Concrete as a Three-Phase System: Complete Guide - Hydration Chemistry, HCP, ITZ, Porosity & Strength

Concrete is a three-phase composite material consisting of: (1) the aggregate phase (coarse and fine aggregate, typically 60 to 75% of total volume), which forms the rigid skeletal framework; (2) the hydrated cement paste (HCP) phase (25 to 35% of volume), the binding matrix produced by cement hydration that fills spaces between aggregate particles; and (3) the interfacial transition zone (ITZ), a thin (10 to 50 micrometre) shell of modified microstructure surrounding each aggregate particle. Despite its thinness, the ITZ can account for 20 to 40% of the total paste volume in typical concrete because aggregate particles have enormous collective surface area. The ITZ has higher porosity (25 to 50% locally vs 10 to 25% for bulk HCP) and is consistently the weakest zone in normal concrete, controlling its failure mechanism.

The interfacial transition zone (ITZ) is the 10 to 50 micrometre thick region of HCP immediately surrounding each aggregate particle. It forms due to two mechanisms: the wall effect, where cement particles cannot pack as efficiently adjacent to a flat rigid aggregate surface as in the bulk, creating higher local porosity; and local water migration (internal bleeding), where water accumulates in the loosely-packed zone adjacent to aggregate, creating a locally higher water-cement ratio. This results in the ITZ having 50 to 100% higher porosity than bulk HCP, more portlandite (Ca(OH)2) in larger, oriented crystals, less dense C-S-H gel, and pre-existing micro-cracks. These micro-cracks initiate and propagate at applied stresses of only 30 to 45% of the ultimate compressive strength, giving concrete its characteristic non-linear stress-strain behaviour and making it consistently 20 to 40% weaker than neat cement paste or mortar at the same w/c ratio.

C-S-H (calcium silicate hydrate) gel, also called tobermorite gel, is the primary hydration product of Portland cement (C3S and C2S reactions) and constitutes 50 to 60% of the solid volume of fully hydrated cement paste. It is a poorly crystalline, nanoscale material with a layered sheet structure (similar to the mineral tobermorite) and an enormous internal surface area of 100 to 500 m2/g. The binding strength of C-S-H comes from van der Waals forces acting across the 0.5 to 2.5 nm water films in its gel pores - these forces generate cohesive pressures of hundreds of MPa at the nanoscale, which translate to the macroscopic HCP strength of 30 to 150 MPa depending on porosity. Two density states exist: low-density C-S-H (LD-CSH, porosity ~37%, stiffness ~20 GPa) forming in the outer product region, and high-density C-S-H (HD-CSH, porosity ~24%, stiffness ~30 GPa) forming in the inner product region. The Jennings (2000) colloidal model quantifies this two-state description.

The Powers model (Powers and Brownyard, 1948) provides a fundamental quantitative relationship between the water-cement ratio, degree of hydration, capillary porosity, and the strength of hardened cement paste. The capillary porosity is: Vcap = (w/c - 0.36 x alpha) / (w/c + 0.317), where alpha is the degree of hydration (0 to 1). Powers extended this to the gel-space ratio X = volume of C-S-H gel divided by total space (gel plus capillary pores), and showed that paste strength follows: fc = A x X^n, where A is approximately 240 MPa (intrinsic gel strength at zero porosity) and n is approximately 3. This framework predicts that: (1) lower w/c gives lower capillary porosity and higher gel-space ratio, hence higher strength; (2) higher degree of hydration (better curing) reduces capillary porosity and increases strength; (3) complete hydration requires w/c >= 0.38 (sealed curing) or >= 0.42 (self-desiccating conditions). The model is the theoretical basis for Abrams law and all modern mix design relationships between w/c ratio and strength.

Portland cement hydration produces four main solid products: C-S-H gel (calcium silicate hydrate, 50 to 60% of solid HCP volume): primary binding phase, responsible for most of the strength and stiffness of concrete; Ca(OH)2 or portlandite (20 to 25%): provides alkalinity (pH 12.4 to 13.5) that passivates steel reinforcement, but contributes little strength due to low surface area and planar crystal cleavage; it is consumed in pozzolanic reactions with silica fume and fly ash, improving ITZ quality; calcium sulphoaluminates, primarily ettringite and monosulphate (15 to 20%): formed from C3A hydration controlled by gypsum; ettringite (needle-like crystals) forms in the first 24 hours and moderates early workability; monosulphate (hexagonal plates) forms when gypsum is consumed and is susceptible to external sulphate attack (expansive re-formation of ettringite causes cracking); unhydrated clinker grains (0 to 10%): act as passive filler and potential future strength source if water later becomes available.

The Bogue calculation (Bogue, 1929) is a set of empirical equations that compute the theoretical phase composition of Portland cement clinker (C3S, C2S, C3A, C4AF percentages) from its oxide analysis (CaO, SiO2, Al2O3, Fe2O3 percentages obtained by X-ray fluorescence). The key equations are: %C3S = 4.071(%CaO) - 7.600(%SiO2) - 6.718(%Al2O3) - 1.430(%Fe2O3) - 2.852(%SO3); %C2S = 2.867(%SiO2) - 0.7544(%C3S); %C3A = 2.650(%Al2O3) - 1.692(%Fe2O3); %C4AF = 3.043(%Fe2O3). The Bogue calculation is used in cement quality control to predict: heat of hydration (C3A generates most heat at 860 J/g); sulphate resistance (C3A content limited to 8% for SRPC per IS 12330); early vs late strength development (C3S dominant for 28 days; C2S for long-term). The calculation assumes complete reaction and no solid solutions, giving results that differ 1 to 5 percentage points from the true composition.

The water-cement ratio is the most powerful single variable in concrete because it simultaneously affects all three phases. For HCP: lower w/c reduces capillary porosity (Powers model); increases gel-space ratio; increases density and stiffness of C-S-H; reduces the total volume of portlandite per unit cement. For the ITZ: lower w/c reduces the local excess water available for bleeding to the aggregate surface; reduces the local water film thickness; produces a thinner, denser ITZ with lower Ca(OH)2 concentration and smaller crystal sizes; the ITZ porosity profile decays more steeply (smaller lambda in the exponential profile). For aggregate: the aggregate itself is unaffected by w/c, but the aggregate-paste bond is improved at lower w/c because the ITZ surrounding it is stronger and less porous. Net effect: reducing w/c from 0.60 to 0.40 approximately doubles concrete strength (from 25 MPa to 50 MPa for OPC), reduces permeability by 2 to 3 orders of magnitude, and dramatically improves durability against chloride, sulphate, and acid attack. IS 456 Table 5 maximum w/c limits (0.40 to 0.55 depending on exposure class) are the codified expression of this relationship.

Bleeding in fresh concrete is the upward migration of free mix water driven by the gravitational settlement of the denser solid components (cement and aggregate). Surface bleeding reaches the top of the concrete element and forms a visible water sheen or laitance layer. Internal bleeding occurs when this rising water is intercepted and trapped beneath horizontal obstacles within the concrete before it can reach the surface - primarily on the underside of coarse aggregate particles and on the top face of horizontal reinforcing bars. Internal bleeding creates permanent water-filled voids at these interfaces, which become air voids after drying. These voids are concentrated in the ITZ on the underside of aggregate particles, creating an asymmetric ITZ that is weakest directly below large aggregate and explains why concrete with 40 mm aggregate is weaker than concrete with 10 mm aggregate at the same w/c: the larger surface area under large aggregate accumulates more bleed water. Internal bleeding cannot be observed visually during casting, making it more insidious than surface bleeding. It is controlled by reducing w/c, using silica fume, reducing maximum aggregate size, and improving paste cohesion.

Hardened cement paste contains several types of voids at different length scales. Interlayer spaces in C-S-H (0.5 to 2.5 nm): inherent structural pores within the C-S-H layered structure; always present regardless of w/c or curing; responsible for shrinkage and creep through capillary tension and disjoining pressure mechanisms; too small to significantly affect compressive strength. Small gel pores (2 to 10 nm): spaces between C-S-H particles; responsible for shrinkage at moderate humidity (RH 50 to 90%); minor effect on strength. Capillary micropores (10 to 50 nm): spaces originally occupied by mix water not filled by hydration products; controlled by w/c ratio and degree of hydration (Powers model); responsible for drying shrinkage and chloride transport. Capillary macropores (50 nm to 10 micrometres): highly detrimental to both strength (act as Griffith flaws, nucleating cracks) and durability (main permeability pathway); dominant when w/c > 0.60 or degree of hydration is low; reduced by lower w/c and better curing. Entrained air voids (50 to 200 micrometres): intentionally introduced by air-entraining admixtures; beneficial for freeze-thaw resistance (pressure relief chambers); each 1% reduces compressive strength by approximately 5%. Entrapped air voids (0.5 to 3 mm): unintentionally present from poor compaction; always detrimental to strength and impermeability.

Silica fume (microsilica, IS 15388) is a highly reactive pozzolan consisting of amorphous silica particles with diameters of 0.1 to 0.3 micrometres - 100 to 200 times finer than cement particles. When added at 8 to 12% replacement of cement by mass, it dramatically improves the ITZ through two mechanisms. First, physical packing: the ultra-fine particles fill the spaces between cement grains in the ITZ that are responsible for the wall effect, reducing the local porosity gradient from aggregate surface into bulk paste. Second, pozzolanic reaction: silica fume reacts with Ca(OH)2 (Si02 + Ca(OH)2 + H2O to C-S-H) consuming the portlandite that concentrates in the ITZ and converting it to additional C-S-H gel, which is 3 to 5 times stronger and denser than portlandite. The combined effect reduces ITZ porosity from approximately 35 to 50% (normal concrete) to approximately 20 to 25%, reduces ITZ thickness from 40 to 50 micrometres to 15 to 20 micrometres, and reduces the Ca(OH)2 to C-S-H ratio in the ITZ from about 2:1 to nearly 1:10. Concrete with 10% silica fume typically achieves 30 to 50% higher compressive strength, significantly reduced permeability, and improved sulphate and chloride resistance compared to plain OPC concrete at the same total cementitious content.

The non-linear ascending portion of the concrete stress-strain curve (from zero stress to approximately 0.85 times ultimate strength) is caused by progressive microcracking that begins in the ITZ at loads as low as 30 to 40% of ultimate. The sequence of events is: at 0 to 30% of ultimate load, pre-existing ITZ micro-cracks are stable; elastic behaviour dominates. At 30 to 45% of ultimate, stress concentrations at aggregate-paste interfaces (due to stiffness mismatch; aggregate 50 to 80 GPa vs paste 10 to 25 GPa) cause tensile ring stresses around aggregate equators, initiating new ITZ micro-cracks in the weakest zone; the secant modulus begins to fall below the initial tangent modulus. At 45 to 75% of ultimate, ITZ cracks grow stably through the paste matrix; acoustic emission activity increases; crack density increases but cracks remain isolated. At 75 to 90% of ultimate, ITZ and paste cracks begin to link across inter-aggregate paste bridges, forming continuous crack networks; this is the point of crack arrest instability or critical stress intensity factor; the curve approaches its peak. Above 90% of ultimate, crack networks span the specimen; post-peak softening occurs as cracks widen (in a displacement-controlled test). This progressive failure sequence is why concrete has a much higher fracture energy than brittle glass (which fails suddenly at a single flaw) despite similar compressive strength per unit weight.

Abrams law (Duff Abrams, 1919) states that the strength of fully compacted concrete is governed solely by the water-cement ratio: fc = A / B^(w/c), where A and B are empirical constants depending on cement type, aggregate type, and curing. For IS 269 OPC at 28 days: A is approximately 97.5 MPa and B is approximately 4.9. Abrams law can be derived from the Powers model by expressing the gel-space ratio X as an explicit function of w/c at a given degree of hydration alpha: as w/c increases, capillary porosity increases (Powers formula), gel-space ratio X decreases, and strength fc = A x X^n decreases. Substituting the relationship between X and w/c and rearranging gives an exponential (power law) relationship equivalent to Abrams form. The Abrams constants A and B absorb the effects of cement chemistry, aggregate mineralogy, and curing conditions, while the Powers model makes these effects explicit through the degree of hydration and the intrinsic gel strength A = 240 MPa. IS 10262:2019 uses a modified form of Abrams law in the mix design procedure to select the w/c ratio needed to achieve the target mean strength f_ck + 1.65 x sigma.

The aggregate modulus of elasticity Eagg strongly influences concrete behaviour through three mechanisms: concrete modulus: the composite modulus Ec lies between the Voigt upper bound (parallel model, Ec = Eagg x Vagg + Epaste x Vpaste) and the Reuss lower bound (series model); the IS 456 formula Ec = 5000 x sqrt(fck) empirically captures the aggregate stiffness effect since harder aggregates (granite 50 to 80 GPa) give higher Ec than softer limestone (30 to 50 GPa). Restraint of shrinkage: stiff aggregate restrains paste drying shrinkage; concrete shrinkage is approximately paste shrinkage times (1 - Vagg)^1.7 (Pickett model); at 70% aggregate by volume, concrete shrinkage is only about 20 to 30% of paste shrinkage, which greatly reduces drying shrinkage cracking risk. Stress concentration at ITZ: larger stiffness mismatch between aggregate and paste creates higher tensile ring stresses at the ITZ under axial compression; very stiff aggregate (quartzite 80 GPa) in weak paste (10 GPa) creates a stiffness ratio of 8, amplifying ITZ stress concentrations and actually reducing concrete strength below what would be expected from paste strength alone. Elastic aggregate (normal: granite, basalt) versus soft aggregate (limestone, sandstone) choices affect all three of these mechanisms simultaneously.

Calcium hydroxide (Ca(OH)2, portlandite) constitutes 20 to 25% of the solid volume of fully hydrated OPC paste. It plays a dual and somewhat contradictory role. Beneficial: Ca(OH)2 maintains the high alkalinity of pore water (pH 12.4 to 13.5), which is essential for the passive oxide film that prevents corrosion of steel reinforcement. Without this alkalinity, steel would corrode rapidly in the presence of moisture and oxygen, compromising structural integrity. Detrimental: Ca(OH)2 is significantly weaker than C-S-H gel (lower surface area, planar cleavage); concentrates in the ITZ in oriented hexagonal crystal arrays parallel to aggregate surfaces, creating directional weakness planes; is highly soluble in soft water (leaching leads to progressive dissolution, increasing porosity, reducing strength and alkalinity); dissolves readily in acidic environments (carbonation by CO2, acid attack); reacts with sulphates to form gypsum (expansion, loss of binding ability); undergoes ASR in the presence of reactive silica. Pozzolanic reactions (with silica fume, fly ash, GGBS) consume Ca(OH)2 and convert it to secondary C-S-H, improving both ITZ quality and durability while maintaining sufficient alkalinity for rebar protection.

Both creep and shrinkage originate in the C-S-H gel component of HCP and are restrained by the aggregate and unhydrated cement grains. Drying shrinkage: occurs when concrete loses moisture to a drying environment. In C-S-H gel pores (0.5 to 10 nm), capillary tension increases as pore water menisci develop, compressing the solid skeleton. In larger capillary pores (10 to 50 nm), disjoining pressure changes as adsorbed water films thin. The aggregate phase is essentially elastic and non-shrinking: it restrains paste shrinkage by a factor of approximately (1 - Vagg)^1.7 per Pickett model. High aggregate volume fraction (70%) reduces concrete shrinkage to 20 to 30% of paste shrinkage but simultaneously generates internal tensile stresses in the paste that can cause micro-cracking. Creep: under sustained compressive load, water migrates from loaded gel pores to less-stressed capillary pores (seepage creep), and C-S-H layers slowly rearrange under shear stress (solid creep). Aggregate and unhydrated clinker grains do not creep; they restrain paste creep by the same composite mechanics model as for elastic modulus. Higher paste volume fraction (lower aggregate fraction, higher w/c) gives higher creep. Loading at young age (low degree of hydration, high capillary porosity) gives higher creep coefficient per IS 456 Table 3 (phi = 4.8 at 7-day loading vs 1.6 at 28-day loading).

Alkali-silica reaction (ASR) is a chemical reaction between the hydroxyl ions (OH-) from alkali oxides (Na2O and K2O) in cement and reactive forms of silica (amorphous, cryptocrystalline, or strained quartz) present in certain aggregates. The reaction produces an alkali silica gel at the aggregate surface and within aggregate particles. This gel is hygroscopic: it absorbs water and expands, generating internal expansive pressure of 5 to 100 MPa around reacting aggregate particles. ASR primarily affects the aggregate-ITZ interface: the expansive gel forms in and around reactive aggregate particles, rupturing the ITZ and causing characteristic map cracking (crazing) on concrete surfaces. Concrete affected by ASR shows reduced tensile and flexural strength, increased permeability, and eventual disintegration. In terms of the three-phase model: ASR transforms previously sound aggregate into an expansive source of stress, converts the ITZ from a tensile-stress zone (normal loading) to a pressure-generated cracking zone (chemical expansion), and progressively destroys the HCP bonding network around affected aggregate. Prevention per IS 456 Annex A: limit alkali content of cement (Na2O equivalent < 0.6%), use low-alkali cement, blend with pozzolans (fly ash, GGBS), or avoid reactive aggregate (test per IS 2386 Part VII).

Maximum aggregate size (MAS) affects all three phases of concrete simultaneously, with complex consequences for strength and workability. Effect on aggregate phase: larger aggregate reduces total surface area per unit volume, requiring less paste for the same workability (reduced cement content for a given strength, hence more economical). Effect on HCP: with less paste for the same mix, a lower w/c can be achieved while maintaining adequate workability, which increases paste strength. Effect on ITZ: larger aggregate creates a larger absolute area of ITZ per aggregate particle; internal bleeding accumulates more water under each large particle (the water film thickness scales with aggregate diameter for the same mix); the Griffith crack half-length a in the ITZ scales with ITZ thickness which increases with MAS; by the Griffith criterion sigma_c is proportional to 1/sqrt(a), so larger MAS gives lower strength for the same w/c. The net result is the so-called optimal aggregate size: for concrete above approximately M30 strength, reducing MAS from 40 mm to 20 or 10 mm consistently increases strength despite the paste volume penalty, because the ITZ flaw size reduction outweighs the w/c increase needed. IS 456 Clause 26.4.2 sets maximum MAS as the smaller of 1/4 of the minimum member dimension, 3/4 of clear bar spacing, and 3/4 of clear cover - primarily structural requirements, but the strength implication of reducing MAS to 10 mm for heavily reinforced sections is also beneficial.