Greenhouse Effect Experiment: Two Jars and a Lamp

A greenhouse effect experiment with two jars under a lamp is the most-copied science demo on the internet, and almost every version of it tells kids something false. The jar with carbon dioxide does get warmer. That part is real and you will measure it. But the reason it warms on your kitchen table is mostly not the reason Earth warms, and a 2010 physics paper proved it with argon. That gap is the best thing here. You run the demo, get the result, then take it apart with your kid.

Key Takeaways

• You will seal two identical jars under one lamp, fill one with carbon dioxide poured from a cup, and log both temperatures every two minutes for 20 minutes.
• Expect the carbon dioxide jar to end 1 to 3 °C warmer. Expect that difference to come mostly from suppressed convection, not from infrared absorption.
• Best for ages 9 to 15, because the payoff is the critique, not the colour change.
• The usual failure: the two thermometers are not actually identical, or the jars sit at different distances from the lamp.
• You get to say the sentence most science articles never say: our model gives the right answer for the wrong reason.

What You’ll Need

Item	How much	Where to get it	If you don’t have it
Identical glass jars with wide mouths	2, about 1 litre each	Pasta sauce or pickle jars, washed and dry	Two identical drinking glasses plus cling film
Thermometers	2, ideally digital with 0.1 °C resolution	Pharmacy, kitchen shop, or cheap aquarium probes	Two identical cooking thermometers; do not mix types
Desk lamp with an incandescent or halogen bulb	1, 40–60 W	Any lamp with an exposed bulb	A 100 W equivalent LED floodlight works but heats less
Baking soda	1 tablespoon (about 15 g)	Kitchen cupboard	Two Alka-Seltzer style tablets
White vinegar	100 ml	Kitchen cupboard	Lemon juice, slightly weaker
A tall cup or small jug	1	Kitchen	Any 300 ml container
Cling film and rubber bands	Enough to seal 2 jars	Kitchen	Freezer bag cut open, taped
Ruler and a notebook	1 each	Home	Phone notes app and a tape measure
Black paper	2 sheets, same colour	Craft drawer	Two identical pieces of dark cloth

Time: prep 10 min, experiment 25 min, cleanup 5 min Cost: everything is probably in your kitchen Mess level: low, with one fizzy moment over the sink

Do the Experiment

Step 1 — Line both jars identically

Cut the black paper so each jar gets an identical piece curled against the inside of the back wall. Dark surfaces absorb more visible light and convert it to heat, which gives you a bigger signal. Both jars must get the same treatment. Any difference you leave between them shows up in your data, and you will not know whether the gas caused it.

Step 2 — Put a thermometer in each jar at the same height

Tape each probe so the tip sits in the middle of the jar, about 5 cm above the base, touching neither glass nor paper. Glass conducts heat differently from air, so a probe against the wall reads the jar, not the gas.

Step 3 — Make carbon dioxide in a separate cup

Put the baking soda in the tall cup. Over the sink, pour in the vinegar. It will foam hard for about 30 seconds. The gas filling the cup above the foam is carbon dioxide, and it is about 1.5 times denser than air, so it stays put like an invisible liquid. Never do this reaction inside a jar you intend to seal. A sealed container with gas still being generated builds pressure, and glass under pressure is not a toy.

Step 4 — Pour the gas into jar B

Tilt jar B slightly and pour the cup’s headspace in the way you would pour water, slowly, with no foam. Repeat with a second batch if you can. Seal both jars with cling film and a rubber band within ten seconds of each other.

Step 5 — Set the lamp dead centre

Measure with the ruler: both jars 25 to 30 cm from the bulb and the same distance from each other. Shift the lamp 3 cm to one side and the nearer jar receives noticeably more light, and the inverse-square law hands you a fake result.

Step 6 — Ask for the prediction, then switch the lamp on

Before the light goes on, have your kid write down which jar will get warmer and by how much. A prediction written afterwards is not a prediction. Then log both temperatures every two minutes for 20 minutes.

Step 7 — Switch the lamp off and keep logging for 10 minutes

This is the step everyone skips and it is the most interesting one. Watch which jar cools faster.

What you should see: Both jars climb steadily for the first ten minutes, then flatten out. The carbon dioxide jar typically ends somewhere between 0.5 and 3 °C warmer. On a good run with digital probes the gap is obvious by minute eight. On a run with two mismatched glass thermometers you may see nothing at all, because the gap is smaller than the instruments can resolve.

What’s Actually Going On in the Greenhouse Effect Experiment

Two separate physical stories are tangled together in this jar, and the famous one is the smaller of the two.

The real atmospheric greenhouse effect is radiative. Sunlight arrives mostly as visible and near-infrared light, passes through the atmosphere, and warms the ground. The warm ground radiates that energy back out as longwave infrared, peaking near 10 micrometres. Carbon dioxide and water vapour have molecular vibration modes that absorb at those wavelengths and re-emit in all directions, including downward. The altitude from which Earth finally radiates to space moves higher, where it is colder, so the surface must run warmer to balance the books. Strip the carbon dioxide out, NASA notes, and Earth’s surface temperature would fall by roughly 33 °C, which is why the natural greenhouse effect holding the planet near a 15 °C average matters (NASA Science).

Now the jar. Paul Wagoner, Chunhua Liu and Roger Tobin tested this exact class of demonstration in the American Journal of Physics. Their conclusion is blunt: the warming comes mainly from carbon dioxide’s greater density suppressing convective mixing with room air, and the genuine radiative greenhouse effect is “very small for most tabletop experiments.” Their control was argon, which nearly matches carbon dioxide’s density but does not absorb infrared. Argon warmed the jar too. The gas was acting like a thicker blanket, not a selective filter.

There is a second irony worth handing your kid. A real glass greenhouse, the kind with tomatoes in it, also warms mostly by blocking convection. So the jar demo models a greenhouse well and the atmosphere badly, even though “greenhouse effect” is the name we use for the atmospheric one. The name was a loose analogy from the start.

None of this makes the atmospheric effect doubtful. Satellite spectrometers measure it directly, looking down at the absorption notch carbon dioxide cuts in Earth’s outgoing infrared, and the gas is tracked continuously: NOAA’s Global Monitoring Laboratory recorded a global annual mean of 425.62 ppm in 2025, up from 399.65 ppm a decade earlier. What the jar cannot do is show you that mechanism. A 20 cm column of gas does not hold enough molecules.

How to Explain It to Your Kid

Ages 5–8: the blanket that only some things can get through

Skip the jars. Shine a torch on their hand so they feel the warmth, then drape a thin blanket over the hand and shine again. The light does not get through, so the hand stays cool. Now warm that hand under the blanket with your own and ask where the heat goes. It stays in. The sky holds the planet’s warmth the same stay-in way, and more of certain gases makes the sky better at holding it.

Ages 9–12: a gas that is picky about light

Introduce the word infrared: light we cannot see, the kind a TV remote uses. Have them picture a window that lets sunlight in but will not let one invisible colour back out. Then show them their own data and ask the hard question: how do we know it was the carbon dioxide and not just that the gas was heavier? If they cannot answer, that is the correct answer, and a real physicist had to run argon to settle it.

Ages 13+: the radiating altitude

Give them the energy balance properly. Earth absorbs roughly 240 watts per square metre averaged over its surface and must radiate the same amount back or its temperature changes. Greenhouse gases do not block that outflow; they move it up to a colder layer, and a colder radiator needs a warmer surface beneath it to emit the same power. Then have them design the control: what gas isolates radiation from density, and where do you get it? Welding suppliers sell argon.

The question to ask: “If we swapped the carbon dioxide for helium, which is much lighter than air, what would happen to the temperature in that jar, and what would that tell us?”

Why Teaching This Early Pays Off

What this experiment pre-loads is not the greenhouse effect. It is the difference between a model and the thing it models, the single most useful idea in secondary-school science and the one most often skipped. The National Research Council’s Framework for K-12 Science Education lists “developing and using models” among eight core science practices, and asks students to identify a model’s limitations, not just its predictions.

There is also a documented reason not to simply correct a wrong idea and move on. Smith, diSessa and Roschelle argued that a learner’s wrong intuitions are productive raw material rather than errors to delete, and that instruction works better refining what a child already believes than replacing it. “Heavier gas, less mixing” is a good intuition. It just answers a different question than the one on the poster.

Be honest about the ceiling. One afternoon with two jars does not make a climate-literate teenager, and no source here claims a single activity moves school performance. What it plausibly does is give your kid a concrete memory to attach the real mechanism to later, plus the habit of asking what else could explain a result.

Before You Start: What Parents Should Know

Best age	9 and up. Below that, the measurable difference is too small and the interesting part of the lesson is about experimental control, which does not land yet.
Supervision	Adult handles the lamp and the vinegar pour. Incandescent and halogen bulbs reach temperatures that blister skin.
Real hazards	Hot bulb; glass jars near a heat source; a fizzing acid-base reaction that can splash into eyes if a face is directly over the cup.
Don’t	Never run the baking soda and vinegar reaction inside a jar you then seal, and never add more gas to an already sealed jar.

When It Doesn’t Work

• Both jars read the same the whole time. Almost always the instruments. Two cheap glass thermometers can disagree by 1 °C out of the box, the size of the effect you are hunting. Put both probes in one glass of water first, write down the offset, and correct for it.
• The gap appears in the first 30 seconds. That is not the gas. The lamp is closer to one jar, or one jar started warmer from sitting in a sunny spot. Re-measure and let both jars equalise for ten minutes first.
• The carbon dioxide jar is cooler. The gas probably came with liquid. A few drops of vinegar in the jar will evaporate and cool it, exactly like sweat. Pour the gas more slowly and never let foam in.
• Nothing climbs at all. LED bulbs emit very little heat, so the air barely warms. Switch to an incandescent or halogen, or move the lamp to 15 cm and run for 40 minutes.
• The cling film balloons out. Your jar is sealed and the air inside is expanding, which is normal and harmless, but it means your jar is not a fixed-volume system. Prick one pinhole in each jar’s film so both behave the same way.

Take It Further

Change the surface, not the gas. Run both jars on plain air, lining one with black paper and one with white. The difference will probably beat your carbon dioxide result, which makes the albedo point vividly. That variable gets its own run in the black cup and white cup albedo experiment.

Measure the cooling curve instead. Warm both jars, switch the lamp off, log temperature every 30 seconds for 15 minutes, and plot both curves on one graph. Cooling depends on how fast the jar loses energy, which is closer to the real atmospheric question than heating is.

Add water vapour. Put an identical damp sponge in both jars, seal, and compare with a dry pair. Water vapour is the dominant greenhouse gas in Earth’s atmosphere, and watching your own data get messier is a legitimate finding. Pair it with making a cloud in a jar.

What to Watch For Over the Next 3 Months

• Week 4: They ask whether a news graph about temperature has a control. “Compared to what?” appearing unprompted is the whole win.
• Month 2: They notice the car is hotter inside than out on a cool day and reach for convection rather than “the glass traps heat rays.”
• Month 3: They want to know how anyone measured the atmosphere’s absorption if a jar cannot. That question leads to satellite spectroscopy. The logging habit also fits building a backyard weather station.

Frequently Asked Questions

Isn’t it dishonest to run a demo I know is misleading?

The opposite. Running it and then dismantling it teaches more than skipping it. What would be dishonest is claiming it proves radiative trapping, which is what most lesson plans do. Say out loud what the jars show and what they do not.

Can I use a candle or a hot plate instead of a lamp?

No. You need radiant light from above, not a hot surface underneath, and an open flame near sealed glass jars adds risk for nothing. A plain incandescent desk lamp is the right tool here.

How much carbon dioxide do I actually get from baking soda and vinegar?

A tablespoon fully reacted produces a few hundred millilitres of carbon dioxide at room temperature, roughly a third of a 1 litre jar. You will not displace all the air, and you do not need to. Partial enrichment is enough.

My kid says this is boring compared to the volcano version. Now what?

Hand them the thermometer and the clock and make them the data logger. Deslauriers and colleagues found that students feel they learn less from methods that actually teach them more, so the complaint is not evidence the activity is failing.

Is a digital kitchen thermometer good enough?

Usually, if you have two of the same model and check their offset against each other first. You want 0.1 °C resolution. A 1 °C thermometer cannot see this effect.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Wagoner, P., Liu, C., & Tobin, R. G. (2010). “Climate change in a shoebox: Right result, wrong physics.” American Journal of Physics, 78(5), 536–540. https://doi.org/10.1119/1.3322738
2. NASA Science. “What is the greenhouse effect?” NASA Global Climate Change. https://science.nasa.gov/climate-change/faq/what-is-the-greenhouse-effect/
3. NOAA Global Monitoring Laboratory. “Global Annual Mean Carbon Dioxide.” Trends in Atmospheric Carbon Dioxide. https://gml.noaa.gov/ccgg/trends/global.html
4. Lueddecke, S. B., Pinter, N., & McManus, S. A. (2001). “Greenhouse Effect in the Classroom: A Project- and Laboratory-Based Curriculum.” Journal of Geoscience Education, 49(3), 274–279. https://doi.org/10.5408/1089-9995-49.3.274
5. Smith, J. P. III, diSessa, A. A., & Roschelle, J. (1994). “Misconceptions Reconceived: A Constructivist Analysis of Knowledge in Transition.” Journal of the Learning Sciences, 3(2), 115–163. https://doi.org/10.1207/s15327809jls0302_1
6. National Research Council. (2012). A Framework for K-12 Science Education: Practices, Crosscutting Concepts, and Core Ideas. National Academies Press. https://doi.org/10.17226/13165
7. Deslauriers, L., McCarty, L. S., Miller, K., Callaghan, K., & Kestin, G. (2019). “Measuring actual learning versus feeling of learning in response to being actively engaged in the classroom.” PNAS, 116(39), 19251–19257. https://doi.org/10.1073/pnas.1821936116