Sample Disturbance PDF
Sample Disturbance PDF
Sample Disturbance PDF
3–4, 2005
Abstract: This paper reviews the causes of sample disturbance and its impact on laboratory meas-
ured design parameters for soft clays. It is the most significant issue affecting the quality and reliabil-
ity of laboratory test data. All key design parameters such as compressibility, yield stress and
undrained shear strength are adversely influenced by sample disturbance. Each stage of the sampling
process, from initiation of drilling to final preparation of laboratory test specimens, causes potential
disturbance to samples and is described in the paper. Several practical solutions are presented, which
if properly implemented can reduce the degree of sample disturbance. These include the use of
weighted drilling mud, special tube sample geometry, fixed piston samplers, and proper sealing and
handling of samples. Qualitative and quantitative methods of assessing sample quality are also pre-
sented together with examples from case studies.
1. INTRODUCTION
Figures 1a and 1b present stress history and undrained shear strength data for
a soft clay site located north of Boston, Massachusetts. The site investigation was tar-
geted at characterizing the thick soft clay deposit for design of a deep foundation sys-
tem to support a highway superstructure. In particular, the key soil parameters re-
quired for design were the preconsolidation stress (σ ′p ) and the undrained shear
strength (su). The σ ′p data were from laboratory one-dimensional oedometer tests
conducted on tube samples. The su data are from field vane (FVT) and torvane (TV)
testing. The trends in the σ ′p and su data in figures 1a and 1b could be used as is for
selection of design profiles. However, a simple inspection of the data in light of poten-
tial sample disturbance can reveal the poor quality and hence unreliability of the
measured data. These data are used in Section 5 of the paper as an example to high-
light the significance of sample disturbance for determination of design parameters in
soft clays and methods that can be used to detect it.
This paper reviews the causes of sample disturbance and its impact on laboratory
measured design parameters for soft clays. The paper first reviews the various stages
of drilling, sampling and specimen preparation that can cause sample disturbance.
Examples of the significance of sample disturbance for laboratory measurement of
design parameters are given. Practical solutions that can help to reduce the degree of
sample disturbance and methods for evaluating sample quality are presented. The data
in figure 1 are examined in the context of the factors that cause sample disturbance
and are evaluated using a simple sample quality assessment method. Data from other
92 D.J. DEGROOT et al.
case studies are also presented. Some of the material presented in the paper is ab-
stracted from LADD and DEGROOT [13].
Ground Surface AB C D E
0
SQD
σ'v0
σ'p
-10
Elevation msl (m)
-20 FVT
FVT
-30
TV
TV
-40
The most important effect of sample disturbance in soft clays is a soil destructur-
ing, which is accompanied by a significant reduction in the sample effective stress
(σ s′ ) . Figure 2 shows an example of how the reality of sampling and testing can vary
unpredictably from the ideal. This figure shows the anticipated stress path experienced
by a low overconsolidation ratio (OCR) clay starting from its in situ state of stress
(Point 1) to its state of stress when ready for laboratory testing (Point 9). While design
is for in situ stress states, disturbance caused by sampling and subsequent storage and
handling can significantly alter the state of stress of samples. Clearly the state at Point
9 is very different from the in situ state at Point 1, and yet much of standard geotech-
nical engineering practice relies on strengths determined from samples starting from
Point 9 (e.g., unconsolidated undrained triaxial compression [UUC] test). Figure 2
further shows the significant difference in potential effective stress paths for soil ele-
ments during undrained compression shear starting at Points 1 (in situ) and 9 (dis-
Sample disturbance – soft clays 93
turbed UUC test in the laboratory). Based on this simple depiction of what can happen
during sampling, it should come as no surprise that there is often a gross mismatch
between design performance based on laboratory derived strengths and field perform-
ance.
There are many factors involved in the process of drilling, sampling and specimen
preparation that cause sample disturbance. The key stages of the process and its poten-
tial contributions to sample disturbance are as follows (with reference made to the
points in figure 2):
Path 1–2. Borehole Drilling: Advancing an open borehole reduces the total verti-
cal stress (σv), and hence subjects the clay at the bottom of the hole to potentially sig-
nificant stress relief. The effective stress path from Point 1 to Point 2 in figure 2
passes through a point for which σv equals the in situ total horizontal stress (σh0). This
stress state represents the theoretical “perfect sampling effective stress” (σ ′ps ) as de-
fined by LADD and LAMBE [14], i.e., the effective stress for undrained release of the in
situ shear stress. However, if no drilling mud is used or its weight is too low, the soil
at the bottom of the borehole can experience undrained extension unloading and in the
worse case scenario can fail before it is even sampled.
Fig. 2. Hypothetical stress path during tube sampling and specimen preparation of center-line element
of low OCR clay (from LADD and DEGROOT [13])
Path 2–3–4–5. Tube Sampling: Many researchers have studied the influence of
tube sampling techniques and sampler design on the quality of soft clay samples (e.g.,
94 D.J. DEGROOT et al.
HVORSLEV [8], BALIGH et al. [2] and CLAYTON et al. [3]). Baligh et al. showed that
for tubes with an inside clearance ratio greater than zero, the centerline soil experi-
ences shear in compression ahead of the tube (Path 2–3), followed by shear in exten-
sion as it enters the tube (Path 3–4), and then compression inside the tube (Path 4–5).
This straining of the soil, which can approach one per cent at the center-line (and even
greater values at the edges) for the standard 76-mm diameter Shelby tube (table 1),
causes destructuring and positive shear induced pore pressures for low OCR clays.
Table 1
Dimensions of sampling equipment
D1 D2 D3 t Angle AR D2/t ICR
Sampler
(mm) (mm) (mm) (mm) (°) (%) (–) (–)
34.9 50.8 34.9† 15.9† 60 111.9 3.2 0†
Standard SPT†
60.3 76.2 60.3† 15.9† 60 59.7 4.8 0†
Shelby tube 72.1 76.2 74.6 1.65 ≈ 44‡ 11.7 46.2 3.5
Modified Shelby
74.6 76.2 74.6 1.65 5 4.3 46.2 0
tube
Sherbrooke – 250 – – – – – –
Notes: See figure 3 for dimensions; †SPT with liners; AR = ( D22 − D12 ) / D12 , ICR = (D3 – D1)/D1;
‡
approximate angle since edge is beveled.
Path 5–6. Tube Extraction: Debris and disturbed soil at the bottom of the borehole
can enter an open tube sampler. After pushing, the strength of the clay at the bottom of
the borehole and suction resists removal of the tube sample. These factors can result in
significant disturbance of soil located at the top and bottom of the sample.
Sample disturbance – soft clays 95
K0) prior to undrained shear. The samples were collected using three types of sam-
plers: the NGI 54 mm diameter sampler, a standard 75 mm diameter fixed piston, thin-
walled tube sampler, and the Sherbrooke block sampler. There are clear differences in
measured stress–strain behaviour and undrained shear strength that would have an
important impact on selection of design parameters.
σ'v0 = 97 kPa
0
5
Vertical Strain εv (%)
10
15
Fixed Piston
20
Free Piston
Fig. 5. CAU triaxial recompression test results for specimens of Lierstranda Clay, Norway,
collected with three different samplers (LUNNE et al. [20])
Sample disturbance – soft clays 97
4. PRACTICAL SOLUTIONS
D = 76 mm
t = 1.6 mm
Cutting
Angle = 5o
Path 5–6. Tube Extraction: The use of a stationary (fixed) piston while tube sam-
pling is essential for high quality sampling. The piston prevents debris from entering
the tube before reaching the bottom of the borehole and controls the entry of soil dur-
ing sampling. The piston head also serves to create suction inside the tube at the top of
the sample to better retain it during tube extraction. Fixed piston samplers use either
actuating rods or hydraulics to control the piston head. The actuating rods are more
cumbersome to use than the hydraulic system, but they do allow direct observation of
the position of the piston head at the ground surface during sampling.
Path 6–7. Transportation and Storage: The ends of tube samples should be sealed
with wax and capped (usually plastic caps that are sealed with duct or electrical tape)
in the field prior to transport. Wax sealing is best done using a 50:50 mixture of paraf-
fin wax and petroleum jelly (LA ROCHELLE et al. [17]). It is important that the wax
mixture be brought just to its melting point and not overheated. Block samples should
be sealed using multiple layers of the same 50:50 mixture together with a couple of
layers of plastic wrap dipped in the wax mixture and a final layer of cheese cloth to
provide additional support. Samples should not be subject to excessive temperature
changes, particularly freezing temperatures. Tube and block samples should be trans-
Sample disturbance – soft clays 99
ported upright in boxes that provide damping of shocks and vibrations. This can be
done by placing the samples on foam padding, surrounding the samples with wood
chips or Styrofoam pellets, and using vibration control mounts for the boxes (e.g.,
figure 7). Once at the laboratory, samples should be stored in a room with humidity
and temperature control.
Fig. 8. MIT procedure for obtaining test specimen from tube sample
(LADD and DEGROOT [13])
100 D.J. DEGROOT et al.
Path 7–8. Sample Extraction: The potential bonding between the inside of tubes
and the soil can result in severe disturbance during laboratory extraction. Therefore,
samples should not be extruded from tubes (except if done immediately after sam-
pling) without first breaking any bonding at the soil–tube interface. Once a location
within a tube is selected for testing, the tube should be cut adjacent to the desired
sample location using a horizontal band saw or by hand (e.g., hack saw). A hypoder-
mic tube can be used to feed a thin wire into the soil/tube interface and rotated several
times around the perimeter to break the soil/tube bond (figure 8). With experience, the
whole process only takes a few minutes.
Path 8–9. Specimen Preparation: Soil within 1 to 1.5 times the tube diameter from
the top and bottom of the tube should not be used for consolidation and strength test-
ing because of greater disturbance near the sample ends (LACASSE and BERRE [10]).
Test specimens should be carefully trimmed in a humid room using sharp cutting tools
and wire saws. Subsampling in the laboratory using tubes can create additional distur-
bance and therefore should not be used. Sample sides should be trimmed during
specimen preparation to also remove potentially disturbed perimeter material. Trim-
ming in a humid room should prevent moisture evaporation from samples but it is also
important to control sources of swelling during specimen set-up (e.g., HIGHT [6]).
5
Vertical Strain εv (%)
10
15
Depth = 7.4 m Depth = 9.8 m
Sherbrooke Block Sherbrooke Block
20
Fixed Piston Fixed Piston
Free Piston Free Piston
25
10 100 1000 10 100 1000
Vertical Effective Stress, σ'v (kPa) Vertical Effective Stress, σ'v (kPa)
Fig. 9. Comparison of CRS data for free piston, fixed piston and Sherbrooke block samples
of Boston Blue Clay, Newbury, Massachusetts
Summary: Figure 9 plots data from CRS tests conducted on samples of Boston
Blue Clay collected with three different sampling methods. The free piston samples
were collected using a standard 76 mm Shelby tube (table 1) in a borehole without
Sample disturbance – soft clays 101
drilling mud, the fixed piston samples were collected using a modified Shelby tube
(table 1) using a fixed piston sampler in a borehole advanced with a barite based drill-
ing mud, and the block samples were collected using the Sherbrooke block sampler
using a barite based drilling mud. The samples were sealed, transported and trimmed
in accordance to the recommendations given above. The difference in quality between
the free piston samples and the fixed/block samples is significant. However, of even
greater significance is the fact that the fixed piston samples produced compression
curves nearly identical to the high quality block samples. These results show that it is
possible to collect very good to excellent quality samples with equipment that is read-
ily available to most drillers (the only custom item for the fixed piston sampling was
the machining of the ends of the standard Shelby tubes to produce the 5° cutting angle
and zero ICR).
different sampling methods were normalized by these values. The decrease in σ ′p with
a decrease in sample quality is large and has obvious consequences for design.
Table 2
Quantification of sample disturbance based on specimen volume change
during laboratory reconsolidation to σ v′ 0
Specimen quality
∆e/e0 criteria
designation (SQD)
(LUNNE et al. [20])
(TERZAGHI et al. [25])
Volumetric strain OCR = 1 – 2 OCR = 2 – 4
SQD Rating*
(%) ∆e/e0 ∆e/e0
<1 A < 0.04 < 0.03 Very good to excellent
1–2 B 0.04–0.07 0.03–0.05 Good to fair
2–4 C 0.07–0.14 0.05–0.10 Poor
4–8 D > 0.14 > 0.10 Very poor
>8 E
1.0
Sherbrooke Block
SPT
Free Piston
0.8
Fixed Piston
0.6
σ'p/σ'p(block)
0.4
0.2 SQD
A B C D E
0.0
0 2 4 6 8 10
εv (%) at σ'v0
Fig. 10. CRS preconsolidation stress normalized by Sherbrooke block sample values
versus volumetric strain at σ v′ 0 for different samplers, Boston Blue Clay, OCR < 4, Newbury, MA
Figure 1 also highlights the value of using simple quantitative methods of as-
sessing sample quality. Figure 1c plots ε vol (= ε v) versus depth corresponding to the
oedometer tests from which the σ ′p values in figure 1a were determined. All sam-
Sample disturbance – soft clays 103
ples below elevation –20 m have SQDs of D and E which based on the recommen-
dation of TERZAGHI et al. [25] would not produce reliable estimates of σ ′p . Cor-
roborating evidence regarding the poor quality σ ′p data is the fact that all values
below elevation –20 m are less than σ v′ 0 . This alone suggests an apparent undercon-
solidated soil. However, the geological history of the region (KENNEY [9]) indicates
the soil should not be underconsolidated and only a slight 1.5 m artesian condition
exists in the underlying glacial till. The apparent underconsolidation is probably
a result of sample disturbance causing a large decrease in the measured σ ′p . For
strength data, it is also useful to plot normalized values (i.e., su / σ v′ 0 ) as is done in
figure 1d for the field vane and laboratory torvane data. The FVT data are highly scat-
tered and many tests had unusually low undrained strength ratios, i.e., su(FVT)/ σ v′ 0
only about 0.1. The su(TV) data show very little scatter and the su(TV)/σ v′ 0 ratios
clearly show a transition from the crust to the softer clay at depth. However, the ratios
for the deep clay are also very low. The poor quality tube samples (hence very low lab
values of σ ′p and su) and unreliable FVT data are believed to be largely due to making
a hollow stem auger borehole without a proper weight drilling mud.
6. SUMMARY
All aspects of drilling, sampling, and laboratory preparation can cause adverse dis-
turbance to soft clay samples. This can result in misleading and unreliable laboratory
measured stress–strain–strength behaviour, which can amount to either unsafe or
overly conservative design. There are, however, a number of practical and relatively
simple steps that can be taken to reduce the adverse effects of sample disturbance.
Drilling should be conducted using an appropriately weighted drilling mud, samples
should be collected using large diameter thin-walled, sharp edged fixed piston sam-
plers. Samples should be protected from shock and excessive temperature changes
during transport and stored under high humidity, controlled temperature conditions.
Laboratory test specimens should be cut from tubes and debonded from the tube prior
to trimming. It is essential to evaluate sample quality for all design parameter tests;
ε vol at σ v′ 0 during laboratory reconsolidation is a simple yet effective indicator of
sample quality.
ACKNOWLEDGEMENTS
Some of the results presented herein are from a research project supported by the U.S. National Sci-
ence Foundation under Grant No. CMS-0219480.
104 D.J. DEGROOT et al.
REFERENCES
[1] ANDRESEN A., KOLSTAD P., The NGI 54-mm samplers for undisturbed sampling of clays and repre-
sentative sampling of coarser materials, Proc. of the Int. Conference on Soil Sampling, Singapore,
1979, 1–9.
[2] BALIGH M.M., AZZOUZ A.S., CHIN, C.T., Disturbances due to ideal tube sampling, ASCE J of Geo-
tech. Eng., 1987, 113(7), 739–757.
[3] CLAYTON C.R.I., SIDDIQUE A., HOPPER R.J., Effects of sampler design on tube sampling disturbance
– Numerical and analytical investigations, Géotechnique, 1998, 48(6), 847–867.
[4] DEGROOT D.J., LUNNE T., SHEAHAN T.C., RYAN R.M., Experience with downhole block sampling in
clays with conventional drilling equipment, Proc. 12th Panamerican Conf. on Soil Mechanics and
Geotech., Eng., MIT, 2003, Vol. 1, 521–526.
[5] HERMANN S., JENSEN T.G., CPTU kombinert med blokkprøvetaking ga kostnadsbesparende løs-
ninger for Nykirke kryssingsspor (CPTU combined with block sampling resulted in cost-saving solu-
tions at Nykirke railway link), Geoteknikkdagen, 2000, Oslo, 32.1–32.23.
[6] HIGHT D.W., Sampling effects in soft clay: An update on Ladd and Lambe (1963), Soil Behavior and
Soft Ground Construction, 2003, ASCE GSP 119, MIT, 86–122.
[7] HIGHT D.W., BÖESE R., BUTCHER A.P., CLAYTON C.R.I., SMITH P.R., Disturbance of the Bothkennar
Clay prior to laboratory testing, Géotechnique, 1992, 42(2), 199–217.
[8] HVORSLEV M.J., Subsurface Exploration and Sampling of Soils for Civil Engineering Purposes, Wa-
terways Experiment Station, U.S. Army Corps of Engineers, Vicksburg, 1949.
[9] KENNEY T.C., Sea-level movements and the geologic histories of the post-glacial marine soils at
Boston, Nicolet, Ottawa and Oslo, Geotechnique, 1964, 14(3), 203–230.
[10] LACASSE S., BERRE T., State-of-the-Art: Triaxial testing methods for soils, Advanced Triaxial Test-
ing of Soil and Rock, 1988, ASTM STP 977, 264–289.
[11] LACASSE S., BERRE T., LEFEBVRE G., Block sampling of sensitive clays, Proc. 11th Int. Conf. on Soil
Mech. and Found. Eng., 1985, San Francisco, 2, 887–892.
[12] LADD C.C., Stability evaluation during stage construction, J. of Geotech. Engrg., 1991, 117(4), 540–615.
[13] LADD C.C., DEGROOT D.J., Recommended practice for soft ground site charaterization: Arthur
Casagrande Lecture, Proc. 12th Panamerican Conf. on Soil Mechanics and Geotech. Eng., MIT,
2003, Vol. 1, 3–57.
[14] LADD C.C., LAMBE T.W., The strength of undisturbed clay determined from undrained tests, Symp.
on Laboratory Shear Testing of Soils, ASTM, 1963, 342–371.
[15] LADD C.C., YOUNG G.A., KRAEMER S.R., BURKE D.M., Engineering properties of Boston Blue Clay
from special testing program, Special Geotechnical Testing: Central Artery/Tunnel Project in Bos-
ton, Massachusetts, ASCE GSP 91, 1999, 1–24.
[16] LANDON M.M., DEGROOT D.J., JAKUBOWSKI J., Comparison of shear wave velocity measured in situ
and on block samples of a marine clay, Proc. of the 57th Canadian Geotechnical Conference, Que-
bec, 2004, in press.
[17] LA ROCHELLE P., LEROUEIL S., TAVENAS F., Technique for long-term storage of clay samples, Can.
Geotech. J., 1986, 23(4), 602–605.
[18] LA ROCHELLE P., SARRAIH J., TAVENAS F., ROY M., LEROUEIL S., Causes of sampling disturbance
and design of a new sampler for sensitive soils, Can. Geotech. J., 1981, 18(1), 52–66.
[19] LEFEBVRE G., POULIN C., A new method of sampling in sensitive clay, Can. Geotech. J., 1979, 16(1),
226–233.
[20] LUNNE T., BERRE T., STRANDVIK S., Sample disturbance effects in soft low plasticity Norwegian
clay, Proc. of Conference on Recent Developments in Soil and Pavement Mechanics, Rio de Janeiro,
1997, 81–102.
Sample disturbance – soft clays 105
[21] MITCHELL T.J., DEGROOT D.J., LUTENEGGER A.J., ERNST H., MCGRATH V., Comparison of CPTU
and laboratory soil parameters for bridge foundation design on fine grained soils: a case study in
Massachusetts, Transportation Research Record, 1999, No. 1675, 24–31.
[22] POIRIER S.E., DEGROOT D.J., SHEAHAN T.J., Measurement of suction in a marine clay as an indica-
tor of sample disturbance, Proc. GeoFrontiers Conference, ASCE, Austin, 2005, TX, in press.
[23] SHIBUYA S., MITACHI T., HWANG S.C., Case studies of in situ structure of natural sedimentary clays,
Soils and Foundations, 2000, 40(3), 87–100.
[24] TANAKA H., Sample quality of cohesive soils: lessons from three sites, Ariake, Bothkennar and
Drammen, Soils and Foundations, 2000, 40(4), 57–74.
[25] TERZAGHI K., PECK R.B., MESRI G., Soil Mechanics in Engineering Practice, John Wiley and Sons,
New York, 1996.