Homogenous Charge Compression Ignition (HCCI) is the latest fashion in trying to eke out that extra bit of efficiency from classical gasoline engines. An HCCI engine does away with the standard spark plug and, instead, mirroring its cousin the Diesel engine, opts to ignite the fuel-air mixture via compression. However, unlike the diesel engine, the HCCI engine does not compress just clean air: it compresses a fuel-air mixture, like normal gasoline engines. Naturally, this means that the single greatest challenge in building an HCCI engine is being able to satisfactorily control the reaction inside the cylinder, as the extremely hot-fuel air mixture is wont to explode on its own accord at an inopportune moment.
Naturally, any engineer worth his (or her!) salt would wonder why bother with such a revolutionary, apparently complex and definitely wacky concept. The most compelling reason, from the outset, is that an HCCI engine can potentially extract much more work from a given volume of fuel than a spark-ignited engine. Since both types of engines approximate an ideal Otto cycle, and since the Otto cycle's efficiency is tied to compression ratio, an HCCI engine running at a higher compression ratio than a regular spark ignited engine is necessarily more efficient.
“But,” you may ask, naively, “don't Diesel engines already operate at much higher compression ratioes than gasoline engines? And aren't they already widely deployed and easy to control?” You would, indeed, be right: however, it is important to note that, at the same compression ratios, the Diesel cycle is less efficient than the Otto cycle. This is the key point: the HCCI engine attempts to run a gasoline engine at Diesel cycle-level ratios, resulting in tremendous improvements in efficiency.
Knocking and Preignition
There seems to be some confusion on the web and in this question about the definition of engine knock. The correct definition (and the one we were taught in class) is that knocking happens when a desired combustion initiates several other undesired combustions simultaneously due to the shock wave from the initial combustion locally creating regions of high enough temperature and pressure to start a new reaction. The popular definition (more correctly called pre-ignition) is when the air-fuel mixture in a cylinder ignites before a spark is even produced, causing the explosion to work against the direction of the piston and really messing up the whole process.
HCCI does, however, influence the effect and severity of both these effects. In the case of true blue engine knock, HCCI does eliminate the problem somewhat by making it irrelevant: as the temperature of the whole gas increases at approximately the same rate, the whole gas will enter combustion at approximately the same time. Thus, once the first combustion starts, there's nothing else left to combust, so the problem of engine knock is effectively eliminated.
Preignition, however, is the number one issue plaguing the HCCI. The fuel-air mixture needs to be brought to the temperature at which it will auto-ignite only moments before it hits TDC in the cylinder. If it happens too early, the combustion will kick back on the cylinder and make a truly ghastly noise; if it doesn't happen before TDC, it'll never happen, since the expanding gas will necessarily cool down from it's high at the top of the cylinder. This illustrates the delicate balancing act that is a running HCCI engine.
Emissions, from the get go, will be reduced by a small but significant margin: as HCCI engines extract more work per unit of fuel, one needs less fuel to accomplish the same task. As you're burning less fuel, you're creating less carbon dioxide. And the carbon dioxide you don't create doesn't get thrown out into the atmosphere.
However, the key point to note with non-carbon dioxide pollution is that it's dependent on two factors: amount of uncombusted reagents in the exhaust and the temperature at which combustion occurs. If the combustion is too hot, more nitrous pollutants (so-called NOx gasses) are created and then released into the atmosphere. If the combustion has more fuel than it has oxygen, then free, unburnt hydrocarbons are released.
The solution, on the surface, seems trivial: keep the combustion cold and the amount of air high. The problem, however, is three fold:
The richer your engine runs (that is, the more excess fuel you have), the cooler the combustion. In fact, slightly-leaner-than-stoichiometric fuel concentrations burn the hottest. As such, since HCCI engines are typically lean, they burn significantly hotter than comparable spark engines.
The colder your engine runs, the less efficient it becomes. This can be gleamed simply by looking at how the Carnot efficiency varies. Furthermore, the HCCI engine has a definite lower bound to the temperature at which it can run: if the temperature is lower than this bound, combustion doesn't happen at all, and no work is extracted.
Furthermore, the colder your exhaust gasses are, the less efficient your catalytic converter is at eliminating the vast majority of pollutants in your exhaust.
These three factors, again, make emission control with an HCCI engine another delicate balancing act. If too much effort is put into reducing one type of emission, all the others may suffer as a result.
Your humble narrator finds it doubtful that HCCI technology would ever be adopted on a commercial scale. Not only is it an entirely new technology with uncertain payoff (in the public's eye), it is, unfortunately, too sensitive to environmental conditions: a slightly unexpected fuel blend could throw off the reaction entirely, making the whole device unable to run; unusually low atmospheric pressures on a cold day could make it nearly impossible to start; a defective sensor somewhere in the engine could stall it hard or could make the engine destroy itself. These problems, as always, could be mitigated by “throwing transistors at the problem” and just installing redundant and highly-sophisticated electronics all over the place. However, this brings up the issue of cost and serviceability, which this armada of electronic devices could very well clobber out of existence.