Reviewed by Er. Inderjit Sharma (Retd.), Punjab PWD
What is grain size analysis? It finds how much of a soil is gravel, sand, silt and clay by shaking it through a stack of sieves with smaller and smaller openings, and weighing what stays on each one.
Why does it matter? The result decides the soil classification (as per IS 1498), whether a soil is suitable for embankment, subgrade, GSB or filter layers, and how it will drain and compact.
This guide follows IS 2720 (Part 4): Grain Size Analysis (sieve method). Silt and clay sizes need the hydrometer method, covered in a separate post.
How Sieving Sorts the Soil
Soil Fractions (IS 1498)
🪨 Gravel
Larger than 4.75 mm (up to 80 mm). Coarse gravel above 20 mm.
🟠 Sand
4.75 mm to 75 µm. Coarse: 4.75–2 mm; medium: 2 mm–425 µm; fine: 425–75 µm.
⚪ Silt
75 µm to 2 µm. Feels like flour.
🔴 Clay
Smaller than 2 µm. Sticky when wet.
Apparatus Required
IS sieves (IS 460)
75 mm, 19 mm, 4.75 mm, 2 mm, 425 µm, 75 µm, with lid and pan.
Balance (15 kg, 1 g)
For weighing the sample and the soil on each sieve.
Trays and bucket
For drying, soaking and washing the soil.
Mortar & rubber pestle
Breaks lumps without crushing the grains.
Oven (105–110°C)
Dries the washed soil before final sieving.
Dispersing agent
Sodium hexametaphosphate: separates clay from sand grains.
How Much Soil to Take
| Largest particle size | Minimum sample |
|---|---|
| 75 mm | 60 kg |
| 40 mm | 25 kg |
| 25 mm | 13 kg |
| 19 mm | 6.5 kg |
| 12.5 mm | 3.5 kg |
| 10 mm | 1.5 kg |
| 6.5 mm | 0.75 kg |
| 4.75 mm | 0.4 kg |
Don’t Overload the Sieves
| IS sieve | Max. on a 450 mm sieve | Max. on a 300 mm sieve |
|---|---|---|
| 80 mm | 15 kg | 6 kg |
| 20 mm | 4 kg | 2 kg |
| 4.75 mm | 1 kg | 0.5 kg |
If there is more soil than this, sieve it in smaller batches.
Procedure
Part A: Coarse Fraction (Dry Sieving)
- Prepare the soil as per IS 2720 (Part 1) and weigh the sample (quantity from the table above).
- Separate it on the 4.75 mm sieve.
- Sieve the retained part through the 75 mm, 19 mm and 4.75 mm sieves, moving each sieve so the soil rolls over the whole mesh.
- If lumps remain, rub them gently in the mortar with the rubber pestle (without breaking particles) and sieve again.
- Weigh the soil retained on each sieve.
- If the moisture is more than 5%, find the water content (IS 2720 Part 2) and correct the weights to dry weight.
- If gravel is more than about 20% and sticky fines cling to it, wash the gravel on the 4.75 mm sieve, using a dispersing agent if needed.
Part B: Fine Fraction (Wet Sieving)
- Oven-dry the soil passing 4.75 mm at 105–110°C, riffle out about 200 g and weigh it.
- Put it in a tray or bucket, cover it with water and add a dispersing agent (commonly about 2 g of sodium hexametaphosphate per litre of water).
- Stir well and leave it to soak for 24 hours.
- Wash the soil over the 75 µm sieve until the water coming out is clear. Do not overload the sieve.
- Oven-dry what is left on the sieve, then dry-sieve it through the 4.75 mm, 2 mm, 425 µm and 75 µm sieves and weigh each fraction.
Calculations
From the grading curve, read D10, D30 and D60 (the sizes that 10%, 30% and 60% of the soil is finer than) and find:
Uniformity coefficient
Cu = D60 ÷ D10
Shows the range of particle sizes.
Coefficient of curvature
Cc = D30² ÷ (D10 × D60)
Shows the shape of the grading curve.
Well graded?
Gravel: Cu > 4 and Cc 1 to 3
Sand: Cu > 6 and Cc 1 to 3
Otherwise: poorly graded.
Worked Example and Grading Calculator
Enter the dry weight retained on each sieve. The table, soil fractions, D-values and the grading curve update instantly. The values filled in are a worked example.
Weights in grams. Total dry weight = sum of all entries.
Precautions
✋ Never push grains through
You may check whether a particle passes a sieve, but never force it through.
⚖️ Don’t overload
Too much soil on a sieve blocks it and gives wrong results. Sieve in batches.
💧 Wash fine soils
Clay sticks to sand grains. Without soaking and washing, fines are counted as sand.
🔍 Check the sieves
Before testing, check the mesh for tears or blocked holes. Clean fine sieves with a soft brush.
Practical Tips for Site Engineers
1. Where You Will Use This Test
🛣️ Earthwork approval
Borrow soil for embankment and subgrade is approved based on its gradation and classification.
🧱 GSB and WMM
Granular sub-base and wet mix macadam must fall within the grading limits given in the MoRTH specifications.
💧 Filters and drains
Filter media behind walls and in drains need the right grading so water passes but soil does not.
🏖️ Sand for concrete
Fine aggregate is checked for its zone (I to IV) as per IS 383 using a similar sieve analysis.
2. Good Practice on Site
- Representative sample: take soil from several points of the borrow area or stockpile, mix it, and reduce it by quartering. A sample from one spot can mislead.
- Check the total: after sieving, add up all the weights. The total should be very close to the starting weight (within about 1%). A big difference means soil was lost or weighed wrongly.
- Watch for changes: when the colour or texture of the borrow soil changes, take a new sample and test again.
- Plot and compare: plot the grading curve on the same sheet as the specification limits; it is easy to see if the material is within the range.
- Quick field check: for a fast idea before the lab result, use the jar (settling) test or the hand tests in our post on Field Identification Tests for Soils.
Key Takeaways
- Sieve analysis finds the % of gravel (> 4.75 mm), sand (4.75 mm–75 µm) and fines (< 75 µm).
- Coarse part: dry sieving; fine part: soak, wash on 75 µm, dry and sieve.
- % passing = 100 − cumulative % retained; report to two decimals.
- Cu = D60/D10 and Cc = D30²/(D10·D60) decide if a soil is well or poorly graded.
- On site: take representative samples, check the total weight, and compare the curve with specification limits.
References: IS 2720 (Part 4): Methods of Test for Soils – Grain Size Analysis; IS 1498: Classification and Identification of Soils for General Engineering Purposes; IS 460 (Part 1): Test Sieves; MoRTH Specifications for Road and Bridge Works.
