Electric alcohol

The re-appearance of the ORNL direct electrocatalytic conversion of CO2 to ethanol and other products in news about ReactWell piqued my curiousity about this process.  The paper describes the yields in terms of "Faradaic efficiencies" which is an unfamiliar term, but appears to relate to Coulombic efficiencies by a direct conversion factor; one Faraday is a mole (6.023*10²³) of electrons, while a Coulomb is an ampere-second.  Ergo, one Faraday is roughly 96485 Coulombs.

I'm interested in the mass yields and energy efficiency of the process, which requires converting from faradaic to regular physical units.  First comes the required charge transfer per reaction.  I calculate the stoichiometry as follows:

2 H3O+ + 2e-	➡	H2 + 2 H2O
CO2 + H2O + 2e-	➡	CO + 2 OH-
CO2 + 6 H2O + 8e-	➡	CH4 + 8 OH-
2 CO2 + 9 H2O + 12e-	➡	CH3CH2OH + 12 OH-

Given the Faradaic conversion efficiencies to various products as given in the paper, I come up with these net yields:

Product	F. yield, %	e-/reaction	mol/Faraday	g yield/mol CO2
EtOH	63.0	12	0.0525	17.31
CH4	6.8	8	0.0085	0.97
CO	5.2	2	0.0260	5.22
H2?	25.0	2	0.1250	n/a

Re-crunching this with an eye toward heat of combustion of the products:

Product	mol/Faraday	ΔHc, kJ/mol	ΔHc, kJ/Faraday
EtOH	0.0525	1367	71.8
CH4	0.0085	891	7.6
CO	0.0260	283	7.4
H2?	0.1250	286	35.8
TOTAL			122.6

Calculating total input energy naïvely, 96485 coulombs times 1.2 volts yields 115.8 kJ.  This is clearly nonsense.  Going back to the electrochemistry, the paper declares that the potential is given "vs. RHE", a reversible hydrogen electrode.  The oxygen evolution reaction is going to occur at a considerably higher potential than this.  The equilibrium potential of an oxygen electrode is +1.23 V vs. RHE, which sets a floor of 2.43 V on the cell voltage.  Using that, 96485 coulombs times 2.43 volts yields 234.4 kJ for a maximum electricity-to-fuel efficiency of 52.3%; only 37% goes toward reducing CO2 and just 30.6% to energy in ethanol.  15.3% goes to hydrogen.

ReactWell appears to be a bio-fuels company previously specializing in biocrude production.  This is a related business, as all the products of the ORNL process can be sold or used at a refinery.  Oxygen can supply anything that would be otherwise fed by air separation, H2 can go straight to hydroprocessing, the CH4 replaces natural gas for SMR or process heat, and the CO can be added to the input of the reverse water-gas shift reactor in the SMR system to make more hydrogen.  Maybe the efficiency is low, but when California has a low-carbon fuel standard and is paying Arizona to take its peak generation from PV, the efficiency is not such a huge factor.

The Engineer's take:

1. This is nowhere near the world-killing advance I thought it was when I read the first reports in 2016.  The energy efficiency is just too low, and it doesn't include any overhead for CO2 capture or separating the ethanol from the aqueous medium.
2. As a dump load for unreliable electric generation (especially wind and PV), this might be just the ticket.  So long as the catalyst is not degraded by voltage swings this process can replace expensive or difficult-to-site storage such as batteries and pumped hydro.  Enough capacity and negative wholesale electric prices would be a thing of the past.  Sure can't complain about that.

The Poet's take:

Electrolytics
Making booze from cee oh two
Amuses me much.

  ¶ 3/09/2019 12:39:00 PM