The amount of fine material available in the soil is significant in Atterberg limits and methylene blue tests. In the context of Atterberg limits, increased amount of clay minerals contained in the soil increases liquid and plastic limit values; however, increasing sand content reduces the moisture content reducing the water retention capacity of the soil which in return reduces the plasticity index (PI) value. In the case of methylene blue test, which is used to specify the quality of the amount of fine material, existence of clay in the medium increases the pollution level of the sand and the amount of methylene solution (V 1) used. In this study, soil classes were determined and pollution rates were identified with Atterberg limits, pycnometer, sieve analysis, hydrometer analysis and methylene blue tests conducted on 11 different natural soil samples collected from different regions. From the data obtained, first the relationship between PI and methylene blue (MB) was examined and was evaluated according to the results of the “single regression” method. Furthermore, the other coefficient of uniformity (C u), coefficient of graduation (C c), unit weight of soils ( γ s) parameters obtained from experimental studies were also subjected to “multiple regression analysis” in order to reveal their impact on the MB and this impact was confirmed taking both statistical analyses into account.
Soil is a material which may also include organic matter and which is generated as a result of the mechanical disruption of the Earth’s crust by the erosive effects of water, ice and wind or by chemical processes which create solid grains and patches of water and/or air between them [
Also known as limits of consistency, “Atterberg limits” were first proposed by Atterberg (1911) [
Limits of consistency are a feature of clay and silts which are classified under fine-grained soil group and as one may know these soils absorb water and they can be found in various phases such as solid or liquid depending on their water content [
Thus, it is possible to explain limits of consistency in terms of the “volume (V)- water content (W) relationship” (
As shown in
A review of the standards for limits of consistency measurements reveals that the measurement is commonly conducted on the soil passing through #40 mesh. However, fine-grained sand is also among the soil group which is able to pass through this sieve [
Gündüz and Dağdeviren (2009), explored the impact of grain size and its amount on some parameters of fine-grained soils. Their findings showed that sand grains available in the soil significantly reduce the liquid and plastic limit values, in other words, amount of water needed for the medium to change its
phase to liquid, and that the soil class itself can be changed according to the plasticity value fine-grains actually have. Nevertheless, during the experiments it was also observed that some of the samples could become non-plastic (semi- solid transition without plastic behavior) with the effect of increased sand content. Therefore, it was claimed that liquid and plastic limit measurements must be conducted on the fine-grained soil which passes through #200 sieve, in order to eliminate some of the negative effects that may occur [
Topçu (2011), reported that miscalculation of the plasticity value of the soil could lead to serious problems in engineering applications. In his study, Topçu showed that liquid and plastic limits of material accumulated under # 200 sieve were significantly higher than the material accumulated under #40 sieve, and suggested that that the reason behind this difference is the increase in the specific surface of the soil [
Also important for soil classification is sieve and hydrometers analyses, as well as Atterberg limits.
The sieves used in sieve analysis usually consist of square meshes and the width of these square meshes are called mesh diameter. The finest sieve used in the soil survey mechanics is the sieve #200. Material passing this sieve are classified under silt and clay, while sand and pebbles are not able to pass the sieve. Therefore, 200 sieve is used to separate fine-grained and coarse-grained soil while 40 sieve is used to separate sand and pebbles. Sieve analysis helps us define the grain size distribution of the grains larger than the mesh size of 200 sieve (0.074 mm).
Hydrometer analysis is also a method used to determine the grain size distribution of the fine-grained soils passing through 200 sieve and to find the weight percentage of the clay-silt group in the soil. Sedimentation analysis is based on the Stokes law, which gives the relation between the rate of sedimentation and the diameter of spherical sediments in a liquid [
Methylene blue point test, on the other hand, is used to determine the clay content of concrete and mortar aggregates which is available between fine grains below 0.063 mm in diameter. It also allows the determination of ion adsorption capacity of the soil with the specification of the amount of methylene blue required to cover the entire surface area of the clay grains available in the soil. This test determines the amount of harmful clay minerals but cannot determine the rate of damage they are able to cause. Presence of excessive amounts of clay minerals in the concrete increases the amount of water needed for concrete and mortar, which leads to impaired hardened concrete and mortar strength, their durability and volumetric stability [
Chiappona et al., (2004) conducted a series of studies on the applicability of the methylene blue test used to define clay minerals as suggested by the standard methods used in France and the US and the findings of these studies showed that the test method defined by the US standards is suitable for fine and homogeneous material, in other words, they only provide information for the determination of the clay content, and that the test method, while the method defined by the French standard must be used on heterogeneous materials in order to provide information representing the material as a whole [
LL, PI, engineering classifications and other engineering properties are parameters used in the ground survey to define the soil. However, identification of LL and PI is not a part of routine soil study definition analysis, and is expensive and time consuming. These are conducted only once, and such data are not generally useful [
In this study, 11 different natural soil samples taken from different regions were used and these regions are shown in
In the pycnometer (specific gravity) experiment; samples were sieved using 40 sieve, then dried at 105˚C in a drying oven for 24 hours and the dried material was ground using a wooden mallet and 50 g of the sample passing 40 sieve was taken. Since the grain sizes of the soil samples used in the experiment were often small, only unit bulk density of fine-grained soil was identified using 500 ml density bottle and the process was conducted in accordance with ASTM D 854 [
Soil class of the samples were defined using liquid limit and plastic limit tests, sieve analysis and hydrometer analysis in accordance with Unified Soil Classification System (USCS).
The liquid limit test can be carried out according to ASTM D 4318 [
Plastic limit test was also performed based on ASTM D 4318 standards using approximately 20 g of the mixture prepared for the liquid limit test, as plastic limit test is conducted in connection with the liquid limit test.
Sample # | Latitude | Longitude |
---|---|---|
1 | 41.404645 | 33.782069 |
2 | 41.430122 | 33.774082 |
3 | 41.428529 | 33.782990 |
4 | 41.433015 | 33.763340 |
5 | 41.384653 | 33.782363 |
6 | 41.434363 | 33.764475 |
7 | 41.431279 | 33.770598 |
8 | 41.428352 | 33.783095 |
9 | 41.432990 | 33.764111 |
10 | 41.432798 | 33.778421 |
11 | 41.361626 | 33.759008 |
The sieve analysis test was carried out on a 500 g sample using 200 sieve in accordance with ASTM C 136 [
During the hydrometer test, it was tried to adhere to the ASTM D 422-63 [
Methylene blue test is conducted based on the principles and regulations of TS EN 933-9 + A1 [
Here;
MB = Methylene blue value (%);
V1 = Total volume of the methylene solution added (ml);
M1 = Mass of the experiment sample (g).
Factor 10 = A factor used to convert the volume of stain solution used to the mass of stain per kilogram of the mass tested (TSEN 933-9+A1) [
Pycnometer test was conducted in order to be used in hydrometer analysis in
MB = (V1/M1) × 10 Experiment # | Amount of methylene solution used (V1) (ml) | MB (%) |
---|---|---|
1 | 1140 | 57.00 |
2 | 625 | 31.25 |
3 | 1373 | 68.65 |
4 | 895 | 44.75 |
5 | 1223 | 61.15 |
6 | 1520 | 76.00 |
7 | 846 | 42.30 |
8 | 1663 | 83.15 |
9 | 640 | 32.00 |
10 | 1784 | 89.20 |
11 | 698 | 34.92 |
accordance with ASTM D 854 and the specific bulk density of the soils tested were found to be in the range of 2.4 - 2.8 g/cm3. The results are shown in
Liquid limit tests were performed until at least 3 - 4 pulses were detected in the range of 10 - 40 while the plastic limit tests were conducted twice paying attention to at least 6g sample available in each container. Then, the test results are plotted in tables created in MS Excel and the number of pulses obtained is plotted logarithmically in x axis while the corresponding water content (W) is plotted in y axis in order to derive a flow curve. In this flow curve, the value of the water content corresponding to 25 pulses gives us the “liquid limit” value of that soil. In all experiments, the WL value was found both with an “estimated” value calculated by the system and a manual drawing on the flow curve. Plasticity index is calculated on the basis of manual calculation from values obtained and these were compared against estimated values. Compatibility mode (correlation) for WL values calculated by the system for each one of 11 tests was in the range between 92.3% and 99.5%. The main purpose here was to establish the correlation between the ratio of fine-grains and the pollution comparing the PI values and the percent of methylene solution used and the results of this comparison are addressed in the methylene test results section. Liquid limit and plastic limit test results are shown in
In this study, the same soil samples are used for all experiments, experiment results of sieve and hydrometer analysis assessed in combination and soil class comparison was made in accordance with TS 1500/1900-1 [
Experiment # | γs | WT (%) | WM (%) | PIT | PIM | Soil Type | Cu | Cc | MB (%) |
---|---|---|---|---|---|---|---|---|---|
1 | 2.65 | 53.50 | 52.00 | 25.01 | 23.51 | CH | 5.70 | 0.85 | 57.00 |
2 | 2.60 | 37.20 | 36.80 | 16.76 | 16.36 | CL | 5.95 | 1.62 | 31.25 |
3 | 2.52 | 58.90 | 58.50 | 25.99 | 25.59 | MH | 5.62 | 0.68 | 68.65 |
4 | 2.73 | 47.60 | 48.00 | 26.25 | 26.65 | CL | 5.00 | 0.85 | 44.75 |
5 | 2.49 | 58.00 | 57.40 | 24.97 | 24.37 | MH | 5.65 | 0.45 | 61.15 |
6 | 2.64 | 58.40 | 58.00 | 40.59 | 40.19 | CH | 6.31 | 0.30 | 76.00 |
7 | 2.62 | 56.7 | 56.5 | 25.46 | 25.26 | MH | 3.38 | 0.46 | 42.30 |
8 | 2.72 | 41.10 | 40.68 | 25.86 | 25.44 | CL | 8.53 | 0.79 | 83.15 |
9 | 2.60 | 44.70 | 44.20 | 16.59 | 16.09 | ML | 4.67 | 0.86 | 32.00 |
10 | 2.65 | 59.00 | 59.00 | 37.47 | 37.47 | CH | 3.12 | 0.49 | 89.20 |
11 | 2.72 | 46.60 | 45.90 | 16.51 | 15.81 | ML | 4.95 | 0.54 | 34.92 |
2487 [
Cu and Cc values were obtained using the granulometry curve, having identified the diameters of D60, D30 ve D10 which corresponds to the material passing the sieve at the percentages of 60%, 30%, 10%, respectively (
Here, soil sample #8 is the only coarse-grained soil sample, and was classified under sand as more than 50% of its content passed the sieve and is represented with an “S” symbol.
Cu and Cc coefficients for pebble and sand are as follows for sand (sample #8);
Cc = 0.79
“SC” symbol was used as the Atterberg limits are above the A line or Ip > 7. Hence, the class of the soil sample #8 is assigned as “bad-graded argillaceous sand”.
Soil classes of the remaining 10 test samples were determined using the plasticity graph. The point where WL and PI values of these soils were intersected was identified as the class of the soil.
Methylene blue test was conducted on 11 test samples with regards to the principles and regulations of TS EN 933-9 + A1: 2013 (2014). With this experiment, pollution levels of fine-grained soils were defined in an attempt to establish the correlation between PI values of the same soil sample.
Two different comparisons were made in the above
MB values corresponding to the “estimated” PI values calculated by the system in accordance with the 11 test results | Experiment # | MB | PI | |
---|---|---|---|---|
1 | 57.00 | 25.01 | ||
2 | 31.25 | 16.76 | ||
3 | 68.65 | 25.99 | ||
4 | 44.75 | 26.25 | ||
5 | 61.15 | 24.97 | ||
6 | 76.00 | 40.59 | ||
7 | 42.30 | 25.46 | ||
8 | 83.15 | 25.86 | ||
9 | 32.00 | 16.59 | ||
10 | 89.20 | 37.47 | ||
11 | 34.92 | 16.51 | ||
MB values corresponding to the PI values obtained from 11 tests with “manually” drawn flow chart | Experiment # | MB | PI | |
1 | 57.00 | 23.51 | ||
2 | 31.25 | 16.36 | ||
3 | 68.65 | 25.59 | ||
4 | 44.75 | 26.65 | ||
5 | 61.15 | 24.37 | ||
6 | 76.00 | 40.19 | ||
7 | 42.3 | 25.26 | ||
8 | 83.15 | 25.44 | ||
9 | 32.00 | 16.09 | ||
10 | 89.20 | 37.47 | ||
11 | 34.92 | 15.81 |
Experiment # | MB | PI | ||
---|---|---|---|---|
“Estimated” test results for 9 tests | 1 | 57.00 | 25.01 | |
2 | 31.25 | 16.76 | ||
3 | 68.65 | 25.99 | ||
5 | 61.15 | 24.97 | ||
6 | 76.00 | 40.59 | ||
7 | 42.30 | 25.46 | ||
9 | 32.00 | 16.59 | ||
10 | 89.20 | 37.47 | ||
11 | 34.92 | 16.51 | ||
Experiment # | MB | PI | ||
“Manually” calculated test results for 9 tests | 1 | 57.00 | 23.51 | |
2 | 31.25 | 16.36 | ||
3 | 68.65 | 25.59 | ||
5 | 61.15 | 24.37 | ||
6 | 76.00 | 40.19 | ||
7 | 42.3 | 25.26 | ||
9 | 32.00 | 16.09 | ||
10 | 89.20 | 37.47 | ||
11 | 34.92 | 15.81 |
and the correlation was found to be y = 1.6148x0.6811. A closer investigation of the plasticity values shows that the effect of the experiments # 4 and #8 diminishes the magnitude of the correlation. Therefore, a second comparison was made using both test results and the resulting “estimated” results gave a correlation of R2 = 0.8544 while “manual” results gave a correlation of R2 = 0.843 which was then translated into y = 1.1097x0.7837 and y = 1.0014x0.8037, respectively.
In other words, it is necessary to investigate the factors influencing the calculation of PI and MB while exploring the relationship between these two parameters. For example, values such as WL and WP which are used in the calculation of PI may result in incorrect results as they are affected by many factors such as the sample preparation method followed and the experience of the operator, etc. (as it was the case in experiments #4 and #8). Moreover, different WL values were obtained from the estimations of the system and the manual drawing on the flow curve using the experiment results. In
Multiple regression analysis is a method used to explain the cause-effect relationship between two or more independent variables which affect a variable and to determine the impact level of these independent variables. Multiple regression model calculations as well as in the establishment of prediction equations made just a single model and calculating the coefficients of variation from the average of the arguments being used. Formula describing the explaining the relationship between dependent and independent variables is as follows;
Here;
xi = Independent variables
y = Dependent variables
ei = Error coefficient
Using the experimental data, such an investigation resulted in PI as the dependent variable and Cu, Cc, γs ve MB as independent variables (x). The modeling effort using these data gave a variety of statistics and their relationship with PI was explored. Using the coefficients of x1, x2, x3 and x4 variables and error coefficients obtained from the regression analysis, the formula gave the P-value, and this value was compared with the actual test results in order to calculate the compatibility coefficient (R2). Results for the 1st variable;
Experiment # | Cu | Cc | MB | γs | PI | |||
---|---|---|---|---|---|---|---|---|
1 | 5.70 | 0.85 | 57.00 | 2.65 | 23.51 | |||
2 | 5.95 | 1.62 | 31.25 | 2.60 | 16.36 | |||
3 | 5.92 | 0.58 | 68.65 | 2.52 | 25.59 | |||
4 | 5.00 | 0.85 | 44.75 | 2.73 | 26.65 | |||
5 | 5.65 | 0.45 | 61.15 | 2.49 | 24.37 | |||
6 | 6.31 | 0.30 | 76.00 | 2.64 | 40.19 | |||
7 | 3.38 | 0.46 | 42.30 | 2.62 | 25.26 | |||
8 | 8.59 | 0.79 | 83.15 | 2.72 | 25.44 | |||
9 | 4.67 | 0.86 | 32.00 | 2.60 | 16.09 | |||
10 | 3.12 | 0.49 | 89.2 | 2.65 | 37.47 | |||
11 | 4.95 | 0.54 | 34.92 | 2.72 | 15.81 | |||
Regression Statistics | ANOVA | |||||||
Multiple R | 0.866388264 | df | SS | MS | F | Significance F | ||
R Squared | 0.750628624 | Regression | 4 | 473.1995599 | 118.29989 | 4.515125003 | 0.05042839 | |
Adjusted R Squared | 0.58438104 | Difference | 6 | 157.2048037 | 26.20080062 | |||
Standard Error | 5.118671763 | Total | 10 | 630.4043636 | ||||
Observations | 11 | |||||||
Coefficients | Standard Error | t Stat | P-value | Low %95 | High %95 | Low 95.0% | High 95.0% | |
Intersection | −7.61135167 | 54.96860867 | −0.138467243 | 0.89440155 | −142.1146917 | 126.8919883 | −142.1146917 | 126.8919883 |
X Variable 1 | −1.46817575 | 1.259648491 | −1.165544005 | 0.288033144 | −4.550424569 | 1.614073075 | −4.550424569 | 1.614073075 |
X Variable 2 | −2.75399562 | 5.732256169 | −0.480438337 | 0.647930675 | −16.78032118 | 11.27232993 | −16.78032118 | 11.27232993 |
X Variable 3 | 0.309516685 | 0.099374299 | 3.114655266 | 0.020725521 | 0.066356534 | 0.552676836 | 0.066356534 | 0.552676836 |
X Variable 4 | 9.567324111 | 20.90733973 | 0.457606 | 0.663338319 | −41.59109326 | 60.72574148 | −41.59109326 | 60.72574148 |
Same procedure was used also for the variables x2, x3 and x4 and it was possible to identify the correlation between R2 values and PI values of each dependent variable and the independent variable. This method was also repeated for 11 and 9 samples as it was the case for single regression method, and the R2 values obtained for the first and second runs were R2 = 0.75 and R2 = 0.85, respectively. Briefly, using this method, it is possible to identify which independent variable(s) influence the dependent variable and to estimate the PI value using these variables.
Liquid limit is an important parameter used to determine many indices and mechanical properties such as compression index, swelling percentage, soil class, liquefaction potential, etc. However, liquid limit and plastic limit test results may differ depending on several factors such as wet or dry sample preparation, the
Experiment # | Cu | Cc | MB | γs | PI | |||
---|---|---|---|---|---|---|---|---|
1 | 5.70 | 0.85 | 57.00 | 2.65 | 23.51 | |||
2 | 5.95 | 1.62 | 31.25 | 2.60 | 16.36 | |||
3 | 5.92 | 0.58 | 68.65 | 2.52 | 25.59 | |||
5 | 5.65 | 0.45 | 61.15 | 2.49 | 24.37 | |||
6 | 6.31 | 0.30 | 76.00 | 2.64 | 40.19 | |||
7 | 3.38 | 0.46 | 42.30 | 2.62 | 25.26 | |||
9 | 4.67 | 0.86 | 32.00 | 2.60 | 16.09 | |||
10 | 3.12 | 0.49 | 89.20 | 2.65 | 37.47 | |||
11 | 4.95 | 0.54 | 34.92 | 2.72 | 15.81 | |||
Regression Statistics | ANOVA | |||||||
Multiple R | 0.920062782 | df | SS | MS | F | Significance F | ||
R Squared | 0.846515523 | Regression | 4 | 531.4000252 | 132.8500063 | 5.515316837 | 0.063441038 | |
Adjusted R Squared | 0.693031045 | Difference | 4 | 96.34986365 | 24.08746591 | |||
Standard Error | 4.907898319 | Total | 8 | 627.7498889 | ||||
Observations | 9 | |||||||
Coefficients | Standard Error | t Stat | P-value | Low %95 | High %95 | Low 95.0% | High 95.0% | |
Intersection | −42.7864391 | 73.7177102 | −0.580409226 | 0.592765934 | −247.4596148 | 161.8867365 | −247.4596148 | 161.8867365 |
X Variable 1 | 0.056581334 | 1.642213845 | 0.034454303 | 0.974165661 | −4.502935258 | 4.616097926 | −4.502935258 | 4.616097926 |
X Variable 2 | −2.30509675 | 5.537969247 | −0.416235022 | 0.698601737 | −17.68096436 | 13.07077086 | −17.68096436 | 13.07077086 |
X Variable 3 | 0.36745354 | 0.103154821 | 3.562155768 | 0.023540512 | 0.081049843 | 0.653857237 | 0.081049843 | 0.653857237 |
X Variable 4 | 18.7467652 | 26.5596245 | 0.705836982 | 0.519228311 | −54.99457423 | 92.48810463 | −54.99457423 | 92.48810463 |
operator errors that occur during experiments, time consumption, and excessive amount of water added. Therefore, alternative solutions are required establishing the relationship between the PI obtained from liquid and plastic limit tests and other parameters.
Higher clay-based material content of the soil increases the liquid limit and plastic limit in the case of limits of consistency while it increases the pollution level of the sand and V1 used in the case of methylene blue test. This shows us that the increase in plasticity index depends on the liquid limit while the increase in liquid limit depends on the fine-grained content. In our experiments it was also found that V1 increases as the PI increases in general, which can be interpreted as a negative relationship and a direct proportion between PI and MB values.
In this study, with this idea in mind, a relationship was observed between the amount of methylene solution used and PI value and “single regression” results showed that this result was statistically significant by 84.3%. Also using “multiple regression” method, the correlation between Cu, Cc, γs and MB parameters and the PI value was found to be very close to the one found using single regression and there was an increase, even if it was small. As it is possible to control more factors influencing the dependent variable (y) using more independent variables (x) in the multiple regression analysis, it is possible to include more variables to the model in order to explain the change in y more efficiently. In the light of this information, it is possible to say that the use of methylene blue test results in combination with a number of parameters can be used for PI estimations, which will result in more reliable results.
Here, the most important independent variable in PI estimation with the highest correlation was found to be MB (P < 0.05) as shown in
Otçu, N.Ü., Uzundurukan, S. and Kaplan, G. (2017) Determınatıon of the Plasticity Index of Soils with Fine-Grained Soils Using Methylene Blue Test. Journal of Geoscience and Environment Protection, 5, 165-181. https://doi.org/10.4236/gep.2017.53012