Significant Figures in Theoretical Yield

A student weighs 5.00 g of a reagent on a balance that reports to two decimal places. The theoretical yield calculator returns 7.500 g from that input. The question is whether the calculator should show 7.50 g, 7.5 g, or 8 g, and at what stage in the stoichiometry that decision gets made.

The answer is that rounding happens only at the end, after the final multiplication, and the number of significant figures in the theoretical yield matches the least precise input measurement. Rounding any intermediate value, such as moles of limiting reactant, before that final step introduces cumulative error that misrepresents the precision the chemistry workflow actually supplies. The rule, drawn from OpenStax Chemistry 2e section 1.5 and IUPAC guidelines, is that a reported yield must reflect the precision of the input measurements, not create the illusion of accuracy that was never there.

Take a worked example. You have 5.00 g of reactant A. That measurement carries three significant figures. Its molar mass is 100.0 g/mol, which carries four. The stoichiometric ratio is 1:1, and the product molar mass is 150.0 g/mol, also four significant figures.

The first step is to find moles of A: 5.00 g divided by 100.0 g/mol equals 0.0500 mol. That intermediate result has three significant figures, matching the least precise input so far. The next step is to calculate the theoretical yield: 0.0500 mol multiplied by 150.0 g/mol equals 7.50 g. That result has three significant figures, matching the limiting input measurement. The reported yield is 7.50 g, not 7.500 g and not 8 g.

If you had rounded the moles to 0.05 mol before the final multiplication, you would have lost one significant figure and calculated 0.05 times 150.0 equals 7.5 g, which is still correct here, but only because the trailing zero was not needed. In a case where the product was 7.04 g, rounding the intermediate to 0.05 mol would give 7.5 g instead of 7.04 g, a distortion of 0.46 g. That is the cost of early rounding.

Consider the failure case. The mass of reactant A is 5.0 g. That is two significant figures. The molar masses remain 100.0 g/mol and 150.0 g/mol, each with four. The least precise input in this workflow is the mass at two significant figures.

Moles of A: 5.0 g divided by 100.0 g/mol equals 0.050 mol. That result has two significant figures. Theoretical yield: 0.050 mol times 150.0 g/mol equals 7.5 g. The reported yield must be 7.5 g, not 7.50 g. Writing 7.50 g implies three significant figures, which the 5.0 g input does not support. The chemistry workflow supplies the precision constraint: the balance that reported 5.0 g can distinguish only to the nearest 0.1 g, so the yield cannot claim accuracy to 0.01 g.

A common mistake is to report the full calculator output of 7.5000 g. That number implies five significant figures, which no input in this example possesses. The result would mislead anyone reading the yield as a valid measurement. The remedy is always to round the final theoretical yield, never an intermediate step, and to count significant figures from the limiting reactant, not from the product molar mass.

The chemistry workflow supplies the precision constraints through the measuring instruments. A balance that reads to 0.01 g gives three significant figures for a mass of 5.00 g. A balance that reads to 0.1 g gives two for 5.0 g. A volumetric pipette that delivers to 0.01 mL gives a certain precision. The reported theoretical yield must never imply greater accuracy than the least precise of these inputs.

A limitation of this rule is that it does not account for systematic errors. A side reaction that consumes 10% of the limiting reactant changes the actual yield, but the significant figure rule only governs how precisely the calculated maximum is reported. The two concerns are orthogonal: the rule preserves the honesty of the number, not the accuracy of the reaction.

The most common failure in rounding theoretical yields is the half-up assumption. A student taught in primary school that 2.5 rounds to 3 (half up) will expect a calculator to do the same. But Python's round() uses round half to even, so 2.5 rounds to 2. JavaScript's Math.round() uses half away from zero, so -1.5 rounds to -2, not -1. The chemistry workflow does not care which tie-breaking rule the software uses as long as the final yield is rounded to the correct number of significant figures, because the rule does not depend on the method of rounding, it depends only on the count of digits.

The floating-point surprise is another failure mode. A number displayed as 0.3 may be stored as 0.2999999999999999 in IEEE 754 binary64. Rounding that stored value to one decimal place can give 0.2 if the tie-breaking rule does not handle the tiny error. In a theoretical yield calculation, this could shift the last digit by one. The fix is to round only the final computed value, not intermediate floating-point results, and to use a consistent rounding method across the workflow.

This treatment of significant figures in theoretical yield suits any student or practitioner who needs to report a stoichiometry result that honestly reflects the precision of their measurements. It suits primary and middle-school students learning to round whole numbers and decimals, high-school chemistry students applying sig-fig rules from OpenStax Chemistry 2e, and parents checking homework. It also suits office and finance users who need to understand how Excel's ROUND, ROUNDUP, and ROUNDDOWN functions behave, and programmers who need to know the difference between Python's round half to even, JavaScript's half away from zero, and IEEE 754's default roundTiesToEven.

It does not suit anyone looking for a philosophical history of approximation, this is a tool and a rulebook, not a historical treatise. It does not suit advanced numerical-analysis work on error propagation in Monte Carlo simulations; that requires a text like Numerical Recipes. And it does not suit currency rounding for tax or legal compliance in a specific jurisdiction, because those rules are jurisdiction-specific and not covered here. The single thing that most often goes wrong is rounding an intermediate value before the final multiplication, which creates a reported yield that does not match the input precision and misleads anyone who reads the result as a valid measurement.