Calibrating a Thermometer with a Suspect Ice Bath
A student calibrates a digital thermometer before a calorimetry lab. The procedure states: prepare an ice-water slurry in a 250 mL beaker, stir for 3 minutes, and record the temperature as the 0.0 °C reference. The student records 0.6 °C. The thermometer's stated accuracy is ±0.2 °C. The student then measures the lab's room-temperature water as 22.4 °C. A classmate suggests simply subtracting 0.6 °C from all future readings. The ice bath was made with tap water and ice from a freezer, and the student admits they did not stir the slurry before inserting the probe.
1. Should the student accept 0.6 °C as a calibration offset and subtract it from later readings? Justify your reasoning using the evidence in the scenario.
Read the worked answer
No - the 0.6 °C result is not trustworthy enough to serve as a calibration offset.
- The reference itself is suspect. A proper ice-point bath requires a well-stirred slurry of crushed ice and water in equilibrium. The student did not stir, so the probe may have sat in a pocket of warmer water or against unmelted ice, neither of which is guaranteed to be at 0.0 °C.
- The deviation is small relative to uncertainty sources. 0.6 °C is only modestly above the ±0.2 °C spec, and an improperly prepared bath can easily produce an apparent offset of that size.
- Subtracting a bad offset corrupts all later data. If the 0.6 °C is an artifact of poor technique, applying it introduces a systematic error in the opposite direction.
Correct action: redo the ice bath with crushed ice, distilled water, continuous stirring, and the probe centered (not touching the beaker walls or ice chunks). Only if a repeatable offset persists across several properly prepared baths should a correction be considered - and even then, the offset's uncertainty must be reported alongside the corrected values.
2. The student wants to report the room-temperature reading with an uncertainty. What should they report, and what reasoning supports it?
Read the worked answer
The student should report 22.4 °C ± 0.2 °C (the manufacturer's accuracy), not a value corrected by the unreliable 0.6 °C offset, and should note that the calibration was inconclusive.
Reasoning:
- Uncalibrated instrument → use the spec limit. With no valid calibration, the best uncertainty estimate is the stated accuracy, ±0.2 °C.
- Do not propagate an invalid correction. Applying an offset derived from a flawed reference would understate uncertainty and bias the result.
- Report the limitation. A defensible report states that the ice-point check gave 0.6 °C under non-ideal conditions and that the calibration was not confirmed, so the reading carries the full spec uncertainty.
This models good scientific practice: distinguish between random scatter and systematic offset, and never claim a correction you cannot justify.