This is one of those statements I frequently encounter, but I've never really seen evidence to support it. The advice often goes even further to say that if your fridge/freezer is relatively empty, you should add something (e.g., bottles of water).

Such statements have appeared here too, as in answers to this question, this question, this question, and most recently this question.

I'm assuming that there may be some minor energy savings in some circumstances, but is it significant enough to actually care about? Moreover, is it enough to offset the energy lost in, for example, cooling or freezing water (or other things) just to fill up the fridge?

Note: I've been doing some of my own research and was going to post it in the question, but I think it will be more appropriate to provide what I've found in an answer below. However, if anyone has reliable sources that give a different answer (and preferably actual measurements or studies to back it up), I'm very interested.

EDIT: After writing my answer, I just found this, where Cecil Adams argues strongly in favor of the full-fridge hypothesis. I rarely disagree with Cecil, and he makes reference to a number of studies, though no citations. So I'm interested if someone can point to some of these studies.


7 Answers 7


SUMMARY: Unless I'm missing something here or you're doing very odd things with your refrigerator, you'd at most save a couple dollars per year by keeping your fridge/freezer full. Moreover, stocking up on water (or other things) to fill up fridge/freezer space won't save you much at all unless you're keeping it stored there for a VERY long time, since it costs so much energy to cool the water in the first place.

There are valid reasons to do this:

  • if you need to survive intermittent power outages (as TFD notes), having a lot of ice or food in the fridge will keep it cool longer
  • a larger amount of food in the fridge/freezer will make it easier to chill or freeze new food more rapidly, which may sometimes aid in food safety/preservation
  • similarly, having more food may help to minimize small temperature fluctuations when opening the door frequently, again perhaps aiding in food safety/quality in some cases
  • if you have a very inefficient fridge that cools unevenly or is not well-insulated, having more food will prevent as much cycling on and off (though packing a fridge too full can also prevent it from cycling correctly)

All of these are perhaps good reasons to have a mild preference for keeping the fridge a little more full. But, from an energy perspective, there isn't a rationale for deliberately filling your fridge with excess food/water, since the energy required to cool any solids or liquids is usually many, many times the amount needed to cool air.

Also, if your main concern is cold air "falling out" of the fridge when the door is open, I'd suggest filling it with empty containers that have only air in them. They'll give you the benefit of not losing that excess cold air, but without the energy expenditure to cool a liquid you don't need to. (But again, the likely benefit is probably a couple dollars per year at most.)

Details below.

I tried searching for some reliable stats, and though I find many, many sources which make this claim, I don't generally see actual numbers regarding energy savings or even a theoretical calculation to support the logic of the practice.

In fact, it sometimes seems to appear on "myths" lists of energy groups, such as here:

  1. MYTH: You can save energy by keeping your fridge full, closing it quickly, and regularly cleaning the coils.

Actually, all three of these actions are not worth your trouble. In the study done by Balsnik it was found:

Total use from ALL fridge door openings adds up to <50 kWh/yr, or about $5.
Putting water bottles in your fridge to keep it full adds up to <0.1 kWh/yr.
Cleaning coils – no actual savings found.

Or from this document (on ultralow temperature freezers' efficiency):

URBAN LEGEND? A full freezer takes less energy to operate: An ostensible reason for this idea is that the thermal mass takes longer to warm up, so the compressor doesn’t have to work as hard. Think about it: while the contents take longer to warm up, it also takes longer to cool down so the compressor works just as long every day. The fundamental insulating factors of wall thickness and gasket integrity do not change with a full or empty freezer, so why should it make a difference to heat transfer? While cycle frequency will go down, cycle duration will go up. The heat entering the cabinet will not change. There is a nominal power spike at the beginning of each compressor cycle, so more cycles could plausibly increase energy use a little bit. Data has not been widely shared, so it remains in urban legend status for now.

The logic of this last reference seems to address the question of whether an unopened freezer/refrigerator will be more efficient when it is full (as is sometimes claimed -- that the fridge will have to "work less hard" somehow). Obviously that doesn't make much sense, as this quote notes.

However, to judge this accurately, we'd need to take into account what happens when you open the fridge/freezer. Here's a report from an energy-savings group that tried a number of tests (including opening the door for various amounts of time). They concluded that top freezer models of fridges did use less energy when full, though they note in their analysis that it doesn't include the additional energy required to cool the extra food in the first place. But once that food is cool and the fridge is full, there is some energy benefit for top freezers. (How much is unknown, since their graphs don't have numbers.) For other types of freezer models, the results of the tests were mixed, so there was no clear benefit of an empty vs. full fridge. Their conclusion: "So our advice is to not worry about keeping the fridge full, and focus more on keeping the door closed."

For a theoretical perspective on fridge opening, let's try some reasonable assumptions:

The average fridge size in the U.S. is around 20 ft3. If we assume that the fridge isn't packed full and half of the air present is replaced by room temperature air when the door is opened, that would be roughly 10 ft3, or about 0.28 m3.

Using stats from here, we can calculate that cooling that 10 ft3 of air down by 20°C (e.g., from "room temperature" of about 25°C to 5°C) would require about 6.8 kJ of energy, or 0.0019 kWh. For a freezer of similar size, the temperature of the air would likely have to be lowered about 40°C, rather than 20°C, so these numbers would be doubled.

If we open the fridge door 20 times per day, over a year that would add up to about 13.8 kWh for a fridge with 10 ft3 of empty space, or 27.5 kWh for a freezer with a similar amount of empty space. The stats in the first quotation above estimate 50 kWh/yr for all fridge door openings, so the numbers seem to be in the right ballpark. Basically, it costs a couple dollars every year in energy loss for opening the fridge.

Now, suppose we instead loaded up that 10 ft3 with water instead of air. (This is a ridiculously large amount of water, but I'm using it to keep the occupied volume the same for a comparison.)

The amount of energy required to cool water from room temperature can be calculated similarly from these numbers. Cooling 10 ft3 of water by 20°C would require approximately 23,000 kJ. Freezing it to -15°C from 25°C would require about 120,000 kJ. (This number is significantly higher, due to the excess energy required to turn liquid water into solid ice.) The effect of adding large amounts of water was shown clearly in the study mentioned above, where the addition of 150 lbs. of room temperature water caused the fridge to spike to about 65°F and take almost a day and a half to return to normal temperature.

To put these numbers in a more useful form:

  • You'd have to cool down air in the fridge about 3500 times to "pay for" the amount of energy expended cooling the same volume of water.

  • You'd have to cool down air in the freezer about 9000 times to "pay for" the amount of energy expended in freezing the same volume of water.

UPDATE: As Joe rightly points out in comments, I assumed dry air here to simplify the calculations. But real kitchen air will be moist, and its effect is not insignificant. (I assumed the error would be less than 50% or so, but under reasonable assumptions, it's probably off by a factor of 1.5-3, depending on your kitchen humidity and how humid your fridge is.)

Anyhow, assuming we start with a relative humidity of 50% in the kitchen at 25°C, and we assume the refrigerator cools to 5°C in the fridge and -15°C in the freezer while maintaining 50% relative humidity at those temperatures (which would obviously require water vapor removal), here are some updated stats:

- You'd have to cool down air in the fridge about 1800 times to "pay for" the amount of energy expended cooling the same volume of water.

- You'd have to cool down air in the freezer about 5500 times to "pay for" the amount of energy expended in freezing the same volume of water.

[See calculations below for details.]

Basically, depending on how often you open your fridge and the room temperature, you will probably need to refrigerate water for at least several months before seeing any energy savings (at all). You'd likely need to keep (the same) water frozen for at least a year to get any energy savings. Even then, for reasonable quantities of water (e.g., a few gallons), it's unlikely that you'd save more than a few dollars per year in energy costs (and likely less).

A final note about full fridges: even supposing you manage to save a few cents per year with a full fridge, my practical experience tells me that I hold the door open much longer when the fridge is full than when it's nearly empty, since I often need to move stuff around or take things out temporarily to get things in the back. So would this theoretical savings ever actually materialize? I don't know.

For those who are interested, here is the "work" for the calculations above. I assume a volume of 10 ft3 = ~0.28 m3. Note that various approximations were used here to get a "ballpark" figure -- in particular, densities and specific heats were assumed to be constant over the temperature range, which might introduce a 5-10% error for the air calculations, and much less for the water calculations.

(1) Cooling (dry) air by 20°C

  • 0.28 m3 of air × density of 1.205 kg/m3 at 20°C from table = 0.337 kg
  • 0.337 kg × 20°C [same as 20 K] × specific heat of 1.005 kJ/(kg K) = 6.8 kJ
  • 6.8 kJ ÷ 3600 = 0.0019 kWh

(2) Cooling (dry) air by 40°C

  • Same weight of initial air
  • 0.337 kg × 40°C × 1.005 kJ/(kg K) = 13.6 kJ

(3) Cooling water from 25°C to 5°C

  • Same volume of 0.28 m3
  • 0.28 m3 × density of about 1000 kg/m3 = 280 kg
  • 280 kg × 20°C × specific heat of 4.18 kJ/(kg K) from table = 23400 kJ
  • NOTE: Obviously one cannot and should not fill a home fridge with ~600 lbs. of water, but I used the same volume here to make the energy required for comparable volumes, since it's asserted that replacing air by an equivalent volume of water will make a difference.

(4) Cooling water from 25°C to -15°C

  • Ice is less dense than water, so to achieve a final volume of 10 m^3, we must start with less water.
  • 0.28 m3 × density of ice of 916.8 kg/m3 = 256 kg
  • Cool to 0C: 256 kg × 25°C × specific heat 4.18 kJ/(kg K) = 26800 kJ
  • Freeze: 256 kg × heat of freezing 334 kJ/kg = 85700 kJ
  • Cool ice to -15°C: 256 kg × 15°C × specific heat of ice 2.108 kJ/(kg K) = 8100 kJ
  • Total cooling energy: 120,700 kJ

(5) Cooling similar amount of water to air in fridge = 23400 kJ ÷ 6.78 kJ = about 3450 times greater

(6) Cooling similar amount of water to air in freezer = 120700 kJ ÷ 13.6 kJ = about 8900 times greater

(7) Cooling air at 50% relative humidity by 20°C:

  • We get weight fractions of water vapor in air at 50% humidity from a Mollier diagram. Here x at humidity of 0.5 is about 0.0098 kg/kg at 25°C and about 0.0026 kg/kg at 5°C.
  • We then follow the calculation of enthalpy (H) of moist air as found at Joe's link here.
  • At 25°C: H = (1.005 kJ/kg°C)(25°C) + (0.0098 kg/kg)[(1.84 kJ/kg°C)(25°C)+(2501 kJ/kg)] = 50.1 kJ/kg
  • At 5°C: H = (1.005 kJ/kg°C)(5°C) + (0.0026 kg/kg)[(1.84 kJ/kg°C)(5°C)+(2501 kJ/kg)] = 11.6 kJ/kg
  • Delta H (change in enthalpy) = 50.1 - 11.6 = 38.5 kJ/kg
  • Moist air is slightly less dense than dry air: using figures from here, moist air is about 1.199 kg/m3 at 20°C.
  • Mass of air using above volume assumption of 0.28 m3 is 0.336 kg
  • Energy required to cool = change in enthalpy × mass = 38.5 kJ/kg × 0.336 kg = 12.9 kJ
  • Note that various numbers here may vary slightly over the temperature change, but as in Joe's link, we can assume they're constant enough that it won't affect the final answer by more than a few percent.

(8) Cooling air from 25°C to -15°C in freezer

  • Using the Mollier diagram linked above, we get a weight fraction of approximately 0.00055 kg/kg for 50% humidity at -15°C
  • Use similar calculations to above
  • H at -15°C = -13.7 kJ/kg
  • delta H from 25°C to -15°C = 63.8 kJ/kg
  • using mass and density as above, the total energy required to cool is = 21.4 kJ

(9) We calculate the ratios as above, ending up with 1800 times more energy to cool an equivalent volume of water in fridge, and 5600 times more energy to freeze it.

(10) Relative humidity can vary in both the kitchen and fridge, so these calculations should only be taken as a ballpark figure, perhaps varying by a factor of 2-3 in either direction in extreme cases. Regardless, the amount of energy needed to cool even moist air is insignificant compared to that required to cool any liquid or solid food.

  • 2
    You assumed dry air. Saturated air (aka 100% humidity) is 76.9 (kJ/kg) at 25°C, which makes your break-even ratios at about 45 and 116. If you have small children, this might be worthwhile after a couple of days in high-humidity areas. And you keep the water in the back of the fridge, so that you can see the stuff you actually want to get to -- if it's slowing you down finding stuff in the fridge, you're doing something wrong.
    – Joe
    Commented Jun 3, 2015 at 15:06
  • 2
    @Joe - Yes, you're right that I used dry air, because I wanted the calculations to be a little easier. I just wanted to get something in the right ballpark, but you're right that I should have been more careful. On the other hand, you've cited a specific enthalpy number, but you tried to use it as a heat capacity. I'll update with more calculations, but basically my numbers are likely off by a factor of 2 or so, not 100. Also, you're right that water in the back isn't a problem, but I'm also considering advice to keep your fridge more full in general, which can slow you down.
    – Athanasius
    Commented Jun 3, 2015 at 17:02
  • The idea about containers filled with cold air sounds quite remarkable. especially since air (in pockets or layers, not running free) is an excellent insulator. With a bit of spatial planning, it may help to insulate some areas, ie, block off a bit of area for something that temperature swings would cause more problems with, or let something warm freeze without effecting quality of surrounding containers' contents.
    – Megha
    Commented Nov 22, 2017 at 3:15

One edge case is if you are on a controlled power plan for your fridge, or have a smart power meter and fridge (both very rare)

With these plans you may save money, but not power directly, you are just saving your country from inefficient peak power generation

There can be many hours going by without power for your fridge, so a well stocked fridge or freezer will have less temperature swing, and this may improve food preservation

  • 1
    It also helps when the power goes out ... a full freezer will defrost less quickly than a half-full freezer. (and it's getting to be hurricane season)
    – Joe
    Commented Jun 3, 2015 at 15:08
  • @Joe I suggest moving to a country without a "hurricane season" :-). Check blitzortung.org/Webpages/index.php?lang=en for a safe place
    – TFD
    Commented Jun 3, 2015 at 20:10
  • or monsoons, blizzards, tornados, tsunamis, questionable electrical infrastructure, etc ... are there any countries that actually have reliable power infrastructure these days? And the US is big enough that not all of it has a hurricane season ... the southwest has monsoons and/or tornados. The north has blizzards. The mountain areas have avalanches. (I'm in a blizzard + hurricane + derachio area ... and earthquakes now, too)
    – Joe
    Commented Jun 3, 2015 at 20:23

There are many factors. The one factor people seldom take into consideration is mass. Below I am considering only that factor. Other factors, such as open doors, may have greater effects, especially when considered cumulatively. I do not know.

If I put a room-temperature bottle of water into a freezer with ten bottles of frozen water, the temperature of the new bottle will descend much faster than in an empty freezer, because there is already more cold mass. However this has a cost; the temperature of the frozen bottles will rise in a direct relationship to the transfer of heat. The refrigeration unit will have to use energy to return the temperature of the ten bottles to their correct temperature. So yes, we have faster freezing. But no, based only on this factor, we do not have less energy expended to do that freezing.

The essential function of a refrigeration unit is to maintain a heat differential between inside and outside. The greater the mass of what must be kept at a lower temperature, the greater the amount of energy needed to do that. Air has very little mass, very little energy is needed to change its temperature. In contrast, a water bottle has much greater mass and much more energy will be necessary to maintain a heat differential.

One reason energy is needed to maintain a constant temperature is because of the transfer of heat from the ambient temperature outside the refrigeration unit and the desired temperature inside (we may also think of this as the cold seeping out in the opposite direction). Otherwise a closed freezer would never need any electricity, once it reached the temperature setting.

The contents of an empty freezer kept at -20° centigrade has a very low mass, whereas that of a freezer filled with water bottles has a much higher mass. The full freezer in fact has "more cold" inside, even when the temperature is the same. It loses more cold to the outside and more energy is needed to keep it cold.

So based only on the factor of cold loss, a full freezer needs more electricity.

How important this single factor is depends on many things, including the efficiency of the heat shielding of the freezer, the ambient temperature (less heat transfer if room temperature is 16°C than 30°C), ventilation of the heat coils, etc.

How this interplays with other factors creates a quite complicated multi-factor equation, and I suspect the answer will not be the same for every refrigeration unit and every usage.


All these calculations instead of a straight bit of common sense. Air has a thermal capacity one quarter of water. That means for every degree to cool it down water requires 4 time as much fridge energy as air. If water is in a warm room, (or subject to an open fridge door) it will take four times as long to warm up than air. So stop talking about energy efficiency, and other such esoteric terms.Just think : If my fridge is full, every time I open the door, and/or put something in, the variation in temperature will be less if it is full than if it is empty. So forget the word "efficient" and replace it by "best". Basically for a fridge to perform well, in other words for the temperature to vary as little as possible, keep it full. End of story.

  • This answer says we shouldn't do calculations but then does a calculation ("4 time as much"). As I calculated explicitly in my answer, the actual factors are more like 1800 times as much for a refrigerator and 5500 times as much for the freezer for water compared to air. If you have found an issue with my calculations (I have actually taught physics and have taken engineering courses in heat transfer), I'd be happy to have errors pointed out.
    – Athanasius
    Commented Oct 24, 2023 at 9:24

Why this issue is still being debated baffles me. Whatever the contents or fill ratio, heat loss is determined by insulation and external surface area. And this equates to energy consumption when the door is kept closed.

The issue seems to be air disturbance when the door is opened and closed. In a partially filled freezer, there is more cold air that is likely to be displaced every time the door is opened. And the warm air that enters needs to be cooled down. So that must be where energy is wasted compared to a full freezer. And it also causes more frozen condensation.

  • The question isn't whether there is some inefficiency caused by opening and closing fridge doors -- obviously there is. The question is whether: (1) that energy loss is substantial enough to worry a lot about, or whether it's something that might only cost you maybe $5-10 per year (as I explicitly calculated in my answer as to show ~50 kWh as a good estimate and quoted sources saying so). And (2) if deliberately filling your fridge with things you may not use actually would even result in meaningful energy savings (probably not).
    – Athanasius
    Commented Oct 24, 2023 at 9:31

There is also one other factor that wasn't considered. Water/ice in a fridge will cool the inside air more quickly when the door closes which reduces the work required by the compressor after each occurrence. So, if the fridge was filled with ice collected after a snow storm, it would decrease the power consumption caused by door being opened and the expense of cooling 150lbs of room temperature water would never occur. Although it's anyone's guess as to how much energy is required to maintain a full fridge vs an empty fridge, it seems to me that these values should be obtained while door remains closed to establish a power consumption baseline, then go from there.

  • "Water/ice in a fridge will cool the inside air more quickly.." Effectively, you raise the thermal mass of the fridge by putting more stuff in it. If that stuff helps cool the air after the door was open, that stuff is getting warmer. The compressor now has to work to cool the stuff back down. To take an extreme case, if you had X amount of stuff, and its temp went up by 5 degrees on door opening, replacing it with 10*X amount of stuff might mean that the stuff only goes up by 0.5 degrees, and air cools faster. But the compressor has to work similar amounts to cool the extra stuff back down.
    – Athanasius
    Commented Oct 31, 2016 at 17:35

Filling the empty space in a refrigerator to make it run more 'efficient' or to save money is just not correct. Think of it this way, your driving a semi truck, is it cheaper to run full or empty. Empty space or full space will be cooled by the refrigerator just the same, it is much cheaper to cool empty space (air). It is as simple as that.

  • 1
    Your analogy doesn't really work -- a full truck is heavier, so more expensive to move more things. In contrast, the theory behind a full fridge is that there is less empty space around that needs to be cool.
    – Erica
    Commented Oct 10, 2015 at 16:24
  • Also, cold air in the amount a fridge produces is useless outside the fridge, cold beer is not :) Commented Oct 31, 2016 at 11:36
  • And a truck has to come to stop from time to time to avoid hitting things. After that it costs energy to accelerate a heavy load up to speed. A fridge doesn't ever need to come to room temperature.
    – bdsl
    Commented Jun 29, 2019 at 14:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.