Hot Coke: the Impact of Temperature on Ebullition

Most of you have seen the explosive reaction that occurs when dropping Mentos into a bottle of soda. There are many hours of footage on YouTube showing many variations of this reaction. We can deduce from these videos that the reaction is optimized with warm Diet Coke, 5-7 Mentos and that it is a great way to spend a warm summer day with the kids. It is also a great opportunity to explore the scientific method: primarily how to design and execute a scientific experiment. We mainly wanted to see how the thrust varied with increasing temperature of the Diet Coke.

We had several technical glitches, and had to adjust and even restart the whole experiment, which prompted my 10 year old to state "there are NO shortcuts in Science". I don't know where she got that, but I almost shed a tear.

For this project you will require the following items:

1. 6 bottles of diet Coke.
2. 3 packages of Mentos
3. Measuring tape
4. Large measuring cup
5. Thermometer capable of temperatures between -10 and +80 degrees Celcius

These items are optional:

1. Spangler Geyser Tube
2. Toy car
3. Pocket Lab
4. Chalk
5. Camera

All fizzy drinks contain some carbon dioxide (CO2) that is dissolved in the liquid. The CO2 would prefer to be in a gaseous state, but it has trouble finding imperfections within the bottle to nucleate around. Mentos has a surface texture which makes it an ideal medium upon which bubbles can form. The temperature of the Diet Coke should affect the amplitude of "gas oversaturation", or how strongly the CO2 would prefer to be in gas form rather than dissolved in the liquid. This reaction has been played with and studied extensively. There are similar reactions in geology where magma is sooooo ready to crystallize or release some gas, but there are just no singularities to get the process going. This "oversaturation" explains such processes as pegmatites (very coarse grained igneous rocks) and even explosive eruptions such as what we saw in Mount St Helens.

We should therefore expect hotter Coke gives more explosive reaction.

We altered a wrecked Barbie Mini Cooper to hold the 2 liter bottle, then taped the Pocket Lab to the whole rig. The concept that we could measure the distance traveled by the car with the different temperature bottles. Although this worked remarkably well for the hot bottles, it failed to move at all for the cold bottles, and thus invalidated the overall results. There is also the issue of dropping the Mentos with the car upright, then dropping it horizontally in a constant and predictable manner. This was not the most repeatable of experiments...

We did successfully launch the car down the road a few times, and we measured the acceleration with a Pocket Lab strapped on for the ride. However, after review we decided to restart the experiment using a more reliable setup that enabled getting data at the low temperatures.

All this being said, here are the steps we followed for this experiment:

1. Chop the windshield of the car in an semi circle shape to support the bottle.

2. Put one bottle in fridge, one in freezer (no more than 1 hour if it had been in the fridge), one in the sun and one in a pot of hot water. We were concerned that the bottle would fail at high temperatures, so we kept the temperature below 70 degrees celcius.

3. Tape a Diet Coke bottle to the car with the nozzle pointing backward. If you have a Pocket Lab, tape it into place also. There are apps for smartphones which collect similar acceleration data.

4. Carefully unscrew the cap, with the car held vertically. Start recording with your Pocket Lab and camera.

5. Insert the Mentos. We found that the Spangler Mentos gadget was quite useful for this and also had a nozzle that appeared to increase the thrust considerably.

6. Quickly lay the car on its wheels. Stand back!

7. Measure the distance traveled by the car.

8. Pour the remaining coke into a measuring vessel, and take the temperature of the Coke. (we initially took the temperature in the bottle, but this took precious time and caused an additional uncontrolled parameter in the experiment).

9. Take notes of each run, with all the collected parameters (temperature, distance, acceleration ect.).

Given the lack of reliable data from the rocket car attempt, we decided to alter the experiment to get more quantitative. Launching the full pop bottle itself would pose a number of issues, due to weight, and getting the Mentos to stay in the bottle during launch. So instead we decided to launch an empty bottle within which the Pocket Lab would be nestled in some cushion batting. To keep the rocket flying straight, we used a metal rod taped to the Coke Bottle, and attached a straw to the rocket bottle. We glued a small cup at the nozzle side of the rocket which fit over the Mentos releaser, theoretically ensuring a a good launch. What could go wrong? As you can see in the video below, the first few centimeters were good, but the rocket did not stay within the stream to continue its propulsion beyond that. For the second test, we used 2 straws, with, a similar result. There is more friction along the rod than we expected and we decided to dump them altogether, and just let the rocket pop off a few decimeters from the launch.

The remaining volume of Coke, and the distance covered by the spray were far more reliable and provide a good measure of the increasing thrust at higher temperatures.

This experiment was a metric tonne of fun, and resulted in some real data that inspire some very important discussions with your kids to help them understand certain key themes in science:

A. Experiment Design: We had all kinds of problems setting up the experiments in a way that resulted in reliable data. We were able to work out some of the bugs on the fly, others needed some discussion and eventually a major overhaul.

B. Data Quality: The Garbage in Garbage out principle was in full force here. The rocket car was fun, but it gave us a single data point in reality because the cold bottles did not advance at all. The remaining Diet Coke Volume were unreliable because half the liquid was going to pour out no matter what (bottle is on its side).

C. Instrumentation: Our first thermometer didn't cover the range of temperatures we needed, so we had to borrow one from a friend.

D. Outliers: The Pocket Lab was really cool, because it gave some really quantitative data. But because of some mechanical challenges with the experiment, there are some datapoints that clearly don't fit the trend (for example, the maximum acceleration). This brings us to question how do we deal with those outliers? It is really tempting to hide them... There are NO shortcuts in science!

E. Trend lines: With just 4 temperatures, we can see that there is a trend, start by fitting a straight line, not bad but we can do better. The hottest datapoint is not as powerful as we would predict. We can discuss what happens at the extreme temperatures (Coke boils, bottle explodes), and what should happen as we approach that natural limit. How can we use the formula generated from the experiments? Can we predict an outcome?

F. Law of Diminishing Returns: We can add energy as heat the the bottles to get more and more thrust, but we gain less and less punch as we do this. The trendline is going to reach a maximum thrust and become asymptotic to that limit. Is there a practical temperature that is optimal?

This was a great activity on a warm summer day, and we had a blast doing actual science! Feel free to comment below and I can add notes if required.