Improving soil density

March 2007 » Feature Articles
Ground modification is used at construction sites to reduce the possibility of settlement or liquefaction.
Scott Mackiewicz, Ph.D., P.E.

Study evaluates the effectiveness of three ground-improvement methods.

Ground modification is used at construction sites to reduce the possibility of settlement or liquefaction. Therefore, evaluation of the performance of ground-modification processes used to densify granular material becomes critical. With this in mind, Kleinfelder and S&ME recently undertook a study to determine the effectiveness of three ground-modification methods—vibro-compaction, vibro-replacement, and deep dynamic compaction (DDC).

In general, vibro-compaction is used when dealing with relatively clean, cohesionless sands with silt contents generally less than 12 percent to 15 percent and clay contents less than 3 percent. Success of this method depends on the sandy material’s response to vibration. On the other hand, vibro replacement uses a relatively clean stone for backfill and vibration for densification. Vibro-replacement is generally used in marginal soils where vibro-compaction may not be effective, such as soils with 15 percent to 25 percent fines. DDC is also used in relatively clean, granular materials and is not effective in saturated, clayey soils because the degree of densification is a function of the energy input—weight and drop height—as well as the saturation level, fines content, and permeability of the material.

Measuring improvement

Since successful ground improvement depends on how soil conditions respond to a given method of modification, some type of verification testing is usually conducted during construction. Our study was designed to take that evaluation one step further. We set out to evaluate the benefits of the three, commonly used ground-improvement methods by comparing the tip resistance of pre- and post-construction cone penetrometer test (CPT) soundings at several study sites. The effectiveness of the individual procedures can be determined by comparing the measured tip resistance (qc) of pre- and post-construction CPT and determining the improvement index, Id, as described by J.E. Dove, L.E.C. Boxill, and J.B. Jarrett in a 2000 paper, "A CPT-based index for evaluating ground improvement," published in an American Society of Civil Engineers (ASCE) Geotechnical Special Publication.

(Equation 1)

Using CPT tip resistances, V.R. Schaefer and D.J. White in 2004 noted in another ASCE paper ("Quality control and performance criteria for ground modification technologies") that such an improvement index could be used on any in-situ quality control measurement method in which a specific soil property is evaluated before and after the improvement method over a specific zone of interest. The improvement index is useful for comparison purposes but does not explicitly account for factors known to influence the degree of improvement. The CPT soundings also provide an estimate of the amount of fines in each subsurface profile.

It has been proven, for instance, that the degree of densification is related to the spacing of the treatment grid (or the diameter of the stone columns), along with the silt/clay content. In 1987, J.D. Hussin and S. Ali ("Soil improvement at the Trident Submarine Facility," ASCE Geotechnical Special Publication No. 12) also determined that no appreciable improvement was obtained with vibro-technologies when the fines content exceeded 12 percent. To further elaborate, the degree of improvement appears to be more sensitive to the quantity of the clay-size fraction than it is to the silt-size fraction, according to H. Fang’s 1991, Foundation Engineering Handbook.

Accounting for fines content

For our study, we adopted the improvement index as defined by Dove, but also tried to account explicitly for the effects of fines content on the degree of improvement. Specifically, the approximate fines content of the subsurface profiles were estimated directly from the pre-construction CPT results using the soil behavior type index combined with the relationship between fines content (FC) and friction ratio, as presented in a 1997 publication by T. Lunne, P.K. Robertson, and J.J.M. Powell (Cone penetration testing in geotechnical practice). This relationship is simplified to the following:

(Equation 2)  FC% = 1.75 * I3c — 3.7


Ic = [(3.47—log Qt)2 + (log Fr + 1.22)]0.5
Qt = normalized penetration resistance (dimensionless), and
Fr = normalized friction ratio (percent)

In several cases, these approximate fines contents derived from the CPT data were compared with laboratory test results. The CPT fines content computed using Equation 2 generally correlated well with the laboratory test results.

Test results

Our case study sites were located near Charleston, S.C., within the Atlantic Coastal Plain Physiographic Province, which is comprised of Quaternary- and Pleistocene-age sediments, underlain by the Cooper Group of the Tertiary period. The sediments within the Charleston area are generally 30 to 80 feet thick and primarily comprised of interbedded sands, silts, and clays. The Cooper Group is generally an overconsolidated, calcareous silt that is more than 195 feet thick in this area.

The specific subsurface conditions were determined at each of our case study sites before construction using CPT soundings and soil borings. After performing ground-improvement methods, post-construction soundings were obtained at each of the sites. These post-construction soundings were taken near the center of each treatment grid to yield a conservative estimate of the degree of improvement.

Vibro-compaction—The vibro-compaction program generally improved granular materials with less than 10 percent fines at Case Study Sites 1 and 2 (Table 1). At Site 1, the lower improvement index values for a zone with less than 5 percent fines was probably caused by the relatively high pre-construction CPT tip resistances of approximately 15 MPa (Figure 1).

The results from Site 2 demonstrated that densification of the materials was reduced in the zone between 2.7 to 4.3 meters, which has a similar CPT fines content as the remaining profile (Figure 2). However, hydrometer testing on material collected from this zone indicated that the clay content was about 7 percent, versus less than 2 percent in the remaining portion of the profile. An appreciable gain in densification also was noted within the layers that had a greater fines content but lower pre-construction CPT tip resistances, typically less than 5 MPa. These measurements showed that the amount of densification improvement decreased with increasing fines content.

Vibro-replacement—Vibro-replacement densified materials at Case Study Sites 3 and 4 with a fines content of less than 5 percent, according to the modification indices, Id, (Table 1) and the CPT tip resistance (Figures 3 and 4). The degree of densification decreased as the fines contents increased, except within zones that had low pre-construction tip resistance of less than 5 MPa.

Deep dynamic compaction—The results from Case Study Site 5 (Table 1 and Figure 5) indicate that fines content of less than about 15 percent does not influence the densification improvement resulting from DDC. Densification improvement resulting from DDC is influenced more by the position of the groundwater table. As seen in Figure 5, the increase in post-construction tip resistance reduces to almost no improvement below a depth of 4.3 meters, the approximate measured level of groundwater.


The results of our research indicate that densification achieved from vibro-compaction and vibro-replacement is a function of the initial density and fines content of the material. As the fines content increases, the densification improvement generally decreases. Additionally, the data from Case Study Site 2 suggests not only that the fines content affects the densification improvement of vibro-technologies, but also that densification performance is affected by the type of fines—clay versus silt.

Conversely, the DDC process is not influenced by the fines content of the material to the same degree as the other two methods, as clearly shown by the results in Case Study Site 5. However, it must be taken into account that most materials densified by this process in our study had a fines content of less than 10 percent. Further study is necessary at other locations to verify this conclusion.

Our study indicates that evaluating the success of ground improvement technologies based only on an improvement index may be misleading. It is more accurate to measure the effectiveness of the technology based on verification testing, accompanied, of course, by sound engineering judgment.

Scott Mackiewicz, Ph.D., P.E., is a principal professional for Kleinfelder, Inc., in Lenexa, Kan. He can be reached at William Camp, III, P.E., is a technical principal for S&ME, Inc., in Mt. Pleasant, S.C. He can be reached at

Table 1: Improvement index (Id) by compaction method and fines content

Improvement index at indicated CPT fines content

Compaction method Case study site < 5% 5% - 10% 10% - 15% > 15%




0 - 1.5

0.5 - 0.7

1.0 - 3.0*

0 - 2.0*

1.4 - 2.0*

0 - 0.2*

0 - 0.5*




0.3 - 2.8

1.5 - 2.7

 0.2 -0.8

0.4 -1.0*

0 - 0.2

0.2 - 1.0*

0 - 0.2

0 -1.6*

Deep dynamic compaction 5 0.7 - 1.2 0.7 - 1.7* 0.6 - 1.2 N/A

*Improvement index on material with pre-construction tip resistance of less than 5 Mpa

Figure 1: Average CPT tip resistance for pre- and post-improvement conditions at Case Study Site 1.

 Figure 2: Average CPT tip resistance for pre- and post-improvement conditions at Case Study Site 2.

Figure 3: Average CPT tip resistance for pre- and post-improvement conditions at Case Study Site 3.

Figure 4: Average CPT tip resistance for pre- and post-improvement conditions at Case Study Site 4.

Figure 5: Average CPT tip resistance for pre- and post-improvement conditions at Case Study Site 5.

Upcoming Events

See All Upcoming Events