← All articles
chemistrymaterialselectronicstechnologySeptember 17, 20264 min read

Why Add Impurities on Purpose? Making Silicon Useful

By the BrainSnail editorial team. How these articles are written and checked, and how to tell us when one is wrong.

Pure silicon conducts electricity poorly and is nearly useless. Adding a few atoms of something else per million transforms it, and controlling where those atoms go is what makes every chip possible.

What the addition does

Silicon atoms each bond to four neighbours in a regular lattice, using all their outer electrons, which leaves very few free to carry current. Substituting an occasional atom with five outer electrons leaves one electron without a bond to occupy, and that electron moves freely, so the material conducts. Substituting instead an atom with three outer electrons leaves a bond short of an electron, and neighbouring electrons shift to fill the gap, which moves the vacancy through the lattice and carries current as though a positive charge were travelling. The two cases are called n-type and p-type for the sign of the moving charge, and the quantities involved are tiny, with concentrations of roughly one added atom per million or less changing conductivity by many orders of magnitude.

How it is done

Several methods put the atoms where they are needed:

  • Ion implantation, firing the atoms into the surface at high energy, which gives precise control of depth and quantity
  • Diffusion, heating the wafer in an atmosphere containing the element so it migrates in from the surface
  • Adding it during crystal growth, which dopes the whole ingot uniformly
  • Masking with patterned layers so only selected regions receive the treatment
  • Annealing afterwards to repair lattice damage and settle the added atoms into proper positions
  • Repeating the whole sequence many times to build the structures a chip requires

The junction that does the work

Bringing the two types into contact produces the structure everything else is built from. At the boundary, free electrons from one side fill vacancies on the other, leaving a region depleted of mobile charge and carrying a built-in electric field. That field allows current to flow readily in one direction and blocks it in the other, which is a diode. Arranging three regions in sequence so that a small current or voltage at the middle controls a much larger one between the outer two gives a transistor, which is both an amplifier and a switch. Everything from a radio to a processor is built from those two structures repeated, and the reason the industry could scale is that the same photographic patterning process makes one or a billion of them for nearly the same effort.

Other materials that get the treatment

The technique is not confined to silicon and the variants explain several familiar devices. Light emitting diodes are junctions in compound semiconductors chosen so that the energy released when charges recombine falls in the visible range, with the colour set by the material rather than by any filter, and blue required a material that resisted the necessary doping for decades, which is why blue and therefore white versions arrived so much later than red and green. Solar cells are large-area junctions in which absorbed light creates the charges rather than a current supplying them. Thermoelectric materials are doped to tune how they carry heat and charge. And adding elements to glass and ceramics to change colour, refractive index or fluorescence is the same idea applied to materials that conduct nothing, which is where the word originally came from.

Why purity matters so much

Since the deliberate additions are measured in parts per million, anything else present at comparable levels ruins the result, which is why semiconductor silicon is among the purest bulk material made. Production reduces impurities to parts per billion and below through repeated chemical purification and then crystal growth techniques that exploit the tendency of impurities to remain in the melt rather than entering the solid, so drawing a crystal slowly out of molten silicon leaves contaminants behind. The resulting ingot is a single crystal with almost no defects, sliced into wafers and polished flat to a remarkable degree. Contamination control in fabrication plants follows from the same requirement, since a stray metal atom behaves as an unwanted dopant and a particle of dust destroys whatever it lands on.

The takeaway

Substituting an occasional atom with one more or one fewer outer electron leaves a mobile electron or a mobile vacancy, and roughly one added atom per million changes conductivity by orders of magnitude. Joining the two types gives a region that passes current one way only, and three regions give a switch. The additions are so dilute that starting material must be pure to parts per billion.

Practise this

Questions from Environmental Chemistry

Reading about something is not the same as being able to recall it. These are real questions from the Environmental Chemistry unit in our Chemistry track, answers and explanations included. The unit has 120 in total across 20 steps.

  • Choose all that applyLevel 1

    1. Which of these are greenhouse gases? (Select all that apply.)

    • Carbon dioxidecorrect
    • Methanecorrect
    • Water vapourcorrect
    • Nitrogen

    Carbon dioxide, methane and water vapour all trap heat, while nitrogen is not a greenhouse gas.

  • Match the pairsLevel 2

    2. Match each gas to its approximate proportion in clean, dry air.

    Answer: Nitrogen = 78%; Oxygen = 21%; Argon = 0.9%; Carbon dioxide = 0.04%

    Air is roughly 78% nitrogen, 21% oxygen, 0.9% argon and only about 0.04% carbon dioxide.

  • Fill the blankLevel 2

    3. Oxides of nitrogen form inside car engines because the ____ there is very high.

    • temperaturecorrect
    • pressure
    • humidity
    • sulfur level

    The very high temperature in an engine makes nitrogen and oxygen from the air react to form nitrogen oxides.