You’ve probably heard it said that string theory contains two types of strings: open, and closed. Closed strings are closed loops, like rubber bands. They give rise to gravity, and in superstring theories to supergravity. Open strings have loose ends, like a rubber band cut in half. They give us Yang-Mills forces, and super Yang-Mills for superstrings.

String theory has more than just strings, though. It also has branes.

Branes, short for membranes, are objects like strings but in other dimensions. The simplest to imagine is a two-dimensional membrane, like a sheet of paper. A three-dimensional membrane would fill all of 3D space, like an infinite cube of jello. Higher dimensional membranes also exist, up to string theory’s limit of nine spatial dimensions.

So where did these branes come from? Why doesn’t string theory just have strings?

You might think we’re just trying to be as general as possible, including every possible dimension of object. Strangely enough, this isn’t actually what’s going on! As it turns out, branes can be in lower dimensions too: there are zero-dimensional branes that behave like particles, and one-dimensional branes that are similar to, but crucially not the same thing as, the strings we started out with! If we were just trying to get an object for every dimension we wouldn’t need one-dimensional branes, we’d already have strings!

(By the way, there are also “-1” dimensional branes, but that’s a somewhat more advanced topic.)

Instead, branes come from some strange properties of open strings.

I told you that the ends of open strings are “loose”, but that’s just loose language on my part. Mathematically, there are two options: the ends can be free to wander, or they can be fixed in place. If they’re free, they can move wherever they like with no resistance. If they’re fixed, any attempt to move them will just set them vibrating.

The thing is, you choose between these two options not just once, but once per dimension. You could have the end of the string free to move in two dimensions, but fixed in another, like a magnet was sticking it to some sort of 2D surface…like a brane.

In mathematics, the fixed dimensions of end of the string are said to have Dirichlet boundary conditions, which is why this type of branes are called Dirichlet branes, or D-branes. In general, D-branes are things strings can end on. That’s why you can have D1-branes, that despite their string-like shape are different from actual strings: rather, they’re things strings can end on.

You might wonder whether we really need these things. Sure, they’re allowed mathematically, but is that really a good enough reason?

As it turns out, D-branes are not merely allowed in string theory, they are required, due to something called T-duality. I’ve talked about dualities before: they’re relationships between different theories that secretly compute the same thing. T-duality was one of the first-discovered dualities in string theory, and it involves relationships between strings wrapped around circular dimensions.

If a dimension is circular, then closed strings can either move around the circle, or wrap around it instead. As it turns out, a string moving around a small circle has the same energy as a string wrapped around a big circle, where here “small” and “big” are comparisons to the length of the string. It’s not just the energy, though: for every physical quantity, the two descriptions (big circle with strings traveling along it, small circle with strings wrapped around it) give the same answer: the two theories are dual.

If it works with closed strings, what about open strings?

Here something weird happens: if you perform the T-duality operation (switch between the small circle and the big one), then the ends of open strings switch from being free to being fixed! This means that even if we start out with no D-branes at all, our theory was equivalent to one with D-branes all along! No matter what we do, we can’t write down a theory that doesn’t have D-branes!

As it turns out, we could have seen this coming even without string theory, just by looking at (super)gravity.

Long before people saw astrophysical evidence for black holes, before they even figured out that stars could collapse, they worked out the black hole solution in general relativity. Without knowing anything about the sort of matter that could form a black hole, they could nevertheless calculate what space-time would look like around one.

In ten dimensional supergravity, you can do these same sorts of calculations. Instead of getting black holes, though, you get black branes. Rather than showing what space-time looks like around a high-mass point, they showed what it would look like around a higher dimensional, membrane-shaped object. And miraculously, they corresponded exactly to the D-branes that are supposed to be part of string theory!

So if we want string theory, or even supergravity, we’re stuck with D-branes. It’s a good thing we are, too, because D-branes are very useful. In the past, I’ve talked about how most of the fundamental forces of nature have multiple types of charge. One way for string theory to reproduce these multiple types of charge is with D-branes. If each open string is connected to two D-branes, it can behave like gluons, carrying a pair of charges. Since each end of the string is stuck to its respective brane, the charge corresponding to each brane must be conserved, just like charges in the real world.

D-branes aren’t one of the original assumptions of string theory, but they’re a large part of what makes string theory tick. M theory, string theory’s big brother, doesn’t have strings at all: just two- and five-dimensional branes. So be grateful for branes: they make the world a much more interesting place.