Recently I have learned of "installable storage heaters". Those are fairly heavy devices which consist of a ceramic brick stack and electric heater. The brick stack is heated during the off-peak electricity hours and gives back heat the rest of the time.

This made me think: the density of bricks is at best twice the density of water. Heat capacity of bricks, on the other hand, is closer to being 4-6 time smaller than heat capacity of water.

A simple calculation thus shows, that device with exactly same dimensions as those available on the market but featuring a drainable water tank as a heat accumulator instead of bricks:

  1. Will be able to store more energy
  2. Will be light and portable when water is drained
  3. Will have better efficiency transferring heat from primary heater to accumulator

So the question is: why offerings on the market mostly consist of immovably heavy, brick filled devices?

  • Whoever is using bricks has never taken a physics course. Grams don't store heat. Atoms store heat. 238g of U238 stores roughly the same amount of heat as 7g of Li7. The densest thermal storage medium isn't even a question, water is either the highest or rounding error away from the highest, due to packing 3 atoms into 18g/mol and high molar density. Nevermind the advantages of cost, pumpability and toxicity. – Harper - Reinstate Monica Feb 25 '18 at 6:12
  • 3
    @Harper - water makes no sense in this application. The upper temperature to avoid a pressure vessel and safety valves would be somewhat less than 100 deg C. Materials like stone or bricks can be heated to several hundred deg C. If using water, the system has to be drained if there’s any potential of freezing, also unneccessary with stone or bricks. Finally, bricks won’t corrode the vessel. – Mark Feb 25 '18 at 14:19
  • The problem with using hundreds of degrees is now you must insulate it to extremes Heat loss is proportional to temperature differential. High temp requires much higher insulation value. You must also protect a high temp source from precious fingers and from starting fires. It's also precludes heat pumping, forcing you to inefficient resistive/ fuel heat. The bolt-to-wall. Units that replace radiators are just unworkable, as they are too small to insulate. – Harper - Reinstate Monica Feb 25 '18 at 14:49
  • 1
    @Harper - I agree with everything you say, but the fact is that these heaters are built as simple electric resistance heaters to take advantage of lower off-peak electric rates. From what I can find, the bricks are heated to very high temperatures, and very well-insulated. I found a thesis which studied thermal performance of one of these heaters, and it states that the internal temperature reaches the 600 deg C range (see pages 70 and 83, for example): esru.strath.ac.uk/Documents/MSc_2013/Becerril.pdf – Mark Feb 25 '18 at 16:02
  • if it can get 4-6X hotter (over ambient) it can store more heat than water, which can't be heated much. – dandavis Feb 25 '18 at 18:56

The better-made storage heaters will use materials optimised for the task

Feolite has

  • specific heat = 920.0 J·kg−1·°C−1,
  • density = 3,900 kg·m−3,
  • thermal conductivity = 2.1 W·m−1·°C−1.
  • maximum operating temperature 1000 °C.

Water has

  • specific heat = 4184 J·kg−1·°C−1
  • density 1,000 kg·m−3,
  • thermal conductivity = 0.591 W·m−1·°C−1.
  • maximum operating temperature <100 °C unpressurised.


These things are heated up using half-price electricity overnight. You have to have a dual-rate meter and dual-rate tariff (e.g. "Economy 7" in UK). The heater's core is surrounded by insulation so that most of the stored heat is retained until needed. When heat is needed, air is blown through the core to extract the heat.

Storage heater cross-section Image source

| improve this answer | |
  • How do they pump/circulate the feolite? Seems like any pump would suffer abrasion damage. If not pumping, how do they efficiently interchange heat when they want it and insulate it when they don't? Motorized insulation plugs? – Harper - Reinstate Monica Feb 25 '18 at 14:55
  • 1
    So the Feolite has a similar volumetric capacity (3.588 "units") vs water (4.184) at the same temperature - but can be used at 10x the temperature - giving a significantly higher storage capacity? – brhans Feb 25 '18 at 15:10
  • @Harper - there’s no pumping - they are just piles of bricks with heating elements running through them. Air blows across and through the bricks to distribute the heat when needed. – Mark Feb 25 '18 at 16:04
  • @Harper: Good question, I should have made that clearer, answer updated. – RedGrittyBrick Feb 25 '18 at 17:54
  • Nice, dense material this feolite is. – oakad Feb 27 '18 at 1:18

Good observation on the relative merits of water.

However, your key assumption about relative heat capacity is almost certainly wrong: Many common metals have volumetric heat capacity exceeding half that of water. And many ceramics have higher heat capacity than water!

You could still argue that it would be cheaper to produce and ship these storage heaters empty, and then fill them with water once they're installed. But then you would have to consider the unique drawbacks of water:

  1. It is corrosive to many metals
  2. It expands when heated (and also if frozen)
  3. It can't be heated beyond its relatively low boiling point in a non-pressurized system
  4. If its container leaks it causes a lot of collateral damage.

Plumbers tend to view water heaters like time bombs: No matter how exotic the alloys and coatings used in their containers, the effects of expansion and contraction seem to always find a way to break them. I.e., it's just a question of when, not if water heaters will leak.

Granted, it is possible to build water-based heat systems that last for generations without leaking or failing. But in a smaller system it might not be – in fact, as you observe, evidently is not – cost effective.

| improve this answer | |
  • 1
    Here’s a figure from a study of materials for this application. The goal was to find the most cost-effective material with high heat storage capacity and high maximum use temperature. It points to bricks as the preferred solution. grantadesign.com/education/InDepth/html/indepth/… – Mark Feb 25 '18 at 14:41
  • @mark That chart certainly reflects a preconceived notion of using very, very high 1000K-ish temperatures to do the heat storage. Sweet Lily Allen, what do they use for insulation, Space Shuttle tiles? But you need to be that hot to store a meaningful amount of heat at a mass workable for a bolt-up appliance. Problem is those temperatures bet all the marbles on resistive electric heat (even fuel is a challenge at 1000K). It makes insulation challenging (our go-to, styrofoam, is right out). – Harper - Reinstate Monica Feb 25 '18 at 17:09
  • @Harper - You’re close - While looking into this I saw a reference somewhere to Aerogel insulation, which is a very high performance insulation material that I think was developed at NASA! I can’t find the reference now though. Believe me, I don’t think these are a good idea, and would never buy one, for all the reasons you mentioned. But apparently they were developed back in the 1950s as a way to shift electric use out of peak periods to take advantage of lower rates. Heat pumps weren’t an option then. – Mark Feb 25 '18 at 17:55

Fire and air: cheap and bolt-up

These things are sold for a few hundred dollars and are intended to "bolt up" to any wall in the normal place a wall heater goes.

The "cheap" requirement forecloses any possibility of using efficient tech like a heat pump. They use simple (but rather expensive) resistive heating. That's wasteful.

The "bolt-up" requirement requires light weight and small size - so a storage reservoir of meaningful size is impossible - it would collapse the floor. Instead they use very high temperatures to store the heat in a sane-sized mass. The high temps won't play with heat pumps or passive solar, the only option at those temps is appallingly inefficient electric resistive heating.

That is your answer: they can't use water because they need to make the unit Very Hot to store a meaningful amount of heat in small space and mass.

Loss through insulation is proportional to temperature differential / insulation value. So very high temps require very good insulation. Temperatures pushing 1000K preclude easy, cheap styrofoam or fiberglass, leaving the asbestos family of rock wools or other silicates. many of(and very balky insulation since most insulation does not like 1000K temps, leaving you in the asbestos family, yay).

You have to pay the piper, either in huge mass, or potent insulation to contain lots of heat. If you don't pay the piper, the device doesn't work very well.

Looking at the units I see on the Web, indeed, they do leak like a sieve, the convection units losing 80% of their heat via uncontrolled loss through the weak insulation. This makes them rather hard to control.

I remain skeptical that these small units can really time-shift heat the 8-12 hours needed to exploit evening electric rates. Where they show real potential is demand side management: the utility commanding heaters to turn off momentarily rather than spinning up a peaker. This ability greatly improves the resiliency of the grid, as it can effortlessly shed load if overloaded. If these are being pushed in your area, I suspect that is the agenda.

This for sure: because they store energy for such a short time, they are 100% energy efficient. Since there is no escape for heat but the living space, the only way these units fall below 100% is if they time-shift heat to times it is undesired. E.g. By leaking.

But they also cannot exceed 100%. We can do much better.

Water. Do everything, be everything.

Or safe propylene glycol antifreeze. "oh no, you can't run water at pyrophilic temperatures" Darn right. This is a completely different approach, and it is much cooler. The working temps would be 10-70C (in heat mode), so no need to cap the tank. Which goes somewher where leaks won't matter.

The delta-T being so small, we need a MUCH larger thermal mass but insulation losses will be 1/10. We also benefit from square-cube law: cube the volume, only square the envelope size.

You install a large tank, either as a cistern (with proper drainage) or outdoors and make up for it with better insulation. Any common agricultural tank will do. 12"+ of styrofoam would be good. The delta-T is 60C not 600C, so the insulation could be less, but it's easy with materials like styrofoam available, so go for a lot more. You'd use a hefty sized tank right-sized for your load - 1000 gal. would not be excessive if your design engineering called for it.

You have many options to add heat to such a low-temp tank - fuel, solar collectors, resistive electric if you really want to, but the darling will be heat pumping because it will run 200-400% efficiency. In fall/spring, the heat it pumps in at night could be the heat you rejected by day using its air conditioning mode.

Then you'd use heat pumps to transfer the heat into the house - if the tank is 60C and you want your heat pump outputting 40C air, that heat transfer will be VERY efficient because it is "downhill" for the heat pump. You're simply using the heat pump to transport the heat from place X to place Y.

Heat pumps which do this aren't uncommon, they are used in large complexes which circulate "service water" at 70 degrees for both A/C and heat. In summer they use cooling towers, in winter the boiler room heats the water with fuel (gas).

I mentioned "heat mode". Heat pumps reverse into air conditioners, so this system comes with A/C. At night your heat pump chills the tank, dumping to ambient or house heat. If the tank can abide freezing, water wins massively in A/C mode: the enthalpy of fusion lets you store an insane amount of cooling power. (too bad you can't find a substance that freezes at 25C). By day, you pump heat from the house out to the tank - "downhill" again. If dump freon temperature exceeds ambient outside air, you can cool to ambient before dumping in the tank.

Earth: build the thermal storage into the house

In this case, you use an even lower delta-T and even more storage material: the masonry construction materials of the house itself. This is obviously not a bolt-on, but passive thermal design from the outset.

In this case a high-thermal-mass construction material is used, and generous insulation is placed on the outside. As a result the building envelope strongly resists change to interior temperature, and itself stores the heat that we are time-shifting.

Dealing with only a 5-10 degree delta-T, proportionately more mass is needed. Insulation has less work to do, but it is still important.

Implementation is easy: any commercial off-the-shelf HVAC equipment will suffice. Using the thermal storage is just a matter of timing thermostat settings.

If that does not suffice and the homeowner wants something more on-demand, it would combine nicely with a water system, and the "water" could even be an enormous chunk of concrete.

| improve this answer | |
  • i heard the school obama went to in hawaii makes ice at night to cool during the day. – dandavis Feb 25 '18 at 22:54

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