Why Is It So Hard To Solve The World's Energy Crisis?

I'm teaching a class at MIT this semester called "Thermal-Fluids Engineering II."  The class is largely concerned with energy, and how it is generated using machines like internal combustion engines and power plants.  This post will provide some background on the current state of affairs regarding energy.

Recently, former NASA astronaut and current NASA climate scientist Piers Sellers was diagnosed with stage 4 pancreatic cancer.  Two weeks ago he wrote an op-ed in the New York Times where, rather than feel sorry for himself or wax sentimental about his life, he described (in broad terms) what humanity needs to do to avoid the potentially catastrophic consequences of runaway global warming, since it is now likely that he won't be around to see what happens.  After describing the initial challenges associated with making effective energy and environmental policy (which are substantial), he writes (emphasis mine),

Ultimately, though, it will be up to the engineers and industrialists of the world to save us. They must come up with the new technologies and the means of implementing them. The technical and organizational challenges of solving the problems of clean energy generation, storage and distribution are enormous, and they must be solved within a few decades with minimum disruption to the global economy.

Here we'll address the question that is probably on many people's minds: why are the technical challenges associated with generating cleaner energy so enormous?  Let's start by looking at the current ways in which we extract energy.

Consider a conventional gasoline-powered car like the Toyota Camry, one of the most common cars on American roads.  The gas mileage on such a car is, accounting for highway and city driving, about 30 miles per gallon.  Since there are 16 cups in one gallon, this means you get about 2 miles per cup, or one mile out of about this much gas:

Pictured: a one-mile drive in a Toyota Camry...from an engineer's perspective at least.
(SEE NOTE 1)

Let's say you put that much gasoline in the tank of a Toyota Camry and drive a mile.  Then, the car breaks down, and the nearest service station is a mile away.  Now imagine the amount of energy you would have to expend to push the car (which typically weighs about 3,240 pounds) for the entire mile to get there.  There's a LOT of energy in that half-cup of gasoline (which, at today's prices, would currently set you back a whopping 5 cents)!

As it turns out, gasoline has a very high energy density compared with many other substances.  Here's a comparison with a few other notable substances (SEE NOTE 2).

	Food calories per pound	Compared to TNT
TNT	295	1
Chocolate chip cookies	2,269	7.7
Coal	2,723	9.2
Butter	3,176	11
Gasoline	4,538	15
Natural gas	5,899	20
Uranium-235	9,000,000,000	32,000,000

(Yes, 1 pound of chocolate chip cookies has the energy of 7.7 pounds of TNT!)

Now let's compare gasoline with other sources from which we draw energy, like a AA battery.  How much energy is stored in it?

Batteries deliver energy in the form of electricity, which at the most basic level means moving electrons.  Batteries are typically characterized by the amount of current (essentially, number of electrons per second) they can deliver, and the voltage (essentially, the amount of energy each electron carries) at which they can deliver it.  Different applications require different amounts of each.  A typical alkaline AA battery can deliver around 2500 milliamp-hours at 1.5 V.  A milliamp is a unit of current, so in other words, it can deliver 2500 milliamps of current for 1 hour at 1.5 V.  Carrying out the calculations,

Power = Energy per Time = Current x Voltage (see this article for an explanation)

Energy = Current x Voltage x Time

Energy = 3750 Watt-hours, or 13,500 Joules.

Here I'm using the metric unit for energy, called the Joule (after James P. Joule, a pioneer in the area of thermodynamics).

So......is 13,500 Joules a lot of energy?  Let's compare this to the amount of energy in food.

What we call 1 food calorie is actually 1,000 "actual" calories, or 1 kilocalorie (long story - see this article), and 1 kilocalorie turns out to equal 4,184 Joules.  So a AA battery contains a little over 3 calories' worth of energy.  Compare that to a granola bar, which has around 200 calories in it.  So, a granola bar is equivalent to 62 AA batteries in terms of energy stored!

What about the equivalent amount of gasoline?  A granola bar weighs about 56 g.  From the table above and some simple arithmetic, 56 g of gasoline has about 567 food calories, or the equivalent of about 186 AA batteries!

But, you say, there's plenty of energy in the Sun!  We just need to harvest and store it, right?

The rate at which we receive energy (in the form of electromagnetic radiation) from the Sun turns out to be about 1000 Joules per square meter per second.  On a perfect sunny day, a patch of ground with an area of 1 square meter receives about 1000 Joules every second.  Sounds like a lot of energy, right?  Well, setting aside the fact that this is only true in direct sunlight (and so by definition only during daytime), the issue is that most solar cells are only about 10% efficient.  This means we can only extract about 100 Joules per second for every square meter of (generally expensive) solar cells that we build.   

So, let's say we buy a 1-square-meter solar cell and collect sunlight for 1 hr.  In the best case scenario, this will give us about 86 food calories' worth of energy.  This is equivalent to half a granola bar, or about 3/4 of a Tablespoon of gasoline!

Conclusion:

Burning stuff like gasoline or coal releases a LOT of energy, and it's going to be hard to find other energy sources that are as cheap and energy-rich.

But that won't stop us from trying.  In future posts, I will address the steps we are taking at MIT to try to come up with new ways to generate clean energy, and just as crucially, how to store and distribute it.

NOTES!
1. Some of you may be surprised that that much gasoline is actually required.  (I was.)  But the thing to keep in mind is that cars are actually quite inefficient when it comes to extracting energy from gasoline: for a typical auto engine, the average efficiency (amount of useful mechanical energy actually extracted divided by the total extractable amount) is around 18-20%.  In fact, thermodynamics shows us that the best we can ever hope for with this kind of engine is around 37%.  So, the "large" amount of gasoline required to drive one mile is not due to a limited amount of energy in gasoline, but due to the limitations on our capability to extract useful energy from it.

2. Source: Muller, R.A., Energy for Future Presidents.  New York: W.W. Norton, 2012.

Electro-osmosis: Moving Water With Electricity

If you ever find yourself with a glass of water, two drinking straws of different diameters, and a little free time, you might notice that if you drink out of the smaller drinking straw, you don't get as much water as you do with the larger one. In other words, if you want to drink water at the same rate through each straw, you need to exert more effort when drinking out of the smaller straw. In general, the smaller the straw, the more effort (i.e., pressure) it takes to drive fluid at a given rate through it. This principle makes good intuitive sense, and is predicted by the mathematical equations that govern fluid flows. Perhaps a little interesting, but nothing special. Why does it matter?

It matters if you work in microfluidics, which, true to its name, involves fluids flowing through extremely small passageways. These passageways are usually hundreds of times smaller than a drinking straw - usually about as big around as a human hair. So, from what you now know about the exciting world of drinking straw physics, you would suppose that we'd need a lot more pressure to make the water flow through a microchannel than a straw, if we wanted to push water at the same flow rate through the straw and the microchannel. And you would be right. The pressure demands for driving fluid through microchannels are significant (see note 1).

If you want to drink water out of a drinking straw, there is pretty much only one way you're going to get the water to go against gravity and creep upward through the straw toward your lips: exert pressure by creating suction with your mouth. How else could you do it, right? But if you want to drive water through a tiny microchannel, it turns out there is an alternative to using pressure that you don’t have when drinking out of a straw. On small size scales, there is a completely different way to make water flow.

What if I told you that if you put two metal electrodes on opposite ends of a microchannel filled with water and applied some electricity, the water would instantaneously begin to flow from one electrode to the other? This is indeed what happens. It is possible to make water flow from one place to another in a microchannel using two pieces of metal and some electricity. That's all you need. No pumps or any other moving parts are necessary. This technique is very common and people in microfluidics use it every day.

What is going on here?

More than H2O molecules

Let’s take a closer look at that glass of water. A much closer look. If you magnified a glass of water a few million times, you would discover a lot more than just H2O molecules swimming around. Among other things, in a typical glass of water there are also a vast amount of protons (the same protons that make up the nuclei of atoms) "swimming" freely about as well. How many protons? A 1-liter bottle of water contains about 60 million billion (6 times 10 to the 16th power) free protons (although there are many, many more H2O molecules than that - see note 2). Recall also that protons have a positive electric charge. This will be important in a moment.

Now that we know there are charged particles swimming around your glass of water along with the H2O molecules, let's look at what goes on at the interface between the glass and the water.

When glass is exposed to air, the silanol (SiOH) groups stay composed and mind their own business.  However, whenever water is brought into contact with the glass, the SiOH molecules deprotonate (give up their protons), rendering the glass surface negatively charged.  Protons from the bulk water solution flock to the surface to "shield" the charged surface, so that the region between the 2 red lines is electrically neutral.  This layer of protons, along with the charges on the glass surface, constitute the electrical double layer.

Glass is primarily composed of silicon dioxide (SiO2), but the outer surface of the glass, the part that "sees" the water, is made of a chemical compound called silanol. The formula for silanol is SiOH: one silicon atom, one oxygen atom, and one hydrogen atom bound together. A single grouping of these atoms is called a silanol group. The reason I bring this up is that whenever SiOH groups are brought into contact with a fluid that has a pH higher than about 3, they cannot hold on to their hydrogens anymore. Or, more specifically, they cannot hold on to the protons from the hydrogens (remember that a hydrogen atom consists of one proton and one electron), and the protons escape from the glass surface into the water, like a top hat blown off of an old man's head in the wind. This loss of protons does happen for a reason, but it is not really that important to the present discussion. The important thing is that this process happens whenever water and glass are brought into contact, and it is called deprotonation.

Positive Band-Aids

Deprotonation has some interesting consequences in a drinking glass-water system. Once the glass loses protons, its surface becomes negatively charged (see note 3). Nature would rather this didn't happen. It's a little like an open wound exposed to open air. Nature would like to apply a positively charged “band-aid” to the surface so that it becomes electrically neutral, which is the preferred configuration.

Now, remember the quadrillions of positively charged protons swimming around in the water? Recall that protons are positively charged, and as discussed above they are readily available. So it only makes sense that the protons act as the "band-aid" to the charged surface. And this is precisely what happens: some (not all) of the freely floating protons gather very (very) near the glass surface, "shielding" the negatively charged glass (see figure above). So the negatively charged "wound" is now covered with the positively charged "band-aid," and this band-aid is known as the Electric Double Layer (EDL). If all this sounds time-consuming, it's not. The whole process takes less than a millionth of a second. Something to think about the next time you pour yourself a glass of water.

Now consider that microfluidic channels are often made from glass. So whenever we fill a microfluidic channel with water, the glass deprotonates, becomes charged, and attracts protons to its surface. By the way, the electrical double layer is extremely thin. If a microchannel the size of a human hair were magnified to the size of an average-size classroom, the EDL would be as thick as the paint on the walls.

Putting it all together

Now for the reason why I've been rambling about protons in glasses of water all this time. Now let’s say you fill a glass microcapillary (basically a tiny hollow glass tube, about as big around as a human hair) with water. As always, there will be layers of protons shielding the glass walls. Now, take your metal electrodes and place them on either end of the channel and apply a voltage between them, i.e., make one electrode negatively charged with respect to the other. Remember that, as always, opposites attract. The protons in the EDL sense a positive and negative electrode in their midst. The protons, being protons, would much rather head towards the negative electrode, and that's what they do. The transport of protons (and indeed, any ions in an aqueous solution) in response to electricity is called electromigration.

Now, you might think that since protons are so tiny, they don’t exert any influence on the surrounding water when they move. Not so. When a proton moves in a fluid through electromigration, it “drags” the surrounding fluid along with it. Now, since protons are indeed incredibly small, one proton drags a minuscule amount of fluid. But remember that there are billions upon billions of these protons coating the walls of the microchannel. There are enough that you can think of the protons as a “sheath” that “coats” the microchannel edges. This proton sheath is substantial enough to drag the rest of the water in the microchannel along with it. The "micro-paint-thin" layer of protons drags the rest of the water in the "micro-room." Again, this process is extremely quick to get started, fast enough so that when you flip the switch, the flow has effectively reached full speed instantaneously.

Schematic of electro-osmotic flow.  The arrows indicate the direction of the proton migration, and ultimately of the fluid flow.

Thus, it is possible to move water through a microchannel by the simple application of electricity. This type of flow is generally known as ­electro-osmotic flow, or EOF, and it is used every day in microfluidics research and industry. It provides an alternative to using pressure to drive flow. It is a key component of capillary electrophoresis, an extremely useful technique used in analytical chemistry to separate different compounds out of a single sample. It is the basis for electro-osmotic pumps, which use EOF to “pump” water, generating sufficient pressures to do useful work using no moving parts.  At MIT, we discover new uses for EOF almost daily, and the technique will likely continue to be useful for many years to come.

Notes
1) Despite this, pressure-driven flow is still used in microfluidics when the required flow rate is not too high.

2) It is very important to note, though, that even though there are charged particles swimming around in your glass of water, there are (roughly) just enough that they cancel each other out. In a standard glass of water, there are about as many hydroxide (OH-) ions as there are protons (also known as H+ ions), so that you can think of any given droplet of water as being electrically neutral. There are about 60 million billion hydroxide ions in your 1-liter bottle of water as well. They are completely harmless and are present in every glass of water you drink.

3) Incidentally, surfaces other than glass can and do become electrically charged when brought into contact with water as well. It turns out there are many different mechanisms for surfaces to become charged in the presence of liquids.

Doctoral Work, Part 1: Bi-metallic Nano-Swimmers

Let's say you were bored one day, and you had a bottle of hydrogen peroxide and a piece of platinum wire lying around. Let's say you wanted to find out what would happen if you dropped the platinum into the peroxide (you are really bored). What would you see? As soon as you drop the wire into the solution, a chemical reaction will immediately begin to occur that will cause tiny bubbles to form all over the metal surface (see note 1 below). Although bubbles would form, the wire would not move around in the container at all. Why would it, right? It would simply settle to the bottom. Similarly, if you dropped a piece of gold into the peroxide, it would not move.

But what if instead you dropped a microscopic metal cylinder made of half platinum and half gold into the peroxide? What would happen then? It would not be unreasonable to assume that the cylinder would act similarly to the individual platinum and gold pieces and simply settle to the bottom of the container. However, the actual result is quite different.

You see, it is possible to make these mini-cylinders, and they are known as 'bimetallic nanorods.'  When immersed in hydrogen peroxide, bimetallic nanorods move entirely on their own at 10-100 body lengths per second. No outside driving force is required. The rods always move along their axis, and always with the platinum end forward (which is a clue as to how they move).  For comparison, the Space Shuttle moved at a bit over 100 body lengths per second while achieving orbit.

Diagram of a typical nanorod.  The black arrow indicates the direction of motion.

Even though the rods move fairly fast relative to their size, it should be emphasized that "their size" is really quite diminutive. Bimetallic nanorods are typically about 2 micrometers (or "microns") in length and about 300 nanometers in diameter (see note 2). To get an idea of how small that is, consider that an average human hair has a diameter of about 100 microns. You would need to lay 50 nanorods end-to-end to equal the diameter of an average human hair.

If that doesn't convince you that these things are tiny, take a look at the picture below. That container has about 100 billion (1 followed by 11 zeros) nanorods in it (see note 3). And they're not even taking up the entire container - most of them are gathered in that dark smudge at the bottom.

~100,000,000,000 nanorods.  Seriously.

At this point, two good questions have probably occurred to you.

1. Why should we care that these little cylinders can swim in peroxide? What can we do with them that would be useful or helpful?
2. How in the world do they manage to move on their own? Why do they only move in hydrogen peroxide?

To answer the first question, bimetallic nanorods have a wide variety of potential applications in nanotechnology. One especially promising application is the enhancement of drug delivery technology. Traditional drug delivery methods include oral ingestion or injection. When you take a drug using one of these methods, the medication goes to a lot of places in your body besides where it is needed. Thus, in most cases, only a small portion of the medication reaches the intended "target." Targeted drug delivery aims to deliver medication only where it is needed in the body. This would make it possible to deliver a more concentrated dose of medication in the intended location while reducing the relative concentration of the medication elsewhere in the body, improving the efficacy of the drug while reducing its side effects.

So, how do bimetallic nanorods come in? It has been experimentally shown that nanorods are capable of attaching themselves to tiny spheres (1-micron diameter) made out of polystyrene (the same material that foam cups are made of) and tugging them around (see video below). One day we may be able to replace the sphere with a drug, and have the nanorods bind to the drug, seek out a certain site (e.g. a tumor) in the body and deliver the drug there (see note 4). One of the reasons that hasn't happened yet is that nobody has figured out how to make nanorods move in fuels other than peroxide. And the main reason that hasn't happened yet is that nobody knows exactly why they move in hydrogen peroxide.

Nanorods can be steered using external magnetic fields and can also pick up and release cargo at pre-determined locations.  Here, the cargo is a spherical polystyrene particle with a magnetic coating.  When it gets close enough to the cargo, the rod snaps on in much the same way that magnets "snap" onto your refrigerator door.

Which brings us to the second question: how do bimetallic nanorods move in hydrogen peroxide?Until my work, a complete physical theory had not been formulated, but clearly it has something to do with the peroxide itself. Somehow, the nanorods are using it as a fuel - they convert the chemical energy stored in the H2O2 molecules into mechanical energy (motion) (see note 5). We know this because they barely move at all in pure water, and because their swimming speed increases with the concentration of hydrogen peroxide; that is, the more peroxide per unit volume you have, the faster the nanorods go. So the peroxide is definitely the energy source for the autonomous motion. Of course, humans can also move autonomously in water (i.e., swim), and we also do it by converting chemical energy (stored in the body) into motion (using our muscles). But nanorods don't have muscles - they're just microscopic pieces of metal! We know they get their energy from the peroxide, but how do they convert it into the mechanical energy that propels them forward?

That question is the subject of my thesis research. There are several prevailing theories behind the nanorods' motion, although it has not been proven which one, if any, is the correct one. We are currently formulating our own theory, which I am using a computer model to simulate. The goal is to successfully recreate the motion of a nanorod on a computer. We are also conducting our own experiments with nanorods. If the results of my simulations agree well with the experimental results found by our group and by other groups, that will constitute strong support for my model. We will be that much closer to being able to use nanorods in the human body and elsewhere in nanotechnology.

Notes

1. What is happening here is that the platinum is acting as a catalyst to initiate the chemical reaction. Remember that the chemical formula for hydrogen peroxide is H2O2. Roughly speaking, the platinum gives the peroxide molecules a little "nudge" that makes them "want" to break down into water and oxygen gas. The oxygen gas shows up as the bubbles you see.

2. Just as there are 1000 millimeters in 1 meter, there are 1000 micrometers in 1 millimeter and 1000 nanometers in 1 micrometer. The general rule is that if one feature on an object is less than 1 micron in length (in this case, the diameter), the nano- prefix is used, hence the name nanorods.

3. You might ask how we know there are that many nanorods in there. (We didn't count them!) The answer has to do with the method used to make them. Nanorods are actually grown by depositing metal into tiny pores in a membrane made of aluminum oxide. The inner diameter of the pores is about the same as the diameter of the nanorods. We know how many pores there are in a given area of the membrane, so for a membrane of any given size we can have a fairly accurate estimate of how many nanorods we made.

4. Targeted drug delivery was recently realized in a proof-of-concept experiment: some colleagues of ours at UC-San Diego used a different type of self-propelling rod (made of zinc) and administered orally to some rather unsuspecting mice.  The nifty thing about these swimmers is that they use stomach acid as a fuel, and deliver cargo payloads into the stomach lining of some unsuspecting mice.  For the technically inclined, the journal article is here, and you can read a non-technical press release here.  This work is a major step forward toward the goal of using self-propelling particles to deliver drugs in nanomedicine.

5. By the way, there are naturally occurring nanoscale motors that are capable of doing a similar thing. For example, several different types of biomolecules are capable of harvesting energy from energy-rich molecules such as adenosine triphosphate (ATP) to initiate autonomous motion.