Mass Effects

Advancing technology generally leaves a reduced material footprint in its wake - as time goes on, it takes less and less physical stuff to perform some given function. Tiny transistors replace larger vacuum tubes, thin LCDs replace heavy CRTs, oscillating quartz crystals replace complex mechanical clock movements, electric motors replace steam engines, etc. Advancing technology gives us new materials with new properties (allowing, say, substituting a large amount of masonry for a small amount of steel), greater precision and control in manufacturing (allowing smaller and smaller transistors), and new processes that can get the same outputs for fewer inputs (such as the cyanamide process being displaced by the Haber Bosch process). Software and semiconductor technology is the ur-example of this, where reams of paper, vast rooms of hardware, and armies of human calculators (along with the facilities to house them) were gradually replaced by a few atoms of silicon and some carefully arranged electric charges.

We see the same pattern with building technology. Over the course of the 19th and 20th centuries, much of the advances in building technology were systems that allowed buildings to get lighter and lighter [0]:

  • In the late 1800s buildings over a few stories were built with heavy masonry walls, which had to get thicker and thicker as the load on them increased (at the base, they might be 6 to 10 feet thick or more). Iron and steel framing allowed the same building to be built with an almost unbelievably light and airy structural skeleton - the first skyscraper, the Home Insurance building, weighed just 1/3rd of what a masonry building would have weighed.

  • Light framed wood construction replaced more material intensive timber frame construction. 

  • Trusses, originally used only for bridges, gradually found their way into building construction, with components such as open web steel joists and metal plate wood trusses.

  • The dawn of modern structural engineering coincided with the construction of Eiffel Tower, then the tallest structure in the world. It was designed based on calculated wind forces, and was optimized to use as little material as possible. It stands twice as tall as the Great Pyramid of Giza (a previous record holder of world's tallest structure) despite using less than 1/300th the volume of material.

  • The (re)-invention of reinforced concrete, which was placed as a liquid and thus could be molded into virtually any shape, allowed placing building material exactly where it needed to go - allowing thin concrete shell structures and large spanning elements with ribs just where they needed to be.

  • A conventional steel beam shape can be thought of as a rectangle with the unnecessary material carved away. Building components such as hollowcore plank and masonry blocks are based on similar principles.

  • Modern metal buildings, with their extremely thin steel decks, thin CFS infill framing, and precisely engineered steel sections, enclose a given volume of space with as little stuff as possible, which is why they’ve become the default building type for large swathes of commercial and industrial construction.

  • Advancing technology has allowed what were once systems so large that they needed to be part of the building itself to gradually get pulled out, and into separate appliances.

Material Volume and Building Performance

Using less material is almost always a winning move if you’re trying to reduce costs - less material and lighter components means lower transportation and handling costs, simpler assembly, and less risk of injury. And it’s particularly beneficial for building construction - the lower your building’s weight, the lighter structural members you can use, the smaller your foundation gets, and the lower the seismic forces you have to design for.

However, there’s a limit to how far we can extend the process of carving material out of a structure. With most produced goods, we can theoretically extract more and more functionality from a given amount of matter by being increasingly clever with how we manipulate it. A clock will tell time just as well whether it’s based on a movement built with gears and springs or with a microscopic quartz crystal - there’s only a loose coupling between how well it does its job and how much stuff it uses. 

But with a building, the functionality is directly connected to how much stuff it uses. A building is useful to the extent that it creates usable space, and creating usable space - creating a controlled environment - requires wrapping it in a shell of physical material. And though this process hasn’t stopped, a few factors conspire to make it increasingly more difficult over time.

For one, many building performance characteristics are closely coupled to building section geometry, which closely tracks material volume - the thicker and heavier your building elements, the better it will perform on a number of different axes:

  • Floor vibration is a function of the floor system's weight and depth - the thinner and lighter your floor, the bouncier it will be when walked on.

  • In a similar vein, floor and roof deflection is a function of moment of inertia - how deep your section is, and how much material is placed at the outer edges. And because deflection is independent of material strength, using less of a stronger material won’t save you [1].

  • The thicker your wall, the less sound it will transmit - there’s a reason “paper thin walls” is not a positive descriptor. Multifamily buildings with wood construction almost always have “party walls”, a back to back stud wall with a space between to prevent sound transmission between units. 

  • Fire resistance is improved the thicker and heavier your building elements are - thinner elements will either burn directly (if it's something like wood) or lose strength and buckle (if it's something like steel). Firefighter guides for construction types focus much more on the dangers of “lightweight construction” (which includes light framed wood but also things like open web steel joists) rather than simply the combustibility of the material.

  • Green building techniques are overwhelmingly focused on energy performance, which is most easily achieved by having a very thick section that heat has a hard time flowing through. Buildings built to high energy performance standards will often have very thick walls, and might even have something like double stud walls, will have extra layers of panes in the glazing, etc. The most recent version of the International Energy Code, for instance, makes it nearly impossible to use an exterior 4 inch wall in residential construction, instead requiring a 6 inch wall.

Some of these factors are more tightly coupled to material volume than others. There’s ways of getting high fire resistance with less material, via intumescent paint for instance. And some green building techniques, such as advanced framing and reduced jobsite waste, also reduce material. By contrast, floor vibration is very closely coupled to floor mass.

None of this means it's impossible to slash material use to the bone and build a building using, say, thin sheets of fabric - yurts exist, after all. But it puts us in the realm of tradeoffs, rather than an expanding frontier of building capabilities.

The other major factor is that at a certain point, removing material begins to negatively affect the experience of a space. People value the feeling of solid, weighty materials [1], and are willing to pay for them. Many building technology “advances” that remove material are things that also make a building feel cheaper, and harm the experience:

  • Hollow-core doors vs solid wood doors - solid wood doors feel much nicer.

  • Brick exteriors vs fiber cement vs vinyl or aluminum siding - our sense of how nice the material is almost exactly matches how heavy and substantial it is. Same for slate or tile roofs vs asphalt shingles vs (gasp) metal standing seam.

  • Heavy ceramic tile showers feel luxurious, while one piece fiberglass shower enclosures feel cheap.

  • Same with wood flooring vs the much lighter luxury vinyl tile (people weren’t fooled by ‘luxury’ in the name).

  • Cheap paneling as used in manufactured homes or in 70s construction, vs thicker and heavier drywall (which itself probably feels cheap and insubstantial compared to plaster).

  • One thing that has harmed the use of advanced framing is that drywall over 24 inch stud spacing will have more give when pushed against than 16 inch stud spacing.

These preferences persist even when the lighter material actually performs better - luxury vinyl tile is basically indestructible and can be made to look like anything, but it still feels cheap compared to wood. A fiberglass shower stall is less likely to leak, and is lower maintenance (since you don’t have to clean or repair the grout), but people want that tile bathroom. And this isn’t simply a preference for more “traditional” materials - quartz countertops and concrete floors are comparatively modern but don’t come with any sort of perception of flimsiness.

People want their buildings to feel solid and substantial - if you push on something, it shouldn’t move. If you walk on a floor, it shouldn’t bounce. If you give a door a gentle nudge, it should have enough momentum to close and latch. Give people the option and they’ll almost always pick the heavier stuff.

Material Volume, Manufacturing, and Cost

Because modern technology is mostly a game of wringing increasingly large amounts of capability out of increasingly small volumes of matter, we often sort of forget the costs associated with moving large, heavy things around. For large swaths of products (say, anything you’re buying on Amazon) the cost of moving something halfway around the world are secondary compared to the potential savings on labor, or from economies of scale, or from being in a dense network of suppliers and manufacturing expertise.

But as we saw with brick, this calculus changes when your product value density drops. $10,000 worth of smartphones can fit in a shoebox; put in your backpack and you might not even know it's there. The cost of physically manipulating the product is almost an afterthought. $10,000 worth of brick, by contrast, equals a cube approximately 9 feet on a side that weighs somewhere in the neighborhood of 100,000 pounds; the costs of physically manipulating brick shapes the entire structure of the industry. 

These costs are difficult to knock down, no matter how much industrialization or mechanization you throw at the problem. The table below contains per-cubic-foot costs for several different goods and materials produced in enormous volumes, with large amounts of optimization pressure [2]:

Moving things around requires infrastructure, and the more stuff you have to move the more infrastructure is required. Coal is inexpensive, and in some ways requires less handling infrastructure than other things (it can be transported in open-topped rail cars, and stored in huge outdoor piles), but it still requires a massive physical infrastructure to get it from where it’s mined to where it’s used - ships and railcars and conveyor belts and massive mining equipment. Municipal water supplies require miles of pipe, pumps, treatment plants, and distribution facilities. Corn requires acres of land, enormous combines, massive grain silos, ships, and trucks to move around.

Alternatively, you can reduce your transportation costs, and instead invest in more distributed production (this is what things like EPS and brick do), trading off the cost of handling infrastructure for more production infrastructure.

Either way, at some point you hit diminishing returns. Costs of physical infrastructure can only be spread so far - at some point your returns to scale stop returning.

We see above that many building materials (lumber, concrete, gypsum) are made basically as cheap as anything else we make or use in large volumes. Things like lumber and drywall are more expensive than something like coal or corn, but not THAT much more expensive - it’s not easy to see how they could be made substantially cheaper, given that they’re already mass produced in huge volumes.

For residential construction, roughly half the cost of construction is the material itself. Barring some massive revolution in material production (large-scale automation that brings costs to near zero? next-generation 3D printers that allow local manufacturing of almost anything?), there’s not much obvious room for building material costs to drop.

Future of Material Use

If we can’t expect building material costs to come down much, that brings us back to looking for ways of using less of it. What sort of building systems does a strategy aimed at maximum material reduction suggest?

For current systems, it suggests something like load bearing cold formed steel. Cold formed steel uses a small volume of material (thin sheets of steel folded into various structural shapes), that can pack very densely in tightly rolled coils until the moment of fabrication. This actually allows for a certain degree of on-site fabrication - you see this with gutter installers or even standing seam roof installers, that can site-fabricate components from a roll of steel. Builders like Scottsdale Construction will drive a roll-former out to the jobsite, and fabricate and build the entire framing of a house in a matter of days.

For future systems, it suggests something like Ultra High Performance Concrete might be promising. Combined with concrete 3D printing, which can economically place concrete only in the areas it’s needed, we can imagine concrete buildings or components with a massively reduced material footprint.

[0] - We might expect this reduction in material to result in lighter buildings. But there’s a sort of Jevon’s Law factor at work here, where a given reduction in material spurs an increase in consumption enabled by the decreased cost. Structural steel didn’t result in less building material use, but instead enabled the construction of unbelievably tall structures which used as much or more.

[1] - Not sure to what extent this is a western or American thing, or if it’s more universal.

[2] - Coal price is based on 50 pounds per cubic foot and $171 per ton. Lumber is based on $400 per 1k board ft. Water price is based on a municipal price 0.29 cents per gallon. Corn is based on $3.2 per bushel.

[3] - Much of the culture around improving manufacturing efficiency is centered around lean practices, which is based on maximizing the value-added portions of the process, eliminating the wasteful portions, and having a continuous “flow” of production. And this sort of thing is important, even for bulky things like lumber or drywall (visit a well-run sawmill and you’ll see an operation based on squeezing as much product flow as possible through the system). But much of lean seems to implicitly assume that the costs of moving between steps don’t need to be considered, and what matters is the value-adding portions of the process. But the bulkier your material, and the lower product-value-density it has, the smaller the value-adding portion is in relation to the costs associated with physically moving a given volume of stuff. At a certain point, economies of scale between processes matter. We see this all the time in logistics and transportation, which rather than continuous flow or small batch size, is moving towards larger and larger batch size, with longer trains and larger container ships that can achieve better economies of scale - In 1920s the largest 13% of the world's ships carried just under half the total cargo. 100 years later, it was down to 5% - cargo carried proportionally more by a smaller number of larger ships.