March 2004 * The Indian Concrete Journal

Point of View

1

Dual role of gypsum: Set retarder andstrength accelerator

These columns of ICJ offer an opportunity to theengineering fraternity to express their views on thecurrent practices in design, construction andmanagement being followed in the industry.To share your opinion with our readers, you may sendin your inputs in about 1500 words via E-mail toeditor@icjonline.com

Gypsum plays a crucial role in cement.Though it is used in a small quantity, inthe range of 2.5-3.0 percent in terms ofSO3, gypsums role in cement issignificant, more predominantly at earlyages. Gypsum renders workability tomortar or concrete by keeping the cementin plastic state at early age of hydration.This is achieved by changing the courseof hydration of calcium aluminate thatmanifests as retardation in cementhydration. This is how gypsum isidentified as a set regulator or retarder,as known popularly . Nevertheless,gypsum also contributes for strengthacceleration in the early stages ofhydration. This dual role of gypsum isdiscussed in the feature.

Gypsum is the set retarder for ordinaryportland cement (OPC). Without gypsum,ground clinker exhibits flash setting in afew minutes, due to the rapid hydration ofcalcium aluminates to form calciumaluminate hydrate (CAH). The hydrationof C3A releases profuse exothermic heatmaking the matrix stiff, minimising thechances for remixing. The CAH, thusformed, does not contribute for strength ofthe matrix and, moreover, hampers the hy-dration of calcium silicate. The sequence of

reactions, in the absence of gypsum, totallyvetoes the commercial use of cement. Thefollowing chemical reaction, in the absenceof gypsum, is explanatory in this regard.

3 CaO. Al2O3 + n H2O reactionsfast

CAH + profuse exothermic heat

Hence, it was found essential to changethe reaction course of C3A, and this wasmet by the use of sulphate salts. Due to itsaffinity with SO3 , aluminate tends to reactreadily with the former and in this processthe reactions of aluminate with water areprevented. Ultimately, gypsum wasidentified as the most effective form ofsulphate to control hydration reactions ofC3A that incidentally resulted in betterworkability for a longer duration.

Chemical reaction in the presence ofgypsum is given below

3CaO. Al2O3 + 3CaSO4. 2H2O + nH2O

3CaO. Al2O3. 3CaSO4. 32H2O

(Ettringite : calcium trisulpho aluminatehydrate) + moderate exothermic heat

Many in the cement and concreteindustry know the role of gypsum as setretarder or set regulator. But thecomplementary role of gypsum, asaccelerator to render high early strengths, isgenerally unnoticed. This knowledge gap

lead to a misunderstanding that theaddition of more gypsum meansadditional retardation in setting, which isnot true. The formation of ettringite atthreshold levels accelerates the hardeningprocess and thus hastens strength gain atearly ages. In view of this behaviour, thelatest European code on cement ENV197 - 1 stipulates higher dosages of SO3(between 3.5 to 4 percent by mass)1. Asagainst this, the Indian specifications onboth 43 and 53 OPC grade (IS 8112 : 19892

and IS 122269 : 19873) specify themaximum SO3 content of 2.5 percent bymass (for C3A< 5 percent) and 3 percentby mass (for C3A> 5 percent). Thespecification on portland pozzolana cement(IS 489 : 19914) and that on portland slagcement (IS 455 : 19895) stipulates SO3 levelof 3 percent by mass.

In the manufacture of blended cements,by virtue of reactive aluminates from flyash and slag that leads to ultimate cement,the chances for existence of resultantcalcium aluminate hydrates are more thanwhat is generally available in OPC, as shownin the Table 1.

Hence, the increasing positive thresholdlevels of gypsum is one of the solutions toovercome the weakness of strengths at earlyages in mortar/concrete with blendedcements. This aspect is also of significancewhen the blending is done in ready mix

N. Bhanumathidas and N. KalidasN. Bhanumathidas and N. KalidasN. Bhanumathidas and N. KalidasN. Bhanumathidas and N. KalidasN. Bhanumathidas and N. Kalidas

2 The Indian Concrete Journal * March 2004

Point of View

concrete plants. While all fly ashes maynot need additional gypsum, care has to betaken not to deprive those fly ashes in needof additional gypsum.

FaL-G technology, developed by theauthors, achieved its breakthrough bytapping the potential of calcium aluminatestowards the formation of ettringite andmono-sulphate; thus changing the pace oflime-fly ash chemistry. Gypsum plays apredominant role as the strength-accelerator in the context of hydratedmineralogy in FaL-G. Moreover, gypsumworks as set-accelerator in lime-pozzolanabinders in contrast to its role as set-retarderin OPC.

Calcium aluminatesulphate chemistryBefore getting into the details of the min-eralogical formations, it is essential to brieflygo through the chemistry and hydration ofOPC, which is a product of four principalmineralogical phases, namely,

3CaO. SiO2: C3S (Tricalcium silicate)

2CaO. SiO2: C2S (Dicalcium silicate)

3CaO. Al2O3: C3A (Tricalcium aluminate)

4CaO.Al2O3.Fe2O3: C4AF (Tetracalciumaluminoferrite)

Upon adding water, the anhydrousmineralogy gets dissociated as CaO, SiO2,Al2O3 and Fe2O3 for associating intohydrated mineralogy. Setting is the interfacefor this transformation. Soroka6 gavesequential illustrations for working period,initial setting and final setting for cementpaste, by graphic representation of themineralogical phase formations as shownin Fig 1.

As explained by Soroka in the firstphase, when water is added to cement, as aresult of the hydrolysis of the calciumsilicates, a super-saturated solution ofcalcium hydroxide is formed. Sulphate andalkali ions, as well as small amounts ofsilica, alumina and ferric oxide are alsopresent in the solution. Calcium hydroxide

and ettringite precipitate out and a denseC-S-H gel coating is formed on the cementgrains. This coating as well as ettringitecoating on C3A grains retard furtherhydration, and this explains the existenceof dormant period, that is, the period ofrelative inactivity lasting for one to twohours. During this dormant period the pasteremains plastic and workable. The end ofthe dormant period can be identified asinitial set. This is attributable to the breakup of C-S-H and ettringite coatings, andthe resultant continuation of hydrationprocess. Due to the osmotic pressure thegel coating gets ruptured, exposing thecement grain, wherein hydration is resumedand setting takes place.

Consequently, as the hydrationproceeds, the hydration products graduallyfill in the spaces of the cement grains. Pointsof contact are formed resulting in stiffeningof the paste, which is identified as thesetting. At somelater stage, theconcentration ofhydration productsand resultantconcentrations ofpoints of contactrestrict the mobilityof cement grains tosuch an extent thatthe paste becomesrigid, reaching thestate of final set.

The volume ofthe hydrationproducts, throughthe process ofcrystallisation andmatrix formation, ismore than twice tothat of theanhydrous cement.The dissociationand association ofmineralogy continueas long as themoisture isavailable andrelease of lime iscontinued in the

matrix. This activity was sustainable forlonger ages in low grade cements, accreditedfor their durability, which is explained interms of heat bank and lime bank by theauthors7. In contrast, accelerated reactionstake place in high grade cements forrelatively shorter periods, leading to thecomplications and decrease in durability.

Ettringite in pre-hardenedand post-hardenedconcrete.It is observed that the ettringite formed inpre-hardening stage is conducive for thematrix formation in comparison to the sameformed in post-hardening stage. The sim-ple reason is its volume expansion by tak-ing almost 32 molecules as water of crys-tallisation. As already mentioned, in a hy-drated cement paste, the level of imperme-ability increases through volume expansionof hydrated phases and the resultantdensification through the process of cur-ing. In the initial ages of hydration this ex-pansion contributes for the internalcompaction of matrix, resulting indensification and strength. These phenom-ena lead ultimately to impermeability ofthe matrix. However, if the same expan-sion occurs after the matrix attains volumestability then the matrix is subjected to in-

Table 1: Need of SO3 corresponding to calcium aluminates formation in the blends

Type of cement C3A CA As monosulphate Need of SO33CaO.Al2O3.CaSO4.12 H2O for final phase

OPC 5.00 9.28 11.50 1.47

OPC(70) + FA(30)* 3.50 16.70 29.40 6.40OPC(50) + Slag(50)** 2.50 38.40 8.35Note: CA - Reactive calcium alumino silicate, *Fly ash with 30 percent reactivity containing 25 percent Al2O3,**Slag with 90 percent reactivity containing 14 percent Al2O3.

March 2004 * The Indian Concrete Journal

Point of View

3

ternal expansive pressures that result incracking of the concrete.

If the hydration chart of Soroka isobserved for the formation of variousmineralogies, the formation ofsulphoaluminate hydrates is predominanttill the final set that is followed by calciumsilicate hydrates. As ettringite descendsfrom its peak of formation by first or secondday, the formation of monosulphate doescommence. What does it mean? It meansthat gypsum gets largely exhausted for theformation of tri-sulphate. At this stagecalcium aluminate hydrates continue to beavailable with affinity to formsulphoaluminate hydrates. Due to theiraffinity for sulphate ion these aluminatehydrates react with ettringite. This ismanifested as the commencement ofmonosulphate formation. These reactionsconcur with the physical status of the matrixwhere porosity of the matrix is largely filledup by ettringite at early age, beyond whichthe reduced porous spaces areprogressively occupied by monosulphate.Thus, the three-day strength is largelyattributable to the formation of calciumsulphoaluminate hydrates.

The transformation of ettringite tomonosulphate is as follows:

3CaO.Al2O3.3CaSO4.32H2O +2(3CaO.Al2O3) + x H2O

3(3CaO.Al2O3.CaSO4.12H2O)

This is a slow reaction that either leavesmonosulphate as the ultimate mineralogyof sulphoaluminates or makes the same intohexagonal plate solid solution with CAH,probably C4AH13, to result in a stable stage.In the sequence of calcium aluminatehydrate chemistry, ettringite (trisulphate)formation is an in-built mechanism or a boonto render high early strengths to the matrix.This can be achieved only in the presence ofgypsum, which also hastens the CSHchemistry8. This is how gypsums role isupheld as early-strength accelerator.

At the early ages, the cementitiousmatrix needs heat for rapid hydration ofchemistry. This is rendered by the hydrationof C3A, which is commensurately availed

as heat of formation. The threshold dose ofgypsum regulates the heat of hydration asalso the ettringite formation for progressivechemistry. In the anxiety of not to retardcement if the gypsum is reduced below thethreshold level, heat of hydration iscommensurately profuse leading to internalthermal stresses and incohesiveness. Thisis where the role of gypsum is evident asretarder.

Maximum amount of gypsum isengaged into ettringite in the first two tothree days. Thus, matrix becomes morecohesive with improved strengthdevelopment. Nevertheless, each clinker hasits own optimum level of demand forgypsum that decides the othercharacteristics. The data in Table 2 elucidatethe same.

In the above studies on OPC, additionof SO3 beyond 2.1 percent is proving adverseboth in strength and heat of hydrationthough the one-day strength is impressive.Hence, it would be highly inappropriate todecide the quality of cement based on one-day strength alone, which unfortunately isprojected as a quality parameter by many.

Strength data of calciumaluminates to calciumsulpho aluminatesGypsum in FaL-GIn FaL-G studies, two formulations weremade for fly ash lime mixes with and with-out gypsum. The predominant differenceis the formation of calcium aluminate hy-drate where there was no gypsum; and theformation of calcium sulphoaluminate hy-drate where gypsum was available. Thesemineralogical formations were substanti-ated by XRD. Notwithstanding these for-mations, the strength data also substanti-ate the observation both at 7-day and 28-day as shown in Table 3. In the light of mod-erate performance by calcium silicate chem-istry in fly ash lime mixes at early ages, thecredit of the early strength goes to ettringite.

Table 3 shows that LT fly ashes show higherreactivity than HT fly ashes. This could be

attributed to the presence of higher reactiveamorphous alumino silicates in LT fly asheswhich produces CAH that can further formettringite with gypsum.

Gypsum in blended cementsThe hydration chemistry of blended ce-ments is a two-phase mechanism. In thefirst phase, the OPC chemistry surfaces andin the second phase pozzolanic chemistrycomes into force. This is represented as fol-lows:

OPC + H fast

Primary hydrated mineralogy + CH

Pozzolana + CH + H slow

Secondary hydrated mineralogy

In pozzolanic reactions, alumina isnoticed but the addition of commensurategypsum is not given much weightage. Eventhe ASTM definition of pozzolanicchemistry is silent on this issue, which says,Siliceous or siliceous and aluminousmaterials, which, though not cementitiousthemselves, react with lime, when in finelydivided form, in the presence of water atordinary temperature, and form stable andinsoluble mineralogical phases, possessingcementitious characteristics.

But the principle behind use ofadditional gypsum in blended cements liesin the capability of SO3 to break the glass ofpozzolana. This occurs due to the affinitybetween SO3 and alumina that facilitatesthe formation of calcium sulphoaluminatehydrates. The indirect benefit is theavailability of reactive silica, freed fromglass of pozzolana, for reactions. Theformation of additional ettringite and CSHmakes the cement matrix more densifiedand impermeable at early ages. The studieson fly ash blended mortars and concretesubstantiate this phenomenon as shown inTable 4.

Table 2: Relation of strength and heat of hydration to increase in SO3 in OPC

SO3, Setting time, min Compressive strength, MPa Heat of hydration, kCal/kgpercent Intial Final 1-day 3-day 7-day 28-day 12-hour 1-day 3-day 7-day 28-day

1.80 115 150 17.7 37.4 53.8 61.6 22.5 33.2 45.4 59.3 67.2

2.10 130 165 20.9 40.2 59.0 65.4 28.2 33.2 44.5 63.0 65.7

2.40 135 180 20.9 32.1 47.8 62.0 46.0 54.2 63.2 80.3 88.5

Surface area of OPC : 320 kg/m2

Table 3: Increase in strength uponaddition of gypsum in fly ash-lime mixesboth in LT and HT fly ashes

Source Compressive strength MPa Fly ash + lime FaL-G

7-day 28-day 7-day 28-day

LT fly ash 1 9.0 17.9 25.0 32.0

LT fly ash 2 11.0 15.8 20.0 25.8

HT fly ash 1 2.6 7.8 8.4 24.0

HT fly ash 2 3.3 4.9 6.5 24.8

4 The Indian Concrete Journal * March 2004

Point of View

Durability aspectsChloride permeability withchange in dose of gypsumWhen gypsum is added as a third compo-nent in lime reactivity studies, the LR valueimproves strikingly that indicates forma-tion of quite cohesive matrix associated withimproved strength as shown in Table 5.

Studies on blended concretes usingSample 3 have brought in interestingphenomena towards improved durabilitydespite higher workability as shown inTable 6.

The rationale of gypsum inEuropean codes:Studies were conducted on OPC at differ-ent doses of gypsum mentioned in terms ofSO3. The heat of hydration diminishes upto a particular level beyond which, it is notedthat, increase of SO3 the increases heat of

hydration also, which is evident from thedata already given in Table 2.

This means, gypsum liberates more heatof hydration on account of rapid chemistry.However, beyond a threshold level of SO3,such rapid chemistry would not contributefor strength gain. This explains the rationalebehind the increase of gypsum for rapidhardening cements vide European code,keeping an eye on the threshold limits.

ConclusionsGypsum works as a double-edged swordin cement chemistry. One needs to have com-prehensive understanding of the cementmineralogy and hydration chemistry beforedeciding on the dosage of gypsum. Asmuch one can get optimum results by judi-cious input of gypsum, so much so onemay get disastrous results too by improperdosing. This is where the holistic knowl-edge on cement chemistry is of significance.

AcknowledgementsThe authors acknowledge the guidance ofProf P.K. Mehta and his wealth of librarymade available. The authors also recordtheir gratefulness to Madras Cements Ltd.,for making available the laboratories to thisresearch.

References1. ______Cement part I : Composition, specifications

and conformity criteria for common cement, ENV,197-1 : 1995, European Committee forStandardisation, Brussels.

Table 5: Improvement in lime reactivitystrength with the addition of gypsum

Fly ash Lime reactivity strengths MPa Without gypsum With gypsum

10-day 90-day 10-day 90-daySample 1 2.8 6.4 5.2 10.8

Sample 2 7.7 11.9 11.9 19.3

Sample 3 9.1 15.9 17.7 21.7

Table 6: Reduction in permeability upon addition of gypsum

Cementitious Slump, mm Compressive strength, MPa Chloride permeabilityContent, percent 3-day 7-day 28-day 90-day 180-day 360-day material, CoulombsOPC Sample 3 28-day 90-day 180-day 360-day

100 14 21.6 31.2 43.3 47.2 51.4 54.7 3852 2451 2251 1912

65 35 80 12.2 17.7 35.1 53.7 58.4 65.9 2529 313 166 155

65 35* 139 13.1 18.8 35.7 54.9 61.3 63.7 2107 234 130 151

*Added with 2.76 as SO3.

2. ______Indian standard code for 43 grade ordinaryportland cement, IS 8112 : 1989, Bureau of IndianStandards, New Delhi.

3. ______Indian standard code for high strengthordinary portland cement, IS 12269 : 1987, Bureauof Indian Standards, New Delhi.

4. ______ Indian standard code for fly ash basedcement, IS 1489 : 1991, Bureau of IndianStandards, New Delhi.

5. ______Indian standard code for blast furnace slagbased blended cement, IS 455 : 1989, Bureau ofIndian Standards, New Delhi.

6. Soroka I: Portland Cement Paste and Concrete;Chemical Publishing Co.,Inc. New York, 1979.

7. BHANUMATHIDAS, N and KALIDAS, N Metabolismof cement chemistry, The Indian Concrete Journal,September 2003, Vol 77, No 9, pp. 1304-1306.

8. MEHTA, P.K. and MONTEIRO, PAULO J.M. Concrete:Microstructure, Properties, and Materials, TheMcGraw Hills Companies Inc, New York, 1993.

Table 4: Impact of gypsum on compressive strength at different inputs of fly ash

OPC Compressive strength of mortar, MPaFly ash 7-day 14-day 28-day 60-day 270-day

(a) (b) (a) (b) (a) (b) (a) (b) (a) (b)50 : 50 24.4 20.0 32.4 22.8 40.8 29.6 50.8 44.4 56.8 48.0(PPC-I)40 : 60 21.2 15.2 29.6 22.0 38.0 34.0 45.2 44.4 56.0 47.8(PPC-II)100 : 0 27.2 35.2 40.4 48.0 54.4(control mortar)(a): Added with anhydrite commensurate to fly ash quantity, (b): Without anhydrite

Dr N Bhanumathidas is thedirector general of Institutefor Solid Waste Research andEcological Balance(INSWAREB), the researchbody dedicated to the utilisa-tion of industrial wastes to-

wards building material. After obtainingher postgraduate degree in physics fromAndhra University, she did her doctoralstudies in chemical engineering (inter-dis-ciplinary). Dr Bhanumathidas has authoredseveral technical papers in collaboration withher associate, Mr N Kalidas, and presentedat several national and international semi-nars. FaL-G is the outcome of this team-work. She is currently focussing on ad-vanced concrete technology, use of indus-trial byproducts as complementary cemen-titious material and promotion of blendedcements.

Mr N Kalidas is engaged inthe pursuit of waste utilisa-tion technologies for the last18 years as a technocrat andby virtue of his assignmentwith certain overseas compa-nies. In order to consolidate

his work on waste utilisation, he along withhis associate, Dr N Bhanumathidas,founded the research body, INSWAREB. Heis the director of INSWAREB. His field ofinterest includes: advanced concrete tech-nology, use of industrial byproducts ascomplementary cementitious material, pro-motion of blended cements. Along with DrN Bhanumathidas, he has authored sev-eral technical papers and presented at vari-ous national and international seminars.