Published online 23 May 2006
Published in Soil Sci Soc Am J 70:1094-1100 (2006)
DOI: 10.2136/sssaj2005.0303N
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
Soil Physics Note
Comparison of American Society of Testing Materials and Soil Science Society of America Hydrometer Methods for Particle-Size Analysis
J. M. Keller* and
G. W. Gee
Pacific Northwest National Lab., Richland, WA 99352
* Corresponding author (jason.keller{at}pnl.gov)
 |
ABSTRACT
|
|---|
Particle-size analysis (PSA) is widely used in both soil science and engineering. Soil classification schemes are built on PSA values and recent developments in pedotransfer functions rely on PSA to estimate soil hydraulic properties. Because PSA is method dependent, the standardization of experimental procedures is important for the comparison of reported results. A study was conducted to compare the American Society of Testing Materials (ASTM) hydrometer method (D422) for PSA with the hydrometer method published by the Soil Science Society of America (SSSA). Tests on soils ranging in texture from sand to sandy clay loam were conducted at temperatures ranging from 20 to 30°C. The main difference between methods is the temperature correction, with the ASTM method relying on an empirical correction and the SSSA method using a blank hydrometer reading. Identical texture estimates for all but one of 48 total samples was observed between methods. Percentage of fines, silt, and clay demonstrated relatively consistent values between methods. The ASTM and SSSA methods were compared at values of D50, D30, and D10 (i.e., effective particle diameter values when the size-distributions have dropped to percentages of less than 50, 30, and 10, respectively). Excellent agreement was found between methods for D50 and D30 values (correlations above 0.99). Less agreement was found for D10 (correlation 0.989) values, but still reasonably good. The results suggest that for the range of soil textures evaluated in this study, ASTM and SSSA methods can be used interchangeably for textural analysis.
Abbreviations: ASTM, American Society of Testing Materials • HMP, sodium hexametaphosphate • PSA, particle-size analysis • SSSA, Soil Science Society of America • USDA, United States Department of Agriculture
|
INTRODUCTION
|
|---|
PARTICLE-SIZE ANALYSIS is one of the most common of all soil physical measurements. It is used in textural analysis to classify soils for both agronomic and engineering purposes (Gee and Or, 2002). It is also used to define particle-size distributions, which in turn can estimate pore-size classes needed in pedotransfer functions for hydraulic properties (Arya and Paris, 1981; Arya et al., 1999a, 1999b). Methods for measuring particle size include sieving, hydrometer, pipette, pressure sensor, x-ray, and more recently laser-diffraction techniques (Gee and Or, 2002). In reality, the particle-size distribution of a soil is seldom, if ever, measured using only one method. Large fractions above 2 mm are sieved while sedimentation techniques are used for size fractions below 2 mm. Unfortunately, particle-size distributions are method dependent (Gee and Bauder, 1986; Gee and Or, 2002), so it is important when reporting distributions to clearly indicate the specific method and pretreatment used to obtain the distribution data. A classic example of method dependency is illustrated by the work of Kubota (1972), who found that clay contents of a volcanic-ash soil varied from 1 to 56% depending on the soil pretreatment (air vs. oven-drying, pH control, etc.). In addition, pedotransfer functions that incorporate particle-size distribution and in some way converts these values to pore-size distributions and subsequently to hydraulic properties, may be rendered inaccurate if the applied particle-size distribution is measured using a different procedure than that used for the pedotransfer function calibration data set. The cautionary note here is that because of the great sensitivity to method, particle-size data must be treated, at best, as an empirical estimate of true pore-size distribution. It should also be understood that these hydraulic property estimates work best for coarse-textured (low clay) structureless soils where method dependency on particle-size distribution is often less pronounced (Gee and Or, 2002).
Due to its simplicity and minimal cost in materials, the hydrometer method is commonly used for measuring particle size of the fine soil fraction (<50 µm, United States Department of Agriculture [USDA]; <75 µm, ASTM). Two hydrometer methods often used are that of the ASTM (ASTM, 2000) and that published by the SSSA (Gee and Or, 2002).
This study explored the difference in the particle-size distribution and size statistics calculated from these two most commonly used hydrometer methods. Both methods were used in the analysis of a textural range of soils, with analysis occurring under two soil suspension temperature conditions to assess the effect of each method's respective temperature correction approach.
A brief description of each method and their calculations follow. The discussion is specific to the 152H hydrometer, as this was the type of hydrometer used in our study. It should be noted that both methods call for treatment of the soil with sodium hexametaphosphate (HMP) to attain complete dispersal of the soil particles, so the chemical pretreatment was the same for the ASTM and SSSA methods.
|
ASTM Method
|
|---|
To measure the true concentration of particles in suspension, the ASTM method calls for the use of a composite correction to be applied to the hydrometer reading. The composite correction is needed to correct for the increased liquid density due to the addition of HMP and inaccuracies resulting from liquid suspension temperatures departing from the hydrometer manufacturer calibration temperature of 20°C. The composite correction can also incorporate a meniscus correction needed to compensate for readings taken at the top of the meniscus rather than at the bottom of the meniscus, as calibrated by the manufacturer. The composite correction is determined by taking hydrometer readings in a sedimentation cylinder containing the same HMP solution to be used in the hydrometer analysis, but without soil. The sedimentation cylinder and HMP solution is placed in a constant temperature water bath and readings taken at two temperatures that encompass the temperatures to be experienced during the hydrometer analysis. A linear composite correction–temperature relationship is assumed to exist between the two measured points.
At each measurement time, the elapsed time, solution temperature, and the hydrometer reading is recorded. The percentage of soil in suspension (P) is calculated as
 | [1] |
where RASTM (g L–1) is the corrected hydrometer reading, computed by subtracting the temperature specific composite correction from the actual hydrometer reading; a is a unitless correction factor to account for particle densities different from the hydrometer calibration particle density of 2.65 g cm–3. Values for a can be found in Table 1 of ASTM D422 (ASTM, 2000). Ms (g L–1) is the dry weight of soil sample per liter of suspension.
The mean diameter of the soil particles (D) in millimeters is calculated as
 | [2] |
where K is a constant based on soil particle density and the temperature of the suspension. A standardized table of K values is available in Table 3 of ASTM D422 (ASTM, 2000). t is the elapsed reading time in minutes. L (cm) is the effective hydrometer measurement depth at time t, representing the depth from the surface of the suspension at which the density is being measured by the hydrometer. Values of L are given in Table 2 of ASTM D422 (ASTM, 2000) and are equivalent to
 | [3] |
where R (g L–1) is the uncorrected hydrometer reading.
View this table:
[in this window]
[in a new window]
|
Table 3. Linear regression and coefficients of determination (r2) from comparison of D50, D30, and D10 statistic from American Society of Testing Materials (ASTM) and Soil Science Society of America (SSSA) hydrometer analysis for subpopulations containing >45% sand and <45% sand. D50, D30, and D10 are the particle diameters when the percentage less than of the particle-size distribution curve is 50, 30, and 10%, respectively.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Measured particle density ( s) and error in calculated percent soil in suspension (P) when not applying the American Society of Testing Materials (ASTM) correction factor a.
|
|
|
SSSA Method
|
|---|
In lieu of a composite correction factor, the SSSA method calls for a hydrometer measurement in a blank HMP solution identical to that used in the soil suspension at each time that a hydrometer reading is taken in the soil suspension. Like the composite correction for the ASTM method, the blank hydrometer reading acts to offset inaccuracies resulting from temperature fluctuations, density increases from the HMP solution, and readings taken from the top of the meniscus rather than the bottom.
As with the ASTM method, at each measurement time the elapsed time, solution temperature, and the hydrometer reading of the soil suspension is recorded, with the addition of the hydrometer reading of the blank solution. The percentage of soil in suspension (P) is then calculated as
 | [4] |
where RSSSA is the corrected hydrometer reading calculated by subtracting the blank hydrometer reading from the hydrometer reading taken in the soil suspension. Again Ms is the dry weight of soil sample used.
The mean particle diameter in suspension at time t (min) is given by
 | [5] |
where 
l (g cm–1 s–1) is the solution fluid viscosity, L (cm) is the effective hydrometer measurement depth, g (cm s–2) is acceleration due to gravity, s (g cm–3) is the soil particle density, l (g cm–3) is the solution density. The calculation of L is identical to that of the ASTM method. The HMP solution density (l) is calculated from
 | [6] |
where Cs (g cm–3) is the concentration of HMP dispersing solution and w (g cm–3) is the density of water at the average temperature across all readings. The solution viscosity, l (g cm–1 s–1), is calculated as
 | [7] |
where w (g cm–1 s–1) is the water viscosity at the average temperature across all readings. For our study, values of w and w were taken from the Handbook of Chemistry and Physics (CRC Press, 2004).
|
MATERIALS AND METHODS
|
|---|
Particle-size analysis was performed in duplicate on 12 samples from five states (Table 1). The samples were selected to provide for a data set that spans various textural designations.
Samples were initially sieved to remove particle-size fractions >2 mm. From the <2-mm fraction duplicate 40-g subsamples were attained for PSA. To facilitate dispersion of the soil particles, 100 mL of 0.05 g mL–1 HMP solution was added to each sample and shaken for 12 h. After shaking, the HMP solution and soil was transferred into a sedimentation cylinder (graduated cylinder) and additional HMP solution added to bring the total volume in the sedimentation cylinder to 1000 mL. To begin analysis, the sedimentation cylinder containing the sample was agitated by turning the cylinder upside down and back for a minimum of 30 s. During the agitation, care was taken to assure that particles were not stuck on the sedimentation cylinder. After agitation the hydrometer was placed on the countertop or in a water bath, signifying time zero. Hydrometer readings were then taken at elapsed times of 0.5, 1, 3, 10, 30, 60, 90, 120, and 1440 min using a 152H hydrometer. At each reading time, the hydrometer was placed in the suspension and then removed after the reading was taken. Temperature of the suspension liquid was recorded at each measurement time. Additionally, at every measurement time a hydrometer reading was taken in a blank solution void of soil. The blank sedimentation cylinder contained 1000 mL of HMP solution at the same concentration (0.05 g mL–1) as that used in the sedimentation cylinders containing soil.
The effect of temperature on PSA was evaluated by performing one set of analyses under ambient laboratory-temperature conditions (20–23°C) and another set under constant 30°C waterbath conditions.
After hydrometer analysis was completed, the sand fraction was collected by wet sieving and then oven dried before being sieved in a nest of sieves comprised of 4, 10, 18, 35, 60, 140, 200, and 270 mesh sizes. Each sample was shaken in the sieve nest for 6 min using an automated shaker.
Percentage of passing and particle-size calculations were performed using the methods and relationships of both the ASTM (ASTM, 2000) and SSSA (Gee and Or, 2002) as outlined in the introduction. Both methods require knowledge of the particle density (s) of the sample to perform the needed calculations. The particle density of each sample was measured using the pycnometer method as described by Flint and Flint (2002).
The composite correction needed for the ASTM method was developed from blank readings taken at 20°C and in the 30°C water bath. A straight line between the 20 and 30°C blank readings represented the composite correction function (Fig. 1
).
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 1. American Society of Testing Materials (ASTM) hydrometer method composite temperature correction function developed from blank 152H hydrometer measurements.
|
|
|
RESULTS AND DISCUSSION
|
|---|
The calculation of percentage of soil in suspension (P) for ASTM differs from SSSA in that the ASTM method introduces the correction factor a to compensate for variations of the soil particle density. Table 2 presents the measured s for the soil samples used in the study and the error in calculated P when the correction factor is not applied. Based on ASTM a values, a s change of ± 0.05 g cm–3 from the calibration s of 2.65 g cm–3 equates in a P change of ±1%. For our data set, not accounting for s variations from 2.65 g cm–3 results in a calculated P error of 1% or less for the majority of the samples, but does produce errors >2% for three of the samples.
The particle diameter, D, is calculated from Eq. [2] for the ASTM method and Eq. [5] for the SSSA method. Setting both equations equal to one another and canceling like terms results in
 | [8] |
The value of K as presented in Table 3 of ASTM D422 (ASTM, 2000) is dependent on temperature and soil particle density. This is consistent with the right side of Eq. [8] in which particle density, s, is present as is solution density, 1, and viscosity, l, the latter two being a function of measurement temperature. For our measurement dataset, values of K and those computed from the right side of Eq. [8] were within 3% of each other. While this difference is minimal, it nonetheless shows that the calculated particle diameter as computed following the ASTM standard will not be equivalent to the SSSA methodology.
Figure 2
shows a comparison of D10, D30, and D50 as determined from ASTM hydrometer method and SSSA hydrometer method data sets. D10, D30, and D50 are the particle diameters determined for a sample when the P value (i.e., percentage less than) of the particle- size distribution curve is 10, 30, and 50%, respectively. Note that D10, D30, and D50 are reported in phi units, defined as
 | [11] |
where d (mm) is the particle diameter at the P value of interest. The 
scale is seldom used in soil science applications, but used rather extensively in sedimentation geology and soils engineering. The scale is used here to simplify statistical calculations and facilitate the reporting of results. Larger values are reflective of smaller diameters (e.g., a value of = 8.966 is equal to a particle size of 0.002 mm). The values in our tests ran from 1.4 to 9.4, equivalent to a range of 0.379 to 0.001 mm. For some fine-textured samples, the determination of D10 (n = 16) and D30 (n = 44) was not possible because the standard 24-h hydrometer analysis time did not capture the particulate suspension below 10 and 30% of the total sample mass for these samples. Regression (y-intercept term forced to 0) of the results of the D50 comparison points show good agreement (y = 0.993x, r2 = 0.999), indicating that the ASTM and SSSA methods provide very similar results. The maximum difference in calculated D50 for the ASTM and SSSA method was 3.5% with a mean difference of 0.7%. Good agreement between methods for D50 can be attributed to this statistic primarily being determined by sieve data for the soils in this study, so influences from the hydrometer analysis method are minimal. D30 points also show good agreement (y = 0.992x, r2 = 0.997) with the resulting maximum difference in calculated D30 being 7.5% and a mean difference of 0.9% D10 data shows reasonably good agreement (y = 0.990x, r2 = 0.989), but data is restricted to coarser soils for reasons described above. Nonetheless, a maximum difference in D10 as calculated from ASTM and SSSA hydrometer data is less satisfactory than D50 and D30, being 10.2% with a mean difference of 2.3%. The D10 statistic includes much of the particle-size data from the hydrometer analysis in its calculation and so the influence of the hydrometer method is more pronounced.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 2. Comparison of D50, D30, and D10 statistic from American Society of Testing Materials (ASTM) and Soil Science Society of America (SSSA) hydrometer analysis under ambient conditions and 30°C water bath. D50, D30, and D10 are the particle diameters when the percentage less than of the particle-size distribution curve is 50, 30, and 10%, respectively.
|
|
To further explore the potential effect of texture on the differences between methods, subpopulations containing >45% sand and <45% sand were evaluated. Results from linear regression analysis are presented in Table 3. For D50 and D30 the texture of the sample does not appear to correlate with the agreement of the hydrometer methods. D10 produced similar results, but the <45% sand population has a much reduced coefficient of determination (r2) compared with the >45% sand population. This suggests that there may be a texture effect on the measurement method, but it is only evident with grain-size metrics that describe the finer range of the particle-size distribution curve, such as D10.
A comparison of regression results for soils exposed to different temperature control methods, ambient (room temperature) air conditions, and 30°C water bath temperature, are presented in Table 4. The results do not reveal systematic error of the temperature correction method in relation to analysis temperature, with both methods in agreement for D50, D30, and D10.
View this table:
[in this window]
[in a new window]
|
Table 4. Linear regression and coefficients of determination (r2) from comparison of D50, D30, and D10 statistic from American Society of Testing Materials (ASTM) and Soil Science Society of America (SSSA) hydrometer analysis under ambient conditions and 30°C water bath. D50, D30, and D10 are the particle diameters when the percentage less than of the particle-size distribution curve is 50, 30, and 10%, respectively.
|
|
Figure 3
shows the soil samples plotted on a USDA textural triangle for both the ASTM method and SSSA method and for the two temperature measurement conditions. USDA texture designations are presented in Table 5. Shown in Table 6 are the percentages of fines, and the percentage of clay and silt values calculated for each test. Sample CD-D1 measured in a 30°C water bath showed a change in texture designation from a sandy loam when classified based on ASTM method data to a loamy sand when using SSSA method data. On the textural triangle, this sample lies on the border between sandy loam and loamy sand allowing a small change in size fraction to result in a change in texture designation. For all remaining samples, no variation in designated USDA soil texture is observed to occur as a result of hydrometer analysis method. The calculated percentages of fines, silt, and clay are relatively consistent between methods, with the largest difference between methods being two percentage points for the fines and clay and three percentage points for the silt. The results support the premise that the SSSA and ASTM hydrometer methods can be used interchangeably when evaluating soil textures similar to those used in this study.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 3. United States Department of Agriculture (USDA) textural triangle plots of soil samples used for the study as determined from the American Society of Testing Materials (ASTM) and Soil Science Society of America (SSSA) hydrometer analysis methods.
|
|
View this table:
[in this window]
[in a new window]
|
Table 5. United States Department of Agriculture (USDA) texture designation of soil samples as determined from American Society of Testing Materials (ASTM) and Soil Science Society of America (SSSA) hydrometer analysis results. a and b indicate sample replicates.
|
|
View this table:
[in this window]
[in a new window]
|
Table 6. United States Department of Agriculture (USDA) percent fines, silt, and clay of study samples as determined from American Society of Testing Materials (ASTM) and Soil Science Society of America (SSSA) hydrometer analysis results. a and b indicate sample replicates.
|
|
|
ACKNOWLEDGMENTS
|
|---|
We thank Karen Waters-Husted of Pacific Northwest National Laboratory (PNNL) for technical support in performing particle-size analysis. We also thank Andy Ward of PNNL for promoting the use of grain-size statistics to discriminate particle-size distributions. This work was supported by the Remediation Decision Support Task of the Groundwater Remediation Program and the Remediation and Closure Science Project sponsored by the U.S. Department of Energy under Contract DE-AC06-76RL01830.
Received for publication September 13, 2005.
|
REFERENCES
|
|---|
- American Society for Testing and Materials. 2000. Standard test method for particle-size analysis of soils. D 422–63 (1998). 2000 Annual book of ASTM standards 04.08:10–17. ASTM, Philadelphia, PA.
- Arya, L.M., and J.F. Paris. 1981. A physicoempirical model to predict the soil moisture characteristic from particle-size distribution and bulk density data. Soil Sci. Soc. Am. J. 45:1023–1030.[Abstract/Free Full Text]
- Arya, L.M., F.J. Leij, P.J. Shouse, and M.Th. van Genuchten. 1999a. Relationship between the hydraulic conductivity function and the particle-size distribution. Soil Sci. Soc. Am. J. 63:1063–1070.[Abstract/Free Full Text]
- Arya, L.M., F.J. Leij, M.Th. van Genuchten, and P.J. Shouse. 1999b. Scaling parameter to predict the soil water characteristic from particle-size distribution data. Soil Sci. Soc. Am. J. 63:510–519.[Abstract/Free Full Text]
- C.R.C. Press, 2004. Handbook of chemistry and physics. 85th ed. CRC Press, Boca Raton FL.
- Flint L.E. and A.L. Flint. 2002. Particle density. p. 229–240. In J.H. Dane and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Series No. 5. SSSA, Madison, WI.
- Gee, G.W., and J.W. Bauder. 1986. Particle-size analysis. p. 383–423. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
- Gee, G.W., and D. Or. 2002. Particle-size analysis. p. 255–293. In J.H. Dane and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Series No. 5. SSSA, Madison, WI.
- Kubota, T. 1972. Aggregate-formation of allophanic soils: Effects of drying on the dispersion of the soils. Soil Sci. Plant Nutr. 18:79–87.
TOP
ABSTRACT
INTRODUCTION
