How to Determine If a Molecule Is Polar or Nonpolar

Water is polar. Carbon dioxide isn't. But both molecules are built from the same kind of bond: a more electronegative atom (oxygen) pulling electron density away from a less electronegative one (carbon or hydrogen).

Same bond. Opposite answer. So the electronegativity difference alone can't be the whole story.

Picture a tug-of-war. Every bonded atom pulls on the shared electrons, and the strength of that pull depends on electronegativity. If everyone pulling on the central atom is arranged so their pulls cancel out, exactly like two equally strong teams pulling from opposite sides, the knot stays put. That's a nonpolar molecule. If the pulls are lopsided, the knot drifts. That's a polar one.

Two questions decide it. Is there a pull at all? And do the pulls cancel? Let's build both up from scratch.

Question one: is there even a pull?

A bond forms when two atoms share electrons. If both atoms want those electrons equally, the sharing is even. Neither atom pulls harder, so there's no lopsidedness to speak of.

But atoms don't want electrons equally. Some atoms hold onto electrons much more tightly than others. Chemists put a number on that tendency and call it electronegativity. Oxygen, nitrogen, and fluorine are electron-hungry. Carbon and hydrogen are more relaxed about it.

When two atoms with different electronegativities bond, the shared electrons don't sit exactly between them. They drift toward the hungrier atom. That atom ends up slightly negative, and the other one ends up slightly positive. We call these partial charges $\delta^-$ and $\delta^+$, and the bond now has a dipole: a tiny built-in lopsidedness.

Drag the slider and watch the electron cloud drift. At a small difference, the cloud barely moves; the bond stays close to nonpolar. Push the difference higher and the cloud piles up almost entirely on one atom, and the bond edges toward fully ionic. There's no hard line between "sharing" and "transferring" electrons. It's a spectrum, and the electronegativity difference is where you are on it.

We write the difference as:

As a rough guide: below $0.4$, chemists call the bond nonpolar covalent. Between $0.4$ and $1.7$, it's polar covalent. Above $1.7$, the pull is strong enough that the bond behaves as ionic instead. These aren't sharp physical boundaries, just a useful way to talk about where a bond sits on the spectrum.

So question one has a simple test: look up the electronegativities and subtract. If $\Delta EN$ is close to zero, that particular bond isn't pulling much of anywhere.

One polar bond doesn't make a polar molecule

Here's where most explanations stop, and it's exactly where the real answer starts. A molecule can be packed with polar bonds and still be completely nonpolar overall.

Carbon dioxide has two C=O bonds, and each one is genuinely polar. Oxygen pulls harder than carbon, full stop. But CO₂ itself has no overall dipole. The reason has nothing to do with electronegativity. It's about shape.

Each bond dipole is a vector: it has a strength and a direction. To find the molecule's overall polarity, you don't just ask "are there polar bonds." You add up the bond dipoles as vectors:

If those vectors point in ways that cancel, the net dipole is zero, no matter how strong each individual pull is. If they don't cancel, there's a leftover pull in some direction, and the molecule is polar.

Switch between the shapes. Linear and symmetric trigonal both cancel completely, even though every individual bond in those diagrams is just as polar as in the shapes that don't cancel. Then look at "trigonal, one different": same three-way symmetry, but one bond pulls harder than the other two. The moment that symmetry breaks, a net arrow appears.

That's the whole reason CO₂ gets away with having two polar bonds. It's linear, so the two dipoles point in exactly opposite directions, and they erase each other completely.

Why water is polar and CO2 isn't, even with the same kind of bond

CO₂ is linear. Water is bent. That's the entire difference, and it turns out to be enough.

Water has two O–H bonds, both genuinely polar, both pulling electron density toward the oxygen. If those two bonds pointed in opposite directions like CO₂'s, they'd cancel too. But the oxygen's lone pairs push the hydrogens together, bending the molecule to about $104.5°$. At that angle, the two O–H dipoles aren't opposite anymore. They both lean the same general direction, and their pulls add instead of canceling.

Hit play and watch the angle sweep. At $180°$, the two arrows point in exactly opposite directions and the net dipole sits at zero. As the angle closes toward $90°$, the two bonds start leaning the same way, and the dashed net arrow grows. Water's actual angle, $104.5°$, isn't some special magic number. It's just bent enough that cancellation fails.

This is the part a lot of people miss: polarity isn't really about individual bonds at all. It's about whether the geometry lets those bonds cancel. A molecule with strongly polar bonds can be nonpolar (CO₂), and a molecule can be polar even though no single bond in it is unusually strong (water's O–H dipole isn't extreme, it's just not canceled).

Putting both questions together

So here's the actual test, in order:

1. Are the bonds polar? Compare electronegativities. If every bond has $\Delta EN \approx 0$, you're done: nonpolar.
2. If yes, does the shape cancel them? Draw the bond dipoles as vectors from the central atom and check whether they sum to zero. Symmetric arrangements (linear with identical ends, trigonal planar with identical corners, symmetric tetrahedral) cancel. Anything that breaks that symmetry, a bent shape, a lone pair, a different substituent, doesn't.

Try CH₄ and CCl₄ first: same tetrahedral shape, same kind of symmetry, both nonpolar, even though C–Cl is a much more polar bond than C–H. The shape is doing all the work. Now try CHCl₃: same tetrahedral shape as methane, but swap one Cl for an H, and the symmetry that caused cancellation disappears. Suddenly it's polar, at $1.04$ D.

Notice NH₃ too. It has the same three-bonds-plus-something-at-the-fourth-position setup as a symmetric shape would, except that "something" is a lone pair instead of a fourth identical bond. A lone pair doesn't create an opposing dipole to cancel the other three, so nitrogen ends up with a real net pull, $1.42$ D, even though ammonia looks deceptively like it should be symmetric.

The short version

A molecule is polar when its bond dipoles don't cancel out. That depends on two separate things: whether the bonds are lopsided to begin with (electronegativity difference), and whether the shape lets those lopsided pulls add up instead of canceling (molecular geometry).

You can't answer the question from electronegativity alone, and you can't answer it from shape alone. You need both. A molecule with strongly polar bonds can still be nonpolar if the geometry is symmetric enough. A molecule with only mildly polar bonds can end up clearly polar if the geometry breaks that symmetry.

Next time you're staring at a Lewis structure trying to decide, skip straight to the two questions: do the bonds pull unevenly, and does the shape let those pulls cancel? That's the whole test.