Sciencemadness Discussion Board

The Bucket Cell Adapter

Swede - 15-10-2009 at 14:43

Some months ago, while flailing away on a very complicated fabricated PVC tank for the (per)chlorate process, I was grumbling to Tentacles about the difficulty in making a reasonably resistyant cell, and we came up with the idea of a bucket cell adapter. Basically, this thing clamps on the lid of a standard HDPE bucket, which, while not 100% proof against electrolyte, should last long enough to make it economical and viable. All you have to do is set a template on a flat bucket lid, drill some holes, and cut a rectangular opening.

This rig clamps on any reasonably flat HDPE bucket lid, and allows two basic electrodes to be mounted. It has four ports, any of them usable for any of the normal tasks; vent, sample, acid, bubbler, temp, etc.

It's first use will be to exercise my data-collection scheme, charting and comparing voltage, current, and temperature throughout a run; the next will be the long-delayed test of the lead dioxide anode to create perchlorate.






I made a blog entry on APC, which shows more detail if anyone is interested:
The Bucket Cell Adapter blog

I think such a device makes sense, rather than expensive, dedicated cells of PVC or some other resistant plastic. Using a bucket cell adapter, I can swap electrodes and cells in minutes, and move from 2 gallons to nearly any size, for almost any electrolytic process compatible with these materials; PVC, Viton, PVDF, and titanium. :P

12AX7 - 15-10-2009 at 16:01

I presume your peripheral bolts are gasketed on the inside?

dann2 - 15-10-2009 at 16:11

Hello,

One piece of grumbling I will do in regard to your BCA is that there is only two slots, one Cathode, one Anode. It would be better to have two Cathodes (small ones) so that current is more evenly distributed on the Anode. With a flat Anode the current will be inclined to be low on the side away from the single Cathode.

How do you seal the lid to the bucket? Does it simple sit on top of it. There is no circle of clamps/bolts or whatever.

Perspex (plexiglass, PMMA) is nice stuff to work with and you can see into the cell too. It is easy enough to pick up in half inch sheets or so. Thats what I used on my glass cell. It consisted of a demi-johns (those glass bottles with two ring handles on top used alot for making wine etc at home) with the top cut off and then inverted and filed dead flat on a sheet of glass with sand. The perspex sheeting (very flat stuff) sat on top of this with little or no leakage when a thin seal of Silicon sealer was placed on top of the glass rim. The bottle is easy enough to cut with a glass cutter and then a hot wire. It cracks around the circumference when a score is made with the glass cutter.



Dann2

Swede - 16-10-2009 at 08:08

The bolts exposed inside are titanium; the only gasketing is the O-ring fitted into the rectangular slot on the upper piece. dann2, I agree that 2 small cathodes would be best, but I did want to make this thing adaptable to 2.5 gallon bucket lids, and it is amazing how much room is rapidly consumed with the 4 ports, the bolts, the O-ring, and the electrode straps. The threading in the PVC is relatively tight, and I don't expect significant creep or gassing through the bolts.

This thing is conceptual. I wanted a way to set up a small cell to test things like LD anodes, rather than fill up my larger 25 liter cell, and I did want to avoid the pitfalls of a flimsy food container.

The sealing of lid to bucket relies 100% on the bucket itself. Some of the better 5 gallon HDPE buckets have a decent gasket that should do OK. I expect the lifespan of the bucket to not be particularly long, but it should be good for several runs.

Something like this could also be fitted to the lid of a PVC tank from Tamco or similar... pretty much anything that is flat, and cuttable, could be a candidate.

watson.fawkes - 16-10-2009 at 08:45

The overall idea is a good one even for a dedicated, permanent tank. What you're doing is isolating all your sealing problems into one place and converting them into a single issue of gasket conformance. This seems better than dealing with a variety of different seal geometries and materials, each of which can fail in different ways.

In a future version, if you were to put the bolt ring just to the inside of the gasket, you could use blind holes in the lower plate and eliminate the need for titanium bolts. You'd need a larger gasket ring, and likely a thicker lower plate, but I'm sure that's cheaper than titanium. I recommend using machine-thread inserts, which can be melted in hot rather than either forced or screwed in.

In this configuration, the upper plate can be turned into an upper clamping ring. If such a clamping ring is made with a small lip on the inner cut-out, protruding downward, the lip would act as a fulcrum for the clamping screw, with the load being the gasket.

Swede - 20-10-2009 at 13:15

Thank you, WF. I've thought of various ways to do this, with (as you noted) a superior version being one not requiring Ti hardware. Blind holes would do the trick.

The overall concept is simple - given a flat surface, by making a cut, one can clamp this device into place and be ready to go in moments. The effort and quality goes into the adapter rather than a tank or container.

Swapping O-ring and bolt hole locations (bolts inside O-ring) would allow a simple rectangular cut. This version does require drilling separate holes for the bolts. I'll call this V 0.9 beta. Lots of room for improvement.

The viton o-rings can be ordered in just about any size (or shape) from these guys among others. Flat gasketing will work too but viton with a 0.23 or 6mm cross section is a good size and will seal irregular surfaces better.

patsroom - 28-11-2009 at 20:09

[Thank you, WF. I've thought of various ways to do this, with (as you noted) a superior version being one not requiring Ti hardware. Blind holes would do the trick.]

One other thought that I have would be to use PVC pipe caps placed over the bolt hole with enough length to allow the bolt to go through the bottom plate. Being that the material for the bottom plate is made of PVC, it would easily make a good seal once glued into place.

Swede - 29-11-2009 at 07:28

The Bucket Cell Adapter - preliminary report.

The run started with pure KCl (so as to collect valuable data) rather than used/recharged electrolyte. The dry yield was 2 fat bags of KClO3, both over 1 kilogram, out of an 8.7 liter cell; barely more than 2 gallons.

For those into efficiencies, I have calculated the following numbers that may be helpful with their own systems. This particular run yielded probably the greatest amount of chlorate per liter electrolyte of any chlorate run that I have eve done, mainly because I dissolved (and started) the KCl at an elevated temperature, yielding a higher % chloride ion than all previous runs.

Cell volume:
8.7 liters

KCl:
375 grams per liter, total 3262.5 grams

Avg. Temp:
62 degrees C.

Chloride:
Starting Chloride (Molar-basis): 1552.6 grams
Starting Chloride % (Molar basis): 17.84%
Starting Chloride % (Hach Titration strips): 17.9%
Ending Chloride % (Hach): 5.5%


This shows that the Hach strips are accurate. Given a known amount of chloride in a given volume, there should have been 17.84% chloride in the electrolyte, and the Hach strips confirmed this beautifully, despite the need to dilute 50:1 with care.


Ampere-Hours consumed:
Exactly 4,300

Mass Yield:
Dry Yield KClO3: 2120 grams
Wet Yield KClO3: 870 grams
Total Yield KClO3: 2990 grams, +/- 87 grams

The wet yield is based upon the solubility of KClO3 in electrolyte at 30 degrees C, which is 100 grams per liter. The actual wet yield is impossible to determine, and is only an estimate. It will probably vary by 10%.

Current Efficiency (CE): KCl --> KClO3
CE = 131.22 * grams/Ah
CE = 131.22*2990 / 4300
CE = 91%

That is a WONDERFUL efficiency for a home cell. Frankly I am a bit shocked, but decent pH control, higher temps, and a good anode is going to produce such efficiencies. If I assume that the wet yield is not 870 grams (likely because the remnants are a mix of chlorate with other species, and the amount of KClO3 in solution will not reach theoretical - assume that the wet yield is minus 20%, 696 grams vs 870, the CE would be:

CE = 131.22*2816 / 4300 = 86%

The probable efficiency was between 86% and 91%, still excellent, and I am very pleased. Without pH control, the best that can truly be expected is 66%. A modern chlorate plant with all the bells and whistles will panic if the efficiency drops below 95%, because their product will not be price-competitive due to power costs.

Further notes: The bucket cell hardware performed well. Only a moderate salt creep between the PVC sandwich and the lid. The Titanium fasteners came through untouched.

Yield of chlorate per liter of electrolyte...
Dry: 243 g/l
Dry + Wet: 344 g/l

pH:
pH varied from 6.37 to 7.27, average of 6.6 over the run, based upon manual measurements. Definitely a best estimation but probably pretty close.
HCl consumed: 550 ml of 16% (half-strength) HCl; approximate mass of 88 grams of pure HCl molecule.
HCl consumed per ampere-hour: 0.0205 grams per aH



Probably the most interesting observation (aside from the electric chart-recorder data, to come later after analysis) is the HCl consumed per amp-hour. The classic rule of thumb is as follows:

Published, recommended HCl consumption:
0.057 ml of concentrated (32%) HCl per ampere, per hour
0.114 ml of diluted (16%) HCl per ampere, per hour


Actual HCl Consumption:
16% HCl consumed per ampere-hour: 0.128 ml per aH


Official: 0.114 ml 16% HCl per aH
Experimental: 0.128 ml 16% HCl per aH

Additionally, as the cell chemistry matures, there is some speculation that a buffering action takes place, meaning that less HCl is needed below about 12% chloride, or when KClO3 crytstals begin to form.

Everything went perfectly with the exception of the HCl, too much acid, too soon, resulting in a pH below 6.7. The data stream from this run will be amalyzed and published soon.

Swede - 29-11-2009 at 07:43

Quote: Originally posted by patsroom  


One other thought that I have would be to use PVC pipe caps placed over the bolt hole with enough length to allow the bolt to go through the bottom plate. Being that the material for the bottom plate is made of PVC, it would easily make a good seal once glued into place.


That is one of the joys of using PVC as a cell material. It is decently impervious to the side-effects of the electrolysis, AND it can be cemented with ease, using common hardware store materials. If limited to stainless steel hardware, as you mentioned, small CPVC caps or even disks can be used to cover through holes permanently. But from what I've seen with titanium so far is that without any sort of charge, it shrugs off the nasty cell effects with ease and comes out corrosion-free. The main problem is the expense... decent Ti fasteners in the 6mm range will cost probably $1 each surplus, and new, forget about it, probably $10 apiece.

eBay has been a good source of Ti fasteners. A lot of them have a nylon patch on the threads to lock the fastener, and this can be dissolved off (if desired) with either nitric acid or an organic solvent. I've had a fastener in 70% nitric for a week with no ill effect at all. But the nylon is gone! :D

patsroom - 29-11-2009 at 12:38

Speaking of PVC as a cell material that is used, does PVC stand up very well against HCl at 32% or would it begin to dissolve into the solution.

Swede - 30-11-2009 at 08:00

Here's a good Chemical resistance database.

It shows that CPVC is rated excellent for compatibility with 37% HCl. CPVC and PVC behave closely enough so that you can use PVC without worry with HCl. It is usually the dosing pump or the tubing that may exhibit incompatibilities.

I have been using Tygon tubing (I think R-4603, I can find the specific type of Tygon) with good success with HCl.

patsroom - 30-11-2009 at 10:03

Swede,
Thank you so much for this great site. I have been hunting for one like this for awhile.
Pat

Contrabasso - 30-11-2009 at 12:15

Could you use one of the parameters that you have available to control a bigger dosing pump to add more KCl solution displacing some used electrolyte for harvesting and return?

dann2 - 30-11-2009 at 19:09

Hello,

Great work Swede.
I presume there were no additives in the cell, a green cell!

Why did you decide to stop the cell. Obviously you had to stop it at some stage. Did you reckon that CE was getting too low?
It would be interesting to see how MMO CE goes as the Chloride conc. drops and drops. When using MMO in a Sodium Chlorate cell that is going to go onwards as a Perchlorate cell (by changing Anode) it would be useful to know at what stage it is sensible to put in the Perchlorate Anode.
I will get up off my backside and set up a cell soon!
@Contrabasso. You can simply weigh the daily extracted Chlorate (lifted out with a spoon) from the cell and put in the equivalent molar quantity of Chloride. You will have to extract Chlorate each day and weigh to do this, but if running the system continously you are going to have to extract Chlorate anyways to stop it building up too much.

Since Chlorine gas is sometimes used in industry to controll pH in Chlorate cells, I was wondering if it would be sensible to set up a cell making Chlorine and bubble it into cell. The Anode could be graphite in both cases (not much erosion in either cell). Sounds daft. May suit those with no access to MMO and cannot get easy access to HCl.

Dann2

Dann2

Swede - 1-12-2009 at 07:34

TY dann2, I wanted to start clean with KCl so that all chlorate produced could be attributed to the amp-hours in this given run... no "free" chlorate from a previous run. And correct, there were no additives at all.

The run was halted not due to any physical observation. I wanted to get a feel for the efficiency of the system in the heart of the envelope, not at the extremes. The CE would drop dramatically at a lower chloride concentration. From an earlier run with the 2-celled system, I had seen the results of running MMO down to zero chloride at very high temperatures... remember my so-called "T-Cell?" It was the system with a small electrode chamber and a circulating pump, the concept being to take hot, processed electrolyte and route it to a cooling chamber for harvest. The pump system clogged, and the electrode chamber soared to 95+ C and ran for hours, taking what electrolyte was in there down to zero chloride. The result was the warped cathodes, a layer of chlorate, and an anode that appeared to be untouched.

Before:


After:


The "10% chloride" rule does not apply to MMO. It continues to function well below 10%, but where to stop... I don't know. For potassium chlorate harvest, I would suggest 5%, but it can probably go lower.

For sodium, while the cathode warping is unsightly, I don't think it truly harms anything, and I think in my case it might have been due to heat.

The voltage (at constant temperature and current) was definitely climbing as chloride depleted; it might be possible to extrapolate the 5% down to zero, and at that point execute an anode swap. Or, if CE is of no concern, simply run it well past the point where the chloride will be at a minimum.

I don't know what that minimum is. I wish I had retained some of the electrolyte from the failed T-Cell... I could have titrated it accurately and seen just how low (chloride) the MMO can take a sample of electrolyte. 1%? 0.05%? I'm not sure, and since perc anodes typically don't like chloride, it'd be a good number to know.

I am off work today, and I'll try to play with the digital chart recorder software and try to come up with some presentable images.