Guide
Testing calcium carbide on delivery: sampling sieve analysis, gas yield and when to reject a lot
Calcium carbide is accepted on three IS 1040 tests: sieve analysis for size and dust, a gas-yield test on exactly 100 g at 27 °C, and analysis of the evolved gas for PH₃, H₂S and purity. Sample sealed drums in about 5 minutes each and judge the lot on its average.
- Gas-yield test
- Exactly 100 g crushed to pass 4 mm
- Sampling
- Unopened original packages
- Acceptance
- At least one gas-yield test per sample and three per lot; the lot conforms only if the average meets the tables
- Sample tolerance
- An individual sample may be 5 % below the tabulated yield provided the lot average conforms
- Delivery tolerance, historical
- −5 % British
A delivery of calcium carbide is worth what it will yield in your generator, and the only way to know that before it goes in is to test it. IS 1040, the Indian specification for technical carbide, prescribes three tests: a sieve analysis for size and dust, a gas-yield test on exactly 100 g, and an analysis of the gas evolved for acetylene content and for phosphine, hydrogen sulphide, arsine and ammonia. It also prescribes how to sample, which matters more for carbide than for almost any other bulk material, because the sample starts deteriorating the moment it meets the air. This page sets out the IS 1040:1987 method, the historical methods it descends from, and the acceptance rules a buyer should write into a contract.
Rules that matter
The S. A. Miller (1965) compilation of the world's carbide specifications opens with a complaint about serious confusion in sampling, grading, gas-yield requirements and test techniques. On one point every specification agreed: sampling must be done quickly under dry conditions, and the final samples placed in hermetically sealed dry containers of a size chosen so that the solid fills the container completely. No headspace means no humid air.
IS 1040:1987 Appendix D sets the Indian procedure:
Sampling: quickly, dry, and with no headspace
- A lot is all containers of one quality and graded size from one batch of manufacture.
- Containers are selected at random; the number rises with lot size from about 2 to 10.
- Sample only from unopened original packages, in a closed space, avoiding moisture and bare hands, with clean dry tools and airtight containers, each package sampled within about 5 minutes.
- Packages of 50 kg or more are opened one at a time; smaller packages are opened until about 50 kg has been collected. The material is spread and portions taken from four places, split into a test sample for the purchaser and a referee sample.
- For gas yield, a composite of about 10 kg per container is crushed below 4 mm, mixed and quartered, with opposite quarters forming the referee sample.
The referee sample is the buyer's protection in a dispute. It should be sealed, labelled with the lot number and date, and stored dry until the lot is consumed or the dispute is settled.
Sieve analysis: size and dust
IS 1040 grades with round-hole sieves, following long international practice in this industry, and uses a square-hole 4 mm sieve only for preparing gas-yield and impurity samples. Appendix A stacks the graded sieves above a dust sieve, 1.0 mm for Quality A and B carbide and 0.5 mm for Quality C, weighs to 0.1 % accuracy, shakes the nest 30 times each way inclined at 30°, and reports retained and passing mass as percentages. For the 4–80, 25–80, 50–80 and 80–120 mm sizes the larger pieces are tested singly first. Size grading is accepted on the average of all grading and dust results.
Sample mass matters. Grading is meaningless on a small sample, and the historical standards used large ones: the proposed British revision of the 1960s used 20 kg for sizes above 15 mm and 10 kg below; the German standard used 50 kg for 25–80 mm and a separate dust determination on the whole contents of a drum. A works test on a 2 kg scoop from the top of an open drum tells you mainly about the fines that settled there.
What to look for is set by your generator, not by the standard. IS 1040 recognises graded sizes up to 80 mm and beyond, but a generator's own manual states the size range it is designed to take, and KIGL's plant manuals give recommended sizes by model; buy to that range. Oversize causes bridging at the hopper neck and unreacted discharge; undersize and dust react too fast, foam, carry over and coat. See carbide grades, sizes and gas yield.
The gas-yield test
IS 1040:1987 consolidated earlier alternatives into a single gravimetric method, Appendix B:
| Step | Requirement |
|---|---|
| Preparation | Crush material above 4.0 mm to pass a 4.0 mm square-hole sieve; remix with the rest including the dust |
| Generating flask | About 1,000 ml, in a water bath at or below 27 °C, immersed at least 75 mm |
| Water feed | Separating funnel of about 350 ml with a 1.5 mm jet, sodium chloride solution about 20 % (m/v) |
| Gas train | Anhydrous calcium chloride drying tube; two 350 ml bubblers each with 90 ml of sodium hypochlorite at 1.0–2.0 % available chlorine |
| Sample | Exactly 100 g |
| Rate | 45 ± 5 bubbles per minute until decomposition is complete, about 5 to 6 hours |
| Result | Gas yield, l/kg = 940.3 m / M, where m is the mass of gas evolved in g and M the sample mass in g |
The constant reflects the reference state: 1 g of acetylene measures 0.9403 litres at 27 °C and 760 mm. Older British practice used 0.9087 l/g at 15.5 °C, and the proposed revision 0.9017 l/g at 15 °C, so a yield certificate must state its temperature before it can be compared with another.
Acceptance under cl D-4 requires at least one gas-yield test per sample drawn and at least three tests per lot, and the lot conforms only if the average meets the yield table. Under cl 5.2 an individual sample may be 5 % below the tabulated value provided the lot average conforms.
Why methods disagree
Miller's review of the gas-yield methods identifies four families, each with a known bias:
| Method family | Used by | Bias |
|---|---|---|
| Gravimetric, water onto carbide | BS 642:1931; IS 1040 | Slightly low: the initial temperature rise condenses some acetylene; reproducibility fairly good |
| Gravimetric, carbide into water | Proposed BS revision | Can be very accurate and is quicker; net correction only 0.02 g |
| Volumetric, with pre-saturated water | DIN, NF, US and others | Slightly high: pre-saturation overcompensates; needs accurate temperature and pressure |
| Manometric | Works control laboratories | Fast and needs less skill; low accuracy |
Laboratory repeatability is a separate number from delivery tolerance. The Knapsack-Griesheim microgasification method on 5 g samples (Mewes, Chem.-Ing.-Tech., 1951) requires three determinations within 1 % of one another; the US gas-meter method required three 50 g determinations spread by less than 0.2 cu ft/lb, about 4.4 %. The tolerance below standard yield before a buyer could claim was 5 % in the British specification and 3 % in the German, French and international ones, with none in the Japanese. A contract that says 3 % should say which of these it means.
Gas quality
IS 1040:1987 Table 3 limits the gas evolved from the carbide, by the standard's own gasification method, to at least 99.0 % acetylene; at most 0.15 % hydrogen sulphide; 0.06 % phosphine for Qualities A and C and 0.08 % for Quality B; 0.001 % arsine; and 0.10 % ammonia, all by volume. The test methods are an oleum and silver sulphate absorption for acetylene, gravimetric barium sulphate for sulphur, a volumetric or referee colorimetric method at 420 nm for phosphorus, a modified Gutzeit test for arsenic, and acid absorption for ammonia. Cl 5.3.1 adds that the gas must not contain impurities liable to make it ignite spontaneously in air.
Gasification mode changes the impurity result: the same carbide gives different phosphine, sulphide and ammonia figures when gasified water-to-carbide and carbide-to-water. A supplier's certificate is comparable with your own result only if the method matches. For a plant, the practical phosphine check is the silver nitrate test on the raw gas when a new lot goes in; a sharp rise in purifier loading on a new delivery is a carbide-quality finding. Why phosphine matters so much more than sulphur is explained in phosphine and Indian carbide.
Rules that matter
EIGA Doc 239 calls for a carbide quality audit whenever the carbide supplier changes. A practical acceptance specification covers:
What to reject, and what to write in the contract
- Graded size matching the generator, with maximum oversize and maximum dust stated.
- Minimum gas yield at a stated temperature and method, with the tolerance and whether it applies to individual samples or to the lot average.
- Maximum phosphine in the evolved gas.
- Packing to IS 1040: sealed moisture-proof steel drums, one size per drum, with the packing date and lot number.
- Physical condition on arrival: no dented, holed, rusted or wet drums; no drum that hisses heavily on opening.
- A referee sample retained, and the procedure for disputes.
Reject outright any drum that has been wet. Treat with suspicion a lot that tests well on yield but arrived in damaged drums: the test sample came from sound drums.
Rules that matter
Rules to apply
- Sample unopened drums, fast, dry, with no headspace in the sample jar.
- Grade on large samples, and grade against what your generator accepts.
- State the yield test's temperature and method, and judge the lot on its average.
- Keep a sealed referee sample for every lot.
- Track yield per kg on the shift log by lot, so the plant confirms what the laboratory predicted; see why carbide gas yield falls.
The KIGL acetylene gas plant manuals reference IS 1040 for carbide feed; for buying the finished gas, see buying acetylene — what to check.
Testing calcium carbide on delivery — frequently asked questions
How do you test the gas yield of calcium carbide?
IS 1040:1987 crushes the sample to pass a 4 mm square-hole sieve, takes exactly 100 g, and gasifies it with water fed from a separating funnel into a flask in a bath at or below 27 °C, drying the gas over calcium chloride. The mass of gas evolved gives the yield as 940.3 times the mass of gas divided by the sample mass, in litres per kg at 27 °C. A test takes about 5 to 6 hours.
How should a carbide sample be taken?
Quickly and dry. IS 1040 requires sampling only from unopened original packages, in a closed space, with clean dry tools, no bare hands and airtight sample containers, each package sampled within about five minutes. Every historical carbide standard insists the sample container be chosen so that the solid fills it completely, leaving no headspace for humid air.
What is the delivery tolerance on carbide gas yield?
Under IS 1040:1987 an individual sample may be 5 % below the tabulated yield if the lot average conforms. Historically the British specification allowed 5 % and the German, French and international 3 %. The 3 % figure is a delivery tolerance below which a buyer can claim; it is not the precision of the laboratory test.
Why do two laboratories get different carbide yields?
Because the methods have built-in biases. Gravimetric water-to-carbide tests read slightly low, as the first temperature rise condenses some acetylene; volumetric tests with pre-saturated water read slightly high; manometric works tests are quick but imprecise. Temperature basis also differs: IS 1040 at 27 °C, older standards at 15 °C, about 4 % apart.
When should a carbide lot be rejected?
When the lot average fails the specified yield beyond the agreed tolerance, when the size analysis shows oversize your generator cannot take or dust above the agreed limit, when drums arrive damaged or wet, or when the evolved gas fails the phosphine limit your purifier is sized for. Write the rejection limits into the purchase contract before delivery.
Do I need to test every drum?
No. IS 1040 treats a lot as all containers of one quality and graded size from one batch, and selects containers at random, with sample numbers rising with lot size. Composite the samples, crush, quarter, and keep opposite quarters as a sealed referee sample in case of dispute.
References
- Bureau of Indian Standards. (1987). Calcium Carbide, Technical (scan, Public Resource BIS mirror) (IS 1040:1987, Third Revision).
- Bureau of Indian Standards. (2024). Calcium Carbide, Technical — Specification (IS 1040:2024).
- European Industrial Gases Association. (n.d.). Mechanical Integrity of Generator Systems (carbide quality audit on change of supplier) (EIGA Doc 239).
Enquiries
Request a quotation
State the gas or equipment, the quantity or capacity, and your location. KIGL will confirm the grade, the cylinder or plant size, and whether the location can be served from Noonmati, or through the agent network.