Semisequicentennial Transportation Conference Proceedings
May 1996, Iowa State University, Ames, Iowa

Concrete Deterioration by Deicing Salts: An Experimental Study

Robert D. Cody, Anita M. Cody, Paul G. Spry, and Guo-Liang Gan

Department of Geological and Atmospheric Sciences,
253 Science I,
Iowa State University,
Ames, Iowa 50011-3210.

Concrete with dolomite coarse aggregate was obtained by coring existing Iowa highways. The concretes were of two types, those which were very durable under highway conditions and those which were low durability. Samples were experimentally deteriorated using wet/dry, freeze/thaw, and continuous soak conditions in solutions of magnesium chloride, calcium chloride, sodium chloride, magnesium acetate, magnesium nitrate, and distilled water in order to determine relative deterioration activities. Magnesium chloride was most destructive. Calcium chloride was next, and sodium chloride was relatively benign. Magnesium acetate produced severe crumbling and moderate fracturing, and magnesium nitrate caused moderately severe deterioration by crumbling and discoloration. Low durability concrete was somewhat more affected by distilled water freeze/thaw conditions than more durable material, but generally both types were severely damaged by magnesium and calcium salts. These results suggest that magnesium and calcium deicers may accelerate highway concrete deterioration. Key words: chloride deicers, concrete deterioration, CMA deicer, alkali-dolomite reaction.

Rock salt deicers may accelerate highway concrete deterioration if the concrete has previously developed alkali-dolomite reactions between cement and dolomite coarse aggregate (1–6) . Field observations substantiate salt-induced deterioration by finding that deterioration is more rapid for highways with heavy deicer applications compared to those with lesser applications (7). Field studies of existing highways provide crucial observations, but there are significant problems in relying solely on field observations to evaluate deicer-induced accelerated deterioration. Only rock salt has been used for long-term deicing so that correlations among deicer composition, deterioration modes, and rates of deterioration are ambiguous. In addition, field conditions are so varied in terms of highway use by trucks and cars, temperature variations, precipitation rates and seasonal patterns, and concrete composition including type of coarse aggregate, for example, that definitive conclusions about deicer-induced deterioration are difficult to determine.

In view of the problems associated with field studies only, the present study (1,3) was designed to explore concrete deterioration by different deicers and the effects of dynamic environmental changes such as cycles of alternate wetting/drying and freezing/thawing on deterioration. Only a few detailed experimental studies have recently investigated the effects of different deicers on concrete (8–11), and none of these has provided a comprehensive overview of the relative magnitude of deterioration caused by magnesium, calcium, and sodium deicers.


Experimental Methods

Well aged concrete cores from highways of known service records were used in the experiments. Only concretes containing dolomite coarse aggregates were used in order to limit the potentially complex effects of coarse aggregate composition. The concretes were of two types, those which were very durable (service life of > 40 years) and those of much lower durability (service life < 16 years). Nondurable concretes contained dolomite coarse aggregate from Iowa Devonian rocks, either from the Cedar Valley Formation or from the Wapsipinicon Formation. Durable concretes contained dolomite aggregate from the Silurian Hopkinton Formation or the Ordovician Galena Formation (2) (Table 1).

TABLE 1 Summary of Concrete Experiments
TYPE NaCl MgCl2 CaCl2 H2O
9 CYC.; 3M
4 CYC.; 3M
NC Paste: brown color and crumbing Aggr.: NC Paste: cracking at paste/quartz sand interface. Aggr.: NC NC
6 CYC.; 3M
NC Paste: brown color and severe crumbling Aggr.: NC Paste: cracking at paste/aggregate interface Aggr.: NC NC
222 DAYS; 3M.
Paste: slight brown color Aggr: NC Paste: dark brown color; cracking at paste/aggre. interface Aggr.: NC Paste: brown color; cracking Aggr.: NC NC
16 CYC.;0.75M.
NC Paste: brown rim in paste around aggreg. Aggr.: cracking Paste: decomposition at paste/aggregate interface. Some cracks. Aggr.: cracking NC
16 CYC.;0.75M.
Paste: brown color, severe crumble Aggr.: NC Paste: v. dark brown, corners crumble Aggr.: NC Paste: corners crumble, some cracks. Aggr. NC NC

*NC= no change.

Each core was sliced into small rectangular blocks ( 12 mm x 12 mm x 25 mm in size) and experimentally deteriorated using wet/dry, freeze/thaw, and continuous soak conditions in 3M and 0.75M single salt solutions of magnesium chloride, calcium chloride, and sodium chloride. Experiments with one durable and one nondurable sample were performed using distilled water as a control. A few experiments used magnesium acetate and magnesium nitrate rather than magnesium chloride to determine if deicer anion composition is significant in affecting damage. All experiments contained 0.01% sodium azide to control mold and bacterial growth. The blocks were immersed in 100 ml of solution, sealed in cleaned polymethylpentene containers, and stored in a constant temperature chamber at 60oC. Continuous soaking experiments were for 222 days, and cycle times were 132 hours. Drying was at either 60oC or 90oC, and freezing was at either 0oC or -70oC. Experiments were continued until at least one sample in a set began to show significant deterioration, at which point the entire set of experiments was terminated. Continuation would have caused extreme crumbling or cracking and would have increased difficulties in making sections for study.

Evaluation of Deterioration

Concrete deterioration was visually evaluated using a binocular microscope. Deterioration was described as: (a) crumbling, or loss of cohesion between paste and aggregate and/or softening of the paste, (b) fracturing, and (c) brownish discoloration. Fracturing was subdivided into fractures formed at the interface between coarse aggregate and paste (rim fractures), fractures with essentially random orientation in the paste, and fractures occurring within coarse aggregate particles. Brown discoloration of paste and/or coarse aggregate was usually accompanied by severe crumbling.


The degree of concrete deterioration depended on deicer cation and on experimental conditions (Table 1). The relative magnitude of deterioration under different conditions and deicer composition can be best visualized by comparing the duration of each set of experiments. The most severe deterioration occurred in wet/dry experiments with 3M solutions and 90oC drying. Only four cycles (26 days) were required for observable visible damage. The least damaging conditions were wet/dry cycling with 0.75M solutions and 60oC drying, and freeze/thaw cycling with 0.75M solutions and 0oC freezing. Each required 16 cycles (104 days) to produce significant deterioration.

Magnesium chloride was the most destructive salt with severe deterioration produced under almost all of the experimental conditions (Figure 1). Deterioration was by discoloration, random/rim fracturing, and severe crumbling. Calcium chloride was the next most destructive salt, with deterioration by very severe random/rim fracturing and crumbling. Calcium chloride also produced highly visible dark-colored reaction rims at the margins of coarse aggregates in nondurable concretes under both wet/dry and freeze/thaw conditions; newly-formed aggregate rims were much less abundant in durable concretes. Sodium chloride was relatively benign except under 0oC freeze/thaw cycling in 0.75M NaCl, and 90oC wet/dry cycling in 3M NaCl. Magnesium acetate produced severe crumbling and moderate fracturing, and magnesium nitrate caused moderately severe deterioration by crumbling and discoloration. Distilled water produced little observable deterioration under any environmental conditions. The only experiments that differentiated the two durable and nondurable concretes were those with 3M calcium chloride with -70oC freeze/thaw; 0.75M rock salt with 0oC freeze/thaw; and distilled water with 0oC freeze/thaw. In these three sets of experiments durable concrete was slightly less damaged than nondurable concrete.


The major cause of deterioration by magnesium chloride was the formation of noncohesive magnesium silicate hydrates, MSH (10), produced by the reaction of dissolved magnesium with calcium silicate hydrates of the cement. Evidence for MSH formation is the high degree of correlation between silicon and magnesium after treatments with magnesium chloride. Replacement of calcium by magnesium creates mobile calcium which precipitated as calcite (CaCO3) and/or portlandite (Ca(OH)2) in the cement paste and aggregate pores. The newly-formed materials may exert crystal growth pressures and initiate micro-fracturing. Brucite (Mg(OH)2) crystallization in the alkaline concrete environment probably also contributed to damage by exerting crystal growth pressures on the surrounding concrete. Causes of deterioration by calcium chloride is more difficult to determine from our experiments. Reactions lead ing to development of calcium chloroaluminate were probably most significant (11). Further damage is caused by concentrated CaCl2 solutions which dissolve Ca(OH)2 from concrete, thus leaving a more porous structure. These pores may be filled with new materials produced by reactions involving CaCl2 solutions, cement phase Ca(OH)2, and CO2 derived from the atmosphere or from dissolution of limestone/dolomite aggregates (12). Growth of the newly precipitated salts may produce crystal growth pressures analogous to those produced by brucite precipitation in magnesium chloride experiments.


In laboratory experiment conditions, magnesium chloride, magnesium acetate, magnesium nitrate, and calcium chloride are much more damaging to concrete under several different environmental conditions than rock salt. Anion composition was much less important than cation composition in affecting deterioration. Deicer concentration was found to be significant. These results show that magnesium and calcium deicers may be deleterious to concrete highways, and that the rate of deterioration will depend on the amount and frequency of applications.

There is considerable interest in new deicers, especially magnesium chloride because of its anti-icing properties and its effectiveness at lower temperatures than rock salt, and calcium magnesium acetate (CMA) because it reduces steel corrosion and groundwater chloride contamination. Our experiments document that the substitution of magnesium and/or calcium deicers for rock salt may have unintended consequences in accelerating concrete deterioration. Long-term, carefully controlled field experiments with magnesium and calcium deicers are essential in order to fully determine the effects of long-term use of these deicers under highway conditions and to determine if they are suitable substitutes for rock salt.


This project was chiefly funded by the Iowa Department of Transportation, Project No. HR-355. We especially want to thank Messers. Vernon Marks, Jim Myers, and Wendell Dubberke, as well as Dr. Wallace Rippie for their advice, suggestions, and support of this project. Partial support of Guo-Liang Gan was provided by the Iowa State Mining and Mineral Resources Research Institute.


  1. G.-L. Gan, P.G. Spry, R.D. Cody, and A.M. Cody. Rim Formation on Iowa Highway Concrete Dolomite Aggregate: The Effects of Dedolomitization Reactions. Environmental and Engineering Geoscience, 1996, in press.
  2. P.G. Spry, G.-L. Gan, R.D. Cody, and A.M. Cody. The Formation of Rims on Dolomite Aggregate in Iowa Highway Concrete. Proceedings of the Semisesquicentennial Transportation Conference Celebrating the 75th Anniversary of the Transportation Research Board (this volume), 1996.
  3. R.D. Cody, P.G. Spry, A.M. Cody, and G.-L. Gan. The Role of Magnesium in Concrete Deterioration. Iowa Department of Transportation Final Report, HR-355, 1994.
  4. M.-S. Tang, Z. Liu, and S.-F. Han. Mechanism of Alkali-Carbonate Reaction. Proceedings of the 7th International Conference on the Concrete Alkali-Aggregate Reaction, Noyes, New Jersey, 1994, pp. 275–279.
  5. E.G. Swenson and J.E. Gillot. Alkali-Carbonate Rock Reaction. Highway Research Board Record, Vol. 45, 1964, pp. 21–45.
  6. E.G. Swenson and J.E. Gillot. Alkali Reactivity of Dolomitic Limestone Aggregate. Magazine of Concrete Research, Vol. 19, 1967, pp. 95–104.
  7. W. Dubberke and V.J. Marks. The Effects of Deicing Salt on Aggregate Durability. Transportation Research Record, Vol. 1031, 1985, pp. 27–34.
  8. W.H. Kuenning. Resistance of Portland Cement Mortar to Chemical Attack—a Progress Report. Highway Research Record, Vol. 113, 1966, pp. 43–87.
  9. A. Nedezdin, D.A. Mason, B. Malric, D.F. Lawless, and J.P. Dedosoff. The Effect of Deicing Chemicals on Reinforced Concrete. Transportation Research Record, Vol. 1157, 1988, pp. 31–37.
  10. M. Cohen and A. Bentur. Durability of Portland Cement-Silica Fume Pastes in Magnesium Sulfate and Sodium Sulfate Solutions. ACI Materials Journal, May–June, 1988, pp. 148–157.
  11. A.M. Neville. Behavior of Concrete in Saturated and Weak Solutions of Magnesium Sulfate or Calcium Chloride. Journal of Materials, ASTM, Vol. 4, 1969, pp. 781–816.
  12. S. Chatterji. Mechanism of the CaCl2 Attack on Portland Cement Concrete. Cement and Concrete Research, vol. 8, 1978, pp. 461–468.

CTRE is an Iowa State University center, administered by the Institute for Transportation.

Address: 2711 S. Loop Drive, Suite 4700, Ames, IA 50010-8664

Phone: 515-294-8103
FAX: 515-294-0467


Iowa State University--Becoming the Best