Development of a simulated natural carbonation test and durability of selected CEM II concretes

Abstract

The study reported here investigates the development of a simulated natural test for benchmarkingthe carbonation resistance of hardened concrete. The European Committee for Standardisation(CEN) had devised a draft simulated natural test and although initial Round Robin test results hadshown poor repeatability, the test had potential for development as a reliable benchmarking tool.The current study identified the areas of variability in the test as being (i) non-conformance to the teststorage environmental conditions and (ii) variability in production of concrete specimens. A storageroom with active control over temperature and relative humidity was proposed for the test. An inhouse CO2 injection system was developed and environmental monitoring showed that temperature,relative humidity and atmospheric CO2 concentration in the test chamber were all within the proposedCEN test limits. In addition, a two stage normalisation procedure was established to reducevariability in the production of concrete specimens. A repeatability test demonstrated thatimplementing environmental control and the normalisation procedure considerably reduced thevariability of the test.A selection of CEM I and CEM II concretes, containing condensed silica fume (10%, 15%, 20%) andmetakaolin (10%, 15%, 20%), were benchmarked against a PC/PFA 30% Reference Mix to assessthe effect of cement type on carbonation resistance. All concretes were tested on an equal strengthbasis of 37N/mm2, the minimum grade requirement in BS EN 206-1, using the modified CEN testand a previously developed accelerated carbonation test. A reasonable correlation was foundbetween the 20 weeks accelerated and 2 year CEN carbonation depths indicating that acceleratedmethods may be useful in testing relative carbonation performance. A detailed permeation andhydration study highlighted the importance of both the physical and chemical properties of concretein resisting carbonation. A relationship between water/PC ratio and carbonation was established anda tentative water/PC upper limit of 0.65 proposed for 37N/mm2 concrete.Three additional coarse aggregate types, Thames Valley gravel, crushed limestone and crushedgranite were also tested and were shown to have a minimal effect on the carbonation resistance ofconcrete. The influence off ines (particles <125gm, including cement) on carbonation resistance wasalso studied. Adjusting the fine/coarse aggregate ratio to increase the fines content was found to bedetrimental to the particle packing characteristics of the concrete. Fractional replacement of the fineaggregate with limestone filler was found to increase the packing density of the concretes, reduce thenear surface permeation properties and improve carbonation resistance. A tentative optimum finescontent of 390-400kg/m3 was proposed.A multi-regressive model which considered the near surface physical and chemical properties wasdeveloped to estimate long term carbonation performance. Comparison with published long termdata proved difficult as did establishing which natural exposure environment each of the CENExposurec lassesr epresentedA. long term naturale xposures itew as establishedto replicateo utdoorssheltered conditions. Comparison of the 1 year data with that from the CEN test determined thatoutdoor sheltered conditions were not exactly replicated by the CEN Exposure Classes however,depths of carbonation were generally between those found in Exposure Class I and Exposure Class 2of the enhanced CEN test.

Date of Award	2001
Original language	English
Awarding Institution	University of Dundee
Supervisor	Rod Jones (Supervisor)