Moisture transfer and change in strength during the construction of rammed earth walls

Authors: Horst Schroeder , Thomas Schnellert, Thomas Heller, Thomas Sowoidnich

Bauhaus University Weimar, D-99421 Weimar, Germany

Presented at kerpic 05 – Living in Earthen Cities, Istanbul Technical University, 6 - 7 July 2005

Abstract

A number of rammed earth projects constructed in recent years in Germany testify to the high level of architectural interest in this material in our country. Rammed earth has been “rediscovered” because of its unique materiality and fascinating and individual surface aesthetics, particularly by young architects. In connection with the realisation of two rammed earth projects realised in Thuringia in 2003/4 some questions arose concerning the processes of moisture transfer and changes in strength properties during construction. The earthen building standards document only very rough estimates of drying times for rammed earth walls.

The idea arose to develop a test program for investigating the question of drying time with regard to the change in material strength in rammed earth walls, as well as for general aspects of testing procedures for rammed earth in standards.

The paper presents first results of a laboratory program trying to approach to this very complex problem. A series of test specimens were produced and the unconfined compressive strength UCS were determined varying the drying times from 7 to 90 days. The moisture content of the test specimens was also varied: at OMC (PROCTOR test), lower and higher OMC.

1. Introduction

A number of rammed earth projects constructed in recent years in Germany testify to the high level of architectural interest in this material in our country. Rammed earth has been “rediscovered” because of its unique materiality and fascinating and individual surface aesthetics, particularly by young architects. In connection with the realisation of two rammed earth projects realised in Thuringia in 2003/4 /1/ (Fig.1) some questions arose concerning the processes of moisture transfer and changes in strength properties during construction. The earthen building standards document only very rough estimates of drying times for rammed earth walls. The Lehmbau Regeln, the German standard for building with earth /2/, recommends a drying period of 4-6 months for a 40cm thick rammed earth wall. However, some wall sections of the aforementioned projects have a thickness of up to 1.70 m. The idea arose to develop a test program for investigating the question of drying time with regard to the change in material strength in rammed earth walls, as well as for general aspects of testing procedures for rammed earth in standards. The general questions were:

  1. The design values of compressive strength for rammed earth walls documented in standards apply to dry material. But at what time is a new built rammed earth wall dry in its core?
  2. The process of drying influences the strength of the wall. The maximum load bearing capacity of a rammed earth wall will only be achieved if the wall core is dry. How long does this take?
  3. Could the bearing capacity of the moist rammed earth material be exceeded at the base of the wall by compaction and the surcharge of the rising wall during construction?
  4. The recommended moisture content of soils for compaction issues in foundation building practice is the Optimum Moisture Content OMC according the PROCTOR Test. Is this recommendation also valid for earthen building?

These questions are very complex. But at least we tried to contribute a small amount to their solution by a testing program.

As part of student graduate work at the Bauhaus University Weimar /3,4/ a test programme was drawn up to try and find solutions for the above questions.

The testing procedure consists of two parts: In the first part a series of cube samples were produced varying the parameters of initial water content, drying time, organic fibre and mineral coarse aggregates. The unconfined compressive strength of the samples was tested after different drying periods. The preparation of the samples was executed according to the real site conditions.

The second part of the testing program was the construction of a rammed earth wall section in real scale. This section was prepared with electronic moisture measuring devices in order to monitor the drying process of the wall.

2. Testing Programme

2.1 Materials and methods

The natural raw earth material used for testing was a typical loess-type clay from Kleinfahner in Northern Thuringia, the same kind used for the aforementioned rammed earth projects. It should be noted that loess is inappropriate for rammed earth and therefore the addition of sand and gravel as well as straw fibres was necessary.

2.2 Sample preparation and parameters tested

According to the Lehmbau Regeln /2/ test cubes of 20 cm edge length were prepared varying the parameters tested in the following manner:

According to this plan a total of 105 samples were prepared.

The manual compaction of all test cubes was carried out in 3 layers of 15cm compacted to 7cm using a falling weight of 7,8 kg falling 32 times on each layer from a height of 45cm. A comparison between this compaction method and those used on site was not provided and this is a general problem.

The compacted cubes were air-dried in a room with a climate similar to the real site conditions: air temperature about 12°C, relative air humidity about 68% (Fig. 2).

The following parameters of the raw loess-type clay of Kleinfahner were additionally determined: mineralogical analyses by REM, cohesion test by 8-shaped sample (NIEMEYER /2/), plasticity index, and measure of shrinkage (DIN 18952).

2.3 Preparation of the test wall

In order to simulate the real conditions during compaction and the subsequent drying process in the project “Stairway to Heaven” in Nordhausen, Thuringia /1/ a test wall section was built at a scale of 1:1 in the Earth Building Laboratory at the Bauhaus-University Weimar (Fig.3). The size of the wall section was 1.5m long, 1.0m high and 0.5m wide.

The test wall was made with the same soil mixture as in the “Stairway to Heaven” project: a loess-type clay from Kleinfahner modified with coarse aggregates and straw fibres. The moisture content during compaction was near the OMC 12.3 – 13.0%.

To simulate the drying process under environmental conditions the “end” surfaces of the wall section were covered with plastic film, so that drying could only occur through the “real” surfaces.

Electronic measuring instruments were installed in the test wall with the aim of investigating the process of drying from the core to the surface according to Fig. 3. Additionally the degree of shrinkage in both horizontal and vertical directions was measured by measuring the change in distance between two nails in each direction.

3. Test Results

3.1 Soil parameters

Particle size distribution
The particle size distribution of the raw soil (1) was determined according DIN 18122, T1 & T2 and DIN 18123 (Fig. 4).

The mixture used in the real project (2) was modified with coarse aggregates.

The distributions are very different. How can one “compose” an optimal or suitable particle size distribution for rammed earth?

According a number of references this can be achieved with the help of the model of the “ideal” FULLER-distribution which is used in foundation practice. In this model the pores created by the large grains are completely (ideal) filled by the smaller ones. Another general recommendation is that the proportions sand + gravel to clay + silt should be 70:30. The grain size distribution of the natural raw clay (1) is far from this recommendation. The grain size distribution of the modified mixture (2) is much better.

Plasticity

The Plasticity Index of the natural raw clay was determined according DIN 18122-T1:
IP = wL – wP;
IP = 0.121
wL = 0.316 wP = 0.1949

According to /5/ the liquid resp. plastic limit for soils used as rammed earth should preferably be: wL = 0.30-0.35; wP = 0.12-0.22.

Cohesion Test

The cohesion test (NIEMEYER /2/, see also /5/) is a wet tensile test using 8-shaped samples. The cohesion, which in this test is designated as “binding force”, is classified in low (“poor”) and high (“rich”).

The cohesion of the natural raw clay was determined and classified according /2/:
100 g/cm2;
“poor” 80 – 110 g/cm2

The cohesion or “binding force” of raw clays used for rammed earth must be at least “poor”.

Shrinkage Test

For clays used as raw materials for rammed earth it is also necessary to know their shrinkage parameters in order to avoid or reduce cracking after drying. The shrinkage criteria is < 2% after 72h air drying and using samples with OMC as initial moisture content. The shrinkage measure was determined according to DIN 18952.

The determined shrinkage of the natural raw clay was 4.6%. This is too much, and coarse mineral and straw fibre aggregates are necessary to reduce the shrinkage to an acceptable measure.

PROCTOR-Test

The PROCTOR compaction test was carried out according DIN 18127 and exhibited the following results: Maximum dry density MDD at optimum moisture content OMC

Raw natural clay
MDD = 1.784 g/cm3; OMC = 0.152

Raw clay with coarse aggregates
MDD = 1.973 g/cm3; OMC = 0.113

Raw clay with coarse aggregates and straw fibres
MDD = 1.804 g/cm3; OMC = 0.134.

3.2 Cube sample tests

The following moisture contents ( / ) are used in the discussion of the results of the sample tests:
first value: moisture content used during compaction for cube sample production
second value: moisture content after 7 days drying
(1) raw natural clay, w ~ OMC (0.1472/0.1307)
(2) raw natural clay, w > OMC (0.2032/0.1717)
(3) raw clay with coarse aggregates, w ~ OMC (0.0992/0.0611)
(4) raw clay with coarse aggregates, w > OMC (0.1242/0.1064)
(5) raw clay with coarse and straw aggregates, w ~ OMC (0.1228/0.1005)
(6) raw clay with coarse and straw aggregates, w > OMC (0.1552/0.1277)

The test results are discussed according to the following topics:

Moisture content during compaction as a function of drying time
Fig. 5 shows that the moisture content during compaction has only little influence on the remaining moisture after 90 days of drying. The range of the absolute values is between 3.8% for raw clay at w > OMC (2) and 0.71% for raw clay with coarse aggregates at w = OMC (3). This also corresponds with the highest 20.32% (2) and lowest 9.92% (3) moisture content during compaction.

All mixtures with moisture contents during compaction w > OMC (2),(4),(6) also exhibit higher absolute values of remaining moisture content after 90 days of drying in comparison to those with w = OMC initial moisture content (1),(3),(5). On the other hand, the percentage ratios of remaining moisture content after 90 days of drying versus initial moisture content confirm this picture only for the “raw clay with coarse aggregates” mixtures at w = OMC (3) and at w > OMC (4).

The speed of drying shows a similar quality for all mixtures: After rapid drying in the first week there process slows in the second week in all except the “raw clay with coarse aggregates” and “raw clay with coarse and straw aggregates” mixtures at w > OMC (4), (6). For mixture (6) this interruption in the drying process is only “displaced” in the third and fourth week. The drying process of mixture (4) is continuous without interruption.

After 28 days of drying the remaining moisture content of all mixtures in relation to their initial moisture content differs between 53.3% for “raw clay with coarse and straw aggregates” at w = OMC (5) and 26.2% for raw clay with coarse aggregates at w = OMC (3). It is evident, that this ratio for the mixtures “raw clay with coarse and straw aggregates” at w > OMC (6) with 39% is lower in comparison with 53.3% at w = OMC (5). For the mixtures with coarse, but without straw aggregates at w = OMC (3) with 26% and with 29.5% at w > OMC (4) this ratio is reversed, likewise for the “raw clay mixtures (1) and (2) with 39.9% and 45.1%. The mixtures (3) and (4) also exhibit the lowest values. This means that straw and coarse aggregates accelerate the process of drying.

Unconfined compressive strength as function of drying time
According to Fig. 6 all mixtures with moisture content w > OMC (2),(4),(6) achieve higher unconfined compressive strengths UCS than those with w = OMC (1),(3),(5) after 90 days of drying.

This difference in strength is most significant for the natural raw clay: 3.89 and 1.83 N/mm2 for w > OMC (1) and w = OMC (2) respectively (= 53%). These values are also the highest and lowest absolute values of all tests. The respective differences of the modified mixtures are lower. The lowest difference was observed for the “clay with coarse and straw aggregates” mixtures (5) and (6) with 2.43 and 2.09 N/mm2 (= 14%). The respective values for “raw clay with coarse aggregates” (3) and (4) are 2.8 and 2.21 N/mm2 (= 21.1%).

The speed at which strength increases is also very different. The natural raw clay at w = OMC (1) already achieves 80% of the maximum UCS at 90 days after 28 days of drying. After 28 days the strength decreases, turning into an increase after 45 days again. But if w is > OMC (2) after 28 days of drying only 30.3 % of the maximum value at 90 days will be achieved. Both modified mixtures with w > OMC (4) and (6) show only a little more than the half (51.4 and 56%) of the maximum value of UCS at 90 days. The “raw clay with coarse aggregates” mixture at w = OMC (3) achieves a strength after 90 days being even 5.9% lower than that measured after 28 days of drying. The function has the same quality as raw clay at w = OMC (1) with two turning points.

Shrinkage of tested cube samples

The shrinkage of the raw natural clay used at w =OMC (4.6%) was higher than the 2% limiting criteria (see above). Therefore it was also of interest to evaluate the influence of aggregates on possible shrinkage reduction. For this purpose the shrinkage was determined as percentage reduction of edge lengths in direction of compaction at all cubic samples before testing the UCS.

The shrinkage measured for the tested cube samples according to the aforementioned mixtures (1) – (6) was very different. After 90 days of drying the maximum shrinkage was measured as 3.51% for raw clay with w > OMC. This was to be expected. The lowest value (~0) was exhibited by the “raw clay with coarse aggregates” mixture at w = OMC (3); 1.51% at w > OMC (4). The values of the “raw clay with coarse and straw aggregates” mixture were significantly reduced in comparison to the raw clay: 0.28% at w = OMC (5) and 0.6% at w > OMC (6).

3.3 Moisture profile of the test wall section

Fig. 7 shows the moisture profile of the test wall section over a period of half a year according the results of the electronic measuring.

The drying process of the wall results from a diffusion of the moisture from the core to the surfaces. Fig. 7 shows that the moisture in the core after six months drying is unchanged and the same as the levels during compaction. By comparison, the drying of the wall surfaces takes only a short time.

It seems that during the first two months the drying of the lower part of the wall is more effective than the upper part. After six month of drying the moisture distribution in the core seems to be homogenous, but now the drying process also intensifies from the base in a vertical direction.

The moisture profile in Fig. 7 can only show a general trend. Further investigations are necessary to fully understand the drying process of rammed earth walls. Under real conditions the air temperature and relative air moisture as well as the wind velocity influence the drying time and the moisture
profile of a rammed earth wall.

4. Conclusions

Investigations were undertaken for a better understanding of the process of moisture transportation and change in material strength of rammed earth walls after compaction as a function of time, material and initial water content.

The test results can only offer some general conclusions limited to the clay mixtures tested. Further investigations are necessary:

  1. The moisture content w = OMC during compaction of cube samples is not generally the “optimum” level for a quick drying of rammed earth walls. Initial moisture content levels of w > OMC can accelerate the drying process. Coarse and straw fibre aggregates influence the quality of drying process: straw fibre aggregates accelerate the drying, coarse aggregates intensify it. The most intensive drying is in the first week after compaction with an ‘interruption’ in the drying process in the following one or two weeks. The remaining moisture after a drying period of 90 days for all tested mixtures differs comparatively little.
  2. All mixtures with initial moisture contents w > OMC achieve higher unconfined compressive strengths UCS as those with w = OMC after 90 days of drying. This is most evident for the unmodified raw loess-type clay used which also exhibited the highest and the lowest absolute USC values. Coarse and straw aggregates reduce the absolute UCS values in comparison to unmodified mixtures.
  3. The increase in strength over the time is very different: The raw clay and raw clay with coarse aggregates mixtures achieve almost the maximum UCS value after 28 days of drying at w = OMC initial moisture content. The other mixtures at this time exhibit only around 30.3 – 56% of their UCS maximum after 90 days of drying.
  4. Coarse and straw fibre aggregates significantly reduce the measured shrinkage. So the addition of coarse and/or straw fibre aggregates to raw clay that is inappropriate for rammed earth can modify its properties such that it becomes appropriate for rammed earth.
  5. It can be recommended that
    • For rammed earth the moisture content during compaction should be about 10% higher than w = OMC,
    • The drying process of clay mixtures for rammed earth takes more time than for concrete. Therefore the determination of UCS should be carried out at samples after 90 days of drying.
  6. The moisture profile determined on a test wall section at 1:1 scale over a period of half a year confirmed the very long drying time necessary for compacted rammed earth constructions. The drying of the surfaces takes place comparatively quickly. However, the drying out of the core to the surface takes a considerable amount of time.

References:

1. Schroeder, H.; Bieber, A.
Neue Stampflehmprojekte in Thüringen
LEHM 2004, 4. Int. Fachtagung für Lehmbau Dachverband Lehm e.V., pp. 190-201
Leipzig 2004

2. Dachverband Lehm e.V. (Hrsg.)
Lehmbau Regeln – Begriffe, Baustoffe, Bauteile
Vieweg-Verlag Wiesbaden, 2. Aufl. 2002

3. Heller, T.; Schnellert, T.; Sowoidnich, T.
Ermittlung von Parametern zur Bestimmung der Festigkeit von Stampflehm
Studienarbeit WS 2003/04; Bauhaus-Universität Weimar 2004

4. Schnellert, T.
Untersuchung von Transportprozessen der Einbaufeuchte in Baukonstruktionen aus Stampflehm während der Austrocknung
Diplomarbeit Fak. Bauingenieurwesen, Bauhaus-Universität Weimar 2004

5. Houben, H.; Guillaud, H.
Earth Construction – a Comprehensive Guide
IT Publ., London 1994