clearly_not_atara
International Hazard
Posts: 2787
Registered: 3-11-2013
Member Is Offline
Mood: Big
|
|
Reflections on the vanadium-chlorine water-splitting process
A number of papers have come out describing a vanadium-chlorine thermochemical water-splitting cycle which has some pretty incredible properties. For
example:
https://www.sciencedirect.com/science/article/pii/S019689042...
contains the reactions
2 VCl3 >> 2 VCl2 + Cl2 (798 K)
2 Cl2 + 2 H2O >> 2 HCl + O2 (373 K; 100 C)
Such papers often conclude that the vanadium-chlorine cycle is highly promising and more efficient than other thermochemical cycles:
https://www.sciencedirect.com/science/article/abs/pii/S03603...
These claims can be traced to a 2006 patent by Amendola, who would later found the battery company EOS:
https://patents.google.com/patent/US8287838B2/en
However, Amendola's claims don't seem to be well-justified. He suggests using an "FeCl3 catalyst" for VCl3 decomposition, because as Knoche et al
determined in 1984 (attached), VCl3 decomposes to VCl4 and VCl2, not producing Cl2 directly. But FeCl3 decomposes at about 150 C (423 K), so this
would have to be FeCl2, and this implies a difficult solid-separation step to follow. Worse, there are no papers reporting an FeCl2 catalyst for VCl3
decomposition.
Amendola also claims that a "Co catalyst" can be used to achieve the reverse Deacon reaction in solution, a possibility which he traced to work by
Ulrichson and Powers in 1975 (attached), but they were actually performing the reaction 2 Cl2 + 2 MgO >> 2 MgCl2 + O2. Furthermore, they claimed
that the use of a Co catalyst to perform this reaction in solution was not practical, because there was still significant production of a chlorate
byproduct, and the Co catalyst loses its activity with cycling.
Overall, it appears that these key claims of Amendola's patent are not realistic. His company, EOS, also seems to have some issues with intellectual
honesty:
https://iceberg-research.com/2021/01/14/eos-energy-fake-cust...
Side note: while Iceberg Research calls out the low round-trip efficiency (<80% RTE) of EOS's zinc battery, Gelion, the competing zinc battery
company, claimed an 87% RTE in their IPO filing. So Iceberg may have been too pessimistic about the technology, but their analysis of the fake
customers seems strong.
The other reactions in the vanadium-chlorine cycle, which are considered well-established, are:
2 VCl2 + 2 HCl >> H2 + 2 VCl3 (353 K; 80 C)
4 HCl + O2 >> 2 Cl2 + 2 H2O
Anyway, the collapse of Amendola's claims leads us back to the old vanadium-chlorine cycle. Knoche 1984 (same paper as before) mentions that
VCl4 decomposes at low temperatures as follows:
2VCl4 >> 2VCl3 + Cl2 (423 K; 150 C)
2 H2O + 2 Cl2 > 4 HCl + O2 (1173 K; 900 C)
but this decomposition is very slow, with a half life of around six hours. He suggests pressurizing the reaction, which was demonstrated successfully
at small-scale, but the scalability of high-pressure processes involving chlorine at around 200 C is dubious. Possibly, Amendola's suggestion of an
"FeCl3 catalyst" could work at this temperature, but to avoid a difficult solid separation, a better choice would be SbCl3:
SbCl3 + 2 VCl4 >> 2 VCl3 + SbCl5
SbCl5 >> SbCl3 + Cl2 (decomposes with boiling at 140 C)
SbCl3 is volatile and will boil away at about 200 C, making the separation trivial. The rationale is that SbCl3 may react more readily with VCl4 than
the self-reaction of the latter, because the antimony atom in SbCl3 is less effectively "screened" by the surrounding chlorines than the vanadium in
VCl4. VCl4 is a liquid which is soluble in nonpolar solvents, so it is more likely to dissolve SbCl3 than it is to dissolve ionic FeCl2, and
homogenous catalysis would be highly desirable.
The second reaction, the reverse Deacon reaction, has been extensively investigated since Knoche's work and has been demonstrated in two steps using
magnesium, following the strategy of Ulrichson and Powers:
2 MgCl2 + O2 >> 2 MgO + 2 Cl2 (500 C)
MgCl2 + H2O >> MgO + 2 HCl (500 C)
which can be implemented efficiently with magnesium supported on zeolites as demonstrated by Simpson et al 2006 (attached) among others. These
reactions constitute half of the "magnesium-chlorine hybrid cycle" which finishes with the electrolysis of HCl, but this requires roughly half of the
energy of separation to come from electricity, which impacts the overall energy efficiency, since the conversion of thermal energy to electric energy
is necessarily inefficient, and likewise when the energy source is solar, photovoltaics top out at a disappointing 25% efficient for the panel, and
closer to 20% for the system. But with the vanadium reactions included we have a purely thermochemical cycle with a maximum temperature of 550 C:
2 VCl2 + 2 HCl >> 2 VCl3 + H2 (80 C)
2 VCl3 >> VCl2 + VCl4 (g) (550 C)
2 VCl4 >> 2 VCl3 + Cl2 (150-200 C)
Cl2 + 2 MgO >> MgCl2 + O2 (500 C)
2 H2O + 2 MgCl2 >> 2 MgO + 4 HCl (500 C)
It remains to be seen whether this cycle can be run with an efficiency high enough for thermochemical water-splitting to be desirable. Usually the
goal is to beat the electric conversion efficiency by a significant margin. An advantage is that the maximum process temperature of 550 C is just
barely within the working temperature of molten salts for thermal storage, which may simplify reactor design.
Attachment: knoche1984.pdf (669kB)
This file has been downloaded 213 times
Attachment: ulrichson1975.pdf (5.4MB)
This file has been downloaded 195 times
Attachment: simpson2006.pdf (221kB)
This file has been downloaded 194 times
|
|
|
j_sum1
Administrator
Posts: 6320
Registered: 4-10-2014
Location: At home
Member Is Offline
Mood: Most of the ducks are in a row
|
|
I will read this more thoroughly later.
Clarify for me though:
This process is being proposed as an alternative to electrolysis of water: presumably to get better energy efficiencies.
VCl3 is effectively a catalyst: first releasing Cl2 by thermal decomposition. Then Cl2 reacts with water at 100C to release O2 and produce HCl. (I
have not heard of this reaction before.) The HCl is then sequestered by VCl2 to release H2 and regenerate the VCl3.
And all of this is in an aqueous state at or below 100C. But further catalysts are needed to improve the kinetics.
My first reaction is that it does not sound particularly feasible from a practical standpoint.
|
|
|
clearly_not_atara
International Hazard
Posts: 2787
Registered: 3-11-2013
Member Is Offline
Mood: Big
|
|
| Quote: | | Then Cl2 reacts with water at 100C to release O2 and produce HCl. (I have not heard of this reaction before.) |
When you finish reading the post, you will realize that the reason you haven't heard of the reaction before is because it doesn't work  A patent author misread an old paper.
| Quote: | | This process is being proposed as an alternative to electrolysis of water: presumably to get better energy efficiencies. |
Only when using a thermal energy source. The goal is to make thermal -> hydrogen more efficient than thermal -> electric ->
hydrogen, and usually the efficiency limiting step in the latter process is the thermal -> electric step. Electric -> hydrogen conversion has
been demonstrated with efficiency around 80%, which is very good.
There is an even lower-temperature process, but the metal catalysts are uranium and europium instead of vanadium and magnesium, which creates new
problems, particularly since this "uranium-europium-bromine" process has not been shown to run without significant losses.
| Quote: | | My first reaction is that it does not sound particularly feasible from a practical standpoint. |
An analysis of the complete process I proposed, all of which are experimentally demonstrated reactions, has not been performed. I just wanted to put
it all in a form that was not made up. It seems that Dr. Amendola created quite a bit of confusion.
|
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
