GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon

Community and Q&A

Thermal mass floor tile conduction

thermal | Posted in Energy Efficiency and Durability on

Newbie, DIYer, trying to finish a direct gain passive solar residence by installing quarry tile with unknown (at the moment) thinset/mortar/adhesive over a very dense, 4″ concrete slab that has a machine-steel-troweled, nonabsorbent, slightly rough (but not fine broomed) profile. I’ve got two issues to solve sequentially;
1.) efficient thermal conductivity of the whole thermal mass sandwich (ie tile to thinset to slab) is in question because there appears to be no testing of the various thinset/mortar/adhesive conductivities, only contradictory opinions as to such.
One passive author states:  a slab’s coverings (ie quarry tile in my case) and its thinset/mortar/adhesive that are poor conductors of thermal heat will clearly interfere with the flow of heat in and out of the heat storage mass. In extreme cases the heat storage material can be rendered useless. The thermal conductivity of both a tile and its thinset/mortar/adhesive should be at least as high as that of the slab. Tile should be bonded with no-air-pocket workmanship to assure a good thermal connection or continuity to the underlying slab…
The problem arises because the types of thinset/mortar/adhesive can include latex-modified Portland cement, acrylic-modified Portland cement, emulsified epoxy Portland cement, 100% epoxy mortar, 100% epoxy adhesive, furan, organic adhesive, and standard Portland cement thinset. Each has a different thermal conduction coefficient, with latexes and acrylics being suspiciously low (see this thermal conductivity chart for specific materials like acrylics, epoxy, dense concrete, cement mortar, rubber (latex is often called synthetic rubber?), porceline etc @, but I can’t figure out how to use the formula to compute each approximate conductivity. ?
Other consultants say: the thinset/mortar/adhesive type doesn’t matter, none of them will interfere with your efficient conductivity.
Or others yet: definitely stay away from organic adhesive and latex-modified.
The National Renewable Energy Lab and the Dept of Energy’s office of Energy Efficiency both have no info re my question.
Both the concrete slab and quarry tile components of the thermal mass sandwich are well known to have acceptable conductivities.

2.) A better bond than standard Portland cement thinset (psst–it, naturally, has acceptable conductivity) is recommended by concrete technicians because this is a ‘cold joint’ application and not an ideal fine-broomed surface profile on the slab. They say that an epoxy adhesive would be an excellent choice if scarification of the surface is not an option for some reason (it isn’t an incredibly messy option!) and if epoxy adhesive can pass the conduction test (some say at least 10 BTU/hr-ft2-degree F) at its 1/16th” application thickness (only 1/16th” because standard Portland cement thinset can be applied over it if mortar sand is broadcast onto the wet epoxy to create excellent ‘tooth’ after curing)

Unimportant details: slab is well insulated from ground and perimeter stem wall foundation with 2″ extruded polystyrene and moisture proof with a tight vapor barrier. The gorgeous Sun-Glo quarry tile has already been purchased for the 700 square feet interior and I really want to use it if I can avoid slipping into stupid mode. The glass-to-thermal-mass ratio has been carefully calculated and built, control joints sawed appropriately.

Any furtive conductivity tests, quasi-engineer inputs, and general advice commentary would all be extremely welcome.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member
    Malcolm Taylor | | #1


    A couple of random comments:
    - You don't need to do anything special to prepare a power-trowelled slab for tile. No broomed, no roughening.
    - The type of mortar used (modified/unmodified) depends on the permeability of the two materials.
    - Relying of control joists in the slab has it's risks under tile. When those joints do their job, there is a good chance the tile will crack. Most builders now use a de-coupling membrane under tile on slabs - which I realize defeats your whole conduction strategy, bur just so you know what you may be getting into.

  2. thermal | | #2

    But the permeability of the slab is zero. The test of putting two drops of water on the slab and seeing how long it takes to absorb revealed no adsorbtion at all after 30 min! Many concrete techies say that the bond will fail if I tried to use Portland cement thinset on such a 'cold' surface because of both temperature and moisture differentials between the slab mass and the thinset in the future. Or am I misreading your comments?
    Control joints in the slab are matched with control joints in the tile field and in the thinset, eliminating differential expansion and contraction stresses/cracks, no?
    Yes, membranes probably inhibit conductivity continuity and therefore destroy the glass-to-mass ratio and invite overheating the air temps in the residence.

  3. Tom May | | #3

    You could attach some form of wire mesh or lath to the floor and use regular mortar. A masonry bonding agent may also be incorporated first if true bonding to the slab is a concern.

    1. thermal | | #4

      Tom May,
      Yes, I have a very highly recommended masonry bonding agent--Dural 452 MV, from Euclid Concrete Chemical Co. techies, which is an 100% epoxy adhesive providing extraordinary bonding capacity all by itself. But again, I only need someone to clue me in on its thermal conductivity firstly because my thermal mass efficiency is my very first priority, without doubt. Successful bonding is second.

      1. Tom May | | #6

        I don't think a thin layer of anything used to bond tile to cement will have any drastic effect on thermal conductivity. I would be more concerned about the color of the tile.

        1. thermal | | #7

          Tom, I'm thinking you may be onto a truth here, but I'm still struggling with the formula for thermal conductivity. Even if I'm not still clear with it, it is indicating a reciprocal relationship that the conductance has with the thinness of the bonding material. Meaning that the conductivity does, at least, rise upon lessening the thickness of the thinset. I wish a quasi-engineer would chime in with it.
          "Construction epoxy is actually an insulator, working to retard the flow of thermal heat into the thermal mass slab" says all epoxy professionals in the field. Is it significant to my application is my further inquiry.

          1. Tom May | | #9

            thermal, if it helps any, I am a mechanical engineer with 40 years in all fields of construction as well as 20 years in solar.. Conductivity (k) isn't a formula. It is a property of the material. The inverse relationship is considered resistance, but is relatively the same thing, low resistance = high conductivity. The formula connected to it is Q = - k A (T2 - T1), the heat rate, or amount of time it takes to transfer heat, through a given area, loosely associated with the thickness, and mostly with the type of material, with a difference of temperature, eg top of tile at T1 and the concrete at T2.. Even insulation has conductivity, it just moves at a slower rate. As mentioned, such a thin layer should not have much of an effect. The only real concern is the space between layers, eg an air gap, where air has a low conductivity. If all materials are in close contact there shouldn't be an issue. That's why they have such things as thermal grease as used on heat sinks to decrease any possible air gaps on uneven surfaces.

        2. thermal | | #21

          Tom May, The old thinking on the correct color of the tile was that it should be a darkish color to suck up as much thermal heat as possible in a location that is pleasant to the whole human body --the soles of the foot, and also where the mass was usually located in a conventionally built house. The new thinking is that a medium color (reflecting 40% of its total light radiation) is advantageous in two different ways: 1) it reflects (sometimes several times if it strikes my multiple white colored wall and ceiling surfaces) a portion of the radiation to areas further into the house to even out the eventual heat distribution throughout the house's mass surface area instead of concentrating it in one specific area and inviting that area's uncomfortable overheating. The idea being that the whole body receiving radiant heat evenly, head to toe, is just as good, or better, than just toasty soles, and 2) much superior daylighting throughout the house, esp to freakos like me who have no other lighting. The northern interior areas are too dark in a nonelectric house if they are more than about 15' away from the main sunny wall/window direction, so more reflected light there makes a difference. I guessed and chose a light brown color because I was swayed by how the new thinking would impact my multiple eccentricities.
          Have I erred, or, by now, you wouldn't even dare to begin to say?
          Please see an 'engineering' post I put to DCContrarian earlier this evening re a further clarification of my main question. Please do relay your view.

  4. DCContrarian | | #5

    I'm going to throw out some back-of-the envelope numbers.

    Let's say your house warms up to 80F during the day and cools to 60F at night, and the slab warms up to 75F and cools to 65F. Four inches of concrete is about 30 pounds per square foot. Concrete has about one quarter of the specific heat of water so a square foot of concrete has the heat capacity of 7.5 pounds of water. So with a daily temperature delta of 10F that square foot is absorbing and releasing 75 BTU.

    The units of R-value are ft2·°F·h/BTU. If we say it takes six hours for each transition we get 1ft2*10F*6h/75= 60/75=R0.8 is the highest R-value you can have.

    If those temperature swings are too extreme, let's look at what happens with more moderate ones. Let's say the indoor temperature varies from 68 to 72 and the slab goes from 69 to 71. So you're at 15 BTU/day. R-value is 1ft2* 2F*6hr/15=0.8, it stays the same. Since tile is typically 3/8 of an inch you're fine as long as the R-value per inch of the materials you're looking at aren't more than R2 per inch or so, which for ceramic tile would be a lot.

    This does show the problem with relying on thermal mass. In order for that slab to store and release heat, it's temperature has to change. In order for it to store and release significant amounts of heat the temperature has to change a lot. Just picking a random northern latitude city, Hartford CT gets average insolation in January of 1.88 kWh/m2 per day, or about 600 BTU per square foot. If you're capturing 10% of that your slab has to fluctuate by 8F each day. In order for heat to flow there has to be a temperature delta, so the house temperature has to fluctuate even more to drive that heat flow. Even with very little insulation over the slab the temperature would have to go from around 62 to 78 each day.

    1. GBA Editor
      Martin Holladay | | #24

      Bingo! DC Contrarian, you have the best answer on the page. Thermal mass only helps in a house with very wide temperature swings between late afternoon and early morning -- and those very wide temperature swings are unacceptable to almost any American homeowner.

      1. Robert Opaluch | | #34

        Thermal mass is one of the components of a passive solar building. "Wide" temperature swings of 68F to 78F might not work for some people. But most people tolerate 78F settings in the summer AC season, and 68F all day during the winter season. So no, they aren't unacceptable to "almost every" American homeowner.

        Martin, you pick examples of poorly designed buildings that are overglazed, leaky, and poorly insulated, which would produce temperature swings that are far greater than 10 degrees per day, and claim these are typical "passive solar homes". They are poorly designed homes. You could just as easily claim that "superinsulated" homes don't work, because air infiltration defeats the insulation. If Thermal wants to build passive solar, please help him out rather than discourage him because passive solar doesn't work in your climate or isn't your choice.

        Thermal, it seems to me that you are overthinking or worrying too much about heat transfer to and from tile and slab. I used ceramic tile, thinset, and a 4" slab, and had a 10 degree difference in interior temperature year-round (68F to 78F), using no auxiliary heating or cooling. But I did do a lot of modeling of heat loss, solar gain, and paid attention to air-sealing and pretty good levels of insulation. (I'd use more insulation and air-sealing with products available today, and less glazing area would be needed.)

        Where is your home being built? I'd worry more about issues such as your climate, making sure the space heating load for each room/space in the home is collecting the needed BTUs in January, and the home is airtight and well-insulated for a low heat load.

        1. Expert Member
          Malcolm Taylor | | #35


          Disregarding the passive solar potential of a design is a lost opportunity, but that's a lot different than trying to design a house around large facing windows and thermal mass to provide the main heat supply.

          1. thermal | | #46

            Malcomb Taylor, Sorry, I sometimes don't have a 'reply' at the bottom of the forum's posts in order to respond in the correct response sequences. I, newbie, don't know why yet. I'm responding here to a question you have asked elsewhere.
            Yes, my main design guide is 1979, but it promotes extremely simple and basic design materials and is therefore somewhat independent of the modern tech materials. Cordwood updates are pretty petty and since I've already constructed two buildings with it I do like familiarity to increase my chances of success. New research re direct gain systems is something I've followed in Bruce Anderson's (1996) work. But I don't always care for some (maybe lots) research because it depends on using high-costing manufactured materials and techniques to make it work. Another factor is that I'm deep in the wilderness working as a low-income DIYer alone, tending to pick doable stuff that requires no assistance or machinery as much as possible. Other new research I can use to further my construction education is indeed missing, but couldn't modern techies also be so focused on modernity that they tend to errantly dismiss timeless materials and techniques in the frantic hurry to get their stuff done or to look acceptable to middle class American values? I don't have time for fitting in (ie my choice for cordwood) or being subject/slave to the public's fickle opinion. Am I necessarily lost in my ancient craftmanship? Is there nothing at all to be said about 'timeless' building styles?

          2. Expert Member
            Malcolm Taylor | | #48


            There are hundreds of ways to build that make sense given the individual circumstances of the occupants. The Nearings, who I have immense respect for, lived in a stone house, and consequently threw huge amounts of wood at it for heat, and tolerated interior temperatures that would phase most people, as part of wider lifestyle choices.

            The responses you get on GBA tend toward a generic answer to what makes sense for most people, or for most people in most circumstances. If you have found that your house works for you that's great.

            Circling way back to my initial answer, what I was getting at was that the permeability of the materials to an extent dictates the options you have for choosing a mortar. If you pick an appropriate one for your situation based on adherence, I doubt the difference in the conductivity of the mortar makes any appreciable change in the loss or gain the slab experiences. And all tile exposed to direct sunlight experiences differential temperatures from the slab below. It is a very common situation, and doesn't seem to cause problems in the absence of some other defect in the tile installation.

          3. thermal | | #50

            Malcolm Taylor, Yes, I've got that opinion a few times now--- that I'm fretting over nothing at all. Thank you for being patiently insistent with me. Apparently I'm not hearing it deeply enough due to a desire not to make a critically irrevocable mistake at this late stage of this important project.

        2. thermal | | #45

          Robert Opaluch and others interested, I'm in eastern Ky's mountains, 5,000 degree days, good solar access to the south, down in a hollow that is an unusual microclimate. I studied the Ky mountain climate and the specific microclimate conditions off and on for 15 years, if you can believe that, before I decided on the house site for direct gain passive solar with an 'open' floor plan and a series of large south-facing windows covered on winter nights with R10 shutters. I totally adopted Edward Mazria's book Passive Solar Energy Book as a design guide because it was 'for dummies' as he put a ton tests, assessments, analyses, and calculations into thorough, but easy, to follow 24 'rules of thumb' for this dummy to follow. I followed his rules to a tee and wound up with decently low temp fluctuations (so far, I'm not totally finished yet) despite a less than ideal insolation of 50% during the winter, which, in turn, yields only a 40% heat savings. Woodstove for the remainder. Mazria was excruciatingly adamant on proper design controls to limit troublesome overheating, excessive fluctuations, etc by his rules: "the location, quantity, distribution and surface color of the masonry that will determine the indoor temp fluctuation over the day. To minimize indoor temp fluctuations, construct interior walls AND floors of mostly masonry with a minimum of 3 to 4in in thickness. Diffuse direct sunlight over the surface area of the masonry by using a translucent glazing material, by placing a large number of small windows so that they admit sunlight in patches, or by reflecting direct sunlight off a light-colored interior surface first, thus diffusing it throughout the space. His general design temp fluctuation of 13 deg F (74-61 deg F), comfortable for most building interiors, was compared to the exact same space but with lightweight materials (wood frame with a 1/2in gypsum board finish) that would have yielded a temp fluctuation of 38 deg F!, demonstrating the dampening effect of thermal mass on temp fluctuations. His surface area of concrete or mortar exposed to direct sunlight is 9 times the area of the glazing. Because of mean radiant temp (mrt) in high thermal mass interiors a significant advantage is attained over just relying air temp for human comfort. The major problem associated with passive systems is one of control. Since each system has a large heat storage capacity which is an integral part of the building's structure, its ability to respond quickly to changes is greatly impeded. Also, storing heat requires a change in the temp of a material, and since storage materials are an integral part of the living space, the space will also fluctuate in temp. Excessive space temp fluctuations can lead to unsatisfactory comfort conditions if the system is not properly designed......" Blah blah blah as you have also so similarly and excellently posted.
          I live 60% of the daytime, even at cold temps, outside with landscaping, logging, gardening, splitting wood, etc in a kind of self-sufficient, old fashioned way, so that the interior fluctuations I'm "passively accepting" don't seem at all excessive except sometimes in Sept. But I hear many others don't feel that way and/or don't have my thick skin.
          I also took his advise to also conserve heat by installed R40 insulation in the ceiling below a radiant barrier attached to the underside of the rafters. My 14.5in thick cordwood walls are only an average of R22 though. I took the time to seal the house up, esp the areas around the windows and doors with a ton of caulk and foam. I also installed 6 large ultra-violet transmitting (high energy wavelength) 'full-spectrum' acrylic windows that likely transmit a better solar fraction heat %, but I don't know the specs offhand re that. I kept the slab's edges from losing conducted heat to the perimeter foundation wall with 2in extruded polystyrene on both sides of the perimeter wall and 2in beneath the slab's underside also.
          I hear you that I'm fretting over the thermal conductivity of the epoxy adhesive but also note you have seemingly used a nicely conducting Portland cement thinset, which I've been cautioned against using. Please see Chris Ingwersen post I submitted today if interested.

  5. thermal | | #8

    DCContrarian, Please bear with me, because I've got to assess if we are really on the same page here with my fundamental question: is usage of, for a good example epoxy thinset, going to retard thermal conductance I'm receiving from solar radiation striking the surface of a quarry tile downward into the thermal mass slab? It is possible it may not be significant if I can keep the thickness of the epoxy at 1/16", but unfortunately I can't get a good read on that truth. After all, to gain the max amount of heat from the sun I can't just bypass that thermal conductivity question/function if the direct gain design is to be efficiently made use of.
    I followed your first paragraph but then got lost when you calculated R-value instead of sticking with conductivity. Also dense concrete is at least, if not more, 150lbs per cubic foot, no? Therefore 4" thickness will be 150/3=at least 50lbs,no?
    Can you do another back-of-the-envelope calculation if the quarry tile is a full 1/2" thick and I live near Lexington Ky, if you must think in terms of R-values?
    Please bear with me, I'll get back to you as I contemplate your numbers for relevance.

    1. DCContrarian | | #10

      Conductivity is the inverse of R-value. I used R-value because the units are familiar. Sometimes you'll also see U-value, which is a direct measure of conductivity. U=1/R

      1. thermal | | #14

        DCContrarian, Here is the issue: a 1980 study by direct gain passive solar researchers concluded that any floor covering or its thinset whose thermal conductance exceeded about 10 BTU/hr-ft squared-degree F will not impair the thermal storage capability of the concrete slab. The formula for thermal conductivity according to in an article entitled --thermal conductivity of selected materials and gases-- is this:
        q/A= (k/s)delta temp where q is heat conductivity in BTU/hr units, A is the surface area in square feet, k is the coefficient of thermal conductivity of the material (see the chart for Imperial units), s is the thickness in ft, and delta temp is the temperature difference in degree F.
        When I plug in epoxy's k value of .202 and a possible thickness of 1/16in (.0052) and just leave delta F at 1, I get q/A = 40. Sounds good? When I plug in 1/4in for the thickness instead of 1/16in I get 9.7. According to the researchers above mentioned that is not an acceptable thermal conductance! No? Have I done this rightly? If so, it reveals to my simple mind (bear with me a little longer can you?) that to maximize the thermal heat getting into the slab (exactly where I want it in order to store as much heat as possible for an overnight time period and satisfy the glass-to-mass ratio design recommended to me by Edward Mazria's rules-of-thumb to avoid overheating during the day) I should somehow keep my epoxy thickness to 1/16in if I can, and also avoid any thickness 1/4in or more.
        If I plug in the numbers for dense concrete at the same 1/16in I come to approx 200, which is 5 times the value of the epoxy, proving what epoxy pros tell me; "construction epoxy is an insulative material, not very conductive in general unless you get a special kind of epoxy called conductive epoxy (used in the manufacture of electronic circuit boards).
        But what the hell, I've got to trust the passive solar experts saying that epoxy's heat conductivity @ 1/16in is 4 times the minimum they recommend. As Tom May has suggested "that even though epoxy is insulative, if you keep it as thin as possible it won't matter a whole lot. Would you agree with that truth by Tom May and my back-of-envelope figures? Or have I missed the mark?
        As you and the formula have pointed out to me: delta temp drives the thermal heat conductivity. If I, in fact, have an average in-the-radiant-sunshine slab surface temp during a sunny winter's daytime of 74 deg and at the bottom of the slab @ 60 deg, am I driving the conduction sufficiently? After installing 2" polystyrene shutters over the large south-facing windows in an average 30 deg overnight, the following morn at 8am the slab's surface temp is 61 deg with the bottom of the slab's temp @ 56 deg. Do your calc's say it is at least partially working?

        1. Tom May | | #27

          Are you sure you have your units correct. If your k is in imperial units and you are using btus, your calculations are going to be off....
          When calculating the over all gain, if you are using thickness you are only calculating one dimensional. When using area, it is two dimensional. You cannot combine the two. You have to add the individual conductivity of each material to get an overall conductivity and then use that over the area. eg. a wall with wallboard, wood, insulation etc will have a single overall conductivity in the assembly. It is that value that you use over the area in your calculations. If you think of it as in electronics, several resistors in series will add up to a single resistance.

          1. thermal | | #49

            Tom May, Please go to and type in the search box the following: thermal conductivity of selected materials and gases. Click on that article that appears at the top of the list. Then you will scroll down only a little and find a k chart. Go to the k column heading and click on the small blue box that reads BTU/hr-ft-F. That action will put blue-colored Imperial unit values just below all the SI unit values for all the materials down the list. Then scroll down past the end of the list to the q/A formula and all the correct Imperial units in the explanations of the formula's meanings. If I've got you there correctly, this is exactly where I'm unsure of myself to read how this formula plays out as I calculate what the correct thermal conductivities of the various materials (ie epoxy, dense concrete, cement mortar, porcelain etc) are. Can you assist me in 1) making sure I'm using the right formula to fit what I'm really searching for, and 2) filling out the formula correctly, esp the units, to get answers to the question: what do I have as thermal conductivities of these and other materials, esp epoxy, that I can compare with what the passive solar researchers say I should have, which is at least 10 BTU/hr-ft squared in order not to interfere with the max (or near max) thermal conductivity to the slab surface.
            This a lot harder for me, the nonengineer, to explain than it will take you to see and understand, of that I have zero doubt. Then your statement that "the epoxy adhesive in a 1/16in thickness I'm thinking of using immediately atop the slab is insignificant to the good thermal conductivity to the slab surface" will be understood better by me (good God, hopefully),

          2. Tom May | | #51

            Thermal, in that formula, they are only calculating the heat transfer through the thickness of the aluminum pot. q /A is the heat rate per unit area. The area is not considered or used in the calculation. So as I mentioned earlier, it is only one dimensional.
            In the calculation table where you enter the values, that I assume you used, it even says it is generic and to be sure your units are consistent and agree with the required input. eg 1/16" in =0.0052 ft
            The "conductive heat transfer calculator" only considers one layer.
            Take a look at the "Conductive Heat Transfer through a Plane Surface or Wall with Layers in Series" which shows how you need to use the individual material conductivity of each layer which are added together, not just one layer, using "fouriers law". Which is basically the overall k of the assembly of which it does not have a calculator for but it does have an example.
            q = dT A / ((s1 / k1) + (s2 / k2) + ... + (sn / kn))

          3. thermal | | #52

            I'm giving up. Either you are too intelligent or I am too dumb to ask even the right question. I suspect the latter. I'm very sorry to have pestered you in my search for a calculated answer to my dilemma.

    2. Trevor Lambert | | #11

      Even materials specifically intended to be insulation rarely exceed R5 per inch. I'd be willing to bet that thinset epoxy is less than that, so if you have a 1/16 inch layer you're looking at well under R0.3. Cement based mortar will be even lower. I'd more concerned about whether your combination of insolation and thermal mass is enough to create the effect you're looking for. This has been well studied and the general conclusion is that most of the time passive solar is usually not even enough to offset the nighttime losses from the windows, let alone have a positive impact.

      1. DCContrarian | | #12

        Yeah, I get dubious every time I read the words "thermal mass," it's not a term used in engineering or science. The relevant quantity is heat capacity. I'll point out here that the rest of the house typically has a lot more heat capacity than the concrete slab. It weighs more and has a higher specific heat. The rest of the house might be 50 lbs/sf and it's mostly lumber and drywall. Drywall has the same specific heat as concrete, but lumber is about 0.6, or 2.4 times as high. So the rest of the house is capable of holding and releasing more heat than the slab.

        The analysis I've read is that the only case where trying to maximize the heat capacity of your house makes sense is a climate where you have dramatic temperature swings and part of the day is warmer than your indoor temperature and part of the day cooler. In that scenario the heat capacity acts to buffer the swings. This might be the case in a high desert climate. In most of the country the downside of high heat capacity is that it makes your house less responsive to mechanical heating and cooling, if you have an excursion away from your desired temperature it takes a lot more energy to get back.

        The problem with passive solar is there's no way to control it. The sun is exactly the same in September as it is in March, and pretty much anywhere you have a heating season September is the cooling season and March is the heating season.

        1. thermal | | #18

          DCContrarian, My unusual walls are 14.5 " thick cordwood, which is, they say, 40 to 50% masonry. The wood component does not have the necessary properties that makes good thermal mass but the masonry component, of coarse, does. Not much of that mass gets in the direct sun alot except for reflected light, so it is considered only "radiantly" valuable as a heat storage if it can "see" the sunlight striking the floor and only "convectively" valuable if it cannot see the solar radiation. Those two catagories have less value towards achieving the glass-to-mass ratio Edward Mazria insists on, but he claims that every bit counts toward getting the mass maxed out so as to buffer extremes. Seems to work so far, but I'm not finished yet.
          Of coarse, the thermal mass architecture really shines in summer when you open all the many vents at floor level in the evening when the outdoor temp is below the interior temp and allow storage of cold into the whole mass all night, then close the house up at 9am and be comfortable enough throughout the outdoor 90+deg temp day. In my air conditionless house (no electricity!) it saves the day. I happen to be in a Ky "hollow" that traps heat during the day, but cools much more than the surrounding landscapes because heavy cool air comes pouring down the heavily forested mountain slopes soon after the sun sets, resulting in a wide temp swing between day and night, something quite similar to the high desert swings. High humidity in Ky makes a difference though.
          I built calculated overhangs according to the geometries of the seasons for my Ky location. Also bought outdoor shades for several windows to block not only direct sun but also diffusely reflected light off outside surfaces, which the energy efficiency experts say can be a significant heat addition to your interior, even though the windows can't see any sun directly. But yes, Sept is usually a humdinger to beat, and yes, if you don't recharge the mass often enough for some reason you need to run the woodstove a longer time to get interior temps back to a radiantly comfort level. Insolation amounts in Ky's winters only allow a 40% savings in heating costs via passive solar, so a woodstove is a necessary supplement. Most busy people find woodstoves a time-consuming hassle, but me, the eccentric, likes too much the flame in my eye.

          1. DCContrarian | | #19

            What do you believe are the "necessary properties" for "thermal mass"?

          2. Expert Member
            Deleted | | #20


          3. thermal | | #22

            DCContrarian, A material that fulfills the necessary properties for thermal mass in a residential building has to have 1) a goodly heat capacity and 2) a rate of thermal conduction that is neither too slow (it never heats up in time) nor too fast (it adds too much heat to the air temp via its surface excess). Wood, I'm told, has good capacity, but too slow a conduction rate to make it fit the time frame of day's 8 hour heat absorption and then the night's 16 hour heat release. Concrete or masonry, I'm told, has the more ideal rate and capacity. "Of which you can almost never have too much". This assumes you have a well insulated structure to fulfill necessary heat conservation. See esp Bruce Anderson, passive solar author, because, as you can guess, I'm just following and repeating the experts' strong suggestions.
            Your thoughts? Please be simple, if you can, I'm not literate in engineering principles/calculations.

          4. Expert Member
            Malcolm Taylor | | #25


            I don't want to be rude, but are the works you are reading recent? The thermal mass, passive solar strategy, and things like building cordwood walls, are all things I haven't seen discussed since the late 1970s. Building science has seen some significant changes is the past decades based on both research and experience. Not a lot of practitioners still adhere to the ideas you are espousing.

        2. Robert Opaluch | | #36

          Malcolm, I think Thermal is referring to Bruce Anderson's 1996 book "Passive Solar Energy: The Homeowner's Guide to Natural Heating and Cooling"

          (I'm not familiar with the book myself.) Although during the 70's, passive solar became more popular in reaction to high oil prices and shortages, now the pressure to avoid fossil fuel usage and concerns about global warming could lead people to want to exploit passive solar heating, and avoid using auxiliary space heating as much as possible. In your climate area (Vancouver Island?), I'm assuming you have overcast winters, and passive solar would not work well. But in areas with sunny winters, it can. As you know, even with overcast winters, if you have good south-facing solar access during mid-winter, it would help daylighting and reduce heating bills if the building had more glazing area facing southward than other directions. Thermal mass is only needed to reduce interior temperature fluctuations ("store heat") in passive solar buildings. I understand your comment re exploiting solar heat gains vs. passive solar architecture (which requires thermal mass to reduce daily interior heat fluctuations). The key difference is thermal mass. We should help Thermal design his home to work well given the climate (including not building passive solar in climates that wouldn't support it, but not being opposed to passive solar in climates where it can work well).

          DCContrarian, yes passive solar buildings tend to run warm in the Fall and cool in the Spring. Both average outdoor temps and interior solar buildings would show that effect. Outdoor temperatures lag the seasons, because the Earth has solar mass too. However, there's plenty of ways to control interior temperatures by limiting solar gains in late August or September, yet not limit solar gains in late February and March. Interior blinds, adjustable exterior awnings, deciduous trees, ... That's not "the problem" with passive solar. People don't seem to understand how to engineer passive solar well enough, usually by misunderstanding how to design passive solar buildings, or by failing to do the math. Its not for everyone but if that's what Thermal wants to build, lets help him do it well, if his climate and building site are appropriate for passive solar.

          Thermal, I assume you have calculated the position (track) of the sun for your building site, done heat loss and solar gain calculations for each room/space in your proposed home for January at least, are building in a cold but sunny winter climate, etc.

          1. Expert Member
            Malcolm Taylor | | #40


            As I read his posts, Thermal's house is almost complete - save for the tile. It has cordwood walls, which he describes solely in terms of their capacity for thermal mass, and is really only interested in the minutiae of conduction of the mortar over his slab. That's the context for what may appear dismissive comments on his approach to passive solar design. It really does sound like the houses featured in architectural periodicals in the late 1970s. But that in no way should be read as a general interdiction against incorporating passive solar into a new house design.

      2. thermal | | #15

        You agree with most passive solar authors that " without good interior shutters over the large south-facing windows overnight in the winter, the windows lose too much heat". So I took the time to design those R10 shutters fitting into the single-pane, full-spectrum, acrylic window frames from the very start. There is a downside to that though-- if you do not have a very tight fit completely around the frames warm moisture-laden interior air seeps into the space between the window surface and the shutter and condenses heavily, esp at temps below 30 deg outside. At temps of 25 deg and below it is ice. Unfortunately, I am not the carpenter that could make that tight fit with the accurate cutting of the 2" extruded styro. Good thing I made the sills of PVC boards because there is water on those sills lots of mornings when I install the shutters overnight, necessitating a small mop-up job as a penalty.

        1. Expert Member
          1. DCContrarian | | #17

            Short synopsis: superinsulation won.

          2. Robert Opaluch | | #33

            This "classic" article is written by someone who is opposed to passive solar architecture, and who regularly shows examples of poorly engineered passive solar buildings to prove his point. His list of principles have some but not all of the principles used by successful passive solar designers. And he claims that excessive glazing and overhangs are key design criteria. So yes, if you design a passive solar home poorly, it will work poorly. You could just as easily (falsely) claim that superinsulated buildings don't work, because without air-sealing, the insulation fails to keep the building warm during cold weather.

            A passive solar building has to be engineered well, and it has to be built only in a climate with cold but sunny winters. For example:

            For some (but not all) of the quantitative design issues:

            The main "problems" with a well-designed passive solar home is that interior temperatures will vary, and the interior will be sunny during the daytime. For example, 68F at 7AM, 78F at 3PM. That's not for everybody. You could just as easily claim (falsely) that the "problem" with a "superinsulated" home is that the low-U windows facing random directions provide little daylighting, and a depressing grayed out appearance. Note that many people suffer from seasonal affective disorder, the "winter blues", from lack of sunlight.

            The "cost" issue is bogus. For example, most people ignore the cost of a central heating system (that's eliminated in a passive solar home). Another example, a direct gain slab is a floor as well as thermal mass (so its not "extra" cost). Just like Passive House overdoing the cost of insulation, but eliminating the central heating system, which helps compensate the insulation cost (somewhat).

  6. Doug McEvers | | #13

    A passive house does not have to have a slab exposed to direct beam sunlight to take advantage of solar gain. The interior surfaces within the building envelope hold a considerable amount of heat in Btu's. My own house holds about 20k Btu's per degree F according to my calculation. So if solar gain warms the interior above the thermostat setpoint, that stored energy is there to be utilized, offsetting the heat loss for a period of time.

  7. thermal | | #23

    Doug McEvers, But I'm greedy, I want the stored heat to last all night long! Therefore, I want to collect lots of solar energy thru big south-facing windows, efficiently store it in high capacity mass without overheating the interior's air temp, then dribble it back out slowly for 16 hours because I'm almost never in the mood to get out of bed at 2am in the winter and fiddle with a woodstove, even if it is a good one. Yuck! Therefore, I'm pure greed, if you don't want to call it super prissy efficiency.

    1. Trevor Lambert | | #26

      To be blunt, what you are asking for is pure fantasy. As has been pointed out, lots of people tried this back in the 70s and 80s. It doesn't work. Even if you managed to get enough solar gain to theoretically last all night (highly unlikely in the first place), how do you think you're going to control how fast it dribbles back out? It's not going to magically wait until night to start releasing the heat. One of the big failures of these designs was that the houses overheated during the day. I don't even have really big south facing windows, at least compared to what you're probably talking about, and I can tell you that the daytime overheating is a real thing. We almost have to open the windows in the winter sometimes (we don't, we just accept the slightly above ideal temperature). And even with that daytime overheating and superinsulation (more than double code) the heat system still has to run overnight. We don't get to 4 hours after sundown, let alone 16.

      1. DCContrarian | | #28

        To pile on, it's instructive to look at net-zero houses with PV cells and air-source heat pumps. They will have hundreds of square feet of panels with efficiencies around 20%, and ASHP's with COP's around 3. So they are basically using 60% of the solar that hits hundreds of square feet to heat the house. A passive solar house isn't going to have that high of an efficiency, and the area of the windows isn't going to be as great. So there's no way it's going to capture enough heat to meet the heating needs of the house. Increasing the amount of windows doesn't help, because they let out more heat at night than they capture during the day.

        What if you just want the solar to "help out" a conventional heating system? That doesn't work so well either, because active and passive don't mix well, the active tends to want to overwhelm the passive. All that "thermal mass" means that when the conventional system is running it has to work harder to get the building to a comfortable temperature.

        Rooftop solar thermal systems are essentially no longer installed for similar reasons, you just can't get the heat when you need it and PV systems work better, especially if you have net metering. Not that long ago you could go on Ebay and search for evacuated tube solar collectors and there would be hundreds of offerings, today they're just being sold as camping ovens.

      2. Robert Opaluch | | #37

        Trevor your home doesn't have the 4" slab or similar thermal mass to effectively reduce daily temperature variations. That doesn't mean that homes that are designed well for passive solar can't work. See:
        I've done it. It can work if designed well. It can also fail if you build in a climate with overcast winters, don't have a 4" slab or similar thermal mass, don't insulate and air seal well enough, don't have south-facing windows with solar access during January, etc. You must engineer it to work well.

  8. Chris Ingwersen | | #29

    Nothing to say about the conductivity issues for a passive house... but using epoxy to set your quarry tile can become very expensive.

    Epoxy is not usually used to set tile to concrete as it is harder to work with and it is pricier then cement based thinset. Also a 1/16 depth of mortar is probably not realistic thickness. You would not likely have proper coverage on the tiles with 1/16 so 1/8” of mortar under the tile is more likely.

    The cold joint reasoning is true if the concrete is smooth, however a good dustless grinder makes quick work of the concrete for such a small area. Most tile-setters have one these days. The epoxy grout is still a great choice for durability.

    You might want to check out a product made by Laticrete. It is an additive for modified (latex) thinsets that is supposed to help with conducting heat for electric in floor systems.

    Good luck with the project!

    1. thermal | | #41

      Chris Ingwersen, The conductivity of an epoxy thinset/adhesive was my original question, but it has since morphed into a much bigger issue than I thought it was going to (ie is a thermal mass slab and thermal mass wall surface in a direct gain passive solar design smart and/or valid in today's world).
      The epoxy adhesive is not being used as the lone thinset component. It is, partially because of the expense, being used something like a thin bond coat is used in painting. The epoxy's conductivity is an unusual question for the tile industry because not many tile installers have the problem of a questionable conductivity of the this type of thinset in a direct gain passive solar designed house where thermal heat conductance to a thermal mass slab is important to the designed glass-to-mass ratio to avoid excessive temp fluctuations (more about that subject later).
      Here are the details re how it is being used in my specific situation: I have a slightly rough (not finely- broomed as the tiling industry's specs want) and absolutely non-absorptive and very dense slab surface. When I talk to concrete materials pros I have trusted for 30 years they strongly caution that the somewhat too smooth slab's surface is also a 'cold joint' that is notorious for having a weak bonding capability when one is attempting to bond another Portland cement product (ie Portland cement thinset) to it. That it might look and act alright for awhile, they say, but under the loading of foot traffic, desk chairs with rollers, differential expansion and contraction forces due to temp (esp passive solar slabs) and moisture conditions (esp around kitchen sinks, showers and utility sinks if one uses a standard Portland cement thinset and grout) it will slowly deteriorate/delaminate and cause either loose and/or completely de-bonded tiles, unless I'm lucky. I told them luck sucks (love that mantra--- get it right the first time) and I want a more foolproof method that did not entail either very messy scarification or grinding (really dustless?) of the slab surface. Epoxy adhesive used in a 1/16in bond coat was their answer.
      Here is the clever way they suggest I use the epoxy; roll/squeegee on the 1/16in epoxy in a limited section of the floor, then immediately broadcast masonry sand (finer than concrete sand) onto (not into, because the weight of the sand does all the necessary embedment) the wet epoxy surface trying to completely (or mostly) cover the epoxy, let it cure up, vacuum up any loose sand grains, and then apply, at your leisure, a standard Portland cement thinset at the normal 3/16in (quarry tile requires that final thickness via a 3/8in notched size trowel) right over the bonded sand, and immediately lay the tile as normal. As you probably know, you cannot hope to bond a Portland cement thinset to an epoxy unless it is still wet and non-tacky, but the sand provides the permanent 'tooth' on the top of the epoxy that provides a lone DIYer (me) with a no-hurry subsequent step of mixing and applying Portland cement thinset plus accurate tile setting later on. This is the improved surface that is much better that the 'cold' one I had originally. Could I this jumble making any clear sense?
      But....... epoxy is an insulator (not what the thermal mass wants to have atop it maybe) of sorts and whether or not the 1/16in thickess is significant is my priority question. Please see other posts by Tom May, Robert Opaluch, and DCContrarian re that question, if you will.
      Latex-modified Portland cement thinset (some say 30% will be latex) is also very questionable re its conductivity because latex has a probable thermal coefficient worse than epoxy and if applied in the 3/16in could be unacceptable to the passive solar researcher's spec noted in my earlier posts. But then again, maybe I "worry too much". But latex is, no doubt, an adhesion-supplying component to all Portland cement products and may reduce any undesirable air pockets/voids somewhat. ? As far as an additive that is added to a latex-modified thinset to increase conductivity goes, I'm not sure that makes good reasoning--add a conductive additive after adding a mostly insulative latex component. But thank you for the tip, I'll check on it.
      I don't like any grout but epoxy for a whole lot of reasons, despite its difficulty to deal with, esp for DIY newbies. Am I asking for trouble?

  9. Expert Member
    Akos | | #30

    Say you have a very low loss house, 1500sqft, say 10k btu. Concrete is 0.2 btu/lb/degF, so your 4" slab can store 14.5kBTU/degF.

    So in case of this, to make it through 14h night, the concrete temperature needs to swing by 10F. Since it takes some deltaT to get heat in and out of the concrete, the room will need to be hotter than this, you'll probably end up with a 12 to 15F room temperature delta. That is well above the 4F to 5F most people will tolerate.

    Bigger issue is the house will be the hottest when you want to go to sleep, which will make it harder to sleep. It will also be the coldest in the morning, not very pleasant wakeup if you ask me.

    1. Robert Opaluch | | #38

      Akos, in the passive solar main floor of the house I built in Colorado, I ended up with 68F at 7AM and 78F at 3PM or so (not at bedtime) most days in January. (Personally, I'd rather get a warm afternoon than live in a home at 68F all day during winter. Personally I'd rather have a bright sunny interior than the wintertime daylighting in most homes. That's what I liked best about Colorado winters in a passive solar home.)

      In the upstairs bedrooms with only drywall for thermal mass (insufficient), the variations were 15-17F, along the lines you cite (that's solar tempered, not passive solar, since the thermal mass component is lacking). If it was too warm upstairs at bedtime, we opened the window for about 5 minutes (after all, its winter outdoors) and the room cooled off to a comfortable temp for sleeping through the night. In the five years we lived there, we never had problems with bedrooms staying too warm at bedtime using this method. And since the drywall didn't provide enough thermal mass, on about a third of January nights we used backup radiant electric (solar tempered, not passive solar). Our extra winter heating bills were about $25/month.

      Contrast this with the "typical" construction 1950's Cape in New England (with oil furnace forced air central heating) that I grew up in. Upstairs we froze our butts in the winter (while the downstairs was comfortable except the warm-cool cycles from an oversized furnace), and ran huge window fan all night in the summer. That's not comfortable at all, but we put up with it. Martin, many Americans who put up with those temperature variations in many "traditional" American fossil-fuel heated homes.

      And the cost of oil, regular maintenance, repairs and replacements far outstrips free solar gains and zero maintenance insulation.

      Bottom line: Passive solar only works well in climates with sunny but cold winters, with good solar access at the building site (often not available unfortunately), and when the building is engineered well for passive solar space heating. But that doesn't mean it doesn't work ever.

      1. Expert Member
        Akos | | #39

        I don't think anybody here would question the benefits of a bit of passive solar heating.

        The issue is trying to rely on it too much. Colorado is definitely one of the better places for thermal mass, besides a lot of sun in the winter it also has a lot diurnal temperature variation.

        In case of the OP, my back of the envelope calcs say that the 4" slab is not enough. More realistic number should be around 10" to 12". That is a lot of concrete.

        1. Robert Opaluch | | #42

          4" slab has been said to be about the right depth. Deeper than that, I've read that it apparently takes too long to absorb and shed heat when needed from that greater depth or mass. I don't quite get that. But if 4" works, why go with more?

          Yes Colorado is pretty much ideal. Few cloudy days in a row. I'd lose 3 degrees in the AM after each day it was overcast and cold all day long. Only once in 5 years got 3 overcast days in a row. On coldest days (-15F lows, highs below 0F) performance was good, since coldest days are clear (when heat radiates into space without that cloud cover).

          I've designed a passive solar home for coastal New England, which is nowhere near as sunny as east of the Rockies, but not terrible. Modeling seems to show it would work okay. Less glazing, more insulation.

          Pacific Northwest way too overcast for passive solar buildings, as is areas around the Great Lakes. Bitter cold North Dakota or Alaska wouldn't work, not just lack of sun but takes too much insulation (cost) to bring the heating load down to match solar gain potential. And natural gas price there is cheap vs. New England.

          In the past, buildings were designed to exploit the climate. Large shaded porches in the south to avoid direct sunlight in windows, central fireplace in New England to distribute heat toward surrounding rooms, etc. More recently, people choice a home style appearance they like, and put it in any climate, using heating and cooling systems to compensate for designs that don't fit the local environment so well.

  10. Brad | | #31

    I don't have anything to add to the OP's questions, other than if its thin, it probably doesn't make much difference, and you are probably worrying over something that is down in the weeds. If you want to get that detailed, you''ll need to really understand the heat transfer, and model it.
    I do feel the need to respond to a lot of misinformation in the responses. I received a great amount of good advice in this forum when I was building my house, but one area which I think many here are misinformed or regionalist is passive solar.
    There is a good article here:
    This article:
    has a lot of information that is just wrong for the western USA. He quotes a Prosskiw paper, which has a bunch of weak or completely unsubstantiated statements. He quotes Lstiburek who says "“Your house will overheat in the winter. Yes, you heard that right. Even in Chicago" I don't know why that would be true in Chicago, when my house in central Colorado never overheats in the winter.
    In this thread, DCContrarian says "because they let out more heat at night than they capture during the day." This is not true in most climates. Refer to the blog by R Opaluch that I mentioned earlier.
    A test case is my house which has 9% south glass to floor area. An energyplus model shows that 35% to 40% of our annual heating requirement comes from the sun. Real world measurements agree pretty well with the model.

  11. Tom May | | #32

    Yup Brad, most of these newbies probably keep their shades closed all day and believe the "new" science that cost them an arm, leg, head, and foot and that 2 +2 = 3 because the gov. tells them so....Guess you can't teach a new dog old tricks.

  12. Mike Theis | | #43

    Go get three sacks of sackrete and pour three 12*12*4 in squares on and formed in foam.
    Embed a thermometer in the middle of each. Set in front of your windows and record temps till you get bored. Attach tile to each in three different ways and record temps till you get bored.
    Make a choice and move on with it.

    1. Tom May | | #44

      Too funny.....

  13. Trevor Lambert | | #47

    "Trevor your home doesn't have the 4" slab or similar thermal mass to effectively reduce daily temperature variations"

    It doesn't? Ok, that's news to me. Lol

    We do have a lot of overcast days in the winter. That doesn't explain why even on full sunny days and the house is overheating during the day it still doesn't retain enough heat to last overnight. It quite likely reduces the heat needed overnight, but likely not even by half.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |