Jar Test: Procedure, Dose Calculation and Interpretation

A jar test runs a series of 1 or 2 L portions of the same raw water through rapid mix, flocculation and settling side by side, each with a different coagulant dose, so that the best-performing jar sets the plant's chemical and dose (US EPA, 1999). The US EPA protocol takes 91 minutes of mixing and settling per run, so which speeds, stock solutions and calculations turn those jars into a defensible dose?

Each run decides the chemical, its dose with a stated basis, and the coagulation pH, from turbidity, dissolved organic carbon, temperature and alkalinity. The coagulants, flocculants and pH adjusters it compares are three of the families of water treatment chemicals.

Below: equipment, an 8-step procedure with 3 protocols, 5 worked dose calculations, settled-turbidity goals of 1 and 2 NTU, the NSF/ANSI/CAN 60 check and 4 raw-water properties that move the optimum.

Every value on this page is checked against primary sources (US EPA, eCFR, Irish EPA, NHMRC, NSF listings); see our methodology and fact-checking process.

What Is a Jar Test in Water Treatment?#

A jar test in water treatment is a bench-scale simulation of full-scale coagulation, flocculation and sedimentation, run on the actual raw water to find the chemical, dose and pH that remove the most turbidity and organic matter. US drinking-water rules define coagulation as the destabilization of colloidal and suspended material by coagulant chemicals and mixing, and flocculation as gentle stirring that grows small floc into larger, more easily settleable particles (40 CFR 141.2).

What does a jar test measure?#

A jar test measures 5 responses of each jar, each recorded against the dose and its basis. The 5 responses are listed below.

• Settled turbidity, in NTU after the settling period
• Filtered turbidity, the filter's view of the settled floc
• pH after dosing
• Color or total organic carbon (TOC)
• Floc appearance: size, density and clarity between particles

Each response answers a different plant question, so the full set is recorded for every jar.

Why do treatment plants run jar tests?#

Treatment plants run jar tests because no published dose fits every water: the Australian Drinking Water Guidelines (NHMRC) set coagulant doses by laboratory trial, since the right dose depends on turbidity, dissolved organic carbon, temperature and alkalinity. Plants use it to choose a coagulant, set the dose and pH, and run rule-driven tests such as enhanced coagulation Step 2; chemical phosphorus removal relies on jar or pilot testing too (Recommended Standards for Wastewater Facilities, 2014 edition).

How Does a Jar Test Simulate Coagulation, Flocculation and Sedimentation?#

A jar test simulates the 3 plant stages with 3 paddle regimes: a short, fast rapid mix for coagulation, a longer slow mix for flocculation and a quiescent period for sedimentation. Charge neutralization is very rapid, and Al3+ is about 700 times as effective as Na+ in drinking water (Hach coagulation handbook LIT2141), so the rapid mix is short. Velocity gradients run about 1,000 s-1 for coagulation and about 100 s-1 for flocculation (SNF Floerger, 2003). Why is the slow mix longer? Floc growth needs contact time: plant flocculation runs 15 to 45 minutes and grows floc of 0.1 to 3 mm.

Coagulation and flocculation are different steps with different chemistry, and the jar test reproduces both in the same beaker. Their mechanisms and the differences between a coagulant and a flocculant are explained under coagulation and flocculation in water treatment.

The Recommended Standards for Water Works (2022 edition, sections 4.2.3 and 4.2.4) limit rapid mix to 30 seconds at a G of at least 750 s-1, base G and detention time on jar testing, and require 30 minutes of flocculation or more.

The quiescent settling period in each jar stands for the basin described under sedimentation in water treatment.

What Equipment Does a Jar Test Need?#

A jar test needs 7 items: a multi-paddle gang stirrer, 1 or 2 L jars (square preferred), pipettes or a micropipette, fresh stock solutions, a nephelometric turbidimeter, a calibrated pH meter and a sampling point about 10 cm (4 in) below the water surface. Table 1 lists what each item does and its published specification.

Covers matter because CO2 transfer with room air shifts pH during a 91-minute run.

Jar Test Procedure: 8 Steps from Raw-Water Sample to Settled Turbidity#

A jar test runs in 8 steps, from sampling the raw water to repeating the series in finer steps around the best jar; the values below follow the US EPA (1999) protocol, with the Irish EPA values in Table 2. The 8 steps are listed below.

1. Collect a representative raw-water sample and record turbidity, pH, alkalinity and temperature.
2. Fill each jar with the same volume, 1 or 2 L, and cover it.
3. Prepare fresh stock solutions: 20 g/L for alum, no more than 7 days old.
4. Dose a series in 10 mg/L (ppm) steps, stating the basis.
5. Rapid-mix at 100 rpm for 1 minute.
6. Flocculate at 30 rpm for 30 minutes.
7. Settle for 60 minutes, then sample about 10 cm (4 in) below the surface.
8. Measure settled turbidity, pH and color, then repeat in 2 to 3 mg/L steps around the best jar.

Record basis, stock strength, pH after dosing and settled and filtered turbidity for every jar; these values convert directly into a plant dose.

How are jar test stock solutions prepared?#

A jar test stock solution is a weighed dilution whose strength makes each milliliter a round dose: 1 mL of a 1 g/L (0.1 %) stock in a 1 L jar gives 1 mg/L, and 1 mL of a 20 g/L alum stock in a 2 L jar gives 10 mg/L. A 1 % solution equals 10,000 mg/L; US EPA alum stock stays below pH 3.0 and is used within 7 days. The Irish EPA manual gives the 5 shelf lives listed below.

• Alum, 1 %: 1 month
• Ferrous sulfate, 1 %: 1 week
• Polyacrylamide, 0.5 %: 1 week
• Sulfuric acid, 0.1 N: 3 months
• Sodium hydroxide, 0.1 N: 1 month

Which mixing speeds and times does a jar test use?#

Jar test mixing speeds depend on the protocol: the US EPA uses 100 rpm for 1 minute, 30 rpm for 30 minutes and 60 minutes of settling, while the Irish EPA manual uses 200 rpm, then 20 rpm for 15 minutes and 30 minutes of settling. Flocculant tests on industrial wastewater run 250 rpm for 10 seconds, then 40 rpm for 5 minutes, and settle about 10 minutes (SNF Floerger, 2003). Table 2 sets the 3 published protocols side by side; one protocol should run for every jar of a series.

How are polymer flocculants and coagulant aids added in a jar test?#

Polymer flocculants and coagulant aids are added after the primary coagulant at much lower doses: polyacrylamide aids run at 0.05 to 0.3 mg/L (ppm) in drinking water (NHMRC) and cationic polymer aids at 0.1 to 0.5 mg/L (Hach LIT2141). Polymer dilution water should carry less than 4 mg/L of chlorine.

Jar polymer stocks are prepared like a plant's polymer make-down and activation. When 2 jars perform equally, the lower polymer dose wins (Irish EPA manual).

Polymer charge and molecular weight decide which flocculant suits which floc. Polymer types are compared under flocculants for water treatment.

Which starting doses suit each coagulant in a jar test?#

A jar series should bracket the published dose range of each coagulant: 5 to 200 mg/L (ppm) as dry alum, 5 to 100 mg/L as product for polyaluminum chloride, 3 to 100 mg/L as product for aluminum chlorohydrate and 0.2 to 6 mg/L as active polymer for polyDADMAC in drinking water (NHMRC). Table 5 lists the starting doses by chemical.

Every coagulant family and its dose basis is compared under coagulants for water treatment.

Jar Test Dose Calculation: 5 Worked Examples#

Jar test dose calculation uses 3 relations: 1 % equals 10,000 mg/L, dose in mg/L times jar volume in L gives the milligrams to add, and a plant feed rate equals dose times flow divided by purity. Each worked example below states every input and gives metric and US units.

Example 1: Jar doses from a 1 % (10 g/L) stock solution#

Worked example: 7.52 mL of liquid alum (specific gravity 1.33) made up to 1 L gives a 10 g/L (1 %) stock, and each 1 mL of it in a 2 L jar adds 5 mg/L (ppm) of alum as product. Inputs: liquid alum at specific gravity 1.33 (commercial range 1.30 to 1.35; use the certificate of analysis value), stock 10 g/L as product, jar 2 L (0.53 gal), series 10 to 60 mg/L as product.

The calculation is 10 g / 1.33 g/mL = 7.52 mL made up to 1,000 mL. One milliliter of stock carries 10 mg, or 5 mg/L in 2 L, so the series needs 2 to 12 mL of stock in 2 mL steps.

Example 2: Dosing neat liquid coagulant with a micropipette#

Worked example: a 20 mg/L (ppm) dose of neat liquid coagulant (specific gravity 1.33) in a 2 L jar needs 20 x 2 / 1.33 = 30.1 µL, delivered with a micropipette. Inputs: 20 mg/L as product, jar 2 L, specific gravity 1.33. The formula, µL = dose (mg/L) x jar volume (L) / specific gravity, comes from a supplier jar-testing procedure, and neat dosing avoids a stock solution altogether.

Example 3: Converting the optimum jar dose between bases#

Worked example: an optimum jar dose of 30 mg/L (ppm) as dry alum equals 61.9 mg/L of 48.5 % liquid alum as product, 2.73 mg/L as Al and 5.16 mg/L as Al2O3, and it consumes 15.0 mg/L of alkalinity as CaCO3. Inputs: optimum 30 mg/L as Al2(SO4)3·14H2O; liquid alum with 48.5 % dry alum (Chemtrade liquid alum product profile, 2020); per mg/L of dry alum, 0.091 mg/L Al, 0.50 mg/L alkalinity as CaCO3, 0.44 mg/L CO2 and about 0.26 mg/L Al(OH)3 (Minnesota Rural Water Association); Al = Al2O3 x 0.529.

The calculation gives 30 / 0.485 = 61.9 mg/L as product, 30 x 0.091 = 2.73 mg/L as Al, 2.73 / 0.529 = 5.16 mg/L as Al2O3, 13.2 mg/L CO2 and about 7.8 mg/L of solids. One jar, three numbers: the basis travels with every dose.

Example 4: Converting the jar dose into a plant feed rate and dosage pump setting#

Worked example: 30 mg/L (ppm) of dry alum at 10 MGD (37,854 m³/d) needs 2,502 lb/day (1,136 kg/d), which is about 1,760 L/d (465 gal/d) of 48.5 % liquid alum, or a dosing pump set to about 1,222 mL/min. Inputs: 30 mg/L as dry alum; flow 10 MGD (1 MGD = 3,785 m³/d); purity 1.0 on the dry-alum basis; liquid alum 48.5 % at specific gravity 1.33, or 0.645 kg dry alum per liter.

The US formula gives 30 x 10 x 8.34 / 1.0 = 2,502 lb/day; the metric formula gives 30 x 37,854 / 1,000 = 1,136 kg/d. Then 1,136 / 0.645 = about 1,760 L/d, and 1,760 x 1,000 / 1,440 = about 1,222 mL/min (73.4 L/h, 19.4 gal/h).

The chemical dosage calculator converts any jar dose in mg/L into kg/d, lb/day, L/h and a pump setting.

Example 5: Converting an alum jar dose to ferric chloride#

Worked example: 30 mg/L (ppm) of alum equals 16.5 mg/L of anhydrous ferric chloride in metal equivalents (US EPA, 1999, Table 3-1), and both doses consume about 15 mg/L of alkalinity as CaCO3. Inputs: 30 mg/L as Al2(SO4)3·14H2O; 10 mg/L alum = 5.5 mg/L FeCl3 = 9.1 mg/L FeCl3·6H2O (linear); 0.92 mg/L alkalinity per mg/L FeCl3. The result is 16.5 mg/L FeCl3 (27.3 mg/L FeCl3·6H2O) and 15.2 mg/L alkalinity against 15.0 mg/L for alum. The equivalence is chemical, not a performance prediction, so the ferric chloride series still runs.

How Are Jar Test Results Interpreted?#

Jar test results are read in 3 layers: settled turbidity picks the dose range, filtered turbidity and pH confirm it, and color, TOC and residual metal decide between jars that clarify equally well. The potable dose then meets one last check, the product's maximum use level.

Settled water turbidity and the optimum dose#

The optimum jar test dose is the lowest dose that brings settled water turbidity to the plant's goal: below 1 NTU 95 % of the time when raw water is at or below 10 NTU, or below 2 NTU when raw water is at or below 20 NTU (US EPA 815-R-20-004, 2020, Table 6-1). Turbidity is visible from 4 NTU, water looks crystal clear below 1 NTU, and NTU and FNU are considered equivalent (WHO). Table 3 lists the interpretation targets.

The two turbidity units are compared under turbidity in water.

Filtered turbidity and the filterability test#

Filtered turbidity shows whether settled floc will be held by the plant's filters: the US optimization goal for each filter is below 0.1 NTU 95 % of the time, and combined filter effluent must stay at or below 0.3 NTU in 95 % of monthly measurements and never exceed 1 NTU (40 CFR 141.173). A jar that settles well can still pass fine floc.

The filters the jar filtrate stands for are described under filtration in water treatment.

pH, color, TOC and residual metal in the jars#

Four chemistry readings separate jars that clarify equally: pH after dosing, color, TOC or UV254, and residual aluminum or iron. The 4 readings are listed below.

• pH after dosing, which falls as metal coagulants consume alkalinity.
• Color: coagulation for color runs at pH 4.5 to 5.5 with aluminum or iron salts (Veolia Water Technologies Handbook, chapter 5).
• TOC and UV254: SM 5310 B, C or D, SM 5910 B or EPA 415.3 (40 CFR 141.131).
• Residual aluminum or iron, total and dissolved (Health Canada advises daily aluminum monitoring or more often).

Dissolved residual metal rises when coagulant is overdosed to lower pH (Irish EPA Advice Note 15).

Floc appearance: what the best jar looks like#

The best jar shows coarse, dense, granular floc with clear water between the particles, and it is not necessarily the jar with the largest floc (Irish EPA manual). With equal results, the lower polymer dose wins.

Underdose, optimum and overdose patterns#

A jar series shows 3 patterns: underdosed jars stay turbid, optimum jars settle clear, and overdosed jars turn turbid again because excess cationic charge restabilizes the particles. The 3 patterns are listed below.

• Underdose: particles keep their negative charge and fine turbidity stays suspended.
• Optimum: zeta potential is slightly negative at most utilities after coagulation (Hach LIT2141).
• Overdose: a positive zeta potential indicates likely overfeed; overdosed polyDADMAC re-disperses particles (NHMRC).

Does the optimum jar dose fit the NSF/ANSI/CAN 60 maximum use level?#

The optimum jar dose for potable water must stay at or below the maximum use level of the specific product, stated as product: the most common NSF/ANSI/CAN 60 values are 150 mg/L for alum, 250 mg/L for polyaluminum chloride, aluminum chlorohydrate and ferric chloride, 50 mg/L for polyDADMAC and 10 mg/L for polyamine (NSF listings, 2026-10-21). Aluminum listings add a condition of at most 2 mg/L aluminum in finished water, which is not the non-enforceable US secondary level of 0.05 to 0.2 mg/L (40 CFR 143.3).

How maximum use levels are set is explained under NSF/ANSI/CAN 60.

Typical Jar Test Settings and Doses by Application#

Jar test settings change with the application: drinking-water coagulation follows the US EPA or Irish EPA protocol, enhanced coagulation Step 2 adds alum only in 10 mg/L steps, and wastewater phosphorus, metals and flocculant tests start from stoichiometric or supplier ratios. The Step 2 test under 40 CFR 141.135 stops when an extra 10 mg/L of alum removes 0.3 mg/L of TOC or less, at a target pH of 5.5 to 7.5 set by alkalinity, with no acid and no polymer (US EPA, 1999). Table 4 compares the starting points by application.

What Changes the Optimum Jar Test Dose?#

Four raw-water properties change the optimum jar test dose: turbidity, alkalinity and pH, organic matter and temperature, which is why the Australian Drinking Water Guidelines set doses by laboratory trial.

Raw-water turbidity#

Raw-water turbidity sets the chemistry tested in the jars: above 60 NTU a cationic polymer alone normally clarifies, while at 60 NTU and below the jars pair a metal salt with a cationic polymer (Veolia Water Technologies Handbook, chapter 5).

Raw-water alkalinity and pH#

Raw-water alkalinity limits how much metal coagulant a jar can take before pH falls out of range, because each mg/L of dry alum consumes 0.50 mg/L of alkalinity as CaCO3. At 30 mg/L of dry alum, 15 mg/L of alkalinity disappears.

P and M readings are explained under alkalinity of water. The Irish EPA (Advice Note 15) corrects pH with acid or alkali, never with extra coagulant, because overdosing to depress pH raises dissolved residual metal.

The acids and alkalis used to correct jar pH are compared under pH adjustment chemicals.

Organic matter (TOC and DOC)#

Organic matter is one of the four inputs that set the jar test dose, and SUVA (100 x UV254 divided by DOC, in L/mg-m) shows how readily coagulation removes it; a SUVA at or below 2.0 L/mg-m is an alternative compliance criterion under 40 CFR 141.135. Step 1 of enhanced coagulation requires 15 to 50 % TOC removal, set by source TOC and alkalinity.

TOC, DOC and UV254 methods are covered under total organic carbon.

Water temperature#

Cold water shifts the optimum: the best pH for aluminum coagulants moves by about 0.4 units between 2 and 20 °C (35.6 and 68 °F), and water at 0 °C (32 °F) is twice as viscous as at 25 °C (77 °F), so floc settles more slowly. Below 5 °C (41 °F), alum is more affected than iron (Irish EPA). Health Canada puts the alum optimum at pH 6.5 to 7.0 below 10 °C (50 °F) and 6.0 to 6.5 in warm water.

9 Common Jar Test Mistakes#

Nine mistakes make a jar test misleading, and the most common is a dose reported without its basis. The 9 mistakes and their fixes are listed below.

1. Dose without a basis: state as product, as dry alum, as Al or as Al2O3.
2. Old stock: use 20 g/L alum stock within 7 days and 1 % alum within 1 month.
3. Uncovered jars: CO2 transfer shifts pH, so cover every jar.
4. Wrong sampling depth: draw about 10 cm (4 in) below the surface.
5. Picking the largest floc: let settled turbidity decide.
6. Lowering pH with extra coagulant: correct with acid or alkali.
7. Acid or polymer in a Step 2 test: the rule test allows coagulant only.
8. The 0.588 kg/L example density: about 9 % low for commercial liquid alum.
9. A potable dose above the product's NSF/ANSI/CAN 60 maximum use level: check the listing first.

How Does a Jar Test Compare with Online Coagulant Control?#

A jar test sets the chemical and dose on the bench, while online coagulant control tracks the dose between tests with streaming current, zeta potential on grab samples and feed-forward models built on raw-water turbidity, color or UV254. Feed-forward control raises an alarm at about 20 % mismatch with the metered dose (Irish EPA Advice Note 15), and streaming current is disturbed by salinity, conductivity and pH (Hach LIT2141).

Both online signals are compared under streaming current and zeta potential.

How long does a jar test take?#

A jar test takes about 91 minutes of mixing and settling under the US EPA protocol, about 47 minutes under the Irish EPA protocol and about 15 minutes for a flocculant test, plus preparation and analysis.

Is a flocculation test the same as a jar test?#

Yes: a flocculation test is the jar test run to compare flocculants or flocculation conditions, with the same gang stirrer, jars and settled-turbidity readings.

Can a jar test set powdered activated carbon and permanganate doses?#

Yes: drinking-water plants set potassium permanganate doses of 0.3 to 5 mg/L by jar test, against stoichiometric needs of 0.94 mg per mg of iron(II) and 1.92 mg per mg of manganese(II) (NHMRC ADWG). The Recommended Standards for Water Works (2022 edition, 4.11.3.3) ask for carbon feed provision of 0.1 to at least 40 mg/L, with carbon type and dose selected by jar testing or other studies.

Carbon doses and NSF/ANSI/CAN 61 limits are on powdered activated carbon.

Where does the jar test sit among water treatment chemistry concepts?#

The jar test is the bench method behind the coagulation concept, and it sits beside scaling, corrosion and disinfection calculations.

All four groups are listed among the water treatment chemistry concepts of this site.