It's time... The Special Hell of PbO2 Anodes
Entry posted by Swede
925 views
Do you like the new board better? Worse?
65 members have voted
With the huge harvest from T-Cell Junior (over 4 kg dry potassium chlorate) I now have about 15 pounds of this oxidizer, and I am ready to move on to the Lead dioxide experiments that really got me interested in the (per)chlorate process in the first place. There is a particular U.S. patent (4038170) that goes into tremendous detail on plating this substance, which is an ideal material for the electrochemical oxidation of chlorate to perchlorate, with one glaring problem... it doesn't stick worth a damn unless the plating bath is perfect, the substrate is perfect, and the additives are perfect. Failure usually consists of either flaking of the PbO2, or creep of the electrolyte BENEATH the PbO2 layer and subsequent passivation of the substrate. Hundreds, if not thousands of guys have attempted it, and the true successes, unless they are hidden away, cackling over hoardes of perchlorate, can be counted on one hand. It CAN be done, and I am determined to do so.
The first step is in gathering reagents. Most, if not all, are powdered death, especially the large quantity of lead nitrate that must be used. I am not going to be a safety-nag... just know what you are getting into. For a plating bath, at a bare minimum, you will need nitric acid (not a lot), perhaps a kilogram of lead nitrate, lead oxide (PbO; Litharge) or Pb sheet, copper sheet, copper nitrate, and a nonionic surfactant. Bonus chemicals include Bismuth salts, and possibly nickel nitrate.
These chems were ordered. While I waited, I finished construction of something I had started quite a while back: a spinning anode rig. Some digression... almost any plating bath develops gasses, usually hydrogen or oxygen via the standard, sixth grade electrolysis of water lab demo. The bubbles tend to clump on the substrate that you want to plate, creating voids, pinholes, craters, all sorts of nasty features. We don't want that... we want the plated surface to be smooth and even. There are several methods to control bubble sticking. The surfactant helps, as does some form of agitation - vibration, spinning, what have you. Back when I was nickel plating, a spinning rig produced beautiful decorative nickel finishes, so I decided to go with a spin-rig.
The setup consists of an extruded aluminum column, bolted to a hardwood base. The column itself (a Bosch product) has T-slots running the entire length on all sides, and it makes the setup very modular... I can reposition the boom arm, for example. A gearmotor is mounted on an aluminum angle (note the speed control also mounted on the column) and attached with a timing belt and pulleys to a 1/2" copper pipe, which acts as the anode carrier. One of the tricker things about such a setup is maintaining continuity through the rotating pipe. The best way to do this is with a carbon brush, such as you'd find in a DC motor. After a bit of mill work, and I had something that would work well:
With the spin-rig complete, I was ready to tackle task number two - setting up, and heating, a plating bath. The temperature required is typically between 60 and 80 degrees, depending upon the reference source. Many plating attempts have been made using borosilicate glass vessels over a hot plate with perfect function, but early in the project, I decided I wanted to use a larger vat, preferably plastic.
A good source of tanks in an infinite number of sizes: The Tank Depot. Of the inexpensive plastic tanks, the best is polypropylene, rather than polyethylene, due to the former's higher heat tolerance. The vat I ordered (the smallest offered) will be roughly 8 liters when completely full.
Since plastic vats and hot plates don't mix, I needed to come up with an immersion heater of some sort. These come in a nearly infinite variety, and most are quite expensive. Be prepared for sticker shock for a true lab immersion heater! Remember also that a lead nitrate plating bath uses (and creates) nitric acid, so whatever heater you select must fulfill three requirements:
- It must have a wattage adequate for the volume of electrolyte
- It must be resistant to nitric acid, and not shed Fe+ ions
- It must be controllable
Stock fish heaters need not apply, as the thermostat is set too low. Can't be cooking those cichlids at 70 degrees C! It is possible to modify one, but I think I've come up with a better answer... the industrial cartridge heater.
Cartridge heaters are heating elements encased in either stainless steel, or incoloy, a space-aged nickel alloy capable of shrugging off extremes of temperature. Inconel and its family of alloys finds extensive use in jet engines at the turbine core, where temps are inconceivable. Cartridge heaters are normally used to heat molds and platens; they are inserted into bored holes in aluminum and steel molds, and they come in a wide variety of wattages and voltages.
Cartridge heaters come in two flavors, low watt density, and high watt density. The low watt density types are usually sheathed in 304SS, not good for this process due to Fe+ ions. The high watt density types have the incoloy sheath, and come in a huge variety of sizes, voltages, and wattages. Best of all, they are remarkably inexpensive, maybe $30 for a 500 watt heater, 120V. Incoloy is a metal acknowledged as compatible with nitric acid processes. The ends of the sheathe away from the wire appeared to be well-sealed, either via resistance welding, or possibly spin (friction) welding.
I bought one from MSC Direct.
Despite the sealed end, I doubted the water-tightness of this heater. I took Tentacles excellent suggestion and created a simple Cu tube, capped at one end. Drop the heater into the tube, and heat away! It is best if the heater makes contact with the walls of the Cu tube, evenly. I found a pair of telescoping brass tubes that, when combined with the heater and the Cu pipe, made for a perfect fit. other options would be to fill the tube with some sort of heat transfer media, perhaps a silicone oil? Mineral oil? Glycol? Or best, MgO powder, but it would be hard to use the latter to evenly surround the heater. Perhaps a bit of air space around the heater won't hurt things too badly.
Wattage density is important. You do not want the watt density to be too high. But the LOW watt density models have the less desireable 304SS sheath, not the incoloy. Solution: Order 240V high-density heaters, that come with an incoloy sheath, and run them on 120V. Watt density will be cut by a factor of four, so a 1 kilowatt 240V heater will be a 250 watt heater at 120V. Experiments have shown that a 250 watt heater used in this manner will raise a 4 liter water bath from 17C to 60C in less than an hour.
Parameters 1 and 2 of the heater list has been met. Now, how to control it? The simplest option is a lamp dimmer, or variac. I decided I wanted something a bit fancier, and elected to gather components for a killer temperature controller unit that can be used both for this project, and for any future project requiring precision heating of a large liquid bath, a furnace, anything that requires precise heat.
The heart and soul of a temperature controller is the microprocessor controller itself:
These have become freaky cheap due to an influx of Chinese-made controller units... and they're good! They usually come in some fraction of a DIN size, such as 1/4, 1/8, etc. I recommend a 1/4 DIN unit because it is a bit bigger and easier to wire. If you search ebay for temperature controller, you wil get hits for dozens of varieties, NIB, at amazing prices, from $35 on up. Things to look for:
- You want a controler than can accept a wide variety of temperature probes, usually thermocouples and PT-100 RTD's
- The controller should have various alarms, and a big, bright display; most do.
- Most have a variety of outputs. You DEFINITELY want one which will drive a SSR (Solid State Relay). More on that in a bit.
I bought this one (pictured above) from Omega. While I was at it, I ordered a PTFE-encased PT-100 RTD probe, and a type-K thermocouple, also PTFE encased. The controller can make use of either one. One last critical batch of items... a solid state relay that can handle 50 amps, and a heat sink for the relay.
SSR - Without going into too much theory, think of it as a relay with no moving parts. The controller outputs a DC voltage (5 to 15VDC) that "closes" the SSR, and the load end handles all of the high current of the heaters. The controller itself never sees more than a few hundred milliamps.
Here is a 50 amp SSR along with a heat sink, and the PTFE-coated PT-100 RTD temperature probe. If your temps are all below 250C or so, an RTD wil deliver better accuracy than a thermocouple. if youneed high temps, a thermocouple is the way to go.
It was time to put it all together. First thought - bolt the components down to a piece if plywood, and be done with it. It would have taken two hours, and it would have functioned perfectly. But after a bit of thought, I decided to make it a bit nicer, so I used a spare Wegeman enclosure, and began to carve it out. The cover was cut for the controller box, a power switch (which switches ONLY the controller, NOT the heater elements), and a red neon light which energizes when the SSR is ON and providing power to the heaters. Don't skip this simple but very helpful addition if you replicate such a project.
The SSR was bolted to the heat sink, which in turn was bolted to the enclosure cover. Heavy Cu lines would eventually exit the enclosure, and I decided to use a terminal strip for both outputs and inputs. The strip was mounted on an aluminum angle, along with a thermocouple socket, and tentatively positioned on the enclosure. I took note of where the cables would exit from the enclosure to mate with the terminal strip, and decided to cut a slot in the enclosure, rather than a series of holes. This is the result.
The strip, or something similar, is important. You will definitely want to swap heater elements and temperature probes, and the terminal strip makes this a snap. Otherwise, you'd have to open the enclosure and tear everything down each time you wanted to do a heater swap.
One large fuse block for the primary SSR output was mounted, and one smaller one for the control box. I added a 30 amp fuse to the SSR block, and a 1 amp protects the controller. Everything in the box is "downstream" of the 30 amp unit, so if it blows, the whole thing goes dead. 30 amps is overkill... I anticipate, at the most, 1 kilowatt being controlled, and that would equate to less than 15 amps.
I began to populate the lid and the enclosure itself.
Wiring internally was heavy where needed, lighter where indicated, and everything went together fine.
How it all works, nutshell variant: The temperature probe tells the controller what the process temp actually is. The controller then uses PID (Proportional Integral Derivative) control to energize the heater elements so as to bring the bath to the desired setpoint. As the temperature aproaches the desired value, the controller begins to cycle the SSR output, and it is a pretty fascinating thing to watch the controller gently ease the system to the correct temperature, without major overshoots. If given a choice, choose a PID controller rather than a simple ON/OFF controller... PID does the job much smoother, with better accuracy.
Programming it can be a chore (Chinglish! "Set AdOH Alarm so to meet PV - PHyT, happy result") But once set, these things have an intelligence; they can literally learn the best way to control the process bath.
Note the red "heater on" light upper left - Again, I consider this to be essential. It lets you know exactly what is going on, and it is a fascinating thing to watch this light silently blink on and off as the system learns what it needs to do to control the bath.
The next blog will be a challenging test of the system. I will say this much - As a whole, it turned out very well, and I am proud of this particular effort; well worth it. It should serve me for many, many years. A homemade system like this, with eBay materials (NIB controller and SSR, used heater, used probe) can be put together for $80, easy. My total cost was more like $160. Now, go price a completed unit with this capability from Cole Palmer or Spectrum... 4 figures, most likely.
6 Comments
Recommended Comments