Summarization and Comparison of Engineering Properties of Loess in the United States J. B. SHEELER, Associate Professor of Civil Engineering, Iowa State University •LARGE deposits of loess are found in many parts of the United States, but published values of the engineering properties of loess are relatively scarce. The data in this paper were gathered to indicate similarities and compare the properties of loess from one area with another. Loess is composed primarily of rather loosely arranged angular grains of sand, silt, and clay. Silt is usually the dominant size. Calcite is also generally present in amounts ranging from zero to more than 10 percent of the total soil.. The aeolian hypothesis of loess deposition is compatible with the physical charac­ teristics of undisturbed loess masses. This hypothesis states that fine-grained ma­ terial was transported, sorted, and redeposited by wind action and thus became loess. During deposition, moisture and clay minerals are believed responsible for cementing the coarser grains together to form a loose structure. The loess is therefore subject to loss of shear strength due to water softening the clay bonds and to severe consolida­ tion caused by a combination of loading and moisture. Loess is usually thought of as an aeolian material that was deposited thousands of years ago and has remained in place since the time of deposition. Loess that has been eroded and redeposited is often referred to as redeposited loess, reworked loess or more simply as a silt deposit. This implies that the word loess indicates an aeolian soil, undisturbed since deposition. Certain engineering properties of loess, such as shear strength, are quite drastically changed by erosion and redeposition. The data presented in this paper are from samples of loess that have been undisturbed since deposition, to the best knowledge of the author, and the term loess is used in this context. IDENTIFICATION Loess may be identified primarily by its color and particle size. The structure and the unique ability to stand in vertical cuts are also distinguishing characteristics. The color varies from yellow to reddish brown with a buff color being the most com­ mon. The grains are almost all smaller than the upper limit of silt (0. 074 mm), and there is a complete absence of pebbles or rocks. However, it is not uncommon to find fairly soft calcium carbonate concretions about the size of a pea. These concretions have formed since deposition by solution and subsequent precipitation of the carbonate. They can be easily identified by their rather violent reaction with hydrochloric acid. The structure of the loess is typically open and contains many void spaces. Grape­ like clusters of balls of silt are sometimes observable. Most of the grains are coated with thin films of clay, while some of the grains are coated with a mixture of calcite and clay. Individual silt-sized particles of calcite can also be found well dispersed throughout the mass of loess. The clay coatings are apparently the major cementing agent. Some authors believe that calcium carbonate also contributes to the cementa­ tion of particles into the typical open structure. Paper sponsored by Ad Hoc Committee on Loess and presented at the 45th Annual Meeting. l 2 AASHO and ASTM Size Cla11ilication of Soil Particles MECHANICAL ANALYSIS Clay Si It Fine Sand Coaree Sa Gravel 5 The gradation characteristics microns of loess are shown in Figure 1. 100 ,-~~~+r-----.~~~-:F"F'~~-t--:::==:;~=----'------, Aloako All gradation curves fall within 90 the confines of the limiting bound­ 80 co aries, with the exception of the .:.. 70 Alaskan loess. This loess ap­ a.. 60 pears to be slightly more coarse Q. ~o than that found in the other states. All of the United States loess is smaller than 2. 00 mm (No. 10 sieve), except the Alaskan loess, which for several samples has 0.1 1.0 10.0 an upper limit of 9. 53 mm (% in. Diameter of Particles, mm sieve); for most samples the upper limit is 4. 76 mm (No. 4 Figure l. Gradation characteristics of loess in the United sieve). It is thought that the States. particles larger than 2. 00 mm are carbonate concretion::>, al­ though the data ftirnished the author did not mention concretions. The general shape of a loess gradation curve is similar to that of the boundary curves shown in Figure 1. The area within the boundaries is arbitrarily divided into clayey loess, silty loess, and sandy loess, after Holtz and Gibbs (10). These investi­ gators found that for loess from the Missouri River Basin area, 71percent of the samples were classified as silty loess, 21 percent as clayey loess, and 8 percent as sandy loess. SPECIFIC GRAVITY The specific gravity of loess in the United States varies between 2. 57 and 2. 79. The range is much smaller in local areas; for example, in Iowa it is from 2. 68 to 2. 72, in Nebraska 2. 57 to 2. 69 , in Tennessee 2. 65 to 2. 70, in Mississippi 2. 66 to 2. 73, and in Alaska 2. 67 to 2. 79. These local variations are most likely due to the variations in the type and quantity of mineral content. ATTERBERG LIMITS Plasticity data for various locations are given in Figure 2. Clevenger (5) indicates that sandy loess values lie in an area covering the lower end of the A-line below the Cl-ML zone, silty loess lies in an area straddling the A-line up to a LL of about 34, and clayey loess values lie in an area just above the A-line. Plasticity data representative of loess from various parts of the United States are plotted as a function of five-micron clay content in Figure 3. The results are similar to those found by the author for Iowa loess (8, 9, 20). However, the additional points provided by the low clay content Alaskan loess- andthe high clay content Illinois loess indicate a nonlinear relationship. Log-log plots of PI vs five-micron clay content show that the PI varies directly with the square of the clay content. The other mathematical relationships are indicated in the figure. Since these loess data exhibit such a good correlation, irrespective of source, it is concluded that the mineralogy is quite uniform and that the Atterberg limits are pri­ marily dependent upon the amount of clay present. It also might be intuitively reasoned that other properties would follow a similar relationship and that loess of similar gra­ dation would exhibit similar disturbed properties regardless of location. PERMEABILITY The only published permeability data available were found in the Holtz and Gibbs paper (10), which indicates a wide variation in permeability of the loess in Nebraska, Iowa, Kansas, and Colorado. The data were obtained as a part of consolidation testing •o 4() I CH. Illinois ( 7) I CH /1 l J." / I "" 3J Tennessee ( I) 30 J•:1 ..,, Pl 20 /i 20 r PI [ 0 ••• I MHr I • O~ A I ~t I 10 '~ 'e 0 ::; • t ~-~· .., "I ' 0 . 0 g. 0 ::; .. 0 :o •o •O I CH LL• 6-~+25 I I CH Nebraska ( 6) l I /1 30 /1 .. '°l """"'"' ",Wl ' 20 PI cL/VI PI 201 CV. I •• .. .., ... ., ~~ •• • I ~t 5 micron "'Cloy Content. percent I 10 I '. CL·ML ] I ~ I 00 0 10 20 30 40 50 60' 70 10 20- 30 .. 40 50 60 70 - 10 Cl M11t1U1P11' 40 40 'e 0 &losU ::; 4- N•b ...... o CH• Iowa ( Northeaet ) O ilhnoi1 Iowa ( Weetern) I CH OTtnnn.- 1,. X l 10'11. CoC03 1c_ •r- 1-11 30~ X ' 10'11. CaCQi 30 0 • .1-lfrl(I ••+ I 0: •1-1[ i;; ... , . ..u •• / J ~ ""UOll CIOJ COt1tt10. ~catit PI 20 PI 20 CL I r CL . ~/I OH I r t MH I 10 I Q.·ML ~·~ ML I I CL•ML J(' ] / I I /- ML J ••. .g.. •o •O ~ I ~ ~ Iowa ( East Central) Ala5ka X • 10'11. CaC02 30 ~ x 0: CL . 10' PI ] I . PI 20 CL Og: 1 0~ _.jof 10 . 5 micfon Clay conteril, perel!t11 ML i Q.·ML ML Figure 3. Plasticity data as a function 00 10 20 30 40 50 60 70 Oo 10 20 40 50 60 70 of five-micron clay content for loess in the United States. The symbol µ repre- Plasticity data of loess in the United States. Figure 2. sents five-micron clay content in the equations. 4 11 5 no 0 - 0 u " 0 ;;; a. ~ " 100 >- t- in z w 0 0 " >- a:: 90 0 85 80 7S 0.1 VERTICAL PERMEABILITY , fl. per yr. Figure 4. Vertical permeability data from consolidation tests on loess in the Missouri River Basin, from Holtz and Gibbs (10). of samples from nine different sites for earth dams. All values of permeability were determined after consolidation was complete under a given load. The data at low den­ sities represent tests completed with small loads and, hence, give values of perme­ ability nearest to those to be expected of undisturbed samples. Therefore, the in­ place vertical permeability of loess should be expected to be in the range of 10 to over 1, 000 ft per year. The data are shown in Figure 4. Terzaghi (21) describes the permeability of loess as a very elusive property. He bases this observation on the breakdown of the loess structure on saturation that is accompanied by densification and a consequent change in permeability. DENSITY The in-place dry density of loess in the United States varies from 66 to 104 pcf. Studies in Iowa (6) indicate that the in-place density is dependent on the depth (or loess overburden) andon the clay content.