SULFATE2 user guide

What the model calculates, how to use the calculator, and five worked examples you can open with one click.

What it calculates

SULFATE2 simulates a square mortar or concrete prism immersed in a sulfate solution. Sulfate diffuses in through all four faces and reacts with the cement’s aluminates (C₃A). The reaction products first fill part of the capillary pores; once they no longer fit, they build stress that cracks the material from the outside in. Cracked zones let sulfate in faster, which speeds up the attack.

For each day, the program calculates three results:

  • Expansion of the prism along its length, in %.
  • Crack front depth, measured from the surface, in mm.
  • Average stiffness (Young’s modulus) of the section, in GPa.

The simulation ends when the reactive aluminates are used up, or at the Simulation length you choose. It is the Tixier–Mobasher model (Journal of Materials in Civil Engineering, 2003), which NIST distributes as SULFATE2.

How to use the calculator

  1. Start with NIST’s example, which is already loaded, or open a sulfate_model_data2.dat file with Open input file.
  2. Change the inputs in the left column. The results recalculate on their own. Changed fields get a blue bar and show the example value, with a link to restore it.
  3. Read the specimen. The square is the cross-section: each ring is one layer of the model, darker the more stiffness it has lost, and the outline marks the crack front. Press Play or drag the slider to move through time.
  4. Find when a limit is reached. Under When is a limit reached? type an expansion (for example 0.10 %) or a crack depth to get the first day it is reached. Press the day to see the specimen at that moment.
  5. Compare alternatives. Press Pin this run, then change inputs: the charts show both runs and list the inputs that differ.
  6. Save and share. Save input file downloads the inputs; the download buttons give out and info files in the original format, plus a CSV for Excel. The page address holds the inputs, so you can copy it and send it to someone to show them the same run.
  7. Export for a report. PNG beside each chart and Download PNG under the specimen save an image with a white background, whatever the page’s theme. The specimen image shows the selected day. Under Inputs table you can download the inputs as an image, or an XLSX with the inputs and the results.
  8. Ask. If something is unclear or you need the program to do more, use Questions and requests at the bottom of the calculator.

Examples

Each table is calculated in your browser with the same model as the calculator. Specifications commonly use an expansion limit such as 0.10 % in ASTM C1012 tests; check the limit and age that apply to your project.

Reproduce NIST’s example

A 25.4 mm (1 in) mortar prism, like an ASTM C1012 bar, in sodium sulfate solution. This is what the calculator shows when it opens, and it matches NIST’s reference output exactly. Start here to get a feel for the curves before changing anything.

CaseReaches 0.10 % expansionAluminates used upFinal expansionOpen
NIST example………Open in calculator

Compare water/cement ratios

Typical question: a specification allows w/c up to 0.50, the contractor proposes 0.45. How much time does the lower ratio buy? Open one variant, press Pin this run, then open another to see both sets of curves together.

Only the water/cement ratio changes here. In the model it changes the capillary porosity, which absorbs expansive products; it does not change the diffusivity. A real mix with a higher w/c is also more permeable, so raise the diffusivity together with w/c if you want that effect (see the diffusivity example).

CaseReaches 0.10 % expansionAluminates used upFinal expansionOpen
w/c 0.45………Open in calculator
w/c 0.485 (example)………Open in calculator
w/c 0.50………Open in calculator
w/c 0.55………Open in calculator

Choose a cement by its C₃A content

C₃A is what reacts with sulfate to form expansive products, which is why sulfate-resisting cements limit it. ASTM C150 caps C₃A at 8 % for Type II and 5 % for Type V. Take the value from the cement’s mill certificate and enter it as a fraction (8 % is 0.08).

CaseReaches 0.10 % expansionAluminates used upFinal expansionOpen
C₃A 10 %………Open in calculator
C₃A 8 % (Type II limit)………Open in calculator
C₃A 6 % (example)………Open in calculator
C₃A 5 % (Type V limit)………Open in calculator
C₃A 3 %………Open in calculator

Bracket an uncertain diffusivity

The uncracked diffusivity sets how fast sulfate gets in, so it controls the timing more than almost any other input, and it is rarely measured. Run a low, middle and high value and report the range rather than a single date.

CaseReaches 0.10 % expansionAluminates used upFinal expansionOpen
1e-12 m²/s………Open in calculator
2e-12 m²/s (example)………Open in calculator
5e-12 m²/s………Open in calculator

Scale from a lab prism to a thicker section

Thicker sections take much longer, roughly with the square of the thickness, because sulfate has to diffuse further. Doubling 25.4 mm to 50 mm more than triples the time to 0.10 % expansion.

The model is a square prism attacked on all four faces. For a column or pile exposed all round, use the least dimension as a rough indication. A wall or slab attacked from one face is a different case that this model does not represent.

CaseReaches 0.10 % expansionAluminates used upFinal expansionOpen
25.4 mm (example)………Open in calculator
50 mm………Open in calculator
100 mm………Open in calculator

Inputs

The calibration values (reaction rate constant, fillable share of pores, residual stress) come from NIST’s example. Without data of your own, leave them as they are and vary the inputs you do know.

InputNIST exampleWhere to get it
Simulation length200 yearsNot a material property: how long to simulate, for example your design life. 200 years at most.
Prism thickness25.4 mmSide of the square specimen. ASTM C1012 bars are 25 mm.
Gage length279.4 mmLength over which expansion is measured. It does not change the results, because expansion is reported in %.
Cement content388 kg/m³Mix design.
Cement specific gravity3.15About 3.15 for portland cement; lower for blended cements.
Water/cement ratio0.485Mix design.
Degree of hydration0.9Between 0 and 1. Mature paste is typically 0.8 to 0.9.
Sulfate concentration at the surface35.2 mol/m³Sulfate at the exposed surface, per m³ of material. See “Sulfate concentration” below.
Diffusivity, uncracked2e-12 m²/sSulfate diffusivity of the uncracked material. Rarely measured; bracket it.
Diffusivity increase when cracked10 × DHow many times faster sulfate diffuses once fully cracked. Example: 10.
Reaction rate constant1e-7 m³/(mol·s)Calibration constant of the sulfate–aluminate reaction. Example: 1e-7.
C₃A content of cement0.06Mill certificate (Bogue C₃A), as a fraction.
Degree of reaction of C₃A0.9Share of the C₃A available to react. Example: 0.9.
Gypsum content of cement0.06Gypsum in the cement, as a fraction. Example: 0.06.
Fillable share of capillary pores0.45Share of the capillary pores that fills before the paste is stressed. Calibration value; example: 0.45.
Young’s modulus, undamaged28000 MPaTests or code formulas for the material’s modulus.
Tensile strength3 MPaSplitting or flexural tests, or code estimates.
Residual expansive pore stress10 MPaCalibration value; example: 10 MPa. It also sets the expansion ceiling (see Limits).

Sulfate concentration

The model takes the concentration per cubic metre of material, not of solution. NIST’s example uses 35.2 mol/m³, one tenth of the 352 mol/m³ (0.352 mol/L) in ASTM C1012’s 50 g/L sodium sulfate solution.

For soil or groundwater with sulfate reported in ppm (mg/L of SO₄): 1 mol/m³ of SO₄ is 96 mg/L. For example, 3000 ppm is about 31 mol/m³ of solution, and with the same ratio as the example about 3.1 mol/m³ of material. That ratio is an assumption carried over from the example; confirm it against Tixier and Mobasher’s paper before using it in a report.

Limits of the model

  • Geometry. A square prism, immersed and attacked on all four faces, as in the lab. For real structures it is an indication, not a service-life prediction.
  • Thick sections. From about 100 mm, NIST’s original program ends early and gives false results, because its time step grows with thickness. Above 37.5 mm this calculator uses a finer grid and a shorter time step, so its results differ slightly from the original’s. Every run lasts 200 years at most, and a thick section can take about two minutes. Use Simulation length to shorten it, for example to 100 years, and it takes less time.
  • End of the run. The simulation ends when the aluminates are used up. The expansion on that day is not necessarily the final expansion of the real material.
  • Expansion ceiling. When a layer is fully damaged, the original program gives it a fixed modulus of 50 MPa. That caps expansion at about 20 % with the example’s values (at w/c 0.40, for instance). The calculator flags it; treat those results as “fully deteriorated”.
  • High w/c ratios. At w/c 0.6 or more, the pores take up all of the products and expansion comes out as zero. That does not mean the mix is durable: the model does not link w/c to diffusivity.
  • Front at the center. When the crack front reaches the center, the original file records 0. The calculator shows it at the center; downloaded files keep the 0.