Why Ancient Roman Concrete Outlasts Our Own (2017) | Hacker News Reader

Why Ancient Roman Concrete Outlasts Our Own (2017)

sci-hubtw.hkvisa.net

54 pointsNiceWayToDoIT68 comments

In a second year class I took the prof posed us the question: why don't we do this today? Why is this not part of an ASTM standard?

He ended up saying it's too tough to get a hold of proper pozzlanic ash, but I suspect its more a "this is the way we've always done it" difficulty. Does anyone know more to the story?

Maybe industries “planned obsolescence” practice?

But surely some company would see it as an advantage over competitors and sell it if it was easily accessible. But we don’t see that so there has to be more to it. Maybe cost is too high for the average customer?

> Maybe cost is too high for the average customer?

Well yes, and even for the non-average one. Plus the goal is generally to build things which barely stand, a building which can face the vagaries of time for thousands of years would generally be way over-engineered.

Countries where pozzolanic ash is readily available do use it but that’s not usually the case. Fly ash is also being explored for that role.

An other issue is understanding of the material’s properties over time.

Also our modern world makes a lot of concrete.

> Plus the goal is generally to build things which barely stand...

Since when?

Design criteria baked into international building code is quite robust; not just naive static load, but wind, snow, seismic activity, fire/flood/termite resistance, etc. are all considered. That's a far stretch from something that can "barely stand".

> Since when?

Since always

> Design criteria baked into international building code is quite robust; not just naive static load, but wind, snow, seismic activity, fire/flood/termite resistance, etc. are all considered. That's a far stretch from something that can "barely stand".

“Barely stands” in this context is “Barely passes code”, as in “any idiot can build a bridge that stands, it takes an engineer to build a bridge that barely stands”.

The more you go over the more money you’re wasting.

> But surely some company would see it as an advantage over competitors and sell it if it was easily accessible.

This is not a guarantee. Look at manufacturers of washing machines: every company makes more money by having unreliable machines. There is no financial incentive to build a better product, and it's not easy to even prove you've built a better product.

Would that logic suggest that washing machine brands are equally unreliable? If not, why not?

Eg Miele benefits from having a reputation of building reliable products. They can and do charge more for that reputation, but also have to live up to it, if they want to keep it.

I've never heard of Miele. If it's a US brand, it is clearly not doing well.

If it's a brand in the EU, the laws there were fixed to prevent every brand from selling intentionally defective products[1].

1. https://uspirg.org/blogs/blog/usp/you%E2%80%99re-not-crazy-y...

Miele is a German brand with a long history. You can buy them around the world.

https://en.wikipedia.org/wiki/Miele

> Miele (/ˈmiːlə/ MEE-lə; German: [ˈmiːlə]) is a German manufacturer of high-end domestic appliances and commercial equipment, headquartered in Gütersloh, Ostwestfalen-Lippe. The company was founded in 1899 by Carl Miele and Reinhard Zinkann, and it has always been a family-owned and run company.[2]

There might be laws that set some minimum standards. But (a) there are still differences between companies' approaches to quality, and (b) Miele is selling around the world also in jurisdictions were EU rules do not apply, but they still deliver on their reputation for quality.

  It's not that simple.
  The consumers do not want washing machines that last forever. 
  Are you sure that you want a washing machine from 1900? It was probably not electric. 
  Aaah, you want an electric one. How about one from the 1920? What do you mean you want intelligent programs, save electricity, be gentle to the clothes and use little water/soap? These are modern improvements of the washing machine. 

  Consumers vote with their money. They stop buying old washing machines because they don't have the features that they want. So companies stop making old washing machines and make new ones with more features.

Brother printers are low cost, cheap replacing the ink/toner and make good drivers. I definitely buy them only.

I think designing your buildings for more than a few hundred year counts as 'excessive' and more than 250 years should not allow for being disqualified under 'planned obsolescence'.

No. An industry does not make a plan 50+ years in advance to get more business.

They do, however, save on costs by delivering products that don't last as long because their customers also don't (as a general rule) care about 50+ year time horizons.

Concrete is typically made with local materials, because shipping them is rather expensive. Maybe that’s one reason?

> Concrete is typically made with local materials

Not really, especially jot for large contructions. Not everybody has a few thousand to million tons of sands of the right composition in their backyard, and cement comes from a factory.

Yes, and these factories are many, and are not concentrated in one place, because shipping cement is expensive. It’s the same with aggregate: not everyone has sand in their backyard, but there typically is good amount of suitable sand within a few dozen miles.

It's even more interesting that than!

There's actually plenty of drama around sand for concrete. There's talks about 'sand mafias' in India. And Malaysia has banned exporting sand to Singapore.

See eg https://www.bbc.com/future/article/20191108-why-the-world-is...

Similar stuff has happened in Germany, in fairly recent times. You could make your own cement, no problem. But good luck trying to unload it anywhere in Germany without paying a fee to certain individuals.

Interesting. Do you have any links with background information? In German is fine.

Email?

In my profile. Thanks!

First, we only see the part of Roman architecture that lasted. There's an observer bias. The Romans also made plenty of ephemeral stuff, just like we are making plenty of ephemeral stuff.

Second, there's more to engineering than making stuff last long. There are different trade-offs. Cost being one of them, but also different material properties.

Eg re-inforced concrete is awesome for lots and lots of applications that the Romans couldn't even have dreamed of. Alas, it's not economical to make re-inforced concrete that lasts forever. (Not even sure if it's physically possible.)

Trade offs should be covered a lot more during engineering studies. I know that during my time at university, as an industrial engineer destined to work at the interface engineering and economics, that aspect wasn't covered nearly as extensive as it should have. The other question that usually get's ignored is maintenance. The Roman stuff we see lasted millennia without maintenance, we on the other hand can maintain the stuff we want to last a long time.

That we don't do it, e.g. infrastructure with the particularly bad maintained bridges in Germany, is not the materials fault, or the original designs fault.

> we on the other hand can maintain the stuff we want to last a long time. That we don't do it, e.g. infrastructure with the particularly bad maintained bridges in Germany, is not the materials fault, or the original designs fault.

I have to respectfully disagree on this one. Humans in general have become spectacularly bad at maintaining physical infrastructure across longer timespans. Technological advances have enabled a cheap/fast/overbuild culture. This very much includes the design phase.

Look at the evolution of design. The large majority of our infrastructure, depending on the spot on our earth, largely dates from the last 200 to 50 years. Look at infrastructure predating that. Look at evolution. The bond with local communities depending on the infra? You'll inevitably find it cut. You'll find more bloat in the design. You'll find less local involvement.

To some extent, this is progress. Unfortunately, this has an impact on maintainability.

Looking at your example of German bridges. I'll make it even more tangible and look at the Eifel region with so much of its infra recently destroyed by flooding. How do we get excellently maintained bridges, when we know this is very much against human nature?

This means questions like: - Does this bridge really need to be (re)built? To this specification? In this place? - How much does this bridge benefit the local community where it is built? Can we think of ways to increase that? - How much of this bridge absolutely needs to be built out of reinforced concrete? - Wouldn't it make sense to build some infra in now very flood-prone areas out of less durable but cheaper, more quickly replaced materials like wood? - Can we bring the design closer to the layman? Can we for example design a bridge so that it will visually degrade in step with safety degradation? - ...

> How much does this bridge benefit the local community where it is built?

Why does a bridge have to benefit the local community? Why not eg the wider community?

> Can we bring the design closer to the layman? Can we for example design a bridge so that it will visually degrade in step with safety degradation?

What's the benefit?

> I have to respectfully disagree on this one. Humans in general have become spectacularly bad at maintaining physical infrastructure across longer timespans. Technological advances have enabled a cheap/fast/overbuild culture. This very much includes the design phase.

You say this like it's a bad thing. If stuff becomes cheap enough to build that every generation can afford to build their own, that's much better, isn't it?

Your comment emphasis closeness in space a lot, with talks of local community etc. So why not emphasis closeness in time, too? Surely the people living at a particular point in time might be best place to judge what infrastructure they need; instead of having to forecast hundreds of years in advance?

> Why does a bridge have to benefit the local community? Why not eg the wider community?

This was written in the context of maintainability. I'm obviously not opposed to infrastructure benefiting a larger community.

Imagine two equal bridges. One benefits the people nearby a lot, the other not so much. Which one do you think will get maintained best?

>> Can we bring the design closer to the layman? Can we for example design a bridge so that it will visually degrade in step with safety degradation? > What's the benefit? This was again written in the context of maintainability. Imagine two equally unsafe bridges. One to the layman looks "visually ok", the other "a nightmare of fear crossing this one". Which one do you think gets repaired first?

>> Technological advances have enabled a cheap/fast/overbuild culture. This very much includes the design phase. > You say this like it's a bad thing. If stuff becomes cheap enough to build that every generation can afford to build their own, that's much better, isn't it? I was mostly saying this is not an optimal thing. Cheap/fast/overbuilt can be a real burden in the long term in terms of maintainability. Budget and environmental issues are also very closely related.

> Your comment emphasis closeness in space a lot, with talks of local community etc. So why not emphasis closeness in time, too? Absolutely! That's why I mentioned an example of building infra in now very flood-prone areas out of potentially less durable but cheaper, more quickly replaced materials like wood.

Thank you for your comment. I'm grateful for the opportunity to discuss this. It's a subject that almost naturally attracts my attention. Might have something to do with living between Belgium and Latvia. Belgium has lots of physical infra, often not very well maintained. Cheap/fast/overbuild is definitely a thing. Latvia has a lot less infra. Due to history's course, there's very interesting distinctions in terms of infrastructure. Some of it is cleverly minimalist. Most of its new infra is heavily EU subsidised. Many projects are no doubt very beneficial, but often one can almost smell the bloat of needing to spend those sweet subsidies. In many places, the bulk of infra is Soviet era. A significant part of what's still in active use is often badly maintained or not at all. Then there's the enormous visible scars in the landscape of crumbling disused infra. Some fascinatingly sad examples are the giant former agricultural collective farm buildings that are falling apart all over the country.

Are you essentially saying that you don't think there are important material and chemical differences between Roman and modern concrete that might be responsible for orders-of-magnitude differences in durability, aside from the absence of rebar, and that it's basically a matter of building so many bridges, harbours, and aqueducts that some of them end up lasting for thousands of years despite being immersed in running rivers or salt water?

Absence of rebar is a huge difference. From what I understand it's the primary reason most modern concrete structures are expected to last only on the order of decades to a century or two. Some modern structures have been built without much if any rebar, for instance the Hoover Dam, and I've heard that might last thousands of years.

But beside that (whch surely is the main reason), there is also a great difference between pozzolanic and portland, nowadays (and since several years) pozzolanic cement (for whatever reasons[1]) came out of use and anything today (and since several years) is portland.

I have worked with both in massive structures (mainly bridge foundations and tunnel lining) and the differences between the two is staggering, in practice portland cement based concrete is very good (compression resistance) already at 3 or 7 days, just fine at 28 days, but tops around the 60-90 days.

Pozzolanic cement based concrete is barely reaching specs at 28 days, but continues to mature (increasing resistance) for years to follow, expecially in massive layers and in humid environments.

If you prefere Portland is a better cememt because it is fast, but - given time - pozzolanic is way "stronger".

In tunnels (which have all the best requisites for concrete to mature correctly) we had at the time (some 30-40 years ago) specs of 250 Kg/cm2 cubic resistance (at 28 days), we used pozzolanic cement for the lower arch and portland for the vault, while both reached specs, after 2-3 years we made some tests and the vault (portland) reached 300, in some cases 350, the lower arch was never below 450, in some cases 500 and even 600.

[1] essentially because it cures faster, allows thinner layers/slabs and for anything where formwork is involved this makes a huge difference

What about non-volcanic pozzolans like fly ash? From what I understand, fly ash is sometimes substituted for some portion of the portland cement that might otherwise have been used, because works similar to volcanic ash to make the concrete stronger.

Fly ash is similar but still doesn't perform just like basaltic volcanic ash [1] in admixtures.

[1] https://www.mdpi.com/1996-1944/12/16/2603/pdf

Fly ash is - generally speaking - an exceptionally good additive, its use is essentially due to the size of its particles, you can consider it as a very good filler with some added value (and it has some very useful side-effects, namely it makes concrete much easier to pump).

Besides the raw resistance, you have to imagine concrete as being a sort of artificial stone, the idea is to fit into a given volume as much material as you can, and you obtain this by mixing together gravel (usually one, two or even sometimes three sizes), crushed sand (rather big in size) and (where available) natural sand (or more finely crushed sand).

Then you add the cement, which is the finer "powder like" material, and water, BUT in many cases the "granulometric curve" remains "empty" in the lower part, and there are practical limits in the amount of cement you can put in the mix, so you need to add something (a filler) that is fine as or finer than cement, and this is often fly ash, which while not being as powerful as cement as a binder has anyway a pozzolanic effect, that helps in reducing the permeability of the set concrete, the matter is briefly explained in the second part of this:

https://en.wikipedia.org/wiki/Pozzolan

No. That's not what I am saying.

Both us and the Romans are and were building lots of different things for lots of different purposes with lots of different properties and trade-offs.

Some of the earliest re-inforced concrete was used to make better flowerpots. https://en.wikipedia.org/wiki/Reinforced_concrete#History

You can bet your hat that this was a very different type of (re-inforced) concrete than what we are using to build houses these days. We use yet other kinds of concrete for tunnels or underwater. Just like the Romans used different building materials.

I read the person's comment as saying - Romans built a lot of stuff, some high quality and some presumably low. We only see the high quality stuff because the low quality stuff has long since disintegrated. We then assume that everything they made was high quality instead of a mix of qualities because we never see the low quality stuff.

Basically survivorship bias

Not to mention that concrete steel reinforcement is the main reason it doesn't last due to water eventually making its way and the rust taking expansion thus cracking the concrete.

Interesting... we’re maybe a century or two into heavy use of reinforced concrete. Does it mean that eg the iconic buildings of Manhattan will eventually need knocking down and rebuilding?

Depends on construction. Steel core buildings have different issues. But reinforced concrete structures certainly will. I think Manhattan in general is somewhat challenging environment, combination of temperatures going below freezing, being relatively wet at times and being near sea. Ofc, good maintenance can prolong the lifespans.

Would the use of stainless steel address this problem? What other problems would it create?

Stainless of the right grade to address the problem is project-killingly expensive. There is work on evaluating alternative kinds of reinforcement not subject to the same kinds of rust problems, but extremely few organizations today plan, build and maintain over the kinds of timescales where this comes into play. They'll give lip service to those timescales, but watch what they do with their budgets, not what they say.

Well, fiberglass rebar is used more and more.

Its longevity has yet to be proven though.

> First, we only see the part of Roman architecture that lasted. There's an observer bias.

Couldn’t it also possible that the romans had some techniques that they weren’t aware of that would last longer than others?

Yes, of course. I didn't offer a guess why some things lasted longer than others.

Might be sheer blind luck, might be deliberate design, might be side-effects of things done for other reasons (like you suggest), or something else. Or a mixture.

I was more referring to the observation bias you mentioned. If they were making informed decisions on which technique to use for it to last a millennia or not then I’d also think there was some sort of observation bias. But if they didn’t, there wouldn’t be any observation bias since the process is then mostly random because they were not really aware of what technique that worked, no? We’re just seeing the results because they sometimes used a technique that lasted.

> But if they didn’t, there wouldn’t be any observation bias since the process is then mostly random because they were not really aware of what technique that worked, no?

Not sure I understand the argument. Though something a bit related:

We can assume that the Romans didn't have the modern sophisticated understand to know exactly what they were doing. But we can assume that they had enough experience to eg make something last for a hundred years with very high probability.

After all, they had enough history to be able to observe various things that their ancestors built a hundred years ago.

Now my argument is that if you build something to last a hundred years with very high probability, more often than not, it'll have a reasonably high chance to survive a thousand years.

Similar to how eg Nasa Rovers on Mars were engineered to have a very high probability to last their official mission length of a few weeks, and thanks to that (over) engineering, they ended up lasting much longer.

There's a saying that "Any idiot can build a bridge that stands, but it takes an engineer to build a bridge that barely stands."

The Roman buildings that lasted were pretty much over-engineered, because they didn't have the knowledge to make the fine trade-offs we can today.

Something I'm curious about - there seems to be some sort of maintenance of reinforced concrete, where people periodically jackhammer a small patch and expose some rebar, and then cover it up again.

Anybody know what they are doing? I used to park in a garage where it seemed like they were always doing it.

A lot of the answer is that the types of structures you can build with concrete that isn't rebar reinforced is fairly limiting. You're pretty much stuck with arch bridges, dams, and domed buildings. Unfortunately, arch bridges don't scale that well (they take concrete proportional to length cubed), and people like buildings in shapes other than domes. As a result, you often have to have designs that hold together partially under tension, and the cost is it won't stay up for 2000 years (if you use rebar which rusts).

That said, in the past few decades there has started to be experiments with fiber reinforced concrete. This has the significant advantage of not corroding, but costs a good bit more than traditional rebar. That said, it might be the future for buildings where long life is desired.

> Unfortunately, arch bridges don't scale that well (they take concrete proportional to length cubed),

Only of you go for one arch right?

Multiple arches helps in some cases, but gets really expensive if you have to bridge a deep gap (especially over water where de-watering is needed) The massive advantage of a suspension bridge is that you get a really big span without anything in the middle.

Because buildings change constantly.. notice how in the examples, you have building for which this requirement must not be fullfilled. Public show-off buildings, statues, temples and harbour walls, they do not change with every generation.

Now if you want to change it constantly, if a building is for living, durability is actually not that much of a value.

There is an ASTM standard: "ASTM C618 Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan for Use in Concrete (AASHTO M 295)."[1]

Fly ash, like volcanic ash, is a pozzlanic ash. Both work in concrete. Fly ash is pulled out of the exhaust from a coal-fired power plant using electrostatic precipitators. Just like electrostatic air cleaners, but huge, they pull particles out of gases. So fly ash is cheap if there's a coal-fired power plant nearby. Convenient when there is no volcano handy.

There are downsides. The concrete takes longer to cure, which can hold up the next stage of construction. Curing in cold weather is difficult. The Romans didn't have that problem in their Mediterranean climate. Concrete curing requires some air in the mix, and fly ash, for some reason, tends to entrain less air than Portland cement.[2]

[1] https://www.cement.org/docs/default-source/fc_concrete_techn...

[2] https://www.fhwa.dot.gov/PAVEMENT/recycling/fach03.cfm

>but I suspect its more a "this is the way we've always done it" difficulty.

Is the construction industry really known for resisting the adoption of new materials like that?

Yeah. incorporating unknown materials into million dollar projects is a really good way to lose millions of dollars. New materials can get added, but especially for big projects there is a long delay between tech invention and widespread use (think decades).

It depends on what you're building. Keep in mind that one way for politicians in office to control the number of jobs and money flowing through the economy is doing road work and other public projects.

I must be missing something. I’ve never used scihub before, and I can’t find the link to read the paper. All I see on this link is the name of the paper and the authors.

You’re still gonna hate this, but try it in landscape.

Thanks for the tip!

Yeah, same. Came in comments to look for alternative link but none so far.

Right hand panel for me was a clickable download of the pdf.

This looks like the journal page: https://pubs.geoscienceworld.org/msa/ammin/article/102/7/143...

Also has a link to download the PDF for free.

There is a save button right under the sci hub logo next to a Google scholar link. Opened the PDF for me on Android chrome.

I am becoming partial to the ancient Egyptian concrete, that they might have used to mold the pyramid blocks in situ.

It takes just limestone crumble, clay (which was already in their limestone), natron, and water.

It is a fair bet the precursors to the Inka who built with the really big blocks had that, or a similar trick. Local observers report the big blocks do not show embedded marine shells at the surface, unlike native limestone. (I have not had opportunity to verify this.)

And while we're in the Sahara Desert with abundant sunshine, sand and carbon dioxide, let's build a silicon carbide brick factory. Sunlight to provide electrical power and, via focussed mirrors, heat. Sand to supply the silicon. Carbon from carbon dioxide to be extracted from the air at $100 per tonne (well, eventually).

Silicon carbide bricks, emerging gloriously from their tungsten moulds, would possess supreme corrosion resistance and almost double the crushing strength of engineering bricks. High thermal conductivity should reduce cracking and spalling, further increasing lifetime. A short railway journey to the nearest port and water desalination plant whence they can be distributed throughout the world.

We'll beat the Romans! Our public buildings will last for millennia!

It will take a very long time to beat the Romans, and more than twice that long to beat the Egyptians.

Corundum bricks would suffice. The Egyptians knew a way to cut corundum like butter; the method apparently was lost before the pyramids were built.

Solar panels have proven quite a lot cheaper than mirror-concentrated solar heat as a source of electrical power. Concentrated solar has not really been tried as a source of direct industrial heat, where it might yet excel. But choosing the bit of Sahara to site in has proven harder than expected. To make building materials economically useful, the site needs immediate sea access. Pisco, Peru might be a better choice, although Nouakchott, Mauritania is well sited. Broome, Australia might do.

Corundum vs carborundum. It's a conundrum!

Even the pre-dynastic Egyptians didn't turn dishes out of carborundum.

Corundum is aluminum oxide, the material of ruby and sapphire. Making dishes out of it was badassery we haven't matched.

The foot of the bridge over river Danube constructed by Apollodorus of Damascus at the request of roman emperor Trajan is still visible in the water.

cpach edited the year from <none> to 2017

New title: Why Ancient Roman Concrete Outlasts Our Own (2017)

Reply on news.ycombinator.com