DURABILITY OF CONCRETE IN SOIL OR GROUNDWATER This material was compiled in January 2011 as a contribution to proceedings within the Concrete Institute of Australia, developing a series of Recommended Practice documents relating to durability. Concrete having strength grade 50 MPa or above is practically impermeable and highly resistant to carbonation, with the result that for this type of concrete most natural soil conditions do not result in any kind of deterioration either of the concrete itself or the steel reinforcement. Attack on the concrete, if it occurs at all, is due to aggressive groundwater - if the concrete is dry, it does not deteriorate. Potentially harmful materials are identified by analysis of groundwater or water extracts from soil samples. The most common aggressive components of groundwater in natural environments are acids, dissolved carbon dioxide (CO2), sulfates and chlorides. Other types of aggressive, not dealt with here, can result from soil contamination by industrial waste.

Acids and Dissolved Carbon Dioxide

The aggressiveness of water containing dissolved CO2 is usually characterised by the concentration of “aggressive CO2”, calculated from total “free” CO2 and calcium (Attachment 1). This is a measure of the amount of calcium which a given volume of the water will dissolve. To indicate the aggressiveness of water containing acids (other than carbonic acid which is CO2 in solution), it is common to quote the pH value. While this is a useful measure of the threat of attack by acid the rate of deterioration is not uniquely determined by the pH, as other factors such as solubility or otherwise of the calcium salts of the particular type of acid will also have an influence. In Attachment 2, examples are plotted of rates of attack from acid v. the pH, and from aggressive CO2 , v. the concentration, over various periods of time. These apply to high quality concrete and are the basis of limits applicable to concrete having minimum cover determined from other criteria. For higher concentrations of the aggressive, the life can be extended by increasing the cover to reinforcing steel (Attachment 3).

There are no marked differences in the resistance to attack by acid between the various types of portland or blended cements (Attachment 4). Higher acid resistance may be conferred by high alumina cement, supersulfated cement and geopolymer but the use of these cements is outside the scope of this account.

Sulfate

The mechanism of sulfate attack is quite different from that of acid or aggressive CO2. Sulfate reacts with calcium hydroxide and tricalcium aluminate in the hardened concrete, to form reaction products which are greater in volume than the reactants, and, depending on the concentration of sulfate and the amounts of these compounds in the hardened concrete, the concrete is distorted or disrupted.

Good quality concrete made with sulfate resisting cement will withstand very high levels. In Attachment 5, results are described of 20 years’ exposure of cast concrete specimens to water containing sodium and magnesium sulfates, the sulfate content totalling 4%. Specimens made with sulfate resisting cement had suffered only minimal attack. Samples from the test series having results shown in Fig. 3 of Attachment 6, with sulfate concentration 10,000 ppm, continued for 24 years and those containing fly ash showed very little effect at the end of the exposure period. This result was better than that obtained in parallel tests using sulfate

resisting cement but no fly ash; ie there was a significant benefit from removal of the free lime by the fly ash, beyond that obtainable just from a low level of tricalcium aluminate.

Chloride

Chlorides are widespread in groundwaters. In Australia, high concentrations are likely to be found near the sea, in ground periodically flooded by sea water. In regions where water evaporates very quickly from the soil, high chloride concentrations build up and become dangerous when the soil is re-wet.

Immersed in water of neutral pH containing oxygen dissolved from the atmosphere, reinforcing steel will corrode rapidly at low levels of chloride, whereas in the same situation at the pH conferred by cement it will remain passive at chloride levels three times that of sea water (Attachment 7). Similar immunity to very high levels of chloride has been found in concrete (Attachments 7 & 8), though much of the literature also reports corrosion at much lower values. There is no satisfactory explanation of the wide discrepancies, by a factor of more than 10, of threshold concentrations of chloride, though it is acknowledged (Attachment 7) that a very dense cement matrix, achievable only with low w/c and therefore high strength grade, could inhibit the initiation of corrosion. The BRE publication “Concrete in Aggressive Ground” encompasses chloride levels approaching those found in sea water (about 2% chloride) but makes no allowance in its concrete specifications, except as a possible indicator of acid resulting from industrial waste. In this publication resistance to aggressive ground conditions is addressed through specification of the concrete, without reference to cover, and so there is no suggestion of increased cover to promote resistance to chloride.

It can be concluded that reinforcing steel will remain passive in the presence of groundwater having a chloride concentration at least up to that in sea water, provided only that alkalinity is maintained at reinforcement depth.

Unlike above-ground environments it is possible to permanently protect concrete from chloride underground by means of a paint coating or other physical barrier, which may be the most economical option where high levels of chloride in the soil are encountered.

Specifications for Concrete & Cover; Exposure Classifications

Acid and aggressive CO2:

In Attachment 2, the lines drawn to correspond with the data points have rates of attack 10 mm per 100 years for acid at pH 5.0, or aggressive CO2 at (approximately) 75 ppm. On this basis limits for 100 years life at 10 mm cover were proposed as set out in Table 1, which includes a somewhat arbitrary allowance for the effect of the soil in favouring or impeding renewal of the aggressive at the surface of the concrete. These were adopted by the pipe industry in Australia from the early 1980s.

Table 1. Limits for Acidity & Aggressive CO2

Constituent Soil Classification

Clay/stagnant Medium Sandy/flowing

Acid, pH min. 4.5 5.0 5.5

Aggressive CO2 , ppm max. 150 50 15

From the analysis in Attachment 3, the time for the concrete to be affected to a given depth is proportional to a power of the depth at least as high as three, and as the data points from

which the limits of aggressives in Table 1 were set are based on much shorter periods of time than 100 years, they are likely to be conservative. This was tested, including some updated data (Attachment 9) confirming the expectation, and further verification obtained by the conclusion of a long term test, Attachment 10. However by this time it was clear that the Table 1 limits were serviceable in the sense that higher levels were unusual in natural environments of undisturbed soil, and so there was little if anything to be gained by adjusting the values.

The tabulated values may be taken as applicable for concrete strength grade 50 MPa and above, straight portland or blended cement and at least 10 mm cover. The exposure classification is non-aggressive.

Slopes of lines shown in Attachment 2 correspond to rates of attack proportional to the concentration of the aggressive. Where the pH is lower or the concentration of aggressive CO2 higher than the limiting value from Table 1, the required amount of cover can be calculated using this and the depth cubed relationship mentioned previously. For example – in a “sandy/flowing” soil condition with pH 4.0, 1.5 units of pH below the limit and therefore a hydrogen ion concentration 10^1.5 = 32 times larger, the required cover for 100 year life would be 10*32^(1/3) = 32 mm. Depending on the method of construction, 32 mm minimum could require nominal cover in the range 35-40 mm, corresponding to exposure classification B2 per AS 3600. However it’s just as easy and makes more sense to adjust cover per the simple calculation illustrated above, rather than trying to link it to an exposure classification originally designed for a quite different purpose.

The cubic relationship between time and depth of attack depends on affected concrete remaining in place and so slowing access of the aggressive. In the immersion tests which were the subject of the analysis in Attachment 3, the lowest value of pH was 3.5 and in all cases the affected concrete remained firmly in place. However it has been observed that at lower pH (round 2), affected material is lost from the surface. In the extreme case of severe H2S in sewers (pH down to 1), mortar is progressively removed leaving coarse aggregate protruding from the surface, unless this is itself of an acid-soluble type. The coarse aggregate is lost or dissolved as the attack continues.

Sulfate:

Unlike attack by acid or aggressive CO2, there is no basis for attempting to extend the life of a structure subject to sulfate attack by specifying increased cover to reinforcement. High strength grade plus a suitable choice of cement will ensure immunity to practically any naturally occurring level of sulfate. In extreme conditions, for example due to industrial waste, the concrete should be isolated by a coating or other type of physical barrier.

Again limiting the strength grade to 50 MPa or greater, the specification is required to set upper limits for the concentration of sulfate where any cement complying with AS 3972 can be used, where the cement must comply with Type SR, and where this by itself is inadequate and an SCM must be included to remove calcium hydroxide.

Table 2. Maximum Concentration of Sulfate Ion

Cement (AS 3972) Any Type SR Type SR, blended

SO4-

ppm in groundwater, or soil extract 2:1 water to soil by mass

1000 3000 10,000

Cover is determined by other criteria.

Chloride:

Adopting the premise that the high pH in concrete will maintain steel in a passive condition for exposure to chloride up to the concentration found in sea water, it is proposed that for strength grade 50 MPa and above, concentration of chloride ion not greater than 2% and no mechanism which will increase the concentration above this level, the condition should be regarded as non-aggressive and will not influence either the cover or concrete specification.

For higher concentrations either: Isolate the concrete using a coating or other physical barrier. Specify the concrete and cover to limit diffusion of chloride to the reinforcing steel, so

that a critical level is not reached at reinforcement depth within a time interval to related to the required service life of the structure.

Design by Exposure Classification

The above description identifies classes of exposure as follows:

Three levels of pH Three levels of aggressive CO2 Three soil classifications applicable to pH and aggressive CO2 Three levels of sulfate One level of chloride

As pH below 4.5 (the lowest value in Table 1) can be encountered in areas where sulfide oxidises when the soil is disturbed, it is proposed that another two levels, down to pH 3.5, should also be included.

Combinations of the above for acid, aggressive CO2 and sulfate amount to 405 separate conditions of exposure. However if combinations of different types of aggressive are excluded, this number is reduced to 28, and the single condition of chloride less than 2% brings the total to 29. Grouping the classifications to situations requiring a separate specification for the concrete or cover, the number is reduced to 8, as illustrated in the table below.

However I am not proposing confidently that the classifications shown in the table should be adopted. With any grouping of actual conditions of exposure there is inevitably a tendency to lock together different types in a way which may not be appropriate in all circumstances. For example, as pointed out previously, the distinctions among the AS 3600 groupings A1-B1 are relevant for lower grades of concrete but not 50 MPa or above. In the present context, levels of acid and aggressive CO2 can be grouped on the basis of their rates of attack, but it’s at least possible that more data would show the correspondence so made to be in error. The alternative should at least be considered of using a diagnostic approach, with input comprised of parameters defining the prevailing situation, and output consisting of one or more pre-classified combinations of concrete strength grade, cement type and cover.

Soil/groundwater environment Maximum level of aggressive 1 Cement 3 Cover or

coating 4 Exposure

Classification 5 pH CO2 Soil 2 SO4

-

5.5 CS NA

5.0 CS NA

4.5 CS NA

4.0 CS N+5 A5

3.5 CS N+15 A15

5.5 M NA

5.0 M NA

4.5 M N + 5 A5

4.0 M N + 15 A15

3.5 M N + 25 A25

5.5 SF NA

5.0 SF N + 5 A5

4.5 SF N + 15 A15

4.0 SF N + 25 A25

3.5 SF Coat AC

15 CS NA

50 CS NA

150 CS NA

450 CS N + 5 A5

15 M NA

50 M NA

150 M 15 A15

15 SF NA

50 SF N + 5 A5

150 SF N + 10 A10

1000 NA

3000 SR AS

10,000 SR,

blended ASB

Plus one for chloride to 2%, NA. Notes on table: 1. Aggressive CO2 or sulfate ion in ppm. Blank cell indicates concentration less than the

lowest level tabulated. 2. CS = clay/stagnant, SF = sandy/flowing, M = medium.

3. Blank cell indicates any portland or blended cement to AS 3972 4. N = nominal cover, unaffected by the ground exposure classification; N + number is

cover depth in mm. Coat = protective coating eg epoxy, which is always an alternative to increased cover.

5. NA = non-aggressive.

Attachment 1

F M Lea, “The Chemistry of Cement and Concrete”, Third Edition, Arnold, 1970

Attachment 2

Humes report 771124 “External Corrosion & Protection of Buried Concrete”

Attachment 3

Humes report 800222

Attachment 4

ACI Monograph No. 4, “Durability of Concrete Construction”, American Concrete Institute, 1968.

BRE (UK) Special Digest 1:2005, “Concrete in Aggressive Ground”

OI I|,BReference RW/Gw0588

l .SUMMARY

Samples ol concrete pipe and autoclaved pipe (asbestoscement and cellulose), exposed to a variety of acidenv i ronmen ls , have shown s im i l a r res i s tances tochemical attack.

P ipe conc re tes made w i th b lended ( f l y ash ) andunblended cement a lso show s imi lar res is tance tochemicals, even lhough the blended cement concre,ucontains much less free l ime.

It is concluded that lree l ime has no signif icant effect onthe chemical resistance of cement in pipes. Removingthe free l ime does not provide any worthwhile benefit .

2. INTRODUCTION

The binder in asbestos cement and cellulose FRC pipesconsisls of a mixture of cement and f inely ground si l ica,which react together when the pipe is steam cured underpressure (autoclaved). The result ing material containsl i t t le " f ree l ime", which is o f ten c la imed to be thecomponent of ordinary concrete most susceptible tochemical attack. l f this was the case. autoclaved cementwould show better chemical resistance than cementcured in the usualway.There appears to be l i t t le direct evidence to show thatfree l ime is any more susceptible to attack than othercomponents of hydrated cement.Like the ground si l ica in autoclaved cement, a pozzolansuch as f ly ash blended with Port land cement and curedin the usual way consumes lhe free l ime l iberated byhydration of the cement (Ref . 1). Though there may besome diff erence. in chemical resistance compared withordinary cement, there have been no eflects reportedcommensurate with the differences in free l ime content.For instance, in sewers subject to H2S attack, addition olpozzolanic materialto the cement has no effect, and if a

' 'FREE LIME' ' AND DURABILITY.REINFORCED CONCRETE, AC AND FRC PIPE

pozzolan is substituted for part of the c6msnl, the rate ofattack is increased because the alkalinity is decreased(Ref. 2.). The absence of free l ime in autoclavedasbestos pipes does nol prevenl them being attacked(Re f . 3 ) .Weake r g rades o f conc re te , f o r examp le w i t hcharacterist ic strength less than 30 MPa, are permeableto water, and in such concretes the presence of free limecould be o l concern because i t is water so luble.However, in the high strength concrete (at least 50 MPa)typical of Austral ian pipe manufacture, water cannotpermeate through the binder (Ref. 4), and the surfacelayer is almost inevitably carbonaled Or converted tosome other water-insoluble form. In theae SilUatenS,there is no reason to expect that the presence orotherwise of f ree lime in the concrete urill haye any elfecton chemical resistance.

In the tests described in the fol lowing seclions, lheres i s tance o f conc re te p ipe i n a range o f ac idenvironments is compared with the resistances ofautoclaved (AC and FRC) pipes.

A f urther test is described in which ordinary concrete pipeis compared with pipe made from blended cement.

3. COMPARATIVE TESTS ON AC, FRCAND CONCRETE PIPE

3.1 Sewer Environment - HzS

Sewers subject to HzS attack are among the most severeenvironments in which concrete or autoclaved asbestoscement pipes may be required to operate. To investigatethe performance ol these pipes and others against HeSattack, a 375mm diameter test l ine was inslal led atWalsh's Offtake, MMBW Sewerage Farm, Werribee, in1966. After operating for over eight years the line wasdismantled to enable a detailed inspection of the pipesto be made. Both the concrete and autoclaved asbestoscemenl samples were heavily attacked. An indication of

Fig. 1 Profi les ofconcrete and autoclaved asbeslos cement pipes after exposure to HzS condit ions for 81/2 years,

ffi ml uE [@ x{* d=

Non-reinf orced Concrete Pioe Autoclaved Asbestos Cement Pioe

HlJlvfE

OlIBB

the extent of the attack suffered by these pipes is givenin F ig. 1 , showing typ ica l prof i les of the remain ingunaf{ected material. The depth of attack, expressed asa percentage of original wall thickness, is greater for theasbestos cement pipe than for the concrele pipe (a0%and 30% respectively).

With concrete pipes, corroded material was lost f rom thesurface progressively as corrosion proceeded, whereasw i th asbes tos cemen t an expanded , f i b rous ma tremained over the surf ace of the sound material.

3.2 Dilute Sulphuric AcidThe corrosion result ing from HzS attack is in fact due tothe conversion of the HzS into sulphuric acid by bacteriaon the wall ol the pipe. To monitor the eflect of sulphuricacid on concrele, autoclaved asbestos cement andcellulose FRC, beam samples were cut from each typeof pipe and placed in a 5% w/w sulphuric acid solution.At varying t imes one beam of each type was removedand broken. An inspection of the cross-section at thebreak of each beam revealed that the penetration of theacid through the uncut sudaces was to al l intents andpurposes the same for each material after the sameexposure per iods (see Fig. 2) .The asbestos cement and cellulose samples retainedtheir corrosion products while (as with HeS attack) a lossof f ines was apparent with the concrete beams.

3.3 Aggressive GroundwaterA corrosion pool simulating aggressive groundwatercondit ions was buil t for the purpose of testing differentpipe materials and protection systems. Acid water(pH3.5) f lows through a layer of gravel in which the pipespecimens are embedded. These specimens includeboth concrete and autoclaved asbestos cement pipefrom which rings have been cut.

Relative to wall thickness, concrete and asbestos cementsamples are showing similar rates of attack (Fig. 3).ln both cases the corrosion products have remainedattached to the pipes, forming a layer through which theaggressive material must f irst dif fuse before furtherattack can occur.Concrete and autoclaved cellulose FRC beams werealso placed in the pool and have now been exposed totl-1e aggressive groundwater conditions for a period of2' lz years. After this period of exposure the cellulosecement exhibits corrosion to a depth of approximately 1.5mm on al l surfaces, which is sl ightly more than thatexhibited by the concrete, where only the f ines have beenaff ected.3.4 Acid Neutral isation

Sawn beams of concrete and autoc laved asbestoscement were placed in two separate containers andimmersed in waler. The pH in each container wasmaintained at approximately 3.5 using hydrochloric acid.A record of the addit ions was kept which al lowed the acidusage per unit area of surf ace to be calculated. Over alwo year period, lhe concrete neutral ised less acid perunit area than the asbestos cement (Fiq. 4).

An inspection of the beams at this stage revealed a sl ightsurf ace etching of the concrete while the asbestos beamsexhibited softening to a depth of lmm.

3.5 Acid lmmersion, pH2

Beams cut from concrete and autoclaved asbestoscement pipes showed depths of attack of 2mm and 3mmrespectively, after being immersed in an acid solutionmaintained at pH2 fo( 4'12 years.

4. CONCRETE MADE WITH ORDINARYAND BLENDED CEMENT

The tes t p ieces were conc re te beams cu t f romnon-reinforced spun pipes. The tests were not set upspecif ical ly to compare blended and ordinary cement butto determine the effect on durabil i ty ( i f any) of reducingthe cement content in spun pipe. ln the lean mixes, eitherf ly ash or si l ica f lour (to 25"/" ol the cement in each case)was added to lhe mix to enable the bore of the pipe to betrowelled to the usual smooth f inish.

The test pieces were immersed in chemical solutions -acid (pH 3.5 and 5.0), carbon dioxide (> 200 ppm),sulphate (2 000 and l0 000 ppm) - for l ive years. Onlythe ac id and carbon d iox ide so lut ions caused anysignif icant reduction in strength, and there were similarreductions with and without f ly ash in the mix.

5. CONCLUSION

Free l ime has no effect of any consequence on thechemical resistance of high strength concrete.Autoclaving and the consequent absence of "free l ime"in asbestos cement or cel lulose FRC pipe does notconf er on them any greater chemical resistance than thalof normal, high quali ty concrete pipe.

REFERENCES

l. "Use of Fly Ash in Concrete", Commitlee Report,AC I Commi t t ee 226 , AC I Ma te r i a l s Jou rna l ,September-October 1987, pp 3Bl - 409.2. "Process Design Manualfor Sulf ide Control in SanitarySewerage Systems", U.S. Environmental ProtectionAgency, October, 1974, p 6 - 14.3. J . Crennan, J . S impson, C. Parker , " ln f luence ofCement Composit ion on the Resistance of AsbestosCemen t Sewer P ipes to HzS Cor ros ion " , Annua lConference, Australasian Corrosion Association, 1978.4. N.L. Harrison, "Propert ies of High Strength Concrete",CIA/lEAust Technical Meeting , "Concrete at i ts Limits",Melbourne, l5th October 1986.

Further detai ls of the tests described in this Bullet in mavbe obtained lrom Humes Concrete.

Hrrr\rEs-+,f.r,ilifffifiil#jitr#"ffiffi

UIBB

Fig.2

- :: : .::. : ri:: ::t'i:::::l:iiit:i.:t:i:::t=:::i:ijit: ?::::;::r;:.]J:i;::;,::iiti:::;:r::it;:i:;rii:ii::t:::::.::

Cross-seclion of concrete, autoclaved asbesloscemenl and cel lulose FRC beams after lmmerslon ln a5o/owlw HzSOc solution for four weeks.

Fig 3. Autoclaved asbestos cement and concrete pipe, 150mmdiameter, exposed to simulated aggressive groundwater(pH3.5) for approximately 12 years.

Asbestos cementDepth of attack (mm) : 3.5-4

Ce l lu lose FRCDepth of attack (mm) : 3.5-4

ConcreleDepth of attack (mm) : 3.5-4

x;i:lX'

HIJ

UIBB

Acid Usage per Unit Area of Surface

Asbeslos Cemnt

Utru

ucr

U

oI

z

u(IF)

1 2

Registercd OfficeHUMES LIMITED (lnc. in Vic.)Humes ConcreteWorld Trade CentreCnr. Flinders & Spencer StreetsMelbourne, Vrc 3005PO Box 31 Melbourne. Vrc.. 3005Telephone (03) 611 3311Telex 38015 Fax (03) 614 6397

Fig. 4 Acid Usage Per Unit Area of Surface

40 50

PERTO0 OF TMMERSTON (wEEKS)

State Offices

New South WalesPark Road, Fegents Park.2143Telephone (02) U4 2351Fax (02) 645 3104Victoria'17 Raglan Street, South Melbourne 3205Telephone (03) 611 3611Fax (03) 696 28BaQueensland9 Euchanan Street, South Errsbane, 4101Telephone (07) 844 5BB1Fax (07) 844 2963

Tasmania1 1 Lampton Ave, Derwent Park, 7009Telephone (m4731422Fax (002)727M8South Australia39-43 f"4axrvell Fd, Pooraka 5095Telephone (08) 349 4544Fax (08) 349 4992Western Australia89 Salvado Road. Wembley 6014Telephone (09) 3BZ 2311Fax (09) 387 8032Northern Territory1606 Rerchardt Rd, Wrnnellre, 5789.Telephone (089) 84 3388

\+",C

Attachment 5

ACI Monograph No. 4, “Durability of Concrete Construction”, American Concrete Institute, 1968.

ACI Monograph No. 4, “Durability of Concrete Construction”, American Concrete Institute, 1968.

Attachment 6

2

GP 400 kg/m3 , GP+FA 330 kg/m3 total cementitious (20% FA), in all tests.

0

2

4

6

8

10

0 1 2 3

Fle

xura

l S

tren

gth

- M

Pa

Series1

Series2

Type GP

GP+FA

Init 1 mnth 6 mnth 1 yr 4 yr 24 yr (Log scale) Time of Immersion

Fig. 1. Strength of Concrete Beams Immersed in Water

0

2

4

6

0 1 2 3

Fle

xura

l S

tren

gth

- M

Pa

Series1

Series2

Type GP

GP+FA

Init 1 mnth 6 mnth 1 yr 4 yr 24 yr (Log scale) Time of Immersion

Fig. 2. Strength of Beams in pH 5

0

2

4

6

8

10

12

0 1 2 3

Fle

xura

l S

tren

gth

- M

Pa

Series1

Series2

Type GP

GP+FA

Init 1 mnth 6 mnth 1 yr 4 yr 24 yr (Log scale) Time of Immersion

Fig. 3. Strength of Beams in 10,000 ppm Sulfate

Attachment 7

CPAA - Technical Report

Protection of steel in concrete and the effect of chloride The mechanism by which steel in concrete is protected from corrosion, even in the presence of moisture and oxygen, and how the presence of chloride can allow corrosion to take place, is explained as follows (4):

Steel embedded in hydrating cement rapidly forms a thin passivity layer of oxide which strongly adheres to the underlying steel and gives it complete protection from reaction with oxygen and water. - Maintenance of passivity is conditional on an adequately high pH of the pore water in contact with the passivating layer. - However, chloride ions destroy the film and, in the presence of water and oxygen, corrosion occurs.

Passivation depends on high pH provided by the cement, and is destroyed by a sufficiently high concentration of chloride. If the chloride level in the water is below the concentration required to destroy the passivity layer, the steel will not corrode. Relationships between pH and salt (sodium chloride) levels to allow corrosion of reinforcing steel were investigated by Shalon and Raphael (5), and to induce stress corrosion cracking in high tensile steel by McGuinn and Griffiths (6). Both test series involved immersing samples in solutions of sodium chloride and either cement extract or calcium hydroxide, the main agent of alkalinity in concrete. The solutions remained in contact with the atmosphere, providing oxygen if the condition was such as to allow any corrosion. For the corrosion series, assessment was by weight loss, allowing each example to be classified as passive or active, and for the cracking series by stability or extension of previously induced cracks. Results are shown in Fig. 1, and while cracking of high tensile steel is not relevant to concrete pipes used in salt water conditions it is of interest that a similar pattern is shown by both sets of data – for low salt levels a dependence on pH but at higher levels, to the maximum investigated, a threshold pH above which there is no effect. For ordinary reinforcing steel this is below the minimum level of pH 12.6 provided in the concrete by saturated calcium hydroxide. Alkali in cement, in addition to the calcium hydroxide, will usually result in pH above 12.6, giving a further margin,

Fig. 1. Effect on steel of alkaline solutions containing sodium chloride

Page 2

Active - Passive Response of Reinforcing Steel

0

2

4

6

8

10

11.0 11.5 12.0 12.5 13.0

pH

% N

aCl

Active Passive

Fracture of High Tensile Steel

0

2

4

6

8

10

12.2 12.4 12.6 12.8 13.0

pH

% N

aCl

Cracking No cracking

CPAA - Technical Report

and passivity is maintained at salt levels well above the typical value for sea water, which contributes chloride equivalent to sodium chloride at about 3%. The effect is further illustrated by photographs of samples of reinforcing steel used for pipes, set up for the purpose of demonstration, in Fig. 2.

Minimum cover to reinforcement In view of these data it is no surprise that concrete pipes underground exposed to levels of chloride as great as the level found in sea water show no deterioration even after very long periods of service. The rate of chloride penetration, given so much emphasis in investigation of above-ground structures, is just about irrelevant – the reinforcing steel in concrete in equilibrium with environments containing moderate levels of chloride, whether from sea water or groundwater, is not subject to corrosion provided only that alkalinity is maintained at reinforcement depth – salty water is no more able than fresh water to corrode the reinforcement. This puts the onus squarely on maintaining alkalinity at reinforcement depth – ie the concrete at this depth is not carbonated; but in terms of required cover this is no greater than would be necessary in a normal environment. Some extra cover for marine exposure may be justified on the basis that the greater the specified cover, the less the risk of reinforcement accidentally being placed at too shallow a depth. However there is no correspondence with high levels of cover required for the most severe open-air environments.

Fig. 2. Samples of reinforcing steel in salt solution REFERENCES D Baweja, H Roper, V Sirivivatnanon, “Specification of Concrete for Marine Environments”, ACI Materials Journal 96, No. 4, July-August 1999, pp 462-470.

“Durability of Concrete Pipe in a Marine Environment”, Concrete Pipe Association of Australasia, 2000. Available on CPAA website www.concpipe.asn.au.

M Bealey, “Durability Considerations – Precast Concrete Pipe”, The Katherine & Bryant Mather International Conference on Concrete Durability, Atlanta, 1-7 May 1987, pp 493-508.

A M Neville, “Properties of Concrete”, 4th edition, Longman, 1995, p 498 & p 563.

R Shalon, M Raphael, “Influence of Sea Water on Corrosion of Reinforcement”, Journal of the American Concrete Institute, June 1959, pp 1251-1268.

K F McGuinn, J R Griffiths, “Stress Corrosion Cracking of Cold Drawn Eutectoid Steel Wire”, Third Tewksbury Symposium, 1974, pp 274-285.

Page 3 November 2007

Attachment 8

exposure to a marine environment and to investigate the concrete for service life prediction.

Various components of the concrete pool were exposed to various types of environments and thus the effect of environment can be studied. There were two exterior concrete walls, which were exposed partially to the ocean. They were the eastern retaining wall and the southern columns-wall system. The eastern wall was facing the ocean and was subjected to the tidal and splash zone. The southern column-wall system was in the atmospheric zone. There were also a series of pedestals supporting two steel pipes from the pump house leading to the ocean. These pedestals were also exposed to the tidal zone. The swimming pool was a reinforced concrete pool lined internally with white tiles. Seawater was pumped into and filtered through a concrete tank adjacent to the pool. The base and side wall were continuously submerged in seawater during the swimming season.

The chloride penetration characteristics of the field concretes were studied and were correlated to the results of the short-term testing carried out in the laboratory. Also the critical durability properties of concrete were studied and were used in service life prediction. 2.0 DETAILS OF THE CONCRETE The first step in evaluating the quality of concrete was to determine the standard for concrete and cement available in 1937, when the pool was constructed. In 1925 the first draft Australian Standard Specification for Portland Cement was published, and it was adopted in 19261 as Australian Standard A.2. The current standard for portland and blended cement is the AS 3972-19972. The main differences between the portland cement specified in AS A.2 and AS 3972 are that the latter allows it to contain up to 5% of mineral additions. The minimum compressive strength at 28 days and maximum SO3 content are also different as shown in Table 1.

Table 1. Details of Portland Cement Specified in AS A.2 and AS 3972

Type of Setting time Soundness

maximum Magnesiamaximum

SO3 content maximu

m

Compressive strengthminimum, MPa

Cement Min. min

Max. h

mm % % 3 days

7 days

28 days

AS A.2-1926 AS A.2-1937

60 -

12 -

- -

4.00 2.00 - -

- 25

20.0 31.5

Type GP AS 3972-97

45 10 5 - 3.5 - 25 40

The first published standard in Australia for concrete structures is the SAA Code for Concrete in Buildings, CA.2-19373. In this code, mix proportions for different parts of structure were left to be specified by the Engineer. For concrete in alkali water or below the ground line of alkali soils, the minimum cement content was seven bags (658 lb.) of portland cement per cubic yard.

To gain more insight into concrete mix proportions, the first British Code of Practice for reinforced concrete in buildings CP114-1948 was considered4. According to this code, for concrete with uncrushed gravel aggregates, the water/cement ratio (by weight) should not exceed the following values for the corresponding mix proportions;

0.43 for 1:1:2 0.51 for 1:1½:3 0.58 for 1:2:4

Without the original structural drawings and specifications of the pool, and the lack of specific mix proportions in the standard at the time, it is not possible to establish the mix proportions of the concretes used. However, the water/cement ratio is likely to be much higher than 0.43 as the aggregates used were angular crushed aggregates. 3.0 DESCRIPTION OF THE POOL Different components of swimming pool were studied during the field evaluation. These included:

♦ interior of the pool itself (base and side walls) ♦ back of the side walls through the inspection gallery ♦ concourses along the length and width of the pool ♦ unreinforced retaining wall facing the ocean ♦ grandstand portals and columns, and ♦ pedestals supporting the seawater inlet pipes.

Interior of the pool – There were two pools, main pool and a shallow pool for children. Both the side-walls and the base of both pools base were fully tiled so the concrete of the pool itself was protected from direct exposure to sea water.

Back of the side walls and the inspection gallery – As the name suggests, an inspection gallery was built along the side and the back (towards sea) wall of the pool. This gallery was used to take the water which overflows from pool back to sea and also used for the inspection.

Concourses along the length and width of the pool – The walking area for the visitors of the swimming pool.

Unreinforced retaining wall facing the ocean – This mass-concrete retaining wall supported the part of swimming pool closest to the sea. This wall was subjected to various levels of exposure. During high tide the base of the wall was exposed to a limited amount direct contact with seawater but would have been constantly exposed to splashing from large waves.

Grandstand portals and columns – These columns support the seating area of visitors and were exposed only to sea spray and air borne salt.

Pedestals supporting the seawater inlet pipes - Pedestals supporting the twin suction pipes which were used to bring water from sea to the pool.

4.0 METHODOLOGY OF THE FIELD STUDY 4.1 Visual Inspection Visual inspection was performed and the mapping of the cracks was carried out. Extensive photographs were taken and were analysed. The lighting in the inspection gallery was not adequate to carry out visual examination. To overcome this problem, photographs were taken with a sensitive film and the photographs were analysed. 4.2 Compressive Strength Several cores of diameter 100 mm were obtained from various components of the pool to evaluate the strength of the concrete and the procedure outlined in AS 1012.14-19915 was followed. All samples were sulfur capped as specified in the standard. The length of each sample was approximately ((2 x diameter) – 10). This was done to ensure that when capped the samples would have the desired height of 2 x diameter. A correction factor was employed to correct the results of samples that were under the specified 2 x diameter. 4.3 Carbonation Depth The presence of carbonation was detected by spraying the surface of the core with a phenolphthalein indicator solution. If carbonation was present the indicator solution would remain clear. This is due to the change in pH of the concrete caused by the reaction of the concrete with carbon dioxide. Concrete with no carbonation present would turn the indicator solution pink due to its basicity. Measurements of carbonation depth were taken on site from the holes where the cores were taken and again later back at the laboratory from the cores themselves. This was done by measuring the width of the carbonated band at a number of positions on the core and then taking the average. 4.4 Measure of Uniformity The uniformity of quality of concrete was determined by the variation in Schmidt hammer readings. Schmidt hammer measurements were taken on site before the sample cores were drilled and on the cores themselves when the samples were brought back to the laboratory. Some of the cores were taken from rendered concrete and therefore any Schmidt hammer measurements that were taken from the surface would reflect the quality of the render and not the concrete. To obtain the correct measurements the mortar on the surface of the cores was cut off and correct measurements could be taken directly from the surface of the concrete. Schmidt hammer measurements were then taken by pressing the hammer against the end of the core perpendicular to its surface. The results of the Schmidt hammer tests were correlated to the results of the compression tests. 4.5 Chloride Penetration Chloride profiles were established to evaluate the resistance of concrete to chloride penetration. To obtain the chloride profile of each sample, the cores were drilled at various depths from the exposed surface. The cores were drilled from the side surface with a 3 mm drill bit. The powder from the first five millimeters of each drilling was discarded to avoid contamination of the powder of the rest of the drilling. The core was then drilled a further 35mm to a total depth of 40mm and the resultant powder obtained was tested for chloride concentration. Chloride concentrations at various depths were used to establish chloride profile.

5.0 VISUAL INSPECTION The overall conditions of the pool and surrounding structures were considered to be exceptionally good given the severity of exposure and age of the structure. Components, which show clear wear and tear, were:

• some columns supporting the grandstand which showed some cracks, • a number of pedestals showing severe erosion, and • Superficial cracks and scaling of the mass concrete retaining wall.

Perhaps the most serious cracks were those found on the end sidewall (towards the ocean) seen from the inspection gallery. The cracks were hardly observed by bare eyes but could only be clearly seen from photos taken from highly sensitive film.

5.1 Pool Structure

The interior of the pool (both the base and sidewalls) was in good condition, showing few signs of deterioration. The deterioration was limited to loose tiles mainly due to exposure to the heat of direct sunlight when pool was empty. No major structural damage was evident. The back of the pool long sidewall (from the gallery side) was in good condition. Some deterioration was evident around the overflow holes and on the concrete below the holes. These areas were in direct contact with seawater. The bulk of the gallery walls was in good condition. The end sidewall showed some spalling. Some large cracks were evident and the wall displayed rust stains and salt deposits.

5.2 Surrounding Structures The underside of the concrete slab forming the concourse was exposed to condensation and was in good condition. The top surface of the concourse, however, showed deterioration in the form of cracks and worn patches. This damage would have mainly been due to the general wear and tear caused by everyday traffic along this walkway.

The columns supporting the grandstand displayed various levels of deterioration. The columns closest to the sea were in poor condition showing extensive cracking whereas the columns furthest from the sea were in quite good condition showing few cracks.

The suction pipe pedestals displayed significant surface deterioration and extensive cracking. The surface deterioration in these pedestals is caused by the eroding force of the waves.

6.0 UNIFORMITY OF CONCRETE AND CARBONATION 6.1 Uniformity of Concrete The uniformity of the quality of the concrete used at Port Kembla Olympic pool was gauged by a set of Schmidt hammer readings given in Table 2. The pedestals were numbered sequentially with number 1 closest to the sea and number 14 was closer to the swimming pool. Measurements were taken after removing the rendering mortar and the relationship between the rebound number and compressive strength was established and is shown in Figure 1. A poor correlation was found with a coefficient of 74%. The compressive strength (F’c) can be determined from the rebound number (R) as given in an equation below:

F’c = 0.69 R + 7.56

Due to the poor correlation between Schmidt hammer reading and compressive strength, it is difficult to find the compressive strength from Schmidt hammer readings. However the variations in Schmidt hammer readings can be used to estimate the variation in compressive strength or the uniformity of quality of concrete.

Table 2 Schmidt Hammer Measurements of Various Concrete Components

Location

Rebound Number

Mean

Standard Deviation

σn-1

Back of end sidewall 46,44,44,45,45,50,48,50,50,50,48,50 48 2.5

Back of long sidewall 46,50,51,52,48,52,24,48,55,52,54,54 51 2.9

Pedestal No.5 44,56,46,39,56,52,38,49,40 47 7.0

Pedestal No.6 46,42,46,44,44,43,46,44,43 44 1.5

Pedestal No.7 38,44,52,43,44,45,42,48,46 45 3.9

Pedestal No.13 37,33,32,32 34 2.4

Pedestal No.14 38,39,38,39 39 0.57

Mass concrete retaining wall Top portion

33,30,40,30,32,34,38,32,39,41,34,40,40,38,36,41,40,40,38,34,42,31,34,38,28,29,32,39,35,36,40,38,26,41,46,41,52,54,44,41,38,38,45,47,46,52,42,52

39 6.4

Mass concrete retaining wall Lower portion

32,32,34,33,31,28,32,30,34,32,34,44,40,39,36,52,41,40,39,40,40,36,34,42,32,32,32,32,30,26,32,40,46,46,39,42,34,36,38,36,34,48,44,44,38,36,38,50

37 5.8

Mass concrete retaining wall, all data

All measurements on retaining wall 38 6.1

0 10 20 30 40 500

10

20

30

40

50

Average Schm idt Ham m er R eading (horizontal)

Com pressive S trength of Cores (M Pa)

R=74%

Figure 1 - Relationship between compressive strength obtained from cores and rebound

hammer number The only concrete surfaces of the pool that were fully exposed were those of the back of the side-walls. The tests were carried out from the inspection gallery. The other

surfaces were fully tiled. The average rebound hammer reading range from 48-51 with low standard deviation between 2.5 to 2.9 (or an equivalent 1.5-2.0 MPa). The quality was clearly very uniform reflecting good workmanship.

The average readings on five pedestals range from 34-47 and a corresponding wide range of standard deviation from 0.57 to 7.0 (equivalent 0.5-5.0 MPa) reflecting lack of uniformity. However, the surfaces of these pedestals were varied in terms of exposed aggregates due to high abrasion. This could have partially contributed to the apparent wide range of variation. Readings were taken to avoid exposed aggregates.

The mass concrete retaining wall showed a wide variation in surface hardness with standard deviation of the population of 6.1. 6.2 Carbonation of Concrete The concrete in the pedestals and retaining wall showed no carbonation. This is consistent with previous findings for concrete exposed to the tidal zone. Constant wetness prevents carbonation.

Columns supporting the grandstands were exposed to atmosphere. A carbonation depth of around 10 mm were observed for both columns closest and furthest from the ocean.

7.0 RESISTANCE OF CONCRETE TO CHLORIDE ION PENETRATION

Figure 2 shows profiles found in the column Nos. 2 and 12 that were exposed to the atmosphere. Column 2 was closest to the ocean and column 12 was furthest from ocean. Figure 3 displays the various chloride profiles obtained from pedestals Nos 13 and 14 as well as the retaining wall all exposed to the tidal zone. Figure 4 shows profiles obtained from the base slab, side and end wall of the pool that are fully submerged during the swimming seasons.

Chloride Concentration (%)

0.000

0.500

1.000

0 40 80 120 160 200Depth from Surface (mm)

C2A

C2B

C12A

C12B

C2B=28, 26.5 MPa

C2A=27 MPaC12A=21 MPa

C12B=39, 32 MPa

Figure 2 - Chloride profiles of cores in atmosphere. C2 is column 2 and C12 is column

12.

Chloride Concentration (%)

0.000

0.800

1.600

0 40 80 120 160 200Depth from Surface (mm)

P13

P14

P14B

R1

R2

R1=31 MPa R2=24, 27 MPa

P13=38, 21.5 MPa

P14=44, 37.5 MPa

Figure 3 - Chloride profiles of cores in tidal zone. P13 and P14 and pedestal 13 and 14.

R1 and R2 are retaining wall next to the beach.

E 1

S 2 = 4 0 ,3 7 .5 M P a

B 1 = 2 4 ,4 2 .0 M P a

0 5 0 1 0 0 1 5 0 2 0 00

0 . 5

1

D e p t h f r o m C o n c r e t e S u r f a c e ( m m )

C h l o r i d e C o n c e n t r a t i o n ( % b y w e ig h t o f c o n c r e t e )

S 2

B 1

E 1

Figure 4 - Chloride profiles of cores in submerged zone during swimming seasons.

Concrete surface protected by tiles. S2 is side-wall, B1 is the base and E1 is the end wall of the swimming pool.

7.1 Influence of exposure conditions

Despite the varied quality of concrete found in various components of the pool, the ingress of chloride ions into the greater depths of concrete is highly dependent on the exposure conditions. This can be observed by comparing the profiles given in Figures 2, 3 and 4. In the 100-150 mm depth zone, exposure to the tidal zone resulted in chloride penetration to a level of 0.5-0.9 % by mass of concrete. This is significantly higher than the levels, ranging from 0.1-0.6% found in concrete columns exposed to the atmosphere (Figure 2) and concrete in the pool which have been submerged in sea water (Figure 4). Similar trends are observed in the 50-100 mm depth region where most reinforcements are placed. The finding supported the relative classification of exposure conditions in AS 3600 of B2 for atmospheric and submerged zones and C for tidal zone. 7.2 Influence of quality of concrete

The resistance of concrete to chloride penetration depends very much on its quality. For each exposure condition, this trend can be observed and confirmed for the 60-year exposure. The quality of the concrete shown in Figures 2-4 is given in terms of

(rebound number, core strength). Within each exposure, concrete with higher rebound number or strength showed better resistance than companion sample of lower rebound number or strength. 8.0 CHLORIDE CONCENTRATIONS AND IMPLICATIONS ON SERVICE LIFE In the columns supporting the grandstand, the concrete cover measured during demolition was 55-60 mm. The chloride at the cover depths for Column Nos. 2 and 12 are 0.4-0.5% and 0.1-0.3% by mass of concrete respectively. Column No. 2 displayed rust stains and cracking along the steel reinforcement where as there was no sign of steel corrosion in Column No. 12. The chloride threshold is therefore between 0.4-0.5 % by mass of concrete.

In the atmospheric zone (columns supporting grandstand), a chloride ion level of 0.4-0.6% by mass of concrete (2.4-3.7% by mass of cement or 9.2-13.9 kg/m3) was found to result in steel corrosion. In the sidewalls which were fully submerged, a chloride level of 0.4 % by mass of concrete (2.4 % by mass of cement or 9.6 kg/m3) at the cover depths did not result in any detectable corrosion. For pedestals in the tidal zone, a chloride level of 0.4 and 0.6 % by mass of concrete (2.4 and 3.7 % by mass of cement or 9.6-14.1 kg/m3) at the cover depths did not result in any detectable corrosion in the two pedestals investigated. However, there were rust stains found in other pedestals. It would appear that corrosion might have initiated in the pedestals but the extent of corrosion varied. The percentage by mass of cement was estimated by assuming a Portland cement content of 390 kg/m3 (CA.2-1937) and the equivalent chloride-ion contents given in kg/m3 were calculated from the measured density.

The Building Research Establishment has proposed a classification6 for assessing the risk of corrosion in terms of acid-soluble chloride contents by mass of cement: low - less than 0.4%; medium - 0.4-1.0%; and high - greater than 1%. A European state-of-the-art report7 concludes that there is no consensus on the permissible levels of chloride but indicates that corrosion will probably occur when the chloride-ion concentration reaches 0.35-1.0% by mass based on the cement content. The Federal Highway Administration (FWHA) recommends8 that when the chloride level of bridge decks reaches 0.3%, for restoration the concrete must be removed to below the rebar level or the deck must be completely replaced. The chloride-ion levels found at the steel in this investigation were well beyond the range of the critical chloride thresholds reported. The inconsistency in the level of chloride support the importance of recognising that there is no unique value for a chloride threshold but a range of values with various degrees of probability.

9.0 CONCLUSIONS

The portland cement concrete used in Port Kembla Olympic pool and its surrounding structures has served its purpose very well for over 60 years. It is an excellent testimony to the potential of concrete as a durable construction material. Various components of the structure were exposed to a range of marine exposures: atmospheric, submerged and tidal zone. The main deterioration mechanisms were chloride-induced corrosion and erosion. Carbonation occurred to a small degree mainly in concrete exposed to the atmospheric zone and hence did not pose any potential durability problem.

The substantially high level and great depth of chloride penetrations into all concrete components were to be expected. They were the result of a long-term exposure of the relatively low strength concretes to sea water. The chloride-ion levels at cover depths were found to be in a range of 0.4-0.6% by mass of concrete (approximately 2.4 and 3.7 % by mass of cement). It far exceeded the chloride contents associated with the risk of steel corrosion.7,9,10 The absence of corrosion activity in the concrete pool and the pedestals was therefore remarkable. Reinforced concrete structures can therefore last well beyond the period of time when the permissible chloride content was first found at the cover depths.

Given the cover depths found in the structure, current codes and standards such as AS 3600 would have required concrete with strength grades in the range of 32 to 50 MPa (depending on specific exposure conditions) for a design life of 40-60 years. The 60-year old concrete was found to vary in strengths from 21.5 to 42.0 MPa. These are significantly below the current standard yet the structure has performed exceptionally well during its service life. The useful service life of many other reinforced concrete structures may have been prematurely terminated.

The ranking of the severity of exposure, with respect to chloride penetration into concrete of similar strength, is tidal, fully submerged and atmospheric zone. This is similar to and confirms our findings at Iluka wharf.9 This partially support the AS 3600 ranking of exposure classes C for tidal and B2 for fully submerged and atmospheric zone respectively. It may be worthwhile to examine the need to distinguish the submerged and atmospheric zone. In any particular exposure, higher strength concrete gave better resistance to chloride penetration.

As a mass concrete structure, the concrete retaining wall has performed excellently in protecting the pool from the force of the ocean. A few cracks and surface blemishes were the only evidences of the deterioration from 60 years exposure to the tidal and upper tidal zone.

10.0 ACKNOWLEDGMENTS

The authors (RPK and VS) are grateful to Mr. Peter Tobin of Wollongong City Council, who brought to their attention the opportunity to learn from Port Kembla Olympic Pool. The investigation is partly funded by the Cement and Concrete Association of Australia.

REFERENCES

1. Standards Association of Australia, ‘Australian Standard Specification for Portland Cement’, A.2-1937.

2. Standards Association of Australia, ‘Australian Standard for Portland and Blended Cements’, AS 3972-1991.

3. Standards Association of Australia, ‘SAA Code forConcrete in Buildings’, CA.2-1937.

4. British Standard Institution, ‘British Standard Code of Practice CP 114 (1948). 5. Standards Association of Australia, ‘Method for Securing and Testing Cores from

Hardened Concrete for Compressive Strength’, AS 1012 Part 14, 1991. 6. Everett, L. M. and Treadaway, K. W. J., ‘Deterioration due to corrosion in

reinforced concrete’, Building Research Establishment Information Paper IP 12/80, 1980.

7. COMITÉ Euro-International du Beton, ‘Durability of concrete structures - state of the art report’, Bulletin D’Information No. 148, Paris, 1982.

8. Locke, C. E., ‘Corrosion of steel in Portland cement concrete: fundamental studies. Corrosion effects of stray currents and techniques for evaluating corrosion of rebars in concrete’, ASTM STP 906, ASTM, Philadelphia, 1985, 7-8.

9. Driscoll, S., Sirivivatnanon, V. and Khatri, R.P., ‘Performance of a 25-year-old Coastal Concrete Wharf Structure’, Proceedings of the AUSTROADS 1997 Bridge Conference, Sydney, Australia, December 1997.

10. Everett, L. M. and Treadaway, K. W. J., ‘Deterioration due to corrosion in reinforced concrete’, Building Research Establishment Information Paper IP 12/80, 1980.

Attachment 9

Attachment 10

Pages following Page 10 of this report are omitted to reduce the size of file.

INDEX Summary

1

Introduction

1

Experimental Method

2

Results Coatings Plastic Film Wrap Uncoated Concrete Pipe Autoclaved Asbestos Cement

2 2 3 3 4

Acknowledgement

4

Keywords

4

References

4

Appendix 1 - Test Details Solution Samples Table A1.1 Backfills Set-up and Operation Inspection

5 5 5 6 6 7 7

Appendix 2 - Performance Coatings - Pipes Table A2.1 Coatings - Beams Table A2.2 Uncoated Concrete Pipes Table A2.3

8 8 8 9 9

10 10

Photo Pages 1-9

1

SUMMARY An experiment simulating pipe exposed on the outside in highly acidic acid groundwater has been completed after a test period of more than 22 years. Paint coatings of appropriate formulation were found to be effective in preventing attack on concrete pipe. Wrapping with plastic sheet was only partially effective. For concrete pipes without coating or wrapping, backfills designed to inhibit movement of water at the pipe surface reduced the rate of attack. Even with heavy attack on the exterior of pipes, there was negligible effect on rubber ring joints. The tests provide further confirmation that AS 4058 advisory limits of pH for concrete pipe with standard cover are appropriate, though somewhat conservative. Autoclaved asbestos cement demonstrated similar durability to concrete. INTRODUCTION Some ground conditions are potentially aggressive to concrete pipe. It is desirable that such conditions can be readily identified and, where appropriate, protective measures adopted to ensure that the required service life is obtained. Humes commenced experimental investigations in the early 1970s to supplement information which could be obtained from literature and from field examples, concerning effects of groundwater on concrete pipe. At an intermediate stage of the experiments, data from all sources were summarised (Ref. 1). This review contributed information for a table of recommended limits for concentrations of aggressives, applicable to concrete pipe made to the Australian Standard, first published in the Concrete Pipe Association of Australasia bulletin “Designing Permanent Pipelines” and subsequently as an advisory appendix in AS 4058. The experimental investigations included tests on concrete pipe in high levels of sulfate, and a simulation of acidic ground conditions. The acid groundwater simulation was set up to determine not only the rate of attack on unprotected pipe but also effectiveness of protective treatments, influence of different backfills, and the effect of the acid environment on rubber ring joints. Samples of autoclaved asbestos cement pipe were included, to enable the rate of attack to be compared with that of concrete. The acid groundwater test was concluded in November 1997, having run for over 22 years. At an interim stage it provided data for a reassessment of the published limits for unprotected pipe as above, which confirmed that the limits are appropriate though somewhat conservative (Ref. 2). Results obtained progressively from the acid groundwater test are described in reports RC.6240 (1976), RC.7695 (1979), RC.8537 (1980), RC.0355 (1983), RC.0916 (1984) and RC1547 (1988). Final results are recorded in this report.

2

EXPERIMENTAL METHOD The majority of samples under test consisted of short lengths of DN 150 spun concrete pipe, assembled in pairs (one length with a spigot, the other with a socket) to incorporate the standard rolling rubber ring joint. The samples stood vertically in a large, shallow tank, backfilled with coarse river gravel (no sand component) with the spigot end downward, and the lower end of the assembly blanked off with PVC sheet. The liquid environment consisted of dilute hydrochloric acid buffered with sodium acetate, with the acid component adjusted to give a pH of 3.5, circulated continuously by an electric pump via an array of pipes in the base of the tank. The arrangement of samples, backfill and solution in the tank are shown in Photos 1, and further details are set out in Appendix 1. To test various protective treatments, some samples had applied paint coatings or were wrapped with polyethylene film. Some of the pipe samples had deliberate damage to the coating, to determine whether this would lead to accelerated failure as with adhered coatings exposed to H2S conditions in sewers. In order to simulate different backfill environments, some uncoated samples were isolated from the gravel backfill and surrounded instead by a test material (clay, soil or lime sand). Separation from the gravel was maintained by PVC rings, the tops of which can be seen in the photographs. At the lower end, the composite of sample and backfill was exposed to the circulating solution in the same way as the samples embedded directly in the gravel. Many of the samples were exposed for the full term of the test. Others were introduced or removed at intermediate stages. To provide a ready test of new coatings, provision was made to test small concrete prisms (beams) cut from spun pipes, coated with the test materials. These samples were immersed freely in the solution, the space being separated from the gravel by PVC rings as used for the special backfills. RESULTS Coatings See Appendix 2, Tables A2.1 & A2.2, and Photos 2 & 3. Satisfactory performance has been obtained from some but not all of the epoxy coatings tested. Successful coatings include a water-based formulation. The test series included many coatings containing coal tar, a material which is no longer acceptable on account of the risk to operator health. Coal tar epoxies did not exhibit better performance than other types. In the samples whose coatings were deliberately damaged before exposure to the solution, there was no evidence of progressive failure of the coating from the damaged areas.

3

Plastic Film Wrap Two samples of the initial set (uncoated pipes including joints) were wrapped in polyethylene film (0.25 mm thick), and installed in the gravel backfill similarly to the samples with paint coatings. The wrappings covered the concrete surfaces but did not exclude the solution altogether. After 8 years of exposure, areas of the concrete surfaces had significant corrosion, and while the treatment had reduced the rate compared with unprotected samples, it was not nearly as effective as the paint coatings. The test was not continued beyond 8 years. Uncoated Concrete Pipe See Table A2.3 & Photos 4-7. In gravel backfill, the concrete was corroded to 7 mm depth at the end of the test period (22 years 8 months). Clay, black soil and lime sand backfills all reduced the depth of attack over the same period. For all samples, corrosion in the region of the rubber ring seal was negligible and would not have affected joint performance. In any given situation the depth of corrosion is not proportional to time, as adhering corrosion products inhibit access of the aggressive medium to the unaffected concrete (Ref. 3). The net effect approximates to a cubic relationship between depth and time (Ref. 4) - ie for a given pH and set of ground conditions, time for the attack to reach a particular depth is proportional to (depth)3. This relationship makes the effect of (for example) clay backfill, as determined by the experiment, more significant than it would otherwise seem. A representative figure for the depth of attack of pipe in the clay backfill would be 4 mm, compared with 7 mm for the gravel. To reach 7 mm depth in clay backfill, the exposure period would have to be increased by a factor (7/4)3 = 5.4, corresponding to a total period of 120 years. Conversely, concrete in clay backfill, compared with gravel backfill, could tolerate a lower level of pH (ie stronger acid concentration) for the same depth of attack after a given time. pH is a logarithmic scale of acid concentration, a change of one unit of pH corresponding to a tenfold change in the concentration of hydrogen ion. Reduction of pH by one unit corresponds to a tenfold increase in hydrogen ion concentration, while reduction by 0.5 units corresponds approximately to a threefold increase. The change in pH which would be expected to offset the slower rate of attack in clay backfill (as above) is log (5.4) = 0.7. This figure is of interest in relation to the advisory table in AS 4058, Appendix E for different ground conditions, where half a unit decrease of pH is allowed for the transition sandy/flowing to medium, and a further half unit to clay/stagnant. All things considered, the present experiment would support allowances of this order. If the depth of attack of 7 mm reached in this experiment is considered a reasonable limit for a pipe in service, and the pipe is required to perform for 100 years, we would require a rate of attack lower than that experienced in the test (of duration 22 years) by a factor of (approximately) 100/22 = 4.5. Assuming proportionality to acid concentration as above, this would correspond to an environment with pH 0.7 units higher than in the experiment - say between 4.0 and 4.5. AS 4058 advises a pH limit of 5.0 for medium ground conditions, which in the light of these results remains a conservative figure.

4

Autoclaved Asbestos Cement Two assemblies each including a collar joint with a rubber ring seal were installed in gravel backfill in the same way as the concrete pipes. The material was attacked to 4 mm depth after 22 years 8 months (compared with 7 mm for concrete). Considering the greater wall thickness of concrete this represents a similar rate of attack proportionally. The result is consistent with a wider range of tests reported in Ref. 5, demonstrating equivalent durability for the two materials. ACKNOWLEDGEMENT The test series described in this report was set up at the Humes R&D Laboratory at Westall by Mr B Barnhill. Interim reports bear the signatures of Mr L Fudge, Mr R Haupt, Mr W Mutsaers and Mr R Wix. Special acknowledgement is due to Mr W Mutsaers who maintained the test from the early 1980’s until his departure from Humes in 1995. REFERENCES 1. RC.6905, 24/11/77, “External Corrosion & Protection of Buried Concrete”. 2. RC.1525, 6/6/88, “Concrete Pipe Exposed to Acid Ground Conditions or Aggressive

Carbon Dioxide”. 3. A M Neville, “Properties of Concrete”, 4th edition, Longman, 1995. P 506. 4. RC.8212, 22/2/80, “Extension of the Life of Concrete Products by Increased Cover to

Reinforcement”.

5. Humes Technical Bulletin 01/88, “Free Lime and Durability - Reinforced Concrete, AC and FRC Pipe” (May 1988).

12/2/99

5

APPENDIX 1 - TEST DETAILS Solution The solution consisted of hydrochloric acid buffered with sodium acetate. The solution was made up with a concentration of sodium acetate of 0.2 moles per litre and HCl was added to maintain a pH of 3.5. The initial requirement of HCl is approximately 0.2 moles per litre. Note: This solution in effect provides two acids to act on the concrete - hydrochloric and acetic. The calcium salts of both these acids are highly soluble in water, and so cannot provide a layer to inhibit corrosion as might occur with acids producing insoluble calcium salts. pH was monitored & corrected on a regular basis effectively from March 1975 to August 1995. The solution was renewed in 1976, 1979, 1980, 1983, 1984, and 1988, following inspections at those times. No corrections were made from August 1995 and final samples taken at December 1996 had pH 4.5. Samples All concrete samples were spun concrete. DN 150 pipes were obtained from Humes factory stock (Melbourne). While details are not recorded, the following would have been typical of production at the time:

Cement Type A (Type GP) at about 400 kg/m3 Aggregate Basalt coarse aggregate, silica sand W/C < 0.4 (after spinning) No admixture or SCM. Curing - short steam cycle then air.

For the pipe samples, preparation of the concrete surface consisted only of wiping with a cloth to remove dust. Coatings were applied according to manufacturers’ instructions, and allowed a curing period of 7 days before the joints were assembled. Coating details are per following table.

6

Table A1.1

Product Manufacturer Type Thickness (mm)*

Beckothane Bildcote Carbomastic 14 DP80/10 & GY250 Epirez 304 Hydrepoxy 938 K91 Permatar HB200 Polytar Thickchlor Vaporlock

A C Hatrick Dulux Vessey Ciba Geigy Epirez Hardman Australia Ciba Geigy Pioneer (Wattyl) Hardman Australia Dulux Nonporite

Polyurethane Filled epoxy & solvent Coal tar epoxy Coal tar epoxy Coal tar epoxy Water based epoxy Water based epoxy Coal tar epoxy Tar polyurethane Chlorinated rubber Coal tar epoxy

0.05

0.13

NR

NR

0.13

NR

NR

NR

NR

<0.13

0.13

* Figure shown applies to pipe samples only. NR Not recorded Autoclaved asbestos cement samples were obtained from James Hardie & Co. Backfills Most of the tank was filled with gravel consisting of pebbles 15-25 mm diameter, and no fines. The material was determined by geological examination to be quartz or quartzite and was visibly unaffected by the test solution. Material for the clay backfill was obtained from Mt Waverley, Melbourne. The black soil was dug from the R&D laboratory site at Westall. The lime sand was packing sand obtained from David Mitchell Estates, with a stated limestone content of 70%.

7

Set-up and Operation The pool was constructed of concrete lined with plasticised PVC (Humes Plastiline), with internal dimensions 3.0 m diameter and 0.75 m depth. The pipes in the base were UPVC and were covered by broken clay tile (initially) or filter fabric (finally) to keep the dispersion area free of the gravel or other backfill material, and the solution level was maintained above the upper surface of the gravel (about 70 mm) to allow free flow back to the pump. A conical tent of PVC fabric was installed over the tank to minimise evaporation and protect against contamination. Flow through the pump was at the rate of approximately 0.5 litres per second. Taking account of the volume occupied by gravel, samples etc, the effective cross-section of flow was about 3 m2 , giving a mean linear flow velocity (upwards) of approximately 0.6 metres per hour. The surrounds for special backfills consisted of rings of PVC formed from flat sheet without welding, which would have allowed liquid interchange with the circulating solution above the backfill level as well as at the lower surface. Inspection For the final inspection, loose material was removed from samples using fresh water and a soft brush. Samples were sectioned with a rock saw to show profiles of affected and unaffected material. Nominal OD for the DN 150 pipes is 196.8 mm. In the photos showing sections across a pipe diameter (semicircular profile), the spacing between the marks on the support under the sample is equal to this distance.

8

APPENDIX 2 - PERFORMANCE

All descriptions refer to those parts of samples fully surrounded by the test environment (below the surface level of the solution & surrounded by the designated backfill).

Coatings - Pipes Sample A - coating intact throughout at start of test. Sample B - chipped to simulate field damage. Samples not designated A or B had no deliberate damage. Gravel backfill. In no case did the simulated field damage lead to accelerated deterioration of the sample. Table A2.1

Number / Type Period of Exposure

Description Rating*

1A & 1B - Beckothane (polyurethane) 2A & 2B - Epirez 304 (coal tar epoxy) 3A & 3B - Vaporlock (coal tar epoxy) 5A & 5B - Thickchlor (chlorinated rubber) 7A & 7B - Bildcote (epoxy with solvent) Hydrepoxy 938 (water based epoxy). Permatar HB200 (coal tar epoxy) DP80/10 & GY250 (coal tar epoxy)

22y 8m

22y 8m

22y 8m

22y 8m

22y 8m

19y 3m

19y 3m

19y 3m

Some failure of coating socket / socket slope Coating flaking from spigot on 2B Large areas of general failure. Small breaks in coating. Some damage to outside corners of sockets; otherwise coating intact. See Photos 2. No visible deterioration. See Photos 3. Coating intact but many blisters (unbroken) 8-10 mm dia. No visible deterioration

F

P

P

F

G

G

F

G

* G - Good - coating intact. F - Fair - severe blistering or local failure of coating. P - Poor - areas of general failure, concrete surface exposed.

9

Note - further coating types had been included at early stages - Dulux Weathershield (acrylic), Vessey Vepox CC33 (coal tar epoxy), and Nonporite Furox, but all failed within two years. Coatings - Beams Cut longitudinally from spun pipe, wall thickness 50 mm or 25 mm. Length 300 mm, width 50 mm. 50x25 mm section is indicated in the table by an asterisk (*). These samples present three different surface conditions - outside of pipe, inside of pipe, and saw cut. Unless indicated otherwise, descriptions refer to all surfaces. Overall, there was no detectable effect from the type of surface on the degree of adhesion achieved. “Blisters” are surface swellings with no break in the coating, indicating separation of the coating from the concrete. Table A2.2

Coating

Period of Exposure

Description

Carbomastic 14 (coal tar epoxy) DP80/10 & GY250 (coal tar epoxy) Polytar (tar urethane) * Hydrepoxy 938 Grey (water based epoxy) K91 (water based epoxy) Hydrepoxy 938 White (water based epoxy) * Permatar HB200

15y 3m

16y 11m

15y 3m

19y 7m

16y 11m

16y 11m

16y 11m

Coating intact. Very few small blisters. Coating intact but many small blisters, especially at edges of beams. Coating intact but large blisters 8-10mm diameter. Some failure at edges. Coating in good condition. Large blisters, but coating intact. Small blisters mainly inside surface. Two breaks in coating. Coating intact but blisters on all surfaces.

10

Uncoated Concrete Pipes Table A2.3 All 22y 8m exposure

Number Backfill Depth of Attack (mm)

Photos Nos.

12A 12B 9A 9B

10A 10B

Gravel Gravel Clay Clay

Black soil Lime sand

6 7 5

2-3 2 4

4 & 6 6

5 & 7 7