The mill you can
smell from town.

What the cook actually does.

Wood is fibre glued together. The fibre is cellulose and you want it. The glue is lignin, roughly a quarter to a third of the dry weight of the tree, and you want it gone without chopping the fibre up in the process. Grinding wood mechanically gives you high yield and short, weak fibre. Dissolving the lignin chemically gives you long, strong fibre, which is why almost every strong brown box and most printing paper starts here.

The kraft process does it with white liquor: sodium hydroxide and sodium sulfide in water. Chips go into a digester, the liquor goes in with them, and the whole thing sits at about 170 °C under its own steam pressure for a couple of hours. Kraft is German for strength, and that is the whole selling point of the process.

The useful reaction is the sulfide ion attacking the bonds that hold lignin together, breaking it into fragments small enough to dissolve. That is the reaction you are paying for. The smell is the side reaction: the same sulfide also attacks methoxyl groups on the lignin and makes small, volatile, organic sulfur compounds that no longer want to stay in the liquid.

The four gases, and the numbers.

Mills group them under one meter reading, TRS, for total reduced sulfur. Four compounds do essentially all of it. Three of them are made by the cook itself; hydrogen sulfide is mostly made later, in the recovery furnace and wherever green liquor meets acid.

Now the part that explains why you smell it from the road. These are not high concentrations. They are extraordinarily low detection thresholds. Here is where each one crosses the threshold of the average human nose, on a log scale, in parts per billion:

0.07 ppb is 70 parts per trillion. One part in fourteen billion. Put it another way: a single gram of methyl mercaptan, spread evenly, would be detectable through roughly seven million cubic metres of air, which is a cube about a hundred and ninety metres on a side. That is the whole answer to why a mill that is well inside its emission permit is still perfectly obvious from a car two kilometres away.

It is worth being clear that smelling it is not the same as being harmed by it. These thresholds sit thousands of times below the occupational exposure limits. The smell is an intensely sensitive detector attached to a compound that is genuinely unpleasant, not an alarm about the dose. That gap is also why odour complaints and compliance reports can both be honest and still disagree.

The sharp note is another building.

Brown stock out of the digester is still brown. Making it white is a separate plant at the other end of the mill, and it smells completely different: sharp and sour rather than rotten. That is not TRS. That is chlorine chemistry.

Modern mills bleach with chlorine dioxide, ClO₂. It cannot be shipped, because it decomposes, so every mill that uses it makes it on site from sodium chlorate, reduced with methanol in strong sulfuric acid:

And here is the part worth saying plainly. Chlorine dioxide has an odour threshold around 0.1 ppm, and the concentration considered immediately dangerous to life or health is 5 ppm. Those numbers are only a factor of fifty apart. Compare that with methyl mercaptan, whose IDLH is 150 ppm against that 0.00007 ppm threshold: a margin of roughly two million.

That narrow margin is the whole reason the rule in a bleach plant is that you should not be smelling it at all. A kraft mill smelling of rotten cabbage is a mill doing its job slightly imperfectly. A bleach plant you can smell is a leak, and the distance between noticing and being in trouble is short.

The loop nobody smells.

Here is the thing that makes kraft an industrial process rather than a way of ruining a river. Cooking chemical is expensive and the dissolved lignin is fuel. So the mill burns the lignin to make its own steam, and pulls the sodium and sulfur back out of the ash to cook the next batch. The black liquor that leaves the digester comes back as white liquor, and it does it continuously, forever.

Walk it once. Black liquor leaves the digester at roughly 15 percent solids, far too wet to burn. Multiple effect evaporators take it up to 65 to 85 percent solids, at which point it is a thick, tarry fuel. It is fired in the recovery boiler, which is two machines in one: the top half is a steam boiler that drives the mill, and the bottom half is a reducing furnace where sodium sulfate is converted back to sodium sulfide. Molten smelt runs out of the floor, mostly sodium carbonate and sodium sulfide.

Dissolve the smelt and you have green liquor. The sulfide half is already what you want. The carbonate half is not, so you add burned lime, which swaps the carbonate off the sodium and drops out as calcium carbonate:

Add it up and 96 to 98 percent of the sodium and sulfur put into the cook comes back to the cook. That efficiency is not an environmental nicety, it is what makes the process economic at all. And it reframes the smell entirely: what you smell from the road is the two to four percent that gets away. Every improvement a mill makes to its odour is really an improvement to its chemical recovery, which is why the two get fixed together.

It is a flowsheet.

Nothing on that circle is exotic. It is evaporation, combustion, dissolution, a precipitation reaction and a kiln, joined by recycles. It is a flowsheet, and the reason it is hard to model is the reason every recycle is hard to model: the composition going in depends on the composition coming out, so you have to converge the loop rather than march through it.

Pulp and paper is specialised enough that it has its own simulators, tuned for fibre, liquor and the non-process elements that build up around exactly the loop drawn above. That is one reason a process engineer's skills do not transfer between industries as smoothly as the thermodynamics suggests they should: the physics is general, and the tool on the desk is not.

It is also, for what it is worth, my dad's industry. Thirty-two years of it. Which is probably why a mill smells like a place to me rather than a problem.