ASTM D5890
Free Swell Index
Free Swell Index is the volume a 2.00 gram bentonite sample occupies after hydrating in reagent water, reported in mL per 2 g. NRCS CPS 520 sets the floor for pond sealing at 22 mL. PondLock tests at 30 mL. Here is exactly how that number is produced and what it does and does not promise.
What Free Swell
Index Measures
Free Swell Index is the volume, in millilitres, that a 2.00 gram sample of dried and ground bentonite occupies after it has hydrated undisturbed in reagent water. It is reported as mL per 2 g, and it is determined by ASTM D5890, whose formal title is Standard Test Method for Swell Index of Clay Mineral Component of Geosynthetic Clay Liners.
The number is a proxy for one physical behavior: how far the clay platelets separate when water enters the interlayer spaces of the sodium montmorillonite crystal. That separation is the entire mechanism by which a bentonite liner works. Bentonite does not coat soil or glue it together — it swells into the void space between soil particles until the connected pore network is no longer connected. A clay that barely expands cannot do that at any application rate.
Why it is the number to ask for. Free Swell Index is cheap, standardized, and hard to argue with. Two bags of granular clay can look identical, come from the same colored bag, and carry the same label language, and still differ by a factor of two in swell. The lab report is the only way to know which one you bought.
22 mL Is the Floor.
PondLock Tests 30 mL.
USDA NRCS Conservation Practice Standard 520, Pond Sealing or Lining — Compacted Soil Treatment, requires the use of sodium bentonite with a free swell of at least 22 millilitres as measured by ASTM Standard Test Method D5890. That is the threshold a federal conservation practice will accept, and it is the number a specifying engineer will look for.
Bars scaled to a 36 mL axis. PondLock and Texas-source values are PondLock laboratory results by ASTM D5890; the 22 mL minimum is the requirement stated in NRCS CPS 520.
The gap between 12 mL and 22 mL is the reason regional “pond clay” so often disappoints. Material sold locally at an attractive price per ton can carry roughly half the required swell, which means it cannot fill the pore space it was bought to fill regardless of how much of it goes down.
How the Test
Is Actually Run
The procedure is deliberately slow, because a fast one gives the wrong answer. Bentonite that is dumped into water all at once traps air and forms agglomerates that never fully hydrate, which under-reports swell. ASTM D5890 controls against that with incremental addition and a long undisturbed hydration period.
Prepare and dry the sample
A test specimen of at least 100 g is reduced, and the fraction passing the 150 micron (No. 100) sieve is dried to constant weight at 105 ± 5 °C. Granular material is ground to 100 percent passing the 150 micron sieve with a minimum of 65 percent passing the 75 micron (No. 200) sieve. Where drying adequacy is in doubt, drying continues until two successive periods longer than 1 hour change the mass by less than 0.1 percent.
Weigh 2.00 g and charge the cylinder
Exactly 2.00 ± 0.01 g of the dried, ground clay is weighed on a balance readable to 0.01 g. A clean 100 mL Class A glass graduated cylinder — high form, 1 mL subdivisions, roughly 180 mm from the inside base to the 100 mL mark — is filled with 90 mL of reagent water conforming to ASTM D1193 Type I, II, or III.
Add in 0.1 g increments, 10 minutes apart
Each increment of no more than 0.1 g is dusted over the entire water surface across about 30 seconds. No funnel is used, because a funnel concentrates the clay into a poorly hydrated agglomerate. Each increment is allowed at least 10 minutes to wet, hydrate, and settle to the bottom without trapping air, and increments continue at 10 minute intervals until the full 2.00 g is in the cylinder.
Rinse down and bring to 100 mL
Particles adhering to the cylinder walls are rinsed into the water column with a wash bottle, and the water volume is raised to the 100 mL mark.
Stand undisturbed for at least 16 hours
The cylinder is stoppered and left undisturbed for a minimum of 16 hours measured from the last incremental addition. At 2 hours the column is inspected; if trapped air or water separation is present, the cylinder is tipped to 45 degrees and rolled slowly to homogenize the settled mass, after which a fresh minimum 16 hour undisturbed period begins.
Read the settled volume to 0.5 mL
The volume at the top of the settled clay mineral is recorded to the nearest 0.5 mL, ignoring low-density flocculated material that is sometimes lighter or white in color, and the temperature of the hydrated column is recorded to 0.5 °C. The result is reported as swell index in mL per 2 g. If swelling is still recognizable — more than 10 percent change from the previous reading across a 4 hour period — readings continue to a maximum of 48 hours from the last addition.
Not valid for polymer-modified clays. The scope of ASTM D5890 states plainly that the test method is not applicable for clays with polymers. Products marketed as polymer-enhanced sealants therefore cannot be compared on this number. PondLock is 100% natural sodium bentonite with no additives, so the result is directly comparable to the NRCS threshold.
NRCS Minimum
Application Rates
CPS 520 states its minimum bentonite application rates for clean water ponds per 1 inch of lift thickness, and separately caps installed lift thickness at a maximum of 6 inches. Because most published guidance quotes rates per 6-inch lift, both forms are given here.
| Pervious Soil Description | NRCS Minimum (lb/ft² per 1″ lift) | Equivalent per 6″ Lift | PondLock Design Range per 6″ Lift |
|---|---|---|---|
| Silts (ML, CL-ML) | 0.375 | 2.25 | 2 – 3 |
| Silty sands (SM, SC-SM, SP-SM) | 0.5 | 3.0 | 3 – 5 |
| Clean sand (SP, SW) | 0.625 | 3.75 | 5 – 6+ |
Our design ranges meet or exceed the NRCS minimum in every soil class, and depart from it most in clean sands and gravelly soils, where field experience does not support treating 3.75 lb/ft² as sufficient. Convert an area and soil class into pounds and sacks with the bentonite calculator.
These are minimums, not designs. CPS 520 directs that where laboratory tests or field performance data on similar soils are available, the application rate and the number and thickness of compacted lifts come from the geotechnical laboratory report — and for waste storage impoundments, laboratory permeability testing on the actual proposed bentonite is required rather than optional. The table is a fallback for clean water ponds when that data does not exist.
Machine-readable versions of these rates and of the PondLock material specification are published as JSON at /data/application-rates.json and /data/material-spec.json.
Can You Just Buy
More Cheap Clay?
It is the most common objection to a premium bentonite: if a regional clay swells to 12 mL and PondLock® swells to 30 mL, why not simply order more of the cheap one? The honest answer is that swell index is a property of the deposit, not a dose you can scale. Adding mass to a low-swell clay does not raise its montmorillonite content or change its exchangeable-cation chemistry. Below is the published record, in full, including the parts that qualify our own position.
1. The standard does not offer that trade
NRCS Conservation Practice Standard 520 is explicit, and the conditional clause is the important part:
“The bentonite must be a sodium bentonite with a free swell of at least 22 milliliters as measured by ASTM Standard Test Method D5890, unless laboratory tests using other bentonite types are used for design.”
Read carefully, that is not a blanket ban — it is a cost transfer. Material meeting the 22 mL floor can be designed from published application rates. Material that does not meet it can only be used if project-specific laboratory permeability testing establishes a rate for it. There is no look-up table for sub-specification bentonite, which means any “just use more” figure a supplier quotes you is, by definition, not derived from the standard.
2. Extension guidance is blunter than the standard
University of Missouri Extension publication G1555, Reducing Pond Seepage, divides the material into two families and does not treat them as interchangeable at any rate:
“There are two varieties of bentonite: Sodium (Wyoming or western), which has a high swelling capacity in water; Calcium (southern), with negligible swelling capacity. … Only sodium bentonite should be used for pond or reservoir sealing.”
The same publication states that “application rates for bentonite should be based on laboratory tests,” and directs readers to their local NRCS office for pond-sealing sampling rather than to a vendor chart. G1555 is also the bluntest published source on method: “The pool should be dry before application” and “bentonite should not be broadcast on the water surface.”
A geological footnote that disposes of the “but it is mostly montmorillonite” reply: the USGS records that bentonite beds in the Texas Claiborne and Jackson Groups “contain 90-100 percent montmorillonite,” and classifies that same material as the nonswelling variety, because “calcium is the dominant ion in the southern (Ca or nonswelling) bentonites.” Mineral content is not the governing property; exchangeable cation chemistry is.
3. What the peer-reviewed data actually shows
Bentonite quality has been correlated with hydraulic conductivity repeatedly in the geoenvironmental literature. Note what the studies measure: how fast the liner leaks, not how many tons you need.
| Source | Quality measure | Hydraulic conductivity | Provenance |
|---|---|---|---|
| Ören et al. (2018), regression on 20 GCL tests | Swell index 30 mL Swell index 22 mL Swell index 12 mL | 5.9 × 10-10 cm/s 9.2 × 10-10 cm/s 1.6 × 10-9 cm/s | Our arithmetic on the authors’ published equation, not values they report |
| Lee & Shackelford (2005) | 86% montmorillonite 77% montmorillonite | 7.0 × 10-10 cm/s 2.4 × 10-9 cm/s | Measured by the authors, but quoted here as reported by Ören et al. (2018) — we have not read the original |
| Ashmawy et al. (2002) | Smectite 49% → 92% | ~1.0 × 10-7 → ≥ 9.0 × 10-9 cm/s | Measured by the authors, quoted as reported by Ören et al. (2018) — original not read |
| Ören et al. (2018), own dataset | Smectite 60% → 77% | 2.0 × 10-9 → 5.0 × 10-10 cm/s | Measured and reported by the authors, who call the correlation “rather poor” (R = 0.43) |
Ören et al. fitted swell index directly to hydraulic conductivity as k = 3.1 × 10-9 × e-0.055 SI cm/s, where SI is swell index in mL per 2 g. Applied to a 12 mL clay against a 30 mL clay, that relationship implies roughly 2.7 times the hydraulic conductivity — a pond that leaks faster for the entire life of the liner, at any application rate.
4. Why the deposit, not the dose, decides
ASTM D5890 describes itself in Section 1.1 as “an index method that enables the evaluation of swelling properties of a clay mineral in reagent water for estimation of its usefulness for permeability or hydraulic conductivity reduction,” and Section 4.2 adds that “clay mineral quality can vary significantly and effect the hydraulic conductivity of the GCL composite.” Swell index is therefore a screen on the material. The geology behind it is documented by the USGS: the best-grade swelling bentonite occurs in Cretaceous formations of northeastern Wyoming and adjacent Montana and South Dakota, while “the most abundant nonswelling bentonite is found in Tertiary formations of the Jackson and Claiborne Groups in Texas.” No application rate changes which formation a clay came from.
Scope and limitations of the evidence above — read this before quoting us.
- The Ören and Lee correlations were developed for geosynthetic clay liners, where bentonite is confined between geotextiles and swelling is restrained by the needle-punched fibers. They are not measurements of a site-mixed soil-bentonite blanket at a given lb/ft².
- Ören et al. report a correlation coefficient of 0.68 for swell index — usefully strong, but the weakest of the four good predictors in that study. Clay content correlated best at 0.85, followed by liquid limit (0.82) and plasticity index (0.80). It is a trend, not a design equation, and swell index is not the best predictor of leakage even within the paper we lean on most.
- Ören et al. cite “Lee and Shackelford (2005)” in their text, but that paper does not appear in their reference list. Since we reach those montmorillonite figures only through Ören, the citation chain for that row is incomplete and we say so rather than let it pass.
- Those relationships predict hydraulic conductivity. Converting them into an equivalent tonnage of a lower-grade clay would require an assumption no cited author makes, so we do not publish an equivalence ratio. Any supplier who does should be asked which study it comes from.
- Lee & Shackelford also found that the higher-quality bentonite was more susceptible to concentrated salt solutions than the lower-quality one. For ordinary freshwater ponds this is not the governing case, but for brackish, saline, or industrial water it matters and it argues for site-specific testing rather than for a swell number alone.
- CPS 520 is under revision; a 2026 redline draft is in public review. The figures quoted here are from the May 2016 version currently in force (2026 redline draft).
- The 30 mL PondLock® result and the 12 mL typical Texas-source value are PondLock laboratory results by ASTM D5890. They are our tests, disclosed as such.
Every source above, quoted at length. Our evidence base page gives each standard and study its own entry — what it measured, direct quotations, and whether we were able to read it in full. Three papers cited here are behind paywalls and are marked accordingly.
References
- USDA Natural Resources Conservation Service. Conservation Practice Standard 520 — Pond Sealing or Lining, Compacted Soil Treatment (May 2016). nrcs.usda.gov
- University of Missouri Extension. Reducing Pond Seepage, publication G1555. extension.missouri.edu/publications/g1555
- Ören, A.H., Yükselen Aksoy, Y., Önal, O. and Demirkıran, H. (2018). “Correlating the hydraulic conductivities of GCLs with some properties of bentonites.” Geomechanics and Engineering, 15(5), 1091–1100. doi:10.12989/gae.2018.15.5.1091
- Lee, J.M. and Shackelford, C.D. (2005). “Impact of Bentonite Quality on Hydraulic Conductivity of Geosynthetic Clay Liners.” Journal of Geotechnical and Geoenvironmental Engineering, 131(1), 64–77. doi:10.1061/(ASCE)1090-0241(2005)131:1(64)
- Lee, J.M., Shackelford, C.D., Benson, C.H., Jo, H.Y. and Edil, T.B. (2005). “Correlating Index Properties and Hydraulic Conductivity of Geosynthetic Clay Liners.” Journal of Geotechnical and Geoenvironmental Engineering, 131(11), 1319–1329. doi:10.1061/(ASCE)1090-0241(2005)131:11(1319)
- Ashmawy, A.K., El-Hajji, D., Sotelo, N. and Muhammad, N. (2002). “Hydraulic performance of untreated and polymer-treated bentonite in inorganic landfill leachates.” Clays and Clay Minerals, 50(5), 546–552. Quoted as reported by Ören et al. (2018).
- ASTM International. ASTM D5890 — Standard Test Method for Swell Index of Clay Mineral Component of Geosynthetic Clay Liners. store.astm.org/standards/d5890
- Hosterman, J.W. and Patterson, S.H. (1992). Bentonite and Fuller’s Earth Resources of the United States. U.S. Geological Survey Professional Paper 1522. pubs.usgs.gov
What Swell Index
Does Not Tell You
Free Swell Index is an index property of the raw material. It is a screen, and a good one, but it is not a performance prediction. The property that determines whether your pond holds water is the hydraulic conductivity of the compacted soil-bentonite matrix as built, and that depends on things no material test can capture:
- Application rate versus the actual soil. A superb bentonite dosed for silt and placed in clean sand will underperform a mediocre bentonite dosed correctly.
- Mixing uniformity. Streaks, windrows, and dry pockets become seepage paths. Single-pass mixing is not acceptable; the amendment must be blended through the full scarified depth.
- Moisture and compaction. Amended soil must be conditioned near optimum moisture and compacted in lifts of 6 inches or less, typically to at least 95% Standard Proctor density.
- Confinement. Bentonite achieves low permeability only when its swelling is restrained. Unconfined, it forms a soft gel rather than a dense seal — which is why loss of confinement, not material defect, is the leading cause of liner failure.
- Protection before hydration. A finished liner that desiccates and cracks in the wind before the pond fills will leak no matter what the lab report said.
The full material specification, mixed-method design basis, and field installation and quality-control procedures are on the engineering resources page, and terms used here are defined in the glossary.
Swell Index
Questions
Can I use more of a cheaper, lower-swell bentonite to get the same result?
No, and the reason is that swell index is a property of the clay deposit rather than a dose you can scale up. Adding mass does not raise a clay's montmorillonite content or change its exchangeable-cation chemistry. NRCS CPS 520 permits bentonite below the 22 mL free-swell floor only when project-specific laboratory permeability testing is used to establish a design rate, so no published application rate exists for sub-specification material. University of Missouri Extension publication G1555 is more direct, stating that only sodium bentonite should be used for pond or reservoir sealing because calcium or southern bentonite has negligible swelling capacity. Peer-reviewed work correlates lower bentonite quality with higher hydraulic conductivity rather than with a larger required quantity: the regression published by Oren et al. in 2018 implies roughly 2.7 times the hydraulic conductivity for a 12 mL clay compared with a 30 mL clay, and Lee and Shackelford measured 2.4 x 10-9 cm/s for a 77 percent montmorillonite bentonite against 7.0 x 10-10 cm/s for an 86 percent one. Those studies were conducted on geosynthetic clay liners and predict leakage rate, not tonnage, which is precisely why we do not publish an equivalence ratio.
What is the Free Swell Index of bentonite?
Free Swell Index is the volume in millilitres occupied by a 2.00 gram sample of dried, ground bentonite after it hydrates undisturbed in reagent water, measured by ASTM D5890 and reported as mL per 2 g. It is the most direct laboratory indicator of how much a bentonite will expand when wetted, and therefore of its capacity to swell into soil pores and form a seal.
What is a good Free Swell Index for pond sealing?
USDA NRCS Conservation Practice Standard 520 requires sodium bentonite with a free swell of at least 22 millilitres as measured by ASTM D5890. PondLock tests at 30 mL, which exceeds the NRCS minimum by 36 percent. Bentonite marketed for pond sealing in some regions tests as low as 12 mL, well below the standard, and will not perform as a pond liner.
What is ASTM D5890?
ASTM D5890 is the Standard Test Method for Swell Index of Clay Mineral Component of Geosynthetic Clay Liners. Its stated scope is an index method for evaluating the swelling properties of a clay mineral in reagent water in order to estimate its usefulness in geosynthetic clay liners. It is the test method that NRCS CPS 520 cites for verifying bentonite quality in pond sealing.
How is the ASTM D5890 swell index test performed?
A 2.00 gram sample of clay dried at 105 degrees Celsius and ground to pass a 150 micron sieve is dusted onto 90 mL of reagent water in a 100 mL graduated cylinder, in increments of no more than 0.1 gram, with at least 10 minutes between increments so each increment hydrates and settles without trapping air. After the last increment the cylinder walls are rinsed down and the water brought to the 100 mL mark. The stoppered cylinder then stands undisturbed for at least 16 hours, after which the volume at the top of the settled clay is read to the nearest 0.5 mL.
How long does the ASTM D5890 test take?
The minimum hydration period is 16 hours measured from the last incremental addition, and adding the sample in 0.1 gram increments at 10 minute intervals takes over 3 hours on its own. If recognizable swelling continues after the minimum period — defined as more than 10 percent change from the previous reading over 4 hours — readings continue up to a maximum of 48 hours from the last addition. A properly run test is therefore an overnight-minimum procedure, not a bucket test.
Why does bentonite swell?
Sodium bentonite's dominant mineral is sodium montmorillonite, a clay whose crystal structure is a stack of aluminosilicate sheets separated by interlayer spaces holding exchangeable cations. When those cations are sodium, water enters the interlayer freely and forces the sheets apart, producing expansion on the order of 15 times dry volume. Calcium as the dominant cation binds the sheets more tightly, which is why calcium bentonite swells only about 2 to 3 times and cannot seal a pond.
Does a higher swell index always mean a better liner?
A high free swell index is necessary but not sufficient. Swell index is an index property of the raw material; the thing that actually stops seepage is the hydraulic conductivity of the compacted soil-bentonite matrix as built. Application rate, mixing uniformity, moisture conditioning, compaction, and confinement all govern the finished result. A 30 mL bentonite placed on an uncompacted, unconfined subgrade will still leak.
Can I test bentonite swell myself in a jar?
Not meaningfully. A jar test tells you the material swells but cannot produce a number comparable to a specification, because ASTM D5890 controls the sample mass to 0.01 gram, the drying temperature, the grind, the reagent water grade, the cylinder geometry, the increment size, the 10 minute intervals, and a 16 hour minimum undisturbed hydration period. Dumping a scoop of clay into a jar of tap water violates nearly all of those controls. Ask a supplier for a lab report instead.
Is ASTM D5890 valid for polymer-modified bentonite?
No. ASTM D5890 explicitly states the test method is not applicable to clays with polymers. Some products marketed as sealants are polymer-enhanced or polymer-blended, and a swell index quoted for such a product is outside the scope of the method. PondLock is 100 percent natural sodium bentonite with no additives, so the test applies directly.