Structural Analysis of Historic Construction – D’Ayala & Fodde (eds)© 2008 Taylor & Francis Group, London, ISBN 978-0-415-46872-5

The impact of the water/lime ratio on the structural characteristics ofair lime mortars

R.M.H. Lawrence & P. WalkerBRE Centre for Innovative Construction Materials, Department of Architecture and Civil Engineering,University of Bath, UK

ABSTRACT: Lime based mortars are now widely acknowledged to be generally superior to cement basedmortars in the repair of appropriate historic infrastructure. Increasingly the benefits of hydraulic lime mortarsare also being realised in new masonry construction. In order to standardise the expected performance of mortars,designers will specify the type of lime, the type of filler (aggregate), the proportions of each and quantity ofwater or the required workability. Limes can be non-hydraulic (calcium or dolomitic) or hydraulic (natural orartificial).

It is well known that the water/binder ratio has a marked effect on the structural performance of cement-basedmortars. This relationship is known as Abrams’ rule, which states that when a cement mortar is fully compacted,its strength is inversely proportional to the water/cement ratio. Abrams’ rule has also been demonstrated to applyto hydraulic lime mortars. The reason for this is that both cement and hydraulic limes require a minimum quantityof water to produce the chemical set resulting from the hydration of calcium silicates and calcium aluminates.Surplus water eventually dries out, leaving micropores in the matrix which weaken the resulting set mortar.It has generally been assumed that the same relationship applies to non-hydraulic lime (air lime) mortars.

This paper reports on results of tests conducted on air lime mortars at early stages of curing. It is known thatthe form of air lime and the physical and chemical characteristics of the aggregate have a strong impact on thestructural performance of air lime mortars. Results to date show that the water/lime ratio has a minimal impacton the structural performance of air lime mortars compared with the impact of lime and aggregate type. 91 daycompressive strengths for air lime mortars with a water/lime ratio of 0.56 (a stiff mix) are identical to those witha water/lime ratio of 0.875 (a loose slurry). Whereas Abrams’ rule is a key consideration for designers of cementand hydraulic lime mortars, it has been demonstrated that it requires modification in the case of air lime mortars.

A relationship between form of lime, type of aggregate, water/lime ratio, and age of mortar is proposed. Theresultant equation allows the compressive strength of air lime mortars to be predicted taking into account thesefactors. The insights gained from this study will allow practitioners to more confidently design and specify airlime mortars.

1 INTRODUCTION

Many of the mechanical properties of hardened cementare associated with the physical structure of thehydration products, viewed at the level of colloidaldimensions [Neville, 1995]. The pore structure ofhydrated cement paste contains two distinct ranges ofpore size – gel pores of about 3nm in diameter andcapillary pores which are two or three orders of mag-nitude larger. Cement requires sufficient water to fullyhydrate the various constituents, and water in excessof this required amount produces capillary pores. Thusthe greater the water/cement ratio above the min-imum required for complete hydration, the greaterthe amount of capillary pores created and thereforethe higher the porosity of the hardened paste. When

concrete or a cement mortar is fully compacted, itsstrength is inversely proportional to the water/cementratio according to Abrams’ rule.

where w/c represents the water/cement ratio of themix by volume, and K1 and K2 are empirical constants.K1 relates to aggregate strength, particle shape, size,grading and surface texture, and K2 relates to the com-pressive strength of the cement paste [Nagaraj & Banu,1996].

Abrams’ rule could well be valid for hydraulic limemortars, since they also gain at least part of their

885

Figure 1. Relationship between water/lime ratio andstrength (Allen et al. 2003).

strength through hydration products.Allen et al. [2003]have shown this relationship in Figure 1.

The data presented in Figure 1 cannot, however,be taken to be truly representative of Abram’s rule.This is because the data are based on the compres-sive strengths of different binder:aggregate ratios –each mortar requiring a different quantity of water toproduce a specified flow. This means that two vari-ables are present in the graph, and it is not clear whatproportion of the compressive strength is affected bywhich variable. There has been conflicting evidenceabout the applicability of Abram’s rule to air limemortars. Schäfer & Hilsdorf [1993] and Winnefeld &Böttger [2006] present data which show that increasedwater content in air lime mortars does not reducecompressive strength. It has been shown that higherporosity in air lime mortars allows greater accessto atmospheric carbon dioxide (CO2), which pro-motes carbonation and therefore can produce greatercompressive strengths [Lanas & Alvarez, 2003].

This paper describes a systematic evaluation of theimpact of the water/lime ratio on the unconfined com-pressive strength of air lime mortars up to 91 days aftermanufacture.

2 EXPERIMENTAL

50 mm × 50 mm × 250 mm prisms of mortar were pre-pared with 1 part of dry hydrated high calcium lime(CL90) and 3 parts of silicate sand by volume usinga range of different water/lime ratios. 9 specimensof each mortar type were prepared and 2 cubes weretaken from each specimen for testing at the appropri-ate time intervals. For comparison purposes a furtherset of 9 prisms were prepared using 1 part NHL3.5lime and 3 parts silicate sand with a range of differentwater/lime ratios.

The amount of water added to each lime type rangedfrom the minimum quantity needed to make a workablemortar to the amount required to make a loose slurry.The water/lime ratios used were as shown in Table 1.

Table 1. Water/lime ratios used [by volume]

Specimen Water/LimeDesignation ratio

Air LimeA1 0.5A2 0.5625A3 0.625A4 0.6875A5 0.75A6 0.875

Hydraulic LimeH1 0.375H2 0.4375H3 0.5H4 0.5625H5 0.625

It was found that air lime required more water inorder to make a workable mix than hydraulic lime,and could accommodate more water before becom-ing a loose slurry. This was likely to be a function ofthe greater capacity of air lime to absorb water thanhydraulic lime as a result of having finer particles andtherefore a greater surface area.

Both mortars were de-moulded after 5 days andcured in a controlled environment of 60% RH at 20◦Cuntil testing.

Compressive tests on six 50mm cubes were con-ducted after 28, 56 and 91 days from the date ofmanufacture.

3 RESULTS

The results of compressive tests on the mortars areshown in Figures 2–4. Error bars are included showingthe range of results of the six tests used to produce eachdata point.

These data compare well with the data producedboth by Schäfer & Hilsdorf [1993] and Winnefeld &Böttger [2006].

4 ANALYSIS AND DISCUSSION

The data for hydraulic lime mortars appear to followAbrams’ rule with the relationship between strengthand water/lime ratio following an approximate hyper-bolic curve. The data for the air lime mortars, apartfrom the lowest water/lime ratio, show very little varia-tion in compressive strength when the water/lime ratiois varied. It is conceivable that the data points for theair lime mortar are all to be found at the lower end ofthe hyperbolic curve, where there would be very little

886

Figure 2. Compressive test results on specimens 28 daysfrom manufacture and different water/lime ratios.

Figure 3. Compressive test results on specimens 56 daysfrom manufacture and different water/lime ratios.

Figure 4. Compressive test results on specimens 91 daysfrom manufacture and different water/lime ratios.

difference to be seen from an increase in the water/limeratio. The factors involved in the strengthening ofair lime mortars are different from those involved inhydraulic lime mortars. In air lime, after an initialstrength gain achieved from the drying out of the mor-tar, subsequent strength gain is achieved over extendedperiods as a result of carbonation. In hydraulic limemortars, there is also an initial strength gain achievedfrom drying, combined with a gain achieved through

an hydraulic set, which takes up to 28 days, dependingon the hydraulicity of the lime. Subsequent to this thestrength gains are due to carbonation.

Van Balen [1994] has proposed a model for car-bonation represented by a differential equation witha ‘sink’ term (R(w, c)) . The factors involved in theequation include time, the porosity and diffusivity ofthe material, the construction method and the presenceof cracks, and the geometrical shape of the surfaceexposed to air. Carbonation depth (x) is proportionalto the square root of time (t) ± a constant (e) in theform x = k

√t or x = e + k

√t, where k is a factor which

does not necessarily correspond to a property of thematerial.

It has been shown that the compressive strength ofair lime mortars varies in proportion to the extent ofcarbonation (Lawrence, 2006a), and it is therefore rea-sonable to expect that development of compressivestrength will also be proportional to the square rootof time.

The following formula has been developed to modelthe variation in compressive strength of air limes as thewater/lime ratio varies.

where fl is the compressive strength of the air lime mor-tar, Km is an empirical constant which varies accordingto the nature of the aggregate, Kl is an empirical con-stant relating to the form of air lime, and d is the ageof mortar in days since manufacture. The factor of 150is an empirical constant derived to provide a best fitwith actual data.

It is known that differences in the mineralogy andgranulometry of an aggregate will have a significantimpact on the compressive strength of air lime mortars[Lawrence et al. 2006], even at a very early stage aftermanufacture. Km represents this effect. This constantwill not vary for a given aggregate whatever the timefrom manufacture.

As air lime mortar increases in age, so carbonationhas an increasing impact on the compressive strengthof the mortar.This effect occurs across the whole rangeof water/lime ratios, and the expression

√d represents

this effect. The value of this expression will increaseas the time from manufacture increases up to the pointwhere carbonation is virtually complete. This expres-sion appears to be valid up to values of 180 for d,beyond this value – once the mortar has carbonated,the expression would not vary.

Different forms of air lime carbonate to a greater orlesser extent and at a greater or lesser rate, mainlydependent on the size, shape and integrity of port-landite crystal present in the lime. Kl represents thiseffect. This constant will not vary for a given limewhatever the time from manufacture.

887

Figure 5. Day 28 air lime data compared with the proposedequation.

Figure 6. Day 56 air lime data compared with the proposedequation.

Figure 7. Day 91 air lime data compared with the proposedequation.

As carbonation progresses through the depth of themortar so the rate of carbonation will decrease. Theexpression 150∗d represents this effect.As commentedon above, this expression would become a constantonce carbonation has completed.

The proposed equation can only be considered tobe valid only up to 180 days from manufacture for airlime mortars.

Figures 5–7 represent models using the above for-mula for mortars at 28, 56 and 91 days, and compared

Figure 8. Proposed equation applied to mortars made withdifferent aggregates. 4SS3 = silicate sand; 4BN3 = crushedbioclastic limestone (Ham Hill stone); 4ON3 = crushedoolitic limestone (Stoke Ground Bathstone).

with the data shown in Figures 1–3. For these calcu-lations, the value of Km was taken as 0.05, and that ofKl as 20.

At 28 days the equation over estimates the com-pressive strength by up to 25%, improving with higherwater/lime ratios. As the mortar ages, so the fitimproves, until by 91 days, apart from the lowestwater/lime ratio, the equation fits within the error barsof the actual data.

The equation predicts a reduction in compressivestrength in line with increases in water/lime ratiowhich the data do not entirely support, although exper-imental variations may explain small increases instrength at higher water/lime ratios.The equation does,however, go some way towards accounting for the moresignificant influence of the type of aggregate on thecompressive strength.

Figure 8 shows the proposed equation applied to dif-ferent aggregate mortars keeping the water/lime ratioconstant, but varying the time (d).The solid lines showthe relationship between compressive strength for limemortars made with oolitic aggregates (green), bioclas-tic aggregates (red) and silicate sand aggregates (blue),and the curves predicted by the proposed equation.Compressive strength data are taken from Lawrence[2006b]. The water/lime ratios used to make the actualmortars were factored in. The constant Km which rep-resents the impact of the aggregate on the compressivestrength of the mortars was 0.420 for the oolitic mor-tar, 0.220 for the bioclastic mortar, and 0.106 for thesilicate sand mortar.

It can be seen that up to 180 days there is a rea-sonable correlation between actual and predicted com-pressive strengths. The equation requires modificationto take account of the completion of carbonation, butup to 180 days it seems to be able to predict with rea-sonable accuracy the compressive strength of air limemortars as they are affected by water/lime ratio, typeof aggregate and time from manufacture. The data for

888

mortars made with oolitic stone shows greater earlystrength than predicted by the equation. Indeed earlystrength exceeds that of moderately hydraulic limemortars. It has been shown by Lawrence (2006a) thatno hydraulic effects are involved, and this phenomenonis the subject of ongoing research.

5 CONCLUSIONS

The experimental data demonstrate that, unlike cemen-titious or hydraulic lime mortars, the increase inporosity produced by higher water content does notresult in the strength reductions predicted by Abrams’rule. Indeed it can be seen that the choice of aggre-gate has a significantly greater impact on the strengthcharacteristics of air lime mortars.

One key implication of this are that for air-limemortars, water/binder ratios can safely be selected toproduce a mortar with suitable workability charac-teristics as demanded by the particular application,rather than based on the erroneous assumption thatAbrams’ rule applies. Unlike hydraulic limes, withair-lime mortars apart from very stiff mortars madewith water/lime ratios of 0.5 compressive strengthsare shown to be remarkably unaffected by greaterquantities of water in the initial mix.

The proposed equation will allow a first approxima-tion of compressive strength of different air lime mor-tars to be predicted. Continuation of the present studywill allow the creation of empirical constants for typesof lime and aggregate (Km and Kl). More work needs to

be done to understand the mechanisms involved whichproduce such significant differences in compressivestrength. Such an understanding will allow the formu-lation of a more developed equation than that whichhas been proposed.

REFERENCES

Allen, G., Allen, J., Elton, N., Farey, M., Holmes, S.,Livesey, P., Radonjic, M. 2003. Hydraulic lime mor-tar for stone, brick and masonry. Shaftesbury, DonheadPublishing Ltd.

Lanas, J., Alvarez, J.I. 2003. Masonry repair lime-based mor-tars: factors affecting the mechanical behaviour. Cementand Concrete Research 33(11): 1867–1876.

Lawrence, R.M.H., Walker, P., D’Ayala, D. 2006a. Non-hydraulic lime mortars. The influence of binder andfiller type on early strength development. Journal ofArchitectural Conservation 12(2): 7–33.

Lawrence, R.M.H. 2006b A study of Carbonation in non-hydraulic lime mortars, unpublished PhD Thesis, Univer-sity of Bath.

Nagaraj, T.S., Banu, Z. 1996. Generalisation of Abrams’ law.Cement and Concrete Research 26: 933–942.

Neville, A.M. 1995. Properties of Concrete. Harlow,Longman.

Schafer, H.R., Hilsdorf, H.K. 1993. Ancient and new limemortars – the correlation between their composition struc-ture and properties. In M. J. Thiel Conservation of Stoneand other Materials. London, E. & F.N. Spon: 605–612.

Van Balen, K., van Gemert, D. 1994. Modelling lime mortarcarbonation, Materials and Structures, 27: 393–398.

Winnefeld, F., Böttger, K.G. 2006. How clayey fines in aggre-gates influence the properties of lime mortars. Materialsand Structures 39: 401–411.

889