A previous posting introduced ion channels, including the voltage-gated variety where the voltage across the membrane controls whether ions can pass through the membrane. In the posting, such an ion channel was compared with an electronic transistor (transistor=‘transfer resistor’) – for the purposes of using transistors to implement models of ion channels. What wasn’t considered was doing things the other way around – trying to use ion channels as ‘transfer resistors’ in order to do computation. The posting also introduced the Sodium Potassium pump – a ‘bipolar’ circuit, flip-flopping state between collecting 3 sodium ions from inside and collecting 2 Potassium ions from outside. From my non-biological background, I find these ‘components’ astounding – that such precision mechanical engineering should be integral (fundamental even) to the gooey stuff that is biology. Why haven’t people tried to make computers out of this technology? Why not circuits from ion channels and memories from Sodium Potassium pumps?
And what of other neurological technologies? Could not the technology of synaptic vesicles, which move around an axon terminal of a neuron collecting and releasing neurotransmitter proteins, be reapplied as a memory? (They look strangely reminiscent of computer magnetic bubble memories that went nowhere in the 1970s.)
Of course for years, the relentless march of silicon technology under the banner of Moore’s Law has rendered most other technologies irrelevant. But, to take a point in time a while back now, 20 years ago when a transistor’s channel length was about 1um=1000nm, a transistor behaved pretty much like a switch: on or off. Nowadays, with feature sized at a nominal 22nm, those billions of transistors inside a Pentium processor are quite horrid little things. A Sodium-Potassium pump looks quite well behaved compared with these modern transistors. And with the membrane thickness of only about 5nm, they’re considerably smaller. That probably isn’t small enough though: silicon technology should be down to that size before 2020.
What might a technology of computing with ions be called? If electronics is interpreted in the most abstract sense as ‘computing with electrons’, we might expect this technology to be called ‘ionics’. But ionics (specifically ‘solid-state ionics’ ) transpires to be about solar cells, batteries and other fuel cells. (An alternative guess would be ‘cationics’ but this transpires to be more about personal hygiene products!)
But then, this week’s New Scientist (19 June 2012) contains an article ‘Positive Switch for Body Electronics’ (subscription required) that suggested my imagined possibilities are not so fanciful. It is concerned not with the (unrealistic) notion of large-scale computation but with the making of neuron sensors – and specifically of the difficulty of connecting biological systems – neurons – which work by moving (generally positively-charged) ions around – with electronics – which works by moving negatively charged electrons around.
Some references it mentions are:
- Engineered devices that push sodium, potassium and calcium ions into cells (Sensors and Actuators A: Physical, vol 93, p 8)
- The ‘proton transistor’: Marco Rolandi (University of Washington, Seattle) has been able to push positive charges around to allow protons to hop along the fibres of a sugar molecule, maleic chitosan, by applying a controlling voltage (‘A polysaccharide bioprotonic field-effect transistor’) . See also the IEEE Spectrum article.
- Only obliquely referred to but in my view the most interesting is news earlier this year of building CMOS-like invertor and NAND gate circuits from ion transistors which can transport both positive and negative ions. This is described in ‘Logic gates based on ion transistors’ (Nature Communications, vol 3, p 871) by Klas Tybrandt, Robert Forchheimer and Magnus Berggren (Linköping University) . The chemical chips can control the delivery of the neurotransmitter acetylcholine, permitting the control of muscles and potentially much more within the human body. See also this article.
Whilst not quite what I was imagining, these at least show some practical progress in the basic area of biological transistors.