A composite material is one that is made from two or more component materials which remain physically or chemically different. The component materials each give some advantageous effect or effects to the final composite. Concrete, the most widely used of manmade composites in construction, involves a basic mix of cement, sand and aggregate. Steel reinforcement can be added, with the concrete and steel acting as another composite material, the concrete providing strength in compression and the steel in tension among other aspects.
Basic composite wall constructions, involving sticks, have been used globally for thousands of years. The Maasai herdsmen of Kenya build their shelter from a lattice work of sticks, usually Leleshwa branches, to a height of two meters and up to six meters long. Due to the nature of their lifestyle, there is no shortage of animal dung, which ‘provides a ready supply of material for plastering walls, roof and interior partitions. [...] As it dries it provides an effective protective layer around the house and, to some extent, repels insects.’(38) Similar to the Maasai huts were the early wattle and daub constructions described below by Paula Sunshine.
Wattles (predominantly hazel sticks) were woven around evenly spaced vertical wooden posts. [...] Wet clay daub was then smeared on to the wattles, filling in the gaps between the sticks.(39)
Later timber framed buildings used wattle and daub as an infill, with a willow or oak lath wattle providing support for the clay and possibly straw daub. Without the wattle, the infill would undoubtedly crumble away and without the daub, an air tight finish would not be achieved.
Much like wattle and daub, hemp stalk could be constructed into a framework as the wattle with hemp-lime set around it as the daub. This may strengthen the hemp-lime enough to become independent from the timber frame.
Hemp-lime and hemp stalk could work in a symbiotic relationship as they both require certain attributes that the other naturally possesses. The lime, as it does for the shiv, can provide a fire, rot and pest resistant barrier for the hemp stalk to be encased by. The nature of the hemp-lime also creates a continuous membrane around the building, which is air tight and breathable, passively regulating the internal conditions. The rigidness of the cured material could provide lateral support to the stalks that would otherwise be vulnerable to bending and buckling. The stalks themselves could provide the much needed load-bearing capacity that the hemp-lime lacks, and a vertical stiffness to reduce compression of the aerated material. If strong enough elements are created from hemp stalk it may also be feasible to use the hemp-lime off the vertical plane, for example, curving the material in two dimensions with the stalks taking the bending loads.
In designing the structural hemp wall, each of the materials was used in a manner which complements its attributes while also making strong bonds with the other materials. The design attempts to achieve this by creating a space-frame of hemp stalk within a lime-hemp wall.
Weak points often occur at joints or fault lines of materials; therefore it is possible to minimize such places by designing a continuous walling system. The hemp stalk frame could be constructed for the whole building, creating a skeletal form and then the lime-hemp cast around it. In-situ casting like this would mean the hemp-lime would be seamless.
The frame is based on a grid of vertical stalks at 150mm centres, 2 wide, running along the wall. This leaves a covering of 75mm of hemp-lime either side of the frame. Each pair of vertical stalks are joined with lateral and diagonal secondary stalks, stiffening into a space-frame column. The depth of these elements is intended to increase the loading capacity of the wall in bending. All of the columns are then joined by more stalks running along the length of the wall. This ensures equal spacing and verticality of primary elements. No diagonal stalks are required to stiffen along this plane, as the hemp-lime should perform in racking. From the analysis in the previous chapter a pin joint was chosen for all connections. For the purpose of structural testing, samples 750mm high by 450mm wide were taken from a wall of 300mm depth. The sample includes 6 vertical stalks which will be reinforcing the hemp-lime under vertical loading.
Hemp stalk, grown for fibre production, was sourced from a farm in East Anglia. It was cut and dried for 10 days before being constructed into the frame. The lime, both NHL3.5 and Hydrated, were provided by Singleton Birch and the hemp shiv by Hemp Technology.
A fine quality shiv is used with low dust content. The basic mix consists of: 10 parts Shiv to 3 parts Lime with water to achieve correct consistency (advised by Modece Architects). Two types of lime were used, a naturally hydraulic lime (NHL 3.5) and a hydrated lime. These were used in a ratio of 2 parts NHL 3.5 to 1 part hydrated lime.
First, the shiv is measured out and added to the mixer. Here a standard cement mixer was used, however, care has to be taken to reduce the likelihood of balling, as the material can clump together. This can be avoided by a careful mixing procedure or the use of a better quality forced-action mixer. Water is added to soak the shiv. All of this water should be absorbed.
The lime is added and left to mix for a short amount of time. Left too long and balling may occur, whereas, if not left long enough the lime will not have mixed through properly. The final mix should have the consistency of porridge with the shiv fully coated in lime and retaining the moisture around it. If the moisture content is too high this may cause the lime to leech out of the formwork. A mix that is too dry may lead to there not being enough moisture present for lime’s curing process and the centre of the wall may remain as a powder. The completed mix holds its shape, does not slump and is moist to touch.
Sample blocks of the new wall construction were produced as well as some hemp-lime blocks without the internal frame. These were made to provide a comparison from which the new construction may be judged. The slide show below explains the construction process for a hemp-lime block with structural hemp stalk frame.
A frame was produced from hemp stalks and placed in a formwork of plywood. The lime-hemp was mixed and added in layers to the formwork. This process can be seen in the animation above. The blocks were then stored in a well ventilated and dry place with the formwork removed. They were left to cure for two months before being tested for load-bearing capacities.
Seventeen blocks were produced in total. Three with hemp stalk frame and three with traditional lime-hemp were produced for compression tests, intended for a comparison of the two structures after two months of curing. At this time three more hemp stalk blocks underwent three-point bending tests. The extra sample blocks were produced for further testing after 180days (six months) of curing.
The process of constructing the samples also highlighted potential problems that would need to be addressed in further research. Hemp which is usually mechanically harvested, had to be cut by hand and carefully extracted from the field to preserve the stalks. Until the stalks were allowed to dry they were extremely vulnerable. If the material were to ever be in mass demand, the method of its harvesting would have to be re-evaluated.
The joints used to connect the hemp stalks were developed from studying connections in similar materials, largely bamboo. The pin joint used in the final design, although an elegant solution producing a stiff connection, was time consuming to achieve and required high skill, accuracy and forward planning. The question of whether this is appropriate arises as the joinery is effectively hidden in the wall and therefore cannot be appreciated. Such jointing could be celebrated if ever in plain sight but when hidden a less time intensive and highly crafted technique could be used.
The intricacies of the frame’s geometry also proved problematic. With vertical elements based on a grid of 150mm centres and regular lateral supports and cross bracing, the sample space-frames were difficult to fill with lime-hemp. It is important to achieve a consistent density of mix, have access to tamper the material and not to leave any voids or hollows behind elements. With the frame making this a very fickle process it could be easy for standards to drop on a construction site. A more user friendly method of construction may be advantageous in enhancing building speed and quality.
The compression testing is intended as a comparison between traditional lime-hemp (LH) construction and the new walling technique of lime-hemp with hemp stalk frame (HF) as reinforcement. The stalk frame is intended to stiffen lime-hemp and take greater loads.
The LH blocks compressed over 200mm without failing, becoming denser and actually increasing in load bearing capacity. The HF samples failed eventually with the hemp frame buckling. This caused lateral movement, eventually popping out sections of limehemp from the side of the sample. Although this failure seemed more dramatic, in a building situation, the compression of the LH samples would also be considered to be a failure. Movement in a structure due to compression can cause cracks and increase the density of the wall, ultimately reducing its thermal performance. This concurs with previous research; ‘cylinder strengths are typically no more than 1N/mm2, and this resistance is only achieved following very high deformation (strains) under load.’(42)
The results of the analysis show that the additional framework increases the wall’s stiffness and strength in compression up until the frame fails. However, once this happens, they lose their strength at a higher rate than the LH blocks. This is much like a pre-stressed reinforcement bar within concrete. It provides more strength however when it fails it does so more violently. From the graph it would appear that the frames failed after approximately 4mm of compression with sample HF C taking 7kN of load at this point.
Levels of compressive strength needed for a load-bearing wall were not reached. Too small a quantity of hemp stalks were used if such strengths are to be achieved. The results show that the HF samples, each with six vertical stalks, achieved strengths directly proportional to the strength of the six stalks themselves. This would suggest that the load-bearing elements of the construction are the stalks, up until the failure of the frame.
‘Based on the properties of lime renders, indicative flexural strength values of lime-hemp have been suggested at 0.3-0.4 N/mm2 but this may be too high an estimate.’43 A structural framework usually takes all bending loads in a lime-hemp wall. It is therefore important to find whether the hemp stalk frame can offer any strength in this respect.
Visual account and results;
The three samples failed in the same way with a crack appearing in the lime-hemp which slowly widened. The hemp stalks within did not break but some of the frame’s joints did fail. Failure was gradual as the stalk frame held the two segments of broken limehemp from falling to the ground. Only low loads were taken in bending with sample HF D taking the most, reaching 3280N.
Analysis of testing;
It can take a long time for the lime to cure, however it would be expected that after six months the wall would have gained most of its potential strength. For the purpose of a self supporting wall it would need to hold its own weight much earlier. Dissecting the blocks after testing revealed that the two month curing period allowed the outer 40mm of lime-hemp to dry, while the centre of the blocks, including the hemp stalk frame, remained moist. When dry the stalk is rigid with good strength under vertically load, however, when damp it is more flexible, reducing its load bearing capacity. As the hemp stalks were still moist inside the blocks, they did not achieve their full potential strength.
The length of time that moisture is retained also causes problems with mould and rot. The shiv is protected by a thorough coating of the lime, preventing such problems due to its high PH. The stalks however were not fully covered and were subjected to mould at such points. The integrity of the frame was therefore compromised.
Although the sample walls still lacked sufficient strength to take the load of a simple dwelling type, they do show that the addition of hemp stalk strengthened it. The results suggest that the strength of the wall in a particular direction is directly proportional to the quantity of hemp stalks aligned in that same plane.
Similar conclusions can be drawn with the blocks in bending. The framework seemed to add strength yet was compromised by damp and weak connections. The frame did stop the lime-hemp from completely crumbling away.
All of the issues encountered during production and testing are only hurdles, they do not signal the death of the proposal but instead provide obstacles that may be overcome and learnt from through further development.