A research project carried out by the Concrete Technology Unit at the University of Dundee, has developed specifications  and specification clauses  geared towards assisting engineers in the use of powdered glass as a fine aggregate or filler for use in concrete.
Overview of proposed specification
The major requirements for glass cullet used as a fine aggregate or filler defined in the specification are grading and constituents which alter the rate of setting and hardening of concrete. Other than these requirements, the specification contains a general requirement that the material shall conform to BS EN 12620.
The grading of glass cullet aggregate is limited to particle sizes not more than 4mm, a limit set primarily to avoid problems which may be encountered with high flakiness indices at larger particle sizes.
The effect of constituents which alter the rate of setting and hardening of concrete is required to be assessed in terms of stiffening time and compressive strength in accordance with BS EN 1744-1:1998, 15.3. Due to the main potential problem arising from the presence of sugars, the simple screening tests included in BS EN 1744-1 are deemed unsuitable, since they do not detect these compounds.
The standard also gives requirements for evaluation of conformity, designation and description and marking and labelling. It should be noted that testing for constituents which alter the rate of setting and hardening of concrete is required on a more frequent basis than BS EN 12620 normally requires, on the grounds of the potential for sugars to be present.
A study carried out to examine the variation in chemical composition of cullet taken at weekly intervals over a six-month period found sulphate levels in the material to be consistently low . For this reason sulphate content is not covered by the specification.
Overview of specification clauses
Project specifications which are to include the option of using glass cullet in concrete should be based on BS EN 206-1 and BS 8500. Additional specification clauses have also been developed specifically for glass.
The proposed additions to the Constituent Materials clauses are simply statements of established suitability for glass cullet as fine aggregate and filler aggregate in concrete. In the case of fine aggregate, it is stated that glass cullet may be used as part or all of the fine aggregate in a mix. In the case of filler aggregate, the limit to established suitability is 20% by mass of the fine aggregate. This limit represents the largest level of glass cullet filler aggregate use that has been investigated .
A note is also included to stress that suitability has yet to be established for concrete containing the combination of glass cullet fine aggregate and filler aggregate.
Proposed additions to the Requirements for Concrete clauses cover issues relating to damaging alkali-silica reaction and batching.
A testing programme carried out to establish the extent to which concrete containing glass cullet underwent damaging alkali-silica reaction has established that, where used as fine aggregate, the material was classed as ‘highly reactive’ and that precautions appropriate to this level of reactivity should be applied . The ASR testing programme has identified the use of GGBS and metakaolin as effective means of preventing harmful ASR in concrete containing glass.
NOTE The observation that GGBS and metakaolin are effective in preventing harmful ASR should be viewed as being tentative until the completion of the ASR study. It should also be noted that the study has not investigated other materials shown to prevent harmful ASR with other reactive aggregate combinations, such as PFA or silica fume. Neither has it investigated the effectiveness of lithium compounds.
The ASR testing programme has identified significant differences in the ASR expansion behaviour of different colours of glass cullet, with green glass displaying the largest expansion and flint glass the lowest . However, the approach taken has been to consider all colours as being capable of causing as much expansion as green glass. This strategy has been chosen on the grounds that the difference in ASR behaviour between different colours of glass is believed to be related to concentrations of trace metals. These concentrations were found to vary significantly from batch to batch , and so a conservative approach has been adopted.
Concrete containing crushed glass as a filler aggregate has not displayed harmfully expansive ASR behaviour. However, it has been decided to still class the material as being ‘highly reactive’, given the potential for expansive reactions of the coarser material.
The method of establishing the alkali contribution from crushed glass is given as being BS 1881-124, since the usual method for aggregate (given in Annex C of BS 8500-2) infers alkali content from chloride content, which is felt to be inadequate for glass. The method is comparable to BS EN 196-21:1992, clause NA.5.1 in terms of the nature of the extraction procedure.
NOTE This approach to determining alkali contribution should be viewed as being tentative until the completion of the ASR testing programme
A clause is also included stating that batching should always be carried out in terms of weight rather than volume. This is because bulk density measurements carried out on glass cullet fine aggregate have identified a large difference between the uncompacted and compacted bulk density .
Most collectors and reprocessors of glass cullet are able to process cullet to the requirements of the specifications discussed above. A list of UK collectors and reprocessors is held and updated on the Letsrecycle website.
 Concrete Technology Unit, Specification for crushed glass as a fine aggregate or filler aggregate for use in concrete, University of Dundee, Dundee, DD1 4HN, 2003
 Concrete Technology Unit, Specification clauses for the use of crushed glass fine aggregate and crushed glass filler aggregate, University of Dundee, Dundee, DD1 4HN, 2003
 Dhir, R.K., and Dyer, T.D., Realising a high-value sustainable recycling solution to the glass cullet surplus, Final Report for DETR Research Contract No. GW-12.10-108), 2003