ChatGPT & DALL-E generated panoramic image of a heavy, white fluid glugging out of a pipe

# Dense Fluid Pumped Hydro Doesn’t Make Any Sense & A Mea Culpa

A recent online discussion on gravity storage brought the usual suspects out of the woodwork. Proponents of heavy fluid pumped hydro reared their sludgy heads this time. Follow along for why this is a silly idea, as all gravity storage options that aren’t pushing water uphill turn out to be.

As a reminder, gravity storage is very basic stuff. The math is Grade 5. The science is Grade 7. Mass times the acceleration due to gravity times height. Kilograms times 9.8 meters per second squared times height in meters. Joules are the unit of energy this produces, and a million of them is about 0.28 kWh. Yes, Joules are tiny. A ton of mass at 100 meters of height has about that much potential energy, so if you were to suspend it with a crane with an electric regenerating winch, you would get about that much electricity out of it. This is why the mere thought of suspending blocks of concrete or steel in mid-air with a crane would never occur to any rational, STEM-competent person, unless they were venally selling it to STEM-illiterate enthusiasts with open wallets.

What is heavy fluid pumped hydro? Well, let’s start with pumped hydro, then let’s go on to the ‘problem’ that heavy fluid pumped hydro ‘solves.’ We’ve been pumping water uphill to store electricity since 1907. All it takes is a hill, a reservoir at the top, a reservoir at the bottom, a tunnel or some really strong pipes, and some reversible hydroelectric turbines. Pump lots of dirt cheap water up to a really simple upper reservoir through a tunnel through rock with really simple electric turbines. When you need electricity, let the water flow back down through the tunnel and turbines to generate electricity.

Hills have been around since the Earth coalesced out of a ball of space dust. The first reservoir was probably built around 2500 BCE in Egypt. We’ve been building tunnels since about 520 BCE, when a Greek tyrant named Polycrates built a kilometer-long tunnel to bring fresh water for the city’s 20- or 30-thousand residents. The first hydroelectric turbine was probably built in 1882 in Wisconsin, USA, of all places. This is not remotely challenging technology.

Yet some people think it really needs improving on. One of those groups are the heavy fluids folks. Their concerns are some combination of a lack of vertical distance, a lack of water, or a lack of places to put pumped hydro. Let’s start with the last one.

See all of those dots? A few years ago, the Australian National University and a group of people led by researcher Matt Stocks did a geographical information system study. They had a computer look for all the places on Earth for which there was decent data where it was possible to put a couple of reservoirs within a small handful of kilometers. They ignored places with less than 400 meters of vertical distance between reservoirs (remember that height thing in the basic science). They eliminated anywhere there was running water, to avoid messing up streams or rivers. They avoided places that were protected areas, mostly parks and the like. They picked places that were fairly close to existing transmission lines, so it would be easy to bring electricity to and from the reservoirs.

They found a rather absurd number of sites that met these criteria. Many of the big blank spots on the map above are just places where an incredibly small number of people live, and there’s no transmission or even good data sets. They found 100 times the resource potential for pumped hydro as the total amount of energy storage that their study concluded was required for all decarbonization. If only 0.5% of the sites, one in 200, pan out, that’s 50% of the problem solved.

There aren’t a shortage of sites to put closed loop, off-river, pumped hydro electricity storage. That’s two of the objections knocked off the list, places to put it where there is sufficient height. Oh, wait, say the pumped hydro critics. What about the Great Plains of the USA, the Prairies of Canada, and the Northern European Plain? Those are a subset of the blank spots to the north of the map. Apparently these people have never heard that we can transmit electricity, or that there’s this thing called the grid that we can have storage assets on.

This leaves the third criteria, having enough water. Let’s imagine a pretty big pumped hydro facility, one with 30 GWh of storage, about 7,700 Tesla Megapacks, the big grid storage one. Its height difference is 500 meters. Its round-trip efficiency is about 80%. It would require around 28 million cubic meters of water. That sounds like a lot.

How much fresh water does the USA consume daily? About 1,200 million cubic meters. Do pumped hydro facilities consume the water? No, it just goes up and down, with a bit of evaporation requiring topping off. Pumped hydro facilities don’t require a new 28 million cubic meters of water every day, they just play with the water they have. 30 GWh of storage requires about one-fortieth of a single day’s water consumption, and doesn’t consume it. That’s about one sixteen-hundredth of the USA’s annual water consumption. When pumped hydro facilities are built, sometimes the developers just let them fill up with rainwater, although that’s fairly slow. Many are certainly replenished with rain sufficiently to require that excess water be fed into nearby streams or rivers. Others run pipes or build temporary channels from nearby rivers or lakes.

If you want more energy storage, just make the top and bottom reservoirs bigger. Bigger reservoirs are remarkably easy to build. They are the least difficult and least expensive part of pumped hydro. Because of the nature of reservoirs as three-dimensional volumes, expanding them 10 meters in all directions produces non-linear results. Let’s take a simple example. Suppose you have a cube 40 meters on a side. It has a volume of 64,000 cubic meters. Let’s expand it 10% in all directions, turning it into a cube 44 meters on a side.

Does the volume and hence the mass of water it could hold increase by 10% to 70,400 cubic meters? No, it expands to 85,184 cubic meters, a full 33%. Double the sides of the cube to 80 meters and the volume isn’t doubled to 128,000 cubic meters but shoots up to 512,000 cubic meters, a full 8 times as much water and hence mass. Note that expanding volume a few meters in each direction results in an exponential gain, not a linear gain.

It’s a bit difficult, in other words, to figure out what problem the heavy fluid folks think that they are solving. Are they making it cheaper? Are they making it more convenient to find fluids? Are they increasing the number of sites with 400 meters of head height or more? Well, no.

What are they claiming? That by using heavier fluids, they can reduce height or volume of fluids, typically by 60%. Remember, it’s mass times acceleration due to gravity times height. There are no exponents in there, unlike with the volume of reservoirs cubing or wind energy swept area squaring or wind velocity power cubing. The increase of energy for pumped hydro by making the fluid heavier is linear. Double the mass of the fluid and you get double the potential energy storage.

If the heavy fluid folks were playing with something where there was an exponent, like swept area of wind turbines or velocity of the wind, they would be on to something. But they aren’t. They are playing with linear stuff, the mass of fluids.

How are they doing that? One proposed method involves using brine solutions, which are mixtures of water and high concentrations of salts like sodium chloride or calcium chloride and are 10% to 40% heavier for the same volume. Magnetorheological fluids, iron filings in water or oil, are 2.5 to 4 times denser. Fluorinated synthetic oils (40% to 90% denser) and glycerin mixtures (15% to 25%) have also been considered due to their higher densities and stable properties. But no one is trying to commercialize any of the above, as far as I’m aware.

There is one firm out there which is trying to commercialize a different solution, fine-milled solids in a suspension in water. What does that mean? A suspension is a bunch of solid particles floating in a liquid. Leave them alone and they’ll sink to the bottom or float to the top eventually. How dense the resulting heavy fluid is determined by the mixture density formula, which is pretty basic. These aren’t solutions, which are like a full jar of marbles that you add water to without increasing the volume of the jar. The solid particles displace water.

They are asserting that their fluid has 2.5 times more mass per cubic meter as water. As a result, they can use 40% of the water or 40% of the height and get the same energy storage. Given that neither water or height are remotely limited resources or particularly hard to exploit, this is a bit of head-scratcher, but maybe it’s cheaper?

Bigger, heavier particles sink to the bottom faster. Bigger, lighter particles rise to the surface faster. Hence, fine-milled solids, on the scale of micrometers or nanometers. There are 1,000,000 micrometers and 1,000,000,000 nanometers in a meter. They are tiny particles. You might be thinking to yourself, how do we make particles of stuff that small, and how much does it cost? Good question.

Ball milling is a widely used method, where materials are ground in a rotating cylindrical chamber with steel or ceramic balls, achieving particle sizes in the micrometer range through repeated impact and friction. Jet milling employs high-velocity jets of compressed air or gas to accelerate particles and cause them to collide, effectively producing sub-micrometer particles. Cryogenic milling cools materials to cryogenic temperatures using liquid nitrogen before milling, making them more brittle and easier to grind — ideal for heat-sensitive materials. Attrition milling involves grinding material by friction and shear forces in a mill with rotating discs or arms, producing fine and uniform particles for various industrial applications.

Is any of this cheap? Not really. Let’s start with the raw materials. The firm in question is coy about what fine-milled solid they are using in their patented fluid, but they have said in public statements that it’s a common substance used in oral medications, basically a harmless substance that increases volume enough that patients can swallow or swish the active ingredients. The most common of these is microcrystalline cellulose made from wood pulp, but its density is much lower than water.

My first thought was that the secret sauce might be silicon dioxide, also known as colloidal silicon, commonly used as a bulking agent in tablets. The colloidal part is important for this. Colloidal suspensions are different than non-colloidal ones in a few ways, but the significant one is that the particles don’t settle out, being fine enough to stay in the mixture through Brownian motion. Milk is a colloidal suspension of fat globules, casein micelles, and dissolved lactose and minerals, as a common example. If a substance isn’t colloidal, all the heavy bits fall out of the liquid and create a layer of mud on the bottom of whatever they are in.

The maximum for suspended silicon dioxide is about 60%, which coincidentally gives only a 60% increase in weight. Silicon dioxide is the densest substance I was able to find that’s also used in oral medications, so there’s a sniff test problem. Its density by itself is 2.65 times that of water, but you can only mix 0.6 grams of silicon dioxide with 0.4 grams of water, and when you add the densities together it doesn’t come close to 2.5 times the mass of water.

This suggests that they aren’t using silicon dioxide, or if they are, they are getting the end mass wrong. Not getting the basics of physics right is relatively common for people who try to outdo basic pumped hydro, but let’s assume that they are competent.

What else might they be using? Possibly barite, aka barium sulfate, which is 4.5 times denser than water. It is widely used in the oil and gas industry as a weighting agent in drilling fluids. It’s also widely used as a radiopaque contrast agent for X-ray imaging and other diagnostic procedures. That humans are given it to drink suggests that this is what the firm meant, and they or the people reporting don’t know that it’s not a medication, but used for diagnostic imaging.

Remember that bit about colloidal suspensions? Barite doesn’t tick that box by itself. Typically bentonite is added, a clay which increases the viscosity of water so that the bentonite particles remain colloidally suspended. This is commonly used in drilling. We’ll come back to that viscosity later.

A bit of napkin math suggests a mixture of 5% bentonite and 41% barite in water would have a density about 2.5 times that of water, and it would be a colloidal suspension. It’s entirely possible to achieve the 2.5 times mass for the same volume that the company claims, in other words.

What would this cost? What does water pumped from a lake or river cost? Permit and usage costs might be a dollar per cubic meter. Pumping it a kilometer costs 10 to 50 cents. Assuming the water source is ten kilometers away, that turns into per \$2 to \$6 per cubic meter. A cubic meter of water weighs a ton.

Fine milled barite typically costs \$100 to \$300 per ton in bulk per what I was able to find online (not being a person who operates drilling equipment). Bentonite costs about \$100 per ton as well. Picking the averages, that suggests that 5% bentonite and 41% barite would cost about \$87 per ton of ingredients, with 54% water added to make it 100%.

Plain water costs perhaps \$4 per ton. The colloidal suspension costs about \$90 per ton. That’s 22.5 times more expensive for 2.5 times as much energy storage.

What would this do to the capital costs for a reasonably-sized pumped hydro solution, that is, one that was worth building? Let’s take the 30 GWh solution above. The big claim of the firm in question is about height, so let’s take the 500 meters in the example and divide it by 2.5, giving us a nice even 200 meters. The same volume of water is required, 28 million cubic meters.

The basic pumped hydro facility would cost \$3 to \$9 billion to build, an average of \$6 billion. Filling it would cost another \$112 million, a fraction of the construction cost.

Building a 200-meter height facility with the same size tunnels and reservoirs saves a bit of construction costs, but not that much. Call it \$2 to \$7 billion for an average of \$4.5 billion. That’s a saving of \$1.5 billion perhaps. Filling it with 28 million cubic meters of \$90 per cubic meter fluid, however would cost \$2.5 billion. So much for that saving. Add a billion to the capital costs.

Actually a bit more. The bentonite and barite aren’t going to get to the site for free. Let’s assume that they have to travel 500 kilometers, 400 of it by rail and 100 by truck. Freight trucking in the USA costs about \$0.25 per ton kilometer. Rail costs around \$0.04 per ton kilometer. Add another \$700 million to the cost. It’s now \$1.7 billion more, about 33% more expensive, for the same energy storage.

Is this a showstopper? Not necessarily, although it’s a head-scratcher as to why anyone would pay a full third more for the same energy storage.

Is there anything else? Yes, viscous fluids will impact turbine efficiency. A barite suspension with bentonite mud or a surfactant like polysorbate which can also enable a colloidal suspension has a centipoise (cP) — the measure of viscosity — in the 50 to 200 range, typically around 100 cP. That’s a bit more viscous than olive oil and a bit less than maple syrup, but they are solutions, not suspensions. The total system impact in one direction is 6% to 15% — my attempts at calculations ended up with that range — but that’s 12% to 30% roundtrip losses overall. Pumped hydro tends to operate around 80% roundtrip efficiency, so this system is likely to be around 70% or lower.

Anything else? Yes, turbine blade abrasion. Colloidal suspensions of solids still have particles of solids. Hydroelectric turbine blades have high tip velocities and they degrade with too much sand and grit. This is like having constant very fine grit, so turbines won’t last as long.

One more thing. Colloidal suspensions are not as stable as solutions. The solids do tend to drift out of them eventually due to aggregation and a few other things. Suspension additives like bentonite mud and polysorbate degrade over time due to a variety of quite normal things including temperature changes, light, oxygen, and microbes. While turbulence in a heavy fluid pumped hydro system would tend to reduce sedimentation and the company intends to enclose the tanks of fluids to reduce light impacts, unless they make it a sealed system with high quality control and inert gases instead of air — all of which adds more cost — the suspension will break down over time, requiring flushing and replacing of all of the water and additives at great expense.

Much higher cost of fluid. Lower efficiency. Lower lifespan. It’s a real headscratcher.

To be clear, my assessment of barite is based on the firm’s statements about the substance in general, but they have not stated that it’s barite, but a ‘high-tech’ powder additive to water. Other alternatives that have been suggested are lighter, toxic or not, things which can be added as a powder. It’s possible that they have something which forms a stable solution, but it doesn’t align with the articles where it’s asserted to be something used in medical conditions where humans swallow it.

But that said, the firm in question is not aiming to really scale up gravity storage, it’s aiming for 10 MW to 50 MW systems and trying to compete with cell-based batteries which are plummeting in cost and have 85% to 95% round trip efficiencies. Like other attempts to make small-scale gravity storage make the least sense, it falls over of its own weight without delivering much value.

Mea culpa: In a previous version of this article I was completely turned around on the viscosity calculations. The efficiency impacts are not nearly as high as that article stated. The previous version of the article asserted that the dense fluid pumped hydro wouldn’t work, but that’s not true. It will work, it will just be expensive and not particularly long-lived. Part of my enjoyment comes from exploring new aspects of science and engineering, and sometimes that means I get the wrong end of the stick completely. This was one of those times.

Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.

#### Latest CleanTechnica.TV Videos

CleanTechnica uses affiliate links. See our policy here.

CleanTechnica's Comment Policy

#### Michael Barnard

is a climate futurist, strategist and author. He spends his time projecting scenarios for decarbonization 40-80 years into the future. He assists multi-billion dollar investment funds and firms, executives, Boards and startups to pick wisely today. He is founder and Chief Strategist of TFIE Strategy Inc and a member of the Advisory Board of electric aviation startup FLIMAX. He hosts the Redefining Energy - Tech podcast (https://shorturl.at/tuEF5) , a part of the award-winning Redefining Energy team. Most recently he contributed to "Proven Climate Solutions: Leading Voices on How to Accelerate Change" (https://www.amazon.com/Proven-Climate-Solutions-Leading-Accelerate-ebook/dp/B0D2T8Z3MW) along with Mark Z. Jacobson, Mary D. Nichols, Dr. Robert W. Howarth and Dr. Audrey Lee among others.

Michael Barnard has 765 posts and counting. See all posts by Michael Barnard