Approach for Atterberg Limits
The collection of “Atterberg Limit” tests establish the moisture contents at which a fine-grained clay and silt soil sample transitions between behavior as a solid, semi-solid, plastic, and liquid material. Atterberg created a series of test procedures and definitions to provide reproducible results among soil samples. One of the principal uses of the system is to differentiate silty soils from clayey soils. The values obtained using the prescribed procedures can also be used to help characterize shrink-swell [expansive] soils, which are particularly problematic.
These values are important and common index tests in geotechnical engineering and are used to classify soils and compare the behaviors of different soils and relate soil plasticity behaviors to other performance qualities such as strength, stiffness, and permeability through empirical correlations. The tests are especially common for shallow foundations, roadway embankment construction, and pavement support. The tests are somewhat less common for deep foundation design and slope stability.
The tests are conducted on specimens taken from disturbed or undisturbed field samples and are usually conducted in a laboratory environment at a time following field sample acquisition, shipping, and subsequent lab extrusion or opening. The samples are remolded as part of the testing procedures; original intact samples are not required.
There are three (3) limits, each with a distinct corresponding laboratory test to establish the moisture content associated with a specific defined observed behavior of the sample. The tests are:
- The Liquid Limit (LL)
- The Plastic Limit (PL)
- The Shrinkage Limit (SL) *
*The Shrinkage Limit test is less common and is not typically conducted or used in common soil identification and classification systems.
A calculated value, the Plasticity Index (PI), is included in the reporting of the Liquid Limit and Plastic Limit. The PI is the Liquid Limit minus the Plastic Limit, PI = (LL – PL). Values of all parameters are reported in whole numbers.
Importantly, the commonly termed “Atterberg Limits” (generally including the LL, PL, and PI values) are used in several formal soil classification systems to determine classifications associated with plastic and non-plastic clayey and silty soils.
- A plasticity chart is available for the Unified Soil Classification System (USCS) which uses a soil’s Plasticity Index (PI) and the Liquid Limit (LL) to determine its classification.
- The PI and LL are also used in the American AASHTO classification system used for highway construction. This system rates soils for their suitability for highway construction. Transportation and other agencies may have their own specifications with identification and classification systems.
The Liquid Limit (LL) is the water content at which the behavior of a clayey soil changes from the plastic state to the liquid state. Depending on the mineral composition, particle size, and nature of the pore fluid, the transition from plastic to liquid behavior occurs over a range of water contents. The definition of the liquid limit for geotechnical purposes is based on a standard test procedure developed by Arthur Casagrande. His standardized methods for determining the liquid, plastic, and shrinkage limits are used today.
Casagrande standardized Atterberg's original liquid limit test by developing a hand cranked (or motor driven) specialty apparatus with a brass cup and resilient base. A rotating cam mechanism creates a reproducible drop height of 10 mm. A mass of soil is placed into the brass Casagrande cup portion of the device. A groove created using a standardized tool through the center of the specimen. The cup is repeatedly dropped at a standard rate of 2 revolutions per second. During the process the soil groove closes gradually at the base, as a result of the impacts. The number of blows necessary for the groove to close 1/2 -inch [13 mm] at 25 blows of the apparatus is desired.
The US Standard has both a single point and multipoint method to determine the Liquid Limit. Using the multipoint method, the test is progressively conducted at several moisture contents, and the moisture content which requires 25 blows to close the groove is interpolated from the test results. The single point test method also allows the use of a single test, where 20 to 30 blows are required to close the groove. A factor, k, provided in the standard, is applied to obtain the liquid limit from the moisture content. Two specimens are required for the single point test, with the Liquid Limit (LL) reported as the average value of the tests, reported to the nearest whole number.
The Plastic Limit (PL) is determined by hand rolling a small mass of a soil sample, creating a roughly cylindrical “thread” of soil on a flat ground glass plate. Plastic behavior is exhibited if the soil thread retains its shape when rolled progressively thinner to a diameter of 1/8-inch [3.2 mm]. At this diameter the rolling is discontinued, and the same sample is remolded into a larger thread and the test procedure repeated, rolling the newly remolded thread to progressively smaller diameters. As the moisture content lessens due to evaporation and loss of moisture along the technician’s hands and the glass plate, the thread will begin to break apart at larger diameters. At the point the thread can no longer be rolled into 1/8-inch [3.2 mm] thread, after having previously been rolled to that diameter, the test is considered concluded The sample is collected and placed in a container with other tested samples until two specimen containers of at least 6 grams of material are obtained and the moisture content can be evaluated. The water content of the soil is measured using approved test methods. A soil sample is reported as non-plastic (NP) if a thread cannot be rolled out down to 1/8-inch [3.2 mm] at any moisture condition.
The Shrinkage Limit (SL) is the water content where additional loss of specimen moisture will not result in additional soil volume reduction. Determination of the shrinkage limit is a more involved, and far less common process than the other two limits. Information on test procedures is in the associated standards.
Final reported [calculated] properties are typically:
- Liquid Limit (LL)
- Plastic Limit (PL)
- Plasticity Index (PI) and in much less frequent cases,
- Shrinkage Limit (SL)
Values are in whole numbers, unless the test result is that the soils are non-plastic, in which case rather than the numeric values a classification of “NP” is assigned.
Reporting of the final test values*, following application of the defined test procedures, is generally straightforward. Calculated result values are expressed as whole numbers. Unlike the SPT test, values for Plastic Limit (PL), Liquid Limit (LL) and the calculated Plasticity Index (PI) are generally well-behaved numeric data without non-numeric possibilities, other than the potential outcome of a non-plastic (NP) soil. If either the liquid limit (LL) or plastic limit (LL) could not be determined, or if the plastic limit is equal to or greater than the liquid limit, the soil is designated non-plastic (NP).
It is common to have these test results reported in conjunction with other test results from a single extracted soil sample (such as moisture content, direct shear, and unconfined compression) and a soil description- often obtained from grain size analysis and Atterberg Limits testing. Different specimens (subsets of the sample) are prepared for each type of testing.
*Typically, only final calculated values are reported and transmitted in log formats. Original measured values, calculations, and interim testing information is usually preserved in the original test records but not included as part of geotechnical design memoranda (although it could appear in comprehensive report appendices). The information may be included in geotechnical data reports which may contain complete field and laboratory testing data.
The applicable US Standard is ASTM D4318-17e1 Standard Test Methods for Liquid Limit, Plastic Limit, and Plasticity Index of Soils The applicable US Standard is ASTM D4943-18 Standard Test Method for Shrinkage Factors of Cohesive Soils by the Water Submersion Method
To demonstrate how the final results, interim test information and related metadata are stored and organized within the FROST server, we offer the following example Atterberg limits test for a single sample. Relevant data are:
GENERAL TEST INFORMATION
- Test procedure used: ASTM D4318-17, Method A (multipoint method)
- Name of sample tested: 1
- Sample location: Collected from a depth of 1.5 to 3 ft from borehole B-001-0-20
- Borehole is is 41 ft deep and its collar is located at lat/lon 39.47466/-81.796858, elevation 249.50928 meters (WGS84)
- Device used to collect sample: Geoprobe Interlocking Split-spoon sampler
- Length of sample recovered from the hole: 1.08 ft (72% recovery).
- Specimen preparation method : Wet Preparation Method described in Sec. 10.1, ASTM Standard D4318-17
- Weight of specimen prior to testing: 65 grams, wet weight.
CASAGRANDE TEST RESULTS FOR DETERMINING LIQUID LIMIT
| Increment Number | Blow Count | Water content (%) |
|---|---|---|
| 1 | 16 | 35.2 |
| 2 | 22 | 28.6 |
| 3 | 27 | 23.1 |
| 4 | 32 | 17.4 |
PLASTIC LIMIT WATER CONTENT MEASUREMENTS
| Increment/Container Number | Water content (%) |
|---|---|
| 1 | 11.9 |
| 2 | 11.7 |
| 3 | 11.4 |
Note: A third trial was run because of concerns that the second container was not properly sealed before measurement
FINAL REPORTED RESULTS
- Liquid limit: 25
- Plastic limit: 12
- Plasticity Index: 13
Object instances and the associations required to properly expose the example test data with the FROST Geotech Plug-in are shown in the following instance diagram:
The following summarizes the various entities in the diagram:
The Sensor object serves as the observing procedure in STA. One object instance is needed for this example (top center of diagram), and in this example holds the information about the test procedure used. One Sensor instance is needed for Atterberg limits tests and the test procedure is specified in its sensorType property and fully referenced in its metadata property. As constructed, this Sensor instance can be reused for multiple Atterberg tests.
The ObservedProperty object instances identify the properties that are observed by the Atterberg limits test. There are eight ObservedProperty instances (top of diagram, below Sensor):
-
3 properties observed that constitute the final reported results of the test:
-
- liquid_limit
-
- plastic_limit
-
- plasticity_index
-
3 properties observed for the Casagrande (liquid limit) measurements:
-
- manualCasagrandeTrialNumber (the Increment Number column in the above table)
-
- blowCount
-
- waterContent
-
2 properties observed for the plastic limit measurements:*
-
- manualPlasticLimitTrialNumber (the Increment/Container Number column in the above table)
-
- plWaterContent
Note: The plWaterContent observed property is not necessary (and probably should not be used) as it is identical to waterContent, but is included in this example to better distinguish the water contents determined for liquid limit from those for plastic limit
As with Sensor, the ObservedProperty instances can be reused for multiple tests.
All of the object instances in the diagram are linked to the Sensor and ObservedProperty instances via Datastream instances (below the ObservedProperty objects on the diagram), which serve to associate observation results obtained from a feature of interest to its observed property, observing procedure, and the borehole.
8 Datastream instances are needed, one for each ObservedProperty instance.
Note: if the plWaterContent ObservedProperty were not used, there would still need to be 8 Datastream instances. Two DataStreams would link to the waterContent ObservedProperty instance. One of these DataStreams would link to the liquid limit water content Observation instances, the other to the plastic limit water content Observation instances.
The DataStreams all link to the borehole via its BhTrajectoryThing object instance. BhTrajectoryThing (left edge of diagram) represents the borehole's geometry and contains the borehole length and information for linear referencing. The trajectory's geometry is given in the associated Location instance. BhTrajectoryThing is associated with a BhCollarThing instance , which represents the borehole as a whole. All general metadata about the borehole is contained in the BhCollarThing object instance; it's geometry is represented by a point Location object instance.
More detail about properties of BhCollarThing and BhTrajectoryThing can be found in the Borehole log discussion.
The act of collecting a sample from the borehole for testing is represented by the single BhSampling object instance (below and to the right of the BhTrajectoryThing in the diagram). BhSampling holds the sample depths (fromPosition=1.5, toPosition=3) and links to BhTrajectoryThing in order to affix the linear referenced sample positions to the trajectory geometry. In addition, BhSampling also links to a BhSampler object instance which identifies the type of sampler used in the sampling act (eg. the Geoprobe split-spoon sampler).
BhSampling produces a BhFeatureOfInterest object, which represents Sample 1 that is collected from the borehole (below and to the right of the BhSampling instance object). This object holds the sample length and recovery percentage of the sample. That this BhFeatureOfInterest is a physical material sample from the borehole is given by the associated BhFeatureType Core and Segment instances.
As part of the Atterberg limits test procedure, a 65 gram specimen of Sample 1 is prepared using the "Wet method". This is represented by an additional BhFeatureOfInterest object instance linked to Specimen, Core and Segment feature types. The linked BhPreparationStep object instance holds the specimen weight, and the further linked BhPreparationProcedure object carries the procedure method used to obtain the specimen.
The remaining entities on the diagram are Observation instances that provide the results for their associated observed properties. Each Observation instance links to the prepared specimen that was tested and to the Datastream instance associated with the appropriate ObservedProperty.
This Atterberg limits test consists of 18 individual observations but only three of them (results for liquid limit, plastic limit and plasticity index) are the primary reportable results. The other observations are interim observations, made as part of the test procedure, that result in the determination of the primary results. The links shown in the diagram relating observations to each other provide the means for distinguishing among the various types of observations.
Starting with the Observation instances associated with the Casagrande liquid limit measurements, each test increment is a set of three observation results. These results are useless independently - for example, the blow count in a Casagrande trial has no meaning without the associated water content observation, and vice versa. To model the observation set, the increment result is linked to its associated blow count and water content results, and the blow count result is linked to its associated water content result. These three observations, as a set contribute to the determination of the final liquid limit result, and to model that association, the increment observations are linked to the liquid limit observation. In this way, one can traverse from the liquid limit observation to access only those blow count and water content observations that contributed to the liquid limit result.
A similar link structure is also made for the interim water content observations that contribute to the plastic limit result.
Finally, the plasticity index observation instance is linked to both the liquid limit and plastic limit observation instances to demonstrate that plasticity index relies on those values for its results.
Note that the liquid limit oand plastic limit observation instances are correctly not linked to show that those results are derived independently from each other.
To provide the most flexibility for querying, Datastream object instances associated with the Observations are linked in the same manner as their associated Observations as seen in the diagram.
The current STA model does not provide for one-way links where an association role can be assigned. Such capability would be useful in modeling evem more complex geotechnical test results.