Why We Need a Reciprocal Lattice

A thicket of planes, and why real space fights us

In the indices rung you learned to slice a crystal into families of parallel planes and give each family a name (hkl), and you measured how far apart the sheets in a family sit — the interplanar spacing d. That single number turned out to be the bridge to diffraction: Bragg's law says a family of planes throws back a bright reflected beam only at one special angle set by its spacing. Recall the worked case — for d = 2 angstrom and copper K-alpha radiation of wavelength lambda = 1.54 angstrom, the first peak lands at theta = arcsin(1.54 / (2 times 2)) = arcsin(0.385) = 22.6 degrees. Big spacing, small angle; small spacing, big angle.

Now picture actually predicting a whole diffraction pattern. A crystal does not have one family of planes; it has infinitely many — (100), (110), (111), (210), (321), on and on — and each family has two things you must track at once: which way it faces (its normal direction) and how tightly its sheets are stacked (its spacing d). To work out where every reflected beam goes, you would have to juggle all of those directions and all of those spacings together, in three dimensions, by hand. Real space, for this job, is a thicket.

There is also a subtler nuisance you already met. In a skewed (non-cubic) cell, the direction a family of planes faces is generally NOT the like-numbered direction [hkl] — plane normals and directions parted ways the moment the axes stopped being perpendicular. So even the innocent question 'which way does this family point?' is fiddly in real space. What we want is a bookkeeping device that hands us the normal direction and the spacing of every family, automatically, in one clean object.

The clever move: one dot per family of planes

Here is the trick that untangles everything, and it is almost absurdly simple. Instead of drawing a whole family of planes — infinitely many parallel sheets filling all of space — represent that family by a single point. The direction from the origin to the point is the plane normal, the way the family faces. The distance from the origin to the point is 1/d, the reciprocal of the spacing. That one dot is a reciprocal lattice point, and the arrow pointing to it from the origin is the reciprocal lattice vector g_hkl, with length |g_hkl| = 1/d_hkl.

1. Pick a family of planes and label it (hkl); look up (or compute) its interplanar spacing d_hkl.
2. Stand at the origin and face along the plane normal — the direction the sheets of that family are turned toward.
3. Walk a distance of exactly 1/d_hkl along that normal. Note the units: if d is in angstrom, this distance is in 'per angstrom' (angstrom^-1).
4. Drop a dot. That single dot is the reciprocal lattice point (hkl); the vector from origin to dot is g_hkl. Repeat for every family, and a new picture starts to fill in.

Sit with why the recipe uses 1/d, because that inversion is the whole personality of this new space. Widely spaced planes (a large d) map to a point close to the origin; finely spaced planes (a small d) map to a point far out. Take a cubic crystal with edge a = 4 angstrom. The (100) planes are spaced d = 4 angstrom, so their point sits at 1/4 = 0.25 per angstrom. The (200) planes are spaced d = 2 angstrom, so their point sits at 0.5 per angstrom — exactly twice as far out, along the same normal. That far point standing for tighter planes is precisely the second-order reflection you will meet again in the diffraction rung.

The surprise: those dots form their own lattice

Now the beautiful part. When you plot the point for every family (hkl), the dots do not scatter about at random. They land on a perfect, periodic array of their own — the reciprocal lattice, the reciprocal lattice. Just as the direct lattice is generated by three vectors a, b, c, this new lattice is generated by three reciprocal lattice vectors written a*, b*, c* (read 'a-star, b-star, c-star'), and every reciprocal point is simply g_hkl = h a* + k b* + l c*, with h, k, l the same integer Miller indices as before. The next guide builds a*, b*, c* explicitly out of the direct cell; for now, just meet them.

 direct-space planes          reciprocal-space point
 (family, spacing d)          g = 1/d, along the plane normal
 --------------------------   ------------------------------
 (100)  d = 4.00 angstrom     |g| = 0.250 /angstrom   near origin
 (200)  d = 2.00 angstrom     |g| = 0.500 /angstrom   twice as far
 (300)  d = 1.33 angstrom     |g| = 0.750 /angstrom   three times

 origin O ---*-------*-------*-------> a* axis
            100     200     300
            evenly spaced by 1/a = 0.25 /angstrom

The three vectors a*, b*, c* span a reciprocal unit cell, the repeating tile of this shadow world, and it has a lovely property: its volume is exactly the inverse of the direct cell's, V* = 1/V. For our cubic a = 4 angstrom cell the direct volume is V = 4^3 = 64 angstrom^3, so V* = 1/64 = 0.0156 per angstrom cubed, and (because a cube's axes are perpendicular) each reciprocal edge is simply a* = 1/a = 0.25 per angstrom, matching the row above. Small real cell, roomy reciprocal cell — hold that thought, because it is the key to the next section.

One honest bit of small print before we go on. There are two conventions in the wild. Crystallographers (and this ladder) define |g_hkl| = 1/d with no extra factor, so the numbers above are exactly the distances you plot. Solid-state physicists usually fold in a factor of 2 pi, writing the reciprocal vectors so that |g| = 2 pi / d and V* = (2 pi)^3 / V. Both describe the identical lattice, just rescaled; when you read a textbook, check which convention it uses before you trust a number.

The shadow world: a Fourier dual with inverse size

The reciprocal lattice is not just a clever filing system; it has a deep identity. It is the Fourier transform of the direct lattice — its mathematical dual, the same object viewed through the lens of waves rather than positions. That is the theme of guide 4, so here we only plant the flag. The vivid way to hold it: the reciprocal lattice is the crystal's shadow. A diffraction camera does not photograph the atoms directly — it photographs this shadow, and reading the shadow is how we work back to the arrangement that cast it.

The dual relationship carries a rule with real bite: the inverse size relationship. A small real cell makes a large reciprocal cell, and vice versa, because a* = 1/a. Watch it work on two extremes. A tiny metal cell with a = 2 angstrom gives a* = 0.5 per angstrom, so its reciprocal points are flung wide apart — few of them, well separated, easy to tell one from the next. A big protein crystal with a = 100 angstrom gives a* = 0.01 per angstrom, so its reciprocal points are crammed together fifty times more tightly — thousands of closely spaced spots. That single fact is why protein diffraction patterns are dense clouds of dots while a simple metal gives a handful of clean ones.

Why crystallographers live in reciprocal space

Here is the pay-off that makes the whole detour worthwhile: a diffraction pattern is a direct picture of the reciprocal lattice. Every bright spot on the film is one reciprocal point that happened to be caught diffracting. So a crystallographer stops picturing planes reflecting at fussy angles and instead reads the shadow straight off the detector. The bookkeeping tool that decides which reciprocal points are 'caught' at any instant is the Ewald sphere — a sphere of radius 1/lambda drawn in reciprocal space, where a reflection fires whenever a reciprocal point touches its surface. It is guide 5's business, but note the punchline now: Bragg's law and the Ewald sphere are the very same statement, spoken once in real space and once in reciprocal space.

Reciprocal space also quietly hosts an idea you will lean on for electrons and heat. Back in the lattice rung you built the Wigner-Seitz cell — the patch of space closer to one lattice point than to any other. Build that very same cell around a point of the reciprocal lattice instead, and it earns a new name: the Brillouin zone. It is the natural stage on which the physics of electrons and lattice vibrations is written. That is a whole story for later; here, just register that the reciprocal lattice is where it lives.

So this is why we need a reciprocal lattice, and why crystallographers genuinely think in it. It is not abstraction for its own sake: it is the object a diffraction camera photographs, a shadow in which every family of planes shrinks to one well-behaved dot, and the natural home of Bragg's law, the Ewald sphere, and the Brillouin zone all at once. With the motive settled, the next guide rolls up its sleeves and does the construction — building a*, b*, c* explicitly from the direct lattice vectors a, b, c.