How this instrument works
Heat capacity belongs to one object rather than to whatever substance it is built from, and it scales with how much of that object you have — two identical copper blocks swallow twice as many joules as one. One kilogram of water sits at 4186 J/K, a 350 g stoneware mug near 300 J/K, a 300 kg cast-iron engine block around 135 000 J/K, and 150 litres of domestic hot water roughly 628 000 J/K. People who design buildings call this quantity thermal mass and prize it for flattening out afternoon peaks: thick concrete floors absorb sunlight all day for barely two degrees of their own, then give it back overnight.
Calorimeters exist to measure it. Lavoisier and Laplace built the first serious instrument in 1782, nesting one sample chamber inside shells of packed ice and weighing whatever meltwater ran out of its tap. Bomb calorimeters still need the capacity of their entire assembly — steel vessel, water bath, stirrer, thermometer, everything together — before any fuel measurement means anything, and that figure comes from burning pellets of certified benzoic acid whose combustion energy is already known, then dividing by whatever rise it produced. Older laboratories called that result a water equivalent, since 'this rig behaves like 2.4 kg of water' is far easier to picture than 10 000 J/K written plain.
C = Q ⁄ ΔT treats capacity as steady across your interval, which holds only while nothing dramatic happens to the sample. Cross any melting or boiling point and ΔT stalls at zero while energy keeps pouring in, so this ratio runs away — phase changes have to be budgeted separately as latent heat rather than folded into one figure. Head toward absolute zero instead, and capacity collapses: it falls roughly as T³ in crystalline solids and must vanish entirely at 0 K, which is what the third law of thermodynamics demands and why no finite sequence of steps reaches that temperature. Between those two extremes, drift is mild but real.
- Put the energy you delivered into Heat added, which switches between joules, kilojoules, calories, kilocalories and watt-hours.
- Enter the climb you observed into Temperature rise (K or °C) — 18.4 °C becoming 19.4 °C is 1, never 19.4.
- Read Heat capacity (J/K) — the joules that object will demand for every further kelvin you ask of it.
- Working from cooling data, enter both figures negative; capacity is a property and should never come back negative.
- To reach specific heat, divide the result by sample mass in kilograms. To isolate the sample, subtract the vessel's own capacity first.
Worked example — calibrating a bench calorimeter
An immersion heater inside a bench calorimeter runs at 41.86 W for exactly 100 seconds, delivering 4186 J. A thermometer read to a thousandth of a degree shows 20.000 °C becoming 21.000 °C. Heat added = 4186, Temperature rise (K or °C) = 1, so Heat capacity (J/K) = 4186 ⁄ 1 = 4186 J/K.
That answer is a landmark rather than a coincidence. 4186 J/K is exactly what one kilogram of water offers, and that equivalence is how calories were pinned down long before joules took over the job. This apparatus therefore has a water equivalent of one litre, and every later reading taken in it can be scaled against that. Push 8372 J in and watch a 2 K rise, and the answer holds at 4186 — capacity describes the apparatus, not a tally of what you poured through it.
Questions
How do I turn J/K into J/kg·K?
Divide by sample mass in kilograms. Reading 4186 J/K off a 1 kg specimen means 4186 J/kg·K; that identical reading off a 4 kg specimen means 1046. Running it forwards instead, C = m·c predicts any object's capacity from a table value — 2 kg of aluminium at 900 J/kg·K comes to 1800 J/K. Divide by moles rather than kilograms and you land on molar heat capacity, which is the version chemistry tables usually carry.
Should Temperature rise be entered in kelvin or Celsius?
Either, as long as you enter change rather than reading. Both scales climb in identically sized steps, so 1 K of rise and 1 °C of rise mean the same entry here. What wrecks the calculation is typing an absolute value: your sample finishing at 21 °C did not rise by 21 unless it began at precisely 0 °C. Subtract first, then enter. Fahrenheit needs converting — a rise of 1 °F is 5/9 K.
Why does my sample's capacity come out too high?
Usually because the container is being counted alongside its contents. Whatever you warm sits in a vessel, and that vessel, its stirrer and its probe each take a share of the joules. Calorimetry deals with this by characterising the empty rig first, then subtracting that water equivalent from every later total. Leakage is the other habitual culprit: warmth escaping during a slow run makes the observed climb smaller than it ought to be, and small denominators inflate your answer. Short runs and good lagging both help.
Does an object have one heat capacity or two?
Two, strictly — one measured with pressure held constant, one with volume held constant, written Cp and Cv. For solids and liquids they part company by fractions of one percent, because so little expansion occurs, so one quoted figure serves fine. Gases are a different story: sealed in a rigid vessel, a gas takes roughly 30 percent less energy per kelvin than the same gas left free to shoulder the atmosphere aside, since that expansion does work on its surroundings. State which condition applies whenever gases are in play.
How long will it take to warm something up?
Divide C by your heating power. A 4186 J/K vessel driven by a 500 W element needs 4186 ⁄ 500 ≈ 8.4 seconds per kelvin, so 20 K wants about three minutes — in the ideal case where nothing escapes. Real bodies shed warmth as they climb, and the loss grows with the gap they have opened, so the approach is exponential rather than straight. Divide C by your loss coefficient in W/K for one time constant: bodies cover roughly 63 percent of their remaining gap in each of those, and have effectively settled after five.
Is heat capacity ever genuinely constant?
Never quite, and near a phase transition it stops behaving at all. Melting ice absorbs 334 kJ per kilogram while pinned at 0 °C, so ΔT is zero and C formally becomes infinite — one spike on that curve rather than any number you can use. Away from transitions the variation is gentle: most solids ease upward as they warm, then fall away steeply below about 50 K on their journey to zero. Across the everyday span between freezing and boiling, treating C as fixed costs a percent or two, which is well inside most measurement error anyway.