low voltage lithium recovery

[whatever]

if you dont like alot of details dont read on!

( Postscript: this thread details ongoing work on my attempts to measure capacity of Lithium Cobalt Oxide cells with a view to capacity recovery eventually, many variables which I had not predicted make it a complex process, I had may as well put lots of the info up here, it might assist others trying to do same or similar things).

I started a thread to document some experimenting with soft a123 20ah cells a few weeks back
http://endless-sphere.com/forums/viewtopic.php?f=14&t=69402
I needed to relearn how to use an icharger with logview before I tested the a123 20ahr cells, so I started messing with some very old lithium cobalt oxides ( not sure of exact age probably around 8 years or so). The cells were 5ahr when new, metal jacket type approx same size/shape as turnigy cells use in quadcopters etc. Theres no brand on the cells.
So I got sidetracked a bit and started messing with these old cells. The cell i'm testing at the moment has lost its voltage, it was near zero volts ( o.26v). I've heard cells that go to low voltage are unuseable but there may be some usage for these type of cells.
The icharger wont charge the cell if set to Lilo type ( assuming lilo is for lithium cobalt oxides), but it will charge the cell if set to nicad manual charge. Using the nicad manual charge at 1amp the cell has charged up. It doesn't follow the standard lilo charge scenario though. Here is a charge graph:



The charge graph is broken into two parts, due to a hickup using icharger on this laptop. The screensaver was set for 10mins, for some reason this stopped the connection between icharger and computer, so I had to restart the charge.
Setting screensaver time for a long period solved the disconnect problem between icharger and this laptop i'm using.
Inbetween the graphs was a 15min period where charge continued on icharger but I didn't log it.
I'm currently doing another charge graph that i'll post shortly.
What it shows is that a higher voltage is required than the standard lithium charge ( lilo or lipo) on the icharger.
If I try to charge the cell on lilo or lipo the cell goes directly to 4.1 or 4.2 volts and the amps goes very low, I assume the icharger is using internal resistance to gauge when to go into constant voltage stage and lower the amps.
Using nicad charge ( manual ) on the icharger allows the voltage to go higher, it sits from 5.2v at start of charge and goes to
5.7v near end of charge. Once the cell is off the charger it settles back down to 4.2v and holds that voltage.
The cell has a high internal resistance ( 1355mohm....assuming the m is milliohm?)
So the higher voltage seems to overcome that high internal resistance and allows the cell to charge at 1amp constant current,
it also means there is no constant voltage stage in the charge.
I stopped charging when the graph started to get fairly steep instead of a steady increase in voltage, which gives some indication when the cell if getting full. There was no temperature increase in the cell as far as I could tell just touching it by hand.
I have a bit of history on this particular cell as I've charged it previously and recorded the date and voltage:
1/7/14 ( d:m:yr)
0.26v
7/7/14
4.05v
17/8/14
3.89v
22/5/15
4.00v

Before the 1/7/14 the cell had been sitting unused for a long period in my possession, I have no idea if the cell had been
used before I acquired it. All I know is that it is very old.
Between 7/7/14 and 17/8/14 the voltage dropped from 4.05v down to 3.89v due to self discharge.
So I can work out the self discharge rate:
4.05-3.89v/ 1.3months = 0.18/1.3=0.14volts per month
To go from fully charged say 4.1volts to 0.26volts would take 4.1-3.86/ 0.14 = 3.84/0.14= 26 months, which means I've left it sitting for a bit over 2 years and its gone down to 0.26v. Its usage previous to that is not known.

I did a discharge test also on this cell :
( once again I had the issue of the screensaver timeout interferring with icharger, so I only have the second part of the discharge test, I will do another one shortly).



The discharge shows some interesting results, what was previously called a 3.7v nominal cell, could now be called a
2v nominal cell. The cell still holds charge, in total about 3ahr ( compared to when new it was 5 ahr).
The discharge graph was only possible using nicad discharge mode. It was 1amp and lvc set to 1.5v on icharger nicad discharge screen.
I will do some more discharge/charge tests to get better graphs. If time I will try some higher discharge rates.
The tests show a couple of things:
a lithium cobalt oxide cell that has gone to near zero volts may still be useful but its discharge and charge profiles will be altered significantly probably due to a high internal resistance ( perhaps due to sei layer build up and possibly due to pore closure of membrane, being in a metal jacket any swelling of thickness of sei layer on the graphite layer increased pressure
on the porous membrane, which may close the pores).
I doubt the cell will be of any use in an ebike, but considering it can deliver a constant current of 1amp at around 2volts for
a long period, it might still be useful for led lighting or other projects.

 

[dnmun]

i have read that the cathodic matrix shrinks in volume, some say collapses, as the cell voltage drops below the 3V level. this is what i have assumed is the weakness with over discharge of the cobalt oxide chemistry.

i asked the RVD guy to test his pouch that had drained down to .7V by charging it back up and then measuring the self discharge rate but he never did. this is the only piece of data we have on cobalt oxide recovery. too bad we got nothing from his pouch to learn from. it would be interesting to see how fast this can will drain down and if the internal resistance drops as it is cycled several times. it may be dependent on the current during the internal resistance test too.

 

[whatever]

I've had another issue with logview:
if I minimise the logview program while running it starts a new graph when maximised again, solution is dont minimise the screen during use of logview.
The next charge cycle graph below, it looks like the capacity has dropped possibly by up to 1 ahr, but its hard to tell at what point to stop the charge stage ( the graph gets slowly steeper but at what point indicates a full charge I have no idea).
I'm thinking of letting it charge until some heat appears in the cell but its possibly dangerous since cobalts can go into thermal runaway and a nice fireworks show if that happens. I will take internal resistance measurements as I cycle the cell ( forgot to get the IR on last charge/discharge)



the graph above doesn't indicate the start voltage of the cell, it seems since the voltage shoots up at the beginning, the logview
program is missing the actual start voltage point in the graph, it may have been around 1.5v ( guesstimate), I will measure cell voltage before starting tests in future to get more accurate data.

 

[whatever]

next discharge graph:
( note its in two parts due to issue of a new graph page starting when I minimised logview during discharge, interestingly
logview continued to log the accumulating capacity instead of starting from zero again)



Its look like the capacity has dropped considerably down to about 2.1ahr , the nominal voltage of the cell has also dropped down to about 1.6v. At this stage it looks like the cell is quite quickly dying during each charge/discharge cycle, so it looks like
it wont be of much use at 1amp charge/discharge each cycle.
I have plenty of these cells to experiment with, now that i'm getting used to using logview and its idiosyncrasies, I'll be able to give better info as time goes by.

 

[whatever]

next charge cycle ( unfortunately I had to stop the charge and restart and lost the majority of the charge data so I only have the final part of the charge cycle)
I stopped charge at 5.75volts, after a few mins after charge cell volts settled at 4.1volts, IR= 1415mohm ( according to icharger)



no idea of actual capacity stored during this charge as I lost most of the charge data on this one)
One note:
the metal jacket has some minor swelling on it, so internal pressure will be high, some safety concerns at this point of internal shorting occuring.

 

[whatever]

next discharge:
start volts 4.06v ir before=1415mohmn ir after = 1410mohmn
rest volts after test = 3.5v



so capacity has dropped.
I read on the net somewhere a suggestion that puting lithiums in freezer can improve them so
I've put the cell in the freezer to see if it makes any difference for charge/discharge.
Not sure if I should try charge the cell while its in the freezer or after its back to room temp?

 

[whatever]

Just did a quick search for info on lithium battery capacity recovery.
Came across a few interesting things.
There are a couple of articles and a patent that basically remove the lithium salts from the sei layer, which apparently contribute to loss of capacity.
This paper:
http://lithiumbatteryresearch.com/pdf/SEI1.pdf
Has two good links in it, it says:

Recently, Abraham et al. found that when a degraded negative
cell was disassembled and the negative electrode was rinsed in the
DMC ( dimethyl carbonate) the cell capacity recovered to a point comparable to the ca-
pacity measured during the cell’s initial cycles.

(D. P. Abraham, J. L. Knuth, D. W. Dees, I. Bloom, and J. P. Christophersen,
J.Power Sources,170, 465 2007)


Likewise, Brous-sely et al. reported that negative electrodes dismantled after more
than 1 year of storage under a charged state at 60°C, but once rinsed
with water, demonstrated an ability to cycle lithium at the same
initial specific capacity and polarization.

M. Broussely, Ph. Biensan, F. Bonhomme, Ph. Blanchard, S. Herreyre, K. Nechev,
and R. J. Staniewicz,
J. Power Sources,97–98,13 2001

The second link is particularly interesting, as the negative electrode was simply washed with water to remove the sei layer
contaminants.
There is also a patent ( didn't save the link sorry) which also uses carbonates to dissolve the lithiums salts from the sei layer.
So it looks like it is possible to rejuvenate lithium cells, but requires the cell to be opened up ( hazardous and not very practical especially if the graphite ( and sei layer) are separating from the anode sheet).
The patent used a rinsing method to rejuvenate the cell, so not necessary to open up the cell, but probably beyond most backyard diy folks. It also involved puting fresh electrolyte into the cell.

 

[whatever]

this is the patent on rejuvenating lithium cells:
http://www.google.com/patents/US20100068605

 

[whatever]

a bit more searching showed this method to recover some lost capacity ( very interesting!):

The cathode was then discharged against an external lithium electrode to increase the amount of active lithium within the cell. About half of the lost capacity was recovered, and the cell cycled for 1500 more cycles. Active lithium replenishment from a reserve electrode may be an effective method of extending the life of lithium ion batteries.
.

from this article
http://www.researchgate.net/publica...mploying_carbon_and_iron_phosphate_electrodes

I'm not sure but are they saying you can discharge the cell to a strip of lithium that is external to the cell? and that lithium ions will travel inside the cell?
If that is correct, its an extremely simple method to regain some capacity ( no idea of lithium metal is expensive or even available though).

 

[whatever]

and here we have a cheap and readily available source of lithium metal!
http://www.instructables.com/id/Get-Lithium-Metal-from-an-Energizer-Battery-1/

from an Energizer Ultimate Lithium battery
this could be interesting. Since its lithium ions that are required inside the cell, is it possible that lithiums ions will travel through the cathode into the cell, they would be extremely small, I'm guessing its possible, worth a try anyway.

 

[nutnspecial]

Thanks for logging and doing this thread. I (as well as the general user of lithium tech) have a lot of room to grow in understanding the tech.
In the spirit of productivity I can add I store my lipo in the fridge . Maybe I wouldn't 'make a fire' in there, but I can potentially sacrifice the food and fridge for the potential building safety and potential longevity for stored batts.

THANKS!

 

[whatever]

i'd heard storing in fridge was good idea, wish I had done that with these cells.
I would never have thought that metal ions can travel through wires. I always thought it was only electrons travelled through wires, but it seems ions can travel through wires also. Its called electromigration, the electrons can push ions along wires.
Lithium is a cation ( positive), so there might just be some slim possibility you can put lithium ions back into a cell from external source of lithium.
I think its unlikely it would work, hope i'm wrong.
heres a good article on electromigration
http://www.doitpoms.ac.uk/tlplib/electromigration/printall.php

 

[whatever]

a bit more searching on the article that used external lithium source, it was toooo good to be true,
it seems it is a third electrode within the cell made of lithium, when externally shorted to the cathode releases lithium
into the electrolyte ( thats my take on it anyway).
they say:

Our results further illustrate novel designs for lithium ion batteries where a third electrode is built in to act as an active lithium reserve, which will provide active lithium on demand.

]\
from this page
http://www.sciencedirect.com/science/article/pii/S0378775311005155

Have done a charge up of the cobalt oxide lithium cell whilst in the freezer, will post results when the discharge is finished.

 

[whatever]

Did a freezer test, charge/discharge were done whilst cell was in freezer, no increase in capacity, but a significant decrease in capacity.
some interesting graphs though:
charge graphs ( voltage sat at 6v but using nicad charge manual 1amp it would cut off), so switched to life 2s charge which allowed continuous charge), the two graphs are the one charge.



Before charge ( after 12hrs in freezer) volts was 3.65v and ir = 1498mohmn
After charge volts was 4.08v and ir = 1442mohmn
I stopped the charge process at about 5ahr capacity was reached, it seems the cell was probably losing the charge not storing it,
losing it as heat probably but staying cool due to freezer temperature.
Discharge graph showed capacity had dropped to 0.75ahr and nom voltage around 1volt.

 

[dnmun]

charging when cold causes the lithium to plate out away from the electrode. it is then lost to the charge/discharge cycle and capacity is reduced by that amount.

 

[whatever]

yup may have been better to freeze and get back to room temp before charge/discharge.
I found more details on the lithium replacement process, it seems they do use external lithium metal, its not inside the cell,
but its in an electrolyte bath, and the cell sits opened at one end in the electrolyte bath that contains the piece of lithium metal.
There is diagram and a bit more info of how they did it here:http://ma.ecsdl.org/content/MA2010-03/1/812.full.pdf
So it appears they are using diffusion to replemish the lithium into the cell.
Unfortunately it doesn't say on that link what the electrolyte is that they use. That would be interesting to find out.

 

[whatever]

here is someone selling Lithium in solvent ( LiPF6)
(http://www.alibaba.com/product-detail/Electrolyte-LiPF6-for-Lithium-ion-battery_1000039293.html

 

[whatever]

I'm interested in comparing the IR values that I got from the tests above to same type of cells that are in good condition.
So using the icharger I got some values for two packs I'm currently using on an ebike that works very well ( 14s)i.e. cells in good condition.
The IR values for the good cells ranged from 1449mohm up to 1605mohmn. So it would seem that the cell with low voltage
and lost capacity gives an IR very similar to good cells............that is a bit baffling.
I measured the resistance of the wires I was using from the icharger to the cells ( very thin with aligator clips) and it is about 1omhn. So it would mean the values i'm getting from the icharger will be quite alot higher than actual IR.
Also I've heard before on this forum that icharger IR measurements may not be particularly accurate.
It is though quite puzzling that good cells ( according to icharger) are giving roughly same IR values as the bad cell I tested.

 

[whatever]

In one of the discharge tests above I noted the start voltage was 4.06v it immediately dropped down to about 1.7v when the 1amp discharge was started.
Using formula for IR on this page
http://physicsnet.co.uk/a-level-phy.../electromotive-force-and-internal-resistance/
I get an IR of (4.06-1.7)/1= 2.36 ohmn , much higher than the 1415milliohmn given by icharger.
I should do a test with 2 different currents and see what value that gets with alternate formula for IR.

 

[whatever]

I just did a two current test on the bad cell ( graphs above) and a good cell:
using formula ( 4) on this page:
http://www.hk-phy.org/energy/commercial/act_int_resist_e.html
bad cell: ir= 1.95ohmn
good cell: ir = 1.36ohmn
I would have expected something in the milliohmn range??
heres the data I got:
bad cell: no load volts 4.06v
v1=3.76v i1=0.1amp
v2=2v i2=1amp
good cell: no load volts 4.1v
v1=3.8v i1=o.1amp
v2=2.57v i2 = 1amp

am i doing something wrong in my calculations?

 

[whatever]

when i use the data above in the formula on this page
http://physicsnet.co.uk/a-level-phy.../electromotive-force-and-internal-resistance/
which used the no load voltage in the calcs I get different results again:
bad cell:
0.1amp , r= 3ohm
1amp, r= 2.06ohm
good cell:
0.1amp, r= 3ohm
1amp, r= 1.53ohmn
Still getting nothing in the milliohmn range for good cells, seems I get different results depending on which formula is used also.

 

[okashira]

I have two Tesla Model S cells that I shorted for about three weeks with 100ohm. They got down to about 0.01V or less before I removed the load.
I removed them, and they, after a few days, recovered to 0.25V each. They have been sitting at 0.25V each for another month.
I can try two different regimens to bring them back to life and compare performance. Suggestions accepted...

@whatever - Those cells are toast. Even the oldest laptop 18650 should be better then 0.2 ohm. Modern NCA cells are 0.018 - 0.050 ohm.

 

[whatever]

I think theres something wrong with my calculations, since the good cells I'm using have no problems on ebike, yup bad cell is very toasted, I would think very low current charge is good idea for bad cell recovery, seems to work well on nimh.

 

[whatever]

Doing another test, same cells ( lithium cobalt oxide), but this time theres three in parallel ( not sure what capacity these cells have, when new they were supposed to 5ahr, since this pack has three in parallel it should be 15ahr when new).
I measured the IR with icharger = 1274mohmn. The cells show no signs of swelling.

But.....the leads from icharger to cells has resistance of 1 ohmn.
( I think this explains the data I posted above, since I was using the icharger voltage, it will be influenced by the high resistance of the very thin leads i'm using, so the voltage its showing me will not be accurate due to the influence of the leads resistance).
Solution: get new thick leads and good aligator clips ( will do asap)
( that will also allow me to do higher amp discharges).
So taking into account the 1ohmn lead resistance ( using standard resistance formula) 1/rt= 1/r1+1/r2
it gives an IR value of 4.64mohmn ( much better result).
One question if anyone actually happens to read all this!
I measured the resistance of the leads of the multimeter I'm using ( just by shorting the leads together) which I used to measure the resistance of my leads from icharger to cells, the multimeter leads shorted together gave 0.7omhn, should I also take that into account when measuring resistance of my icharger leads?
I'm guessing I should?
If I do that I get 0.4975ohmn resistance of the leads.
Plugging into the resistance formula gives an IR of 357.78mohmn
That would seem to be a reasonable figure. So I'll assume the IR of the pack i'm currently discharging is 360mohmn approx.

 

[whatever]

I'm thinking very short low resistance wires with strong alligator clips will be necessary to get accurate data in future from the icharger.
Lesson learned!