Discussion A discussion of the paper ‘‘Characteristics of pastes from a Portland cement containing different amounts of natural pozzolan’’ by Adnan C ßolak $ Edgardo F. Irassar * Departamento de Ingenierı ´a Civil, Universidad Nacional del Centro de la Pcia de Buenos Aires, B7400JWI Olavarrı ´a, Argentina Received 10 October 2003 The paper by C ß olak [1] examines the influence of natural pozzolan on the physical, mechanical and durability prop- erties of blended Portland cement pastes. I like to celebrate the study on the use of natural pozzolans that are the more extended source of mineral admixtures in the developed countries. I agree with some parts of the paper, but I have some commentaries that I would like to point out. From the experimental procedure, it is understood that the author has used paste specimens to determine the compressive strength on cement, although it is not a usual practice in cement standard around the world (i.e., ASTM C 109, EN 196-1). In addition, the author has tested the specimens in wet and in dry conditions. To determine the compressive strength, the specimens should be saturated in water according to standard. The author reports that addition of natural pozzolan produces an increase on the total porous volume, but he should has taken account that the addition of natural pozzolan produces a shift of large capillary pores by numerous fine pore, called pore size refinement by Mehta and Monteiro [2]. However, my main disagreement with the author is about the sulfate performance of blended cement. The author describes that ‘‘Resistance to chemical attack caused by Na 2 SO 4 decrease considerably when increasing the natural pozzolan content in the blend. Increasing the Portland Cement replacement by natural pozzolan from 20% and 40% leads to decrease the time of failure from 14 to 7 days’’. In the conclusion, the author attributes this worse performance of blended cements to the ettringite formation. The author remarks the importance of C 3 A and C 4 AF on the sulfate resistance of Portland cement, assuming that some hydrated compounds of cement paste react with sulfate ions and cause expansion when they are exposed to sulfate environments. The Portland cement used in this research have 9.1% of C 3 A and 9.4% of C 4 AF, as calculated by Bogue. According to the results presented in the paper, plain cement paste specimens, with or without superplasti- cizer, performed in good condition during the sulfate test. However, blended cements containing a natural pozzolan (S + A + F = 80.6%—Class N, according ASTM C 618) performed worse. Commonly, the causes of worse performance of concrete containing natural pozzolan in sulfate environment are attributed both to the increase of water demand or the presence of reactive alumina in its composition. The results show that the incorporation of natural pozzolan used in this research causes an increase of water demand from 34% to 38.3%, and the use of superplasticizer allows to reduce the water/cementitious ratio for a 60:40 mixture. Then, the worse performance could not be attributed to the increase of water demand. Finally, the author suggests that it could be attributed to the ettringite formation and the high alumina provided by the cement or natural pozzolan into the pore solution. If the cement has 9.1% of C 3 A content and the natural pozzolan has an Al 2 O 3 content of 2.82% (see Table 1 of Ref. [1]), I would like to know, who provides more alumina during its hydration to the pore solution? The dilution effect on Al 2 O 3 content produced by the addition of natural pozzolan cause a decrease of alumina bearing phases in the blended cement. I like to point out that the suggestion proposed by the author is wrong. I believe that the worse performance of blended cement may be attributed to the salt crystallization phenomenon [3,4] or to salt hydration distress phenomenon as called by Hime et al. [5]. In contrary to chemical sulfate attack, which manifest itself by the formation the gypsum or/and ettrin- gite, this type of attack is strictly a physical nature caused by 0008-8846/$ – see front matter D 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.cemconres.2003.12.001 $ Cem. Concr. Res. 33 (4) (2003) 585– 593. * Tel.: +54-2284-451055; fax: +54-2284-450628. E-mail address: firassar@fio.unicen.edu.ar (E.F. Irassar). Cement and Concrete Research 34 (2004) 1265 – 1266