, and uniaxial compression testing. In addition, the sandstone pores were classified and characterized. The 3D reconstruction results visually displayed the effects of stress-induced damage on the internal microstructure of the soft rock. Finally, the correlations between the microstructural and mechanical parameters of the rock samples were analyzed.

The following conclusions were derived from this study: Stress-induced damage significantly affected the spatial distributions of internal pores of the rock. As increased to 19.6 MPa, no cracks appeared in the rock, small pores were uniformly distributed but became increasingly interlinked, and large pores became decreasingly interlinked. As increased beyond 19.6 MPa, noticeable microcracks appeared in the rock, and large pores were concentrated around cracks. Thus, from a microscopic perspective, crack formation was caused by spatial adjustment of the pore structure of the rock. As increased, the value of the rock decreased, albeit nonlinearly. As increased, the E value of the rock first increased and then decreased. As increased to 19.6 MPa, E increased. As increased beyond 19.6 MPa, E decreased considerably. The microstructural parameters of the stress-damaged soft rock were correlated with the mechanical parameters to a certain extent. In the stress-damaged soft rock, and N were strongly positively correlated. E of the soft rock was positively correlated with P_I but negatively correlated with P_III and P_IV. The results of this study can serve as a reference for evaluating rock damage and determining the microstructural mechanism of damage evolution in rocks. Moreover, the experimental design and approach used in this study can provide a basis for studying microstructural changes in and mechanical behaviors of rocks after sustaining damage. Furthermore, this study is valuable for safety evaluations and early-warning provisions for geotechnical engineering projects.

Data Availability

The datasets used or analyzed during the current study are available from the corresponding author upon reasonable request.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgments

This work was supported in part by the National Natural Science Foundation of China (51674149) and Shandong Provincial Natural Science Foundation, China (ZR2019MEE082).

References

J. B. Zhu, T. Zhou, Z. Y. Liao, L. Sun, X. B. Li, and R. Chen, “Replication of internal defects and investigation of mechanical and fracture behaviour of rock using 3D printing and 3D numerical methods in combination with X-ray computerized tomography,” International Journal of Rock Mechanics and Mining Sciences, vol. 106, pp. 198–212, 2018.
View at: Publisher Site | Google Scholar
W. Wang, S. Zhang, H. Li, S. Gong, and Z. Liu, “Analysis of the dynamic impact mechanical characteristics of prestressed saturated fractured coal and rock,” Advances in Civil Engineering, vol. 2019, Article ID 5125923, 10 pages, 2019.
View at: Publisher Site | Google Scholar
Y. Chen and H. Lin, “Consistency analysis of Hoek-Brown and equivalent Mohr-Coulomb parameters in calculating slope safety factor,” Bulletin of Engineering Geology and the Environment, vol. 78, no. 6, pp. 4349–4361, 2019.
View at: Publisher Site | Google Scholar
D. A. Lockner, J. D. Byerlee, V. Kuksenko, A. Ponomarev, and A. Sidorin, “Quasi-static fault growth and shear fracture energy in granite,” Nature, vol. 350, no. 6313, pp. 39–42, 1991.
View at: Publisher Site | Google Scholar
E. Townend, B. D. Thompson, P. M. Benson, P. G. Meredith, P. Baud, and R. P. Young, “Imaging compaction band propagation in Diemelstadt sandstone using acoustic emission locations,” Geophysical Research Letters, vol. 35, no. 15, Article ID L15301, 2008.
View at: Publisher Site | Google Scholar
S.-Q. Yang, P. G. Ranjith, Y.-H. Huang et al., “Experimental investigation on mechanical damage characteristics of sandstone under triaxial cyclic loading,” Geophysical Journal International, vol. 201, no. 2, pp. 662–682, 2015.
View at: Publisher Site | Google Scholar
H. Kawakata, A. Cho, T. Kiyama, T. Yanagidani, K. Kusunose, and M. Shimada, “Three-dimensional observations of faulting process in westerly granite under uniaxial and triaxial conditions by X-ray CT scan,” Tectonophysics, vol. 313, no. 3, pp. 293–305, 1999.
View at: Publisher Site | Google Scholar
Y. Chen, Y. Yang, F. Gao, and X. Zhang, “Researches on damage evolution and acoustic emission characteristics of rocks,” Advances in Civil Engineering, vol. 2018, no. 4, Article ID 3108065, 7 pages, 2018.
View at: Publisher Site | Google Scholar
H. Q. Shuang, S. G. Li, L. Liu, G. F. Chen, and K. I. Song, “Use of acoustic emission for the detection of brittle rock failure under various loading rates,” Advances in Civil Engineering, vol. 2018, Article ID 5735139, 9 pages, 2018.
View at: Publisher Site | Google Scholar
Z. Qin, H. Fu, and X. Chen, “A study on altered granite meso-damage mechanisms due to water invasion-water loss cycles,” Environmental Earth Sciences, vol. 78, no. 14, p. 428, 2019.
View at: Publisher Site | Google Scholar
X. Yang, J. Wang, C. Zhu, M. He, and Y. Gao, “Effect of wetting and drying cycles on microstructure of rock based on SEM,” Environmental Earth Sciences, vol. 78, no. 6, p. 183, 2019.
View at: Publisher Site | Google Scholar
M. Tao, H. Zhao, A. Momeni, Y. Wang, and W. Cao, “Fracture failure analysis of elliptical hole bored granodiorite rocks under impact loads,” Theoretical and Applied Fracture Mechanics, vol. 107, Article ID 102516, 2020.
View at: Publisher Site | Google Scholar
R. A. Ketcham and W. D. Carlson, “Acquisition, optimization and interpretation of X-ray computed tomographic imagery: applications to the geosciences,” Computers & Geosciences, vol. 27, no. 4, pp. 381–400, 2001.
View at: Publisher Site | Google Scholar
L. B. Wang, J. D. Frost, G. Z. Voyiadjis, and T. P. Harman, “Quantification of damage parameters using X-ray tomography images,” Mechanics of Materials, vol. 35, no. 8, pp. 777–790, 2003.
View at: Publisher Site | Google Scholar
Y. Wang, D. Zhang, and Y. Z. Hu, “X-ray computed tomography characterization of soil and rock mixture under cyclic triaxial testing: the effects of confining pressure on meso-structural changes,” Environmental Earth Sciences, vol. 78, no. 6, p. 185, 2019.
View at: Publisher Site | Google Scholar
H. Liu, G. Yang, Y. Yun et al., “Investigation of sandstone mesostructure damage caused by freeze-thaw cycles via CT image enhancement technology,” Advances in Civil Engineering, vol. 2020, Article ID 8875814, 13 pages, 2020.
View at: Publisher Site | Google Scholar
Y. Wang, C. Li, J. Han, and H. Wang, “Mechanical behaviours of granite containing two flaws under uniaxial increasing-amplitude fatigue loading conditions: an insight into fracture evolution analyses,” Fatigue & Fracture of Engineering Materials & Structures, vol. 43, no. 9, pp. 2055–2070, 2020.
View at: Publisher Site | Google Scholar
S. Li, R. Huo, B. Wang et al., “Experimental study on physicomechanical properties of sandstone under acidic environment,” Advances in Civil Engineering, vol. 2018, no. 4, Article ID 5784831, 15 pages, 2018.
View at: Publisher Site | Google Scholar
Z. T. Bieniawski and M. J. Bernede, “Suggested methods for determining the uniaxial compressive strength and deformability of rock materials. 1. Suggested method for determination of the uniaxial compressive strength of rock materials,” International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, vol. 16, no. 2, pp. 137-138, 1979.
View at: Publisher Site | Google Scholar
Z. T. Bieniawski and M. J. Bernede, “Suggested methods for determining the uniaxial compressive strength and deformability of rock materials. 2. Suggested method for determining deformability of rock materials in uniaxial compression,” International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, vol. 16, no. 2, pp. 138–140, 1979.
View at: Publisher Site | Google Scholar
Y. Lang, Z. Liang, D. Duan, and Z. Cao, “Three-dimensional parallel numerical simulation of porous rocks based on CT technology and digital image processing,” Rock and Soil Mechanics, vol. 40, no. 3, pp. 1204–1212, 2019.
View at: Google Scholar
B. Yu, “Classification and characterization of gas shale pore system,” Earth Science Frontiers, vol. 20, no. 4, pp. 211–220, 2013.
View at: Google Scholar
G. R. Chalmers, R. M. Bustin, and I. M. Power, “Characterization of gas shale pore systems by porosimetry, pycnometry, surface area, and field emission scanning electron microscopy/transmission electron microscopy image analyses: examples from the Barnett, Woodford, Haynesville, Marcellus, and Doig units,” AAPG Bulletin, vol. 96, no. 6, pp. 1099–1119, 2012.
View at: Publisher Site | Google Scholar
Z. Wu and N. E. Huang, “Statistical significance test of intrinsic mode functions,” in Hilbert-Huang Transform and Its Applications, N. E. Huang and S. S. P. Shen, Eds., pp. 107–127, World Scientific, Singapore, 2005.
View at: Google Scholar

Copyright

Copyright © 2021 Hao Wang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies