Fun with M & M's

M&Ms in water on a plate

I came across a simple, yet interesting experiment that was first described by Elizabeth Sumner Walter in 2001. She merely had students pour water into a dish containing some Gobstoppers candies.

I showed this experiment to some of my college chemistry students while they were working on a different laboratory experiment. Two of my students, Nathan Ford and Rachel Hubbard captured some video of this experiment, except that they substituted M & M’s candies for the Gobstoppers. You can see this experiment below:

The experiment sparked a lot of interest and discussion. Questions and comments ranged from “what is going on?” to “what happens if you use hot water?” and “Why don’t we try adding some soap?!” The most common question I heard was “why don’t the colors mix?”

A few students got online, and found one website on which it is claimed that the colors don’t mix due to a layer of water insoluble wax on the candies. It is claimed that this wax forms a shield that prevents the water soluble dyes from mixing. Indeed, carnauba wax is listed as one of the ingredients in Gobstoppers candy. However, carnauba is NOT listed in the ingredients in M & M’s.

We decided to test the claim that a waxy candy coating prevents mixing. If this is indeed the wax that prevents the colors from mixing, then its presence should be important in seeing this non-mixing effect. We tested a variety of candies (Gobstoppers, Dubble Bubble Gumballs, Skittles, Spree) that contained carnauba wax. Sure enough, they all showed this non-mixing effect. However, when we tested candies that do not have wax listed in the ingredients (M&M’s) the same effect was observed!

I’m interested in a couple of things regarding this experiment. First, I’m curious as to why this non-mixing effect occurs. I’m wondering if any of you have any ideas why the colors tend to stay in their own particular region in the water. We have noticed that, given enough time, the colors will bleed together.  But even after very long periods of time, semi-distinct regions of colors are observed. Second, I’m wondering if any of you have any ideas for modifications on this experiment, and how these modifications might be used to illustrate chemical principles. For example, I have found one video online which shows that the temperature of the water affects the rate at which the dyes spread out from the candies – what a great way to demonstrate the kinetic-molecular theory!  

Editor's Note: See a follow up post where Tom explores this idea further:



General Safety

For Laboratory Work: Please refer to the ACS Guidelines for Chemical Laboratory Safety in Secondary Schools (2016).  

For Demonstrations: Please refer to the ACS Division of Chemical Education Safety Guidelines for Chemical Demonstrations.

Other Safety resources

RAMP: Recognize hazards; Assess the risks of hazards; Minimize the risks of hazards; Prepare for emergencies


Join the conversation.

All comments must abide by the ChemEd X Comment Policy, are subject to review, and may be edited. Please allow one business day for your comment to be posted, if it is accepted.

Comments 27

Erica Jacobsen's picture
Erica Jacobsen | Fri, 08/16/2013 - 01:08

This would be a great activity to use with students for the 2014 National Chemistry Week theme of candy. I wondered what would happen if the candy pieces, after dissolving a portion of the coating, were transferred to another container and water added again to continue the dissolving process. If there is an outer wax coating that's causing it, perhaps after it's largely gone, the effects would change. Students could ask so many questions about this—a good system to investigate further. A great opportunity for experimental design.

Tom Kuntzleman's picture
Tom Kuntzleman | Mon, 08/19/2013 - 15:33

Hi Erica:

Your experimental suggestion is a good one.  Let me know what happens if you try it.  Your comments bring two questions to mind.  First, I wonder what would happen if this experiment were repeated with a non polar solvent such as  isopropyl alcohol instead of water.  My second question is only tangentially related.  You mention that the 2014 NCW theme is candy.  What is the theme for 2013?  And how does one go about finding out what the theme is for NCW for the current or future years?

Erica Jacobsen's picture
Erica Jacobsen | Mon, 08/19/2013 - 16:52

Interesting idea with the isopropyl alcohol—nice that it's an easy consumer chemical.


The 2013 NCW theme is energy, which ties in really nicely with your "And the Oscar Goes to...a Chemist!" JCE Classroom Activity and the hot dog demo. ACS keeps the NCW themes updated at There's a link on the page for future "NCW Dates and Themes." 2015 is color chemistry.

Tom Kuntzleman's picture
Tom Kuntzleman | Wed, 08/21/2013 - 08:42

Thanks, Erica.  Hey, I guess the M&M's experiment might also go well with the 2015 NCW week theme of color!

Benjamin Barth's picture
Benjamin Barth | Wed, 08/21/2013 - 13:40

This is really neat!  It brings a couple questions to mind:

1) Are the coloring additives truly being dissolved? Or, is the mixture of coloring molecules applied as an emulsion or a colloid of some sort and it is actually bits of that emulsion/colloid that are being "dissolved" by the water solvent? If this were the case, it could be possible that the larger particles would rather not mix and would "bounce off" each other rather than mix together. One additive that could have caused this effect to disappear could have been soap which might be a strong enough surfactant to make everything mix. Emulsions are a common annoyance in working up many reactions and can often be combatted by increasing the ionic strength via addition of a brine solution. It could be interesting to see what happens upon addition of brine.  I keep going back and forth in my head with the logic of whether brine should make them mix better or worse... 

2) What happens to the coloring when the pH of the solution is changed?  Some of the coloring molecules are very similar to pH indicators.  It would be neat to change the color by causing a reaction.

3) Also, what is the time frame for these colors to remain separated? Is it indefinitely?  That time-frame may give some clues as to the type of particles that are mixing.

Interested to hear your thoughts!


Tom Kuntzleman's picture
Tom Kuntzleman | Wed, 08/21/2013 - 16:44

Good to hear from you, Ben.

Actually, I find that I am quite interested to hear YOUR thoughts.  I do not have much experience working with colloids and emulsions, and it appears that you do.  So your thoughts are quite edifying to me.  I think I'll soon be going to the store to buy some candies so I can try some of these ideas.  I hope you do the same and share some results with us.  I'd be very interested to see what we discover.

The colors do not stay separated indefinitely.  They mix quite easily; bumping the plate repeatedly causes the colors to mix.  If I remember correctly, the colors stay separate for 5 - 10 minutes.  How long it takes for the colors to smear depends on temperature.   

Erica Jacobsen's picture
Erica Jacobsen | Sat, 03/22/2014 - 17:28

I've been noodling around with this experiment this week, preparing some NCW materials for the ChemClubs. I hope to put together a video (with my limited equipment) of some of the things I did and write a post about it. I noticed today that the Steve Spangler link now has its carnauba explanation deleted. Interesting.

Tom Kuntzleman's picture
Tom Kuntzleman | Sun, 03/23/2014 - 09:03

I look forward to seeing your post and video.   I am interested to see what you have discovered!  If you haven't already, you should look through the comments section to this post, specifically the comments by Ben Barth.  He has some interesting suggestions.  I will be having one of my General Chemistry students playing around with this experiment for her end of year project.  I hope we can learn some more about this experiment, too.  Thanks for the update on the Steve Spangler link.  That is interesting...

Lowell Thomson's picture
Lowell Thomson | Wed, 04/02/2014 - 13:00

Thanks for the post and the comments to get my brain thinking!

I've got a grade 10 introductory class that is just starting to discuss solutions. So we're going to try some inquiry tomorrow and see what happens. I'm sure we'll be taking some pictures and/or video, which I'll be sure to share. My hope is to have the students develop research questions, which we can then turn into experimental designs for further testing.

I'll keep you posted.


William Farmer | Mon, 04/21/2014 - 09:25



I'm watching this one after an e-mail I've got and see a lot of nice things I want to use in class. But the explantion...

Thinking about the experiment brought me to the idea of other solutions that don't really mix. There is a very common one; Fresh water and salty water.

So thinking like that, I came up with a suggestion like osmotic pressure. In that I reason from what the force behind the going into solution from the colouring agent is.

Is dissolves because the is "room" between the water molecules. It moves to the lower concentration zones. Until it meets a zone in the same concentration. Now the only reason for mixing is the movement of the molecules (kinetic theory, Brownian Movement). The influence of that one with the large molecules of the colours, is relatively low. So that one will take time.

What do you think?


Chemistry Teacher

The Netherlands

Tom Kuntzleman's picture
Tom Kuntzleman | Tue, 04/22/2014 - 06:43

Hi William

You state that the colors first spread out as a result of osmosis or diffusion.  However, when the different dyes meet each other (say red meets green), diffusion becomes limited because both the red region and green region are "filled" with large dye molecules. 

How do you think we could test this idea?  One experiment I have tried is using all M & M's of the same color in the different regions.  When this is done, the separation boundaries are still observed.  I think this experiment is consistent with your proposal.  

You might want to check out Erica Jacobsen's recent post on this topic: .  

I hope that by working together, we will solve the mystery of why the colors don't mix!   

William Farmer | Thu, 04/24/2014 - 14:00

Yes, I would like to find out how it works.

It could be so that the colour travels along with the sugar. That it is the sugar that dissolves. That would be easy to check, by using only a sugarcube (or maybe some aspratame) instead of a candy....

Or otherwise a bit of salt instead of a candy. That might prove something.


Marsilio Mark Langella's picture
Marsilio Mark L... | Sat, 04/26/2014 - 07:12

Erica's  experiment triggered an idea. Williams idea may confirm.

Obviously there is a density gradient issue as the materials from the third candy imediately went to the bottom of the beaker. What would happen if the experiment was performed in a large centrifuge tube. Revove the hard candies once the dyes and obviously initial sugars are dissolved into the water. Now centrigufe the results.


Water stratification occurs when water masses with different properties  form layers that act as barriers to water mixing . These layers are normally arranged according to density, with the least dense water masses sitting above the more dense layers.Would Differential centrifugation be able to seperate the different dyes? Is it Water stratification that creates the barrier.?



ALFREDO TIFI's picture
ALFREDO TIFI | Tue, 06/03/2014 - 15:45

People thought to a diffusion phenomenon, whereas it is stratification indeed.

Diffusion occurs at different and longer times.

Try also to repeat the experiment with a sugar solution instead of simple water.


Tom Kuntzleman's picture
Tom Kuntzleman | Sat, 04/26/2014 - 17:59

Thank you , Mark:

I will look into water stratification, thank you for this information.  I have a student currently studying this for a small project in my General Chemsitry II class.  I think you might like to hear about some of our results.  We notice that an insoluble white solid spreads away from the candy pieces, in addition to the dye.  It seems that the dyes are dissolved, but the white solid is not (at least a good portion of it).  We think that this insoluble white solid might be forming the barrier.  We have repeated this experiment by dripping four different food dyes into water rather than using candy.  In this case, the food dyes tend to mix a bit more.  We have also dripped food dyes into water to which a good bit of Mirilax has been dissolved (Mirilax is comprised of polyethylene glycol).  In this case, the food dyes tend to stay in separate regions.  In any event, I think we are making progress on understanding why the dyes stay in distinct regions.  I hope that by working together, you, me, Erica, Ben, Lowell, William and all of our students can figure this out! 

John VonBokel | Fri, 07/25/2014 - 10:15

Have you tried removing the insoluble coating?  I accidentally left a bag of M&M's in a hot car, and the candy coating held its shape, but once it cooled down to room temperature, there's now a waxy substance that seems to have accumulated where individual M&M's were touching.  I'm thinking if you found the right temperature (probably ~100 degrees F, based on the car circumstances), and laid some M&M's on a paper towel or something, you might be able to remove a substantial portion of that coating, and then could put the uncoated M&M's in water and see if that makes a difference.

Liza Huffman | Sun, 05/25/2014 - 15:01

I work for a children's museum, and I am planning a program called Wonder Wonka: Edible Chemistry, and we are planning on doing this experiment with kids of all ages, and I had the same exact question as you posed in your blog. Did you ever find out a definitive answer?

Tom Kuntzleman's picture
Tom Kuntzleman | Tue, 05/27/2014 - 15:05

Hi Liza:

No definitive answers yet, but I do have a hypothesis.  I currently think that the presence of sugars or other hydroxyl-containing compounds might be responsible for the fact that the dyes don't mix very well.  I hope to be posting a little more about this in the near future.  Do you have any ideas as to what might be happening?

ALFREDO TIFI's picture
ALFREDO TIFI | Tue, 06/03/2014 - 15:14

The velocity of diffusion is largely slower to observe in a few minutes. It could be obtained by mechanically stirring, but it is not caused by laminar flow.

This laminar flow is driven by a dense sugar solution that "falls & flows" in the bottom of the beaker and spreads in a layer carrying with it the colour from the surface.

This is the fast phenomenon we can see from the experiment.

If you substitute water with water-sucrose solution you will see a almost static experiment.

That is the measure and the explanation of the fact that the different coloured solutions don't mix each other.

They could mix only with diffusion, but this would happen in a very long times.

Around one minute and 10 s, in the video, the red-green and red-blue are separated by a colorless line.

These are the frontliness of dense sugar solutions with normal water pressed in between.

The water rises towards the top, by mechanical forces, thus maintaining the two fronts diluted and separated for a while.

After 10 more seconds the pelicule of water is depleted and the two frontlines reach direct contact, but diffusion prevent any quick mixing.

On the other hand from the final part of the video are visible very tiny lines of superposition green (between orange and yellow)

and purple (between red and magenta). A slow diffusion process is occurring.

Tom Kuntzleman's picture
Tom Kuntzleman | Wed, 06/04/2014 - 08:46

Alfredo, your ideas seem to be similar to Mark Langella's (see his comments above).  By using Petri dishes in this experiment, I have noticed that the dyes tend to stay on the very bottom of the dish (or plate).  This observation seems consistent with what you are proposing.  Also, you would likely be intereted in the experiment done by Marion Maw.  You can read about that experiment in the comments section at  I'd like to mention an experiment I did yesterday.  I tried to repeat the experiment using the same colored M & M's (I used blue) in each region.  When doin this, I did NOT notice the non-mixing effect.  This result to me seems inconsistent with my hypothesis that the negative charge on the dyes is responsible for the non-mixing effect.  Is the result I observed yesterday consistent with what you are proposing? 

Tom Kuntzleman's picture
Tom Kuntzleman | Wed, 06/04/2014 - 09:10

I forgot to mention two things.  First, when repeating the experiment using the same colored M & M's (brown instead of blue) in each region, I DID see the non-mixing effect.  Also, when using sugar water instead of water, I also saw the non-mixing effect.

ALFREDO TIFI's picture
ALFREDO TIFI | Wed, 06/04/2014 - 12:29

No. the high density frontline expansion model can't explain the direct mixing without introducing supplementary hypotheses.

One is that for some reason the blue color diffuses more than the brown one, maybe because of its smaller size.

In this case an invisible sugar frontline precedes the color frontline in the approach of blue to brown and of brown to brown,

but not in approaching the two blues, if the blue color diffusion has had enough time (if the two blue M&Ms are enough far away from each other).

Therefore, in the adapted model, you should find the nonmixing effect by putting the two blue M&Ms closer from each other,

or you could as well observe direct mixing by putting the two brown M&Ms far off.


In general, I tried in the past experiments attempting diffusive motion of different size anions and cations that were supposed to meet in some place along the connecting line forming some precipitate or coloured reaction. The outcome was flawed by the density layer effect.  Only by using a liquid with the same density of the two salt solutions you could testify something that could be correlated to the different ion sizes. I learned from the experiments (unfinished) and from calculations that the diffusion proper is not measurable in normal lab class times, because you would need several hours to observe pure diffusive motion of some centimeters in iso-density conditions.


However I'm really more interested in such experiments that try to eliminate the excess of useless variables and driving phenomena, so that a single theoretical source of explanation can be isolated. 

Furthermore, in my view, a real submicroscopical motion, as the diffusive motion of invisible and neutral hydrated ion clusters of different sizes, in a perfectly still and homogeneous liquid, that gives a final macroscopic and visible result that fits some atomic concept, is more interesting than some multiphenomenal coloured candies that serves to capture the curiosity at the beginning, but drives  quicly your students in a mess of magic and fantasy uncheckable pseudo-explanations.

Adolescent students already have an excess of fantasy, but poverty of construction of coherent links between claims, warrants, and data, and this is NOT, or not only because of incomplete knowledge of theories or experience, but also for the limited capacity of elaboration of thought in cases of multiple information, and for limited mastery of attention and voluntary memory. One of the reasons to attend school classes is to help them modifying and developing their high order thinking skills. In this case, it is better if they have only a very few theoretical abstract models, or only one at once, and if the experiment complexity is reduced in advance, because what matters more is that they discover their very limited explanation mechanism engine working well. After that you can add some more variable and complexity.

By following the path of direct complexity and fantasy you give only the illuson that everything's allright and no evolution of thinking skill is demanded: everyone lives in a aethernal brainstorming in which every utterance is good as every other.

ChemEd X Admin's picture
ChemEd X Admin | Thu, 06/05/2014 - 09:03
Thanks for pointing this out, Alfredo. I have fixed the link so that it now works.