It doesn't need to be on every corner, you'd only fill up the ethanol tank every 3 to 10 refills depending on how big your gasoline tank is and how much high-torque operation you use. Once the ethanol runs out it would just limit the turbo boost to a very low level until you filled up the ethanol tank again. It's a very creative way to satisfy both the desire to run good old regular gasoline for economy but still have all the power of a big engine.

The only major downside I see here is that typically highly boosted engines have the most turbo lag since on a low bost engine, it just adds another 30% more atmospheric pressure to the manifold - this implies they're going to run 1 bar or more (100% or more additional manifold pressure) which tends to translate into a strongly felt rubber band effect common to older turbocharged engines.

"every other corner gas station"
Not at every corner, or even at every gas station. Most gas stations are at street corners/intersections. Perhaps it would be more accurate to say "every other gas station" or "every few gas stations", or something like that.

...and in the early 90s Mitsubishi had a 1.6L V-6 in the Mirage Cyborg before they went to the 1.6L MIVEC I-4.

A very creative idea. I could see some usability issues i.e. what happens if you fill the ethanol tank with gas by accident or fill the gas with ethanol. This would need some extra safety controls. I would be fine with the idea. However, I could see non-technical people having a problem with two fuel systems

By the way, could this system work with methanol or buthanol as well?

Methanol and butanol have totally different calorific count, or energy density, or whatever you want to call it.

Methanol: 55,000 btu/gal
Ethanol: 78,000 btu/gal
E85: 82,000 btu/gal
Butanol: 105,000 btu/gal
Gasoline (for reference): 110-114,000 btu/gal

It makes it extra hard to do precise fuel metering if you're not sure what's in the tank. The good news is all the alcohols have reasonably similar octane values, so in that sense, any of them would work reasonably well.

Interesting idea, but couldn't a similar system use high octane gasoline? If there were a way to not destroy the catalytic converter such a system might get much higher MPG if the injector used 120 octane leaded gasoline. Nasty, but might be worth a cost/benefit before we judge it too much.

Methanol would work great in this application because MIT is relying on the Latent Heat of Evaporation (methanol has a way higher LHE than ethanol) and Octane of the fuel to lower the knock limit. Since they are using small amounts, methanol would be even better than Ethanol.

Butanol would not work as well because it has a lower LHE and lower octane than either methanol or ethanol.

Also, the ECU would really not care what fuel(s) are used as long as there is a wide band sensor and the long term and short term trims are able to adjust to a wide variety of AFRs.

This sounds like a good idea, but for an estimated cost increase of only $1000, it is way too optimistic. The turbocharger with intercooler and wastegate control etc. will cost ~$2000 USD extra. And then, there is another set of high-pressure alcohol direct injectors with at least four of them, AND another high pressure fuel pump for these alcohol direct injectors, which will cost at least another $1000 USD. Plus an alcohol reservoir that must be strategically placed and protect from collision-induced fire hazard, since alcohol can burn without a flame that would at times be even more hazardous than gasoline with yellow flame. This is a concern in Indy Car Racing which uses methanol. Plus more careful calibration and higher quality sensors, better programming, etc. to avoid catastrophic engine destruction at high boost. It doesn't take long or much to destroy an engine from preignition at high boost.
Can ethanol injection be done via the same gasoline port injectors? Probably not, due to lead volume inside the injector, the time lag in introducing the ethanol hence the power boost may be longer than the driver would want when instant power boost is needed.

So far, I've tallied over $3000 USD to $3500 USD of extra cost for ~30% gain in fuel efficiency. By comparison, a Camry hybrid at 40mpg is 42% better than a 4-cyl Camry at 28mpg, and costs ~$4000-5000 USD more when compared to a 4-cyl Camry loaded with similar luxury options. The Camry hybrid accelerates much faster than the non-hybrid version, so some of the $4000 cost increase in the Camry hybrid is justifiable based on the increase in performance alone. Now, if the Camry hybrid was made to use the Prius' engine and drive train, the mpg improvement would be even more dramatic, perhaps at 60-70% that of the 4-cyl Camry, at comparable or slightly slower acceleration, but with even less cost differential, given the fact that Prius drive train is smaller hence cheaper and mass-produced at higher number.

My Lord, what a shame for MIT.

Dual-fuel injection is established practice in performance cars for years. Port injection was proven decades ago to yield the best results because of:
1) evaporation of fuel in intake air makes air more dense;
2) cooled air has way lower viscosity, unlike liquids.
The best results are achieved with methanol/ethanol mixed with water injection, not just high-octane fuel. The practice of water injection into intake air is routinely employed on diesel engines. On high-performance gasoline engines it is standard feature even on semi-production performance cars.

Optimization of dual-fuel injection to downsize engine, boost effective compression ratio, and hence yield better fuel efficiency is really something new, but yield could be in the range of 5% gain max.

This hype is surely to compete with my “scum of the year” award with EESTOR and Hy-Drive.

Good old GM tried a version of this Circa 1962. Note the serious performance with only 5psi of boost, a single barrel carb, and no intercooler. The engine mentioned is currently still in use in Land Rovers. As for # of cylinders and small displacement, The Benelli Sei has a 750cc six, and is as smooth as a 3phase induction motor @ 60hz, IMO.

Jetfire Engine
In 1961, the F-85 had a 155 HP version of the 215 ci. (3.5L) engine, and the Cutlass had a 185 HP version. The 1961 versions of the motor were rated at 155 hp, but later years saw increases - up to 200 hp normally aspirated for the Buick version and 215 hp in turbocharged form from Olds (the 62-63 Jetfire).

In 1962, Olds, along with AiResearch, introduced a 'turbocharged' (called Fluid Injection) version of this engine, which put "Turbo Rocket Fluid" (½ distilled water, ½ methyl alcohol) into the carb. Along with a 10.25:1 compression ratio, yielded 0-60 in 8.5 seconds (with the manual tranny). The turbo was a Garrett TO-3 with an integral wastegate, the first. Unfortunately, due to the 10:1 compression ratio, boost was limited to only 5 psi, not the best use of a turbo.

The induction setup itself is fairly sophisticated (especially for 1962), with something like 54 separate connections to the intake system. The turbo has an integral wastegate, being the first mass production turbo application to use a wastegate. This was arguably the most complex induction system build to that time, with something like 50 different hose connections in the intake system (pressure sensors, wastegate, fluid injection, fuel, etc).

Olds attempted to get around the boost lag problem by using a high compression ratio (10:1!), which limited boost to only 5 psi. Fluid injection (Turbo Rocket Fluid) was used (a water/alcohol mix) to suppress detonation. Properly running cars will not go into boost if the "Turbo Rocket Fluid" reservoir is empty. There is an automatic shutoff for this. Parts for this injection system are even harder to get than the turbo parts.

The carb is a rare single barrel Rochester side-draft unit, whose only other application was on the Corvair Turbo. While they resemble the Corvair carb, they are much larger.

The Olds' turbocharged Jetfire was supposedly quicker than the 4V version, but it had maintenance problems due to its complex mechanics for that era. It attained the magic goal of 1 HP per CID. A power boost on the order of 40% was claimed. The automatic Cutlass with 10.75:1 compression gave 195 HP @ 4800 and 235 lb/ft @ 3200. The Jetfire's 10.25:1 compression gave 215 HP @ 4800 and 300 lb/ft @ 3200.

Source: http://www.442.com/oldsfaq/ofjet.htm
Of course SAAB can give you this performance today with 140 ci (2.3L) or less.

Increasing BMEP is the answer, and biohols are the correct fuels for doing this in SI engines. Any serious arguments against MIT's work in this area have pretty much been made moot by Harry Ricardo. One would do well to start their research by reading his works.
http://en.wikipedia.org/wiki/Harry_Ricardo

Roger -

I hope you're not working in purchasing. Your estimates regarding the extra cost are way too high. Turbos have become pretty cheap since European diesels all use them and, GDI is well on its way to becoming a commidity subsystem as well. Don't assume that just because US engine technology generally lags Europe and Japan by at least a decade that it cannot catch up quickly.