Sciencemadness Discussion Board

Sodium metal - Illustrated Practical Guide

len1 - 16-1-2008 at 16:45

This article is part of a series of Illustrated Key Syntheses in Chemistry which I intend to post, with the purposes and guidelines outlined in the thread of the same name in the General Chemistry section.<br> I shall attampt to ensure that the first post in all syntheses contains all the relevant information. If something needs to be added later on, due to comments further down the thread for instance, I shall edit the first post.




Aim

It is the aim here to present an illustrated demonstration of a method for the amateur chemist to produce useful quantities of sodium metal for laboratory use. The method relies on electrolysis of molten NaOH, a process patented by Castner in 1891, which has been found to be the simplest method of sodium production for the amateur. The present method is a modification of the Castner process in both form and scale to enable construction by the amateur using only materials easily obtainable from hardware stores. It is nevertheless capable of producing up to several hundred grams per day. The present appartus is based on a simpler non-reusable cell presented on the German site .de which produced several grams sodium per day., here modified and increased in scale to make the equipment reusable, durable and allow easy replacement of consumable elements, such as the cathode and anode. One should note that many instances of alkali metal production using 'electrolysis in a can' have been previously presented. While these methods can be constructed in a matter of hours they are not capable of reliably producing quantities of more than fractions of a gram, the reason being that the electrolytic bath produces yields over a very narrow temperature range, 320-330C - for the sodium cell. Consequently in a simple cell initial production subsequently redisolves and there is no yield. The present equipment uses the minimum amount of sophistication necessary to achieve a consistent yield in a reusable cell.


Motivation


Sodium is an indispensible element in a serious amateur laboratory. Its primary use is in thoroughly drying organic solvents, such as Grignard reagents, in the formation of alkoxides used in the syntheses of ethers for instance, and in the synthesis of whole groups of inorganic compounds such as cyanides, azides, and metal amides. In the latter case, the disolution of metalic sodium in liquid ammonia produces an intense blue solution with electronic properties similar to metals. In industry sodium is used as a heat transfer agent in situations such as nuclear power plants due to its excelent thermal properties. Despite its plethora of uses, in the past there were no proven methods for the amateur to synthesise his own sodium except the subgram amounts available from electrolysis in a can. This presented a problem since in many locations sodium can not be purchased OTC, for reasons of liability and other considerations. Professional laboratories generally buy their sodium, and if it is required to be made, for reasons of purity for instance, decomposition of NaN3 at 300C is about the only laboratory preparation available. Since NaN3 is extremely poisnous and an explosive precursor, the availablity of sodium azide is generally poorer than that of sodium.

This leaves the amateur with industrial methods, many from the 19th century, for sodium production, methods which have generally not shown results due generally to the difficulty of adapting industrial processes to the laboratory. These include

  1. vacuum reduction of sodium salts with carbon at > 1000C, such as 4NaOH + 2C -> Na2CO3 + CO + 2H2 + 2Na
  2. vacuum reduction with metals such as lead, magnesium or zirconium at > 500C, such as 2NaCl + Pb -> PbCl2 + 2Na
  3. electrolysis of NaCl + CaCl2 at 600C (Na is soluble in the bath at the mp. of NaCl) with a consumable carbon anode
  4. variations of the above with mobile liquid cathodes, such as molten lead or mercury to deal with reactivity of Na at 600C
  5. electrolysis of NaOH (due to its low mp. of 318C) with iron cathode and S/S or nickel anode 2NaOH -> 2Na + H2 + O2 (some H2O also escapes)


The latter method due to its low temperature and evolution of hydrogen gas protective to the sodium, obviates the need for a vacuum and has been found suitable for modification to the amateur scale. The main problem which needs to be solved with this method is the extreme reactivity of molten sodium hydroxide, which degrades almost all materials except a few metals. The basis of this method, electrolysis, was the means used by Humphry Davy the discoverer of sodium who in 1807 electrolysed NaOH with platinum electrodes, but who was not able to produce more than subgram quantities. He in turn owes the discovery to Alexander Volta, who six years earlier invented the battery - the voltaic pile. This new force, was immediately applied by investigators all over Europe in an attempt to produce transformations in all kinds of materials, which previously defied change, and appeared elementary. One such material was lye. Davy used a large pile belonging to the Royal Society, and to make lye conduct dissolved it in water, but observed that water, rather than lye underwent transformation, and a new principal, of which there was a hint of formation at the cathode rapidly reacted with the water. As a result Davy undertook electrolysis of the molten salt - which he cleverly melted and electrolysed with the same current. For the next 80 years all sodium production was based on the high temperature methods above - which resulted in quick degradation of the apparatus, and expensive sodium. In 1891 the original method of Davy was adapted for industry by Hamilton Castner, in a process bearing his name. Castner observed that nickel was largely inert to the action of molten lye, and hence used this metal in th anode, where the positive potential in combination with the corrosive bath acted to dissolve most other materials. If one tries to conduct the electrolysis with an overhead cathode, the molten sodium adheres to it as it floats in the bath - and subsequent molten metal formed then attempts to form a bridge to the anode (see my experiments in the electrochemical sodium thread) due to the potential difference in solution making this the lowest energy configuration. If one attempts to eliminate this by using a non-conducting separator, the sodium will crawl under it, and commence forming in the anode compartment (resulting in explosions). Hence an underside cathode - not in contact with the floating accumulated sodium - is required, and this introduces isolation and hermiticity problems. Castner solved both these problems by using the bath material itself - a long underside stem encasing the cathode has a temperature gradient along it, such that the bath at the bottom is solidified. The Casner cell uses an iron gauze to separate the anode and cathode below the surface, this allows flow of ions, but not the sodium. Some sodium still manages to pass to the anode where it reacts with the oxygen formed there and produces small explosions. The standard Castner cell is a cylinder of about 60cm diameter and 80cm high, holding about 200kg of bath. Our cell has a diameter of 8cm and is 15cm high - a scaling down of about 250 - it holds 1.5kgs of bath, hence explosions are more serious for it - and I have found a procedure for almost eliminating them. The current density at the electrodes is almost the same as in the original.

The only chemical on which the present method relies to produce sodium is sodium hydroxide NaOH, which is freely available OTC. Even if difficulties arise in its purchase it can easily be manufactured from the electrolysis of salt NaCl, using a much simpler electrolysis cell than the present. It can also be easily made using such chemicals as Na2CO3 and Ca(OH)2 in the soda lime process. Since NaCl will always be available, the consequence of the present method is that it guarantees the same for sodium.

Note: A method described in a 1987 US patent 4725311 by metallgesellschaft aktiengesellschaft puportedly employing reduction with magnesium at 200C appeared simpler to execute in an amateur laboratory than any of the above. It has been tried in the unconventional sodium thread for the production of potassium following the procedure in example 1 of the patent closely. Despite repeating the experiment several times with different sources of magnesium absolutely no alkali metal was obtained, all observed indications of a reaction were consistent with the reaction of magnesium with water. This, combined with the obviously incorrect measurements of evolved H2 presented in the patent, suggest the results the patent reports are false.

Findings



  1. 30hrs operation produced 200gms of Na, figure 1, the nickel anode shows no signs of corrosion
  2. The cell operates at 47-49 amps, achieving a current yield of 23% - typical for small cells, compared to 50% max theoretical current yield
  3. Electrical energy cost about AU$7.90/kg Na produced.
  4. A simple purification procedure, which produces shiny pure sodium is presented.
  5. A method for starting and operating the cell is presented, which completely eliminates the small explosions normally accompanying Castner operation.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig1.JPG">
Theory


The construction of a typical Castner cell (not the scaled cell for amateur use presented here) us shown in figure 2.

The following are the major reactions occuring at the cathode

  1. Na+ + e- -> Na(l) (-2.71V in aqueous solution)
  2. 2H2O +2e- -> H2 + 2OH- (-0.83V in aqueous solution)
  3. H2O + Na(l) -> NaOH + 1/2H2
  4. Na(l) + NaOH -> Na2O + 1/2H2

The following are the major reactions occuring at the anode

  1. 2H2O -> O2 + 4H+ 4e- (1.23V in aqueous solution)
  2. Ni + 2OH- -> Ni(OH)2 + 2e-
  3. 2Na(l) + 1/2O2 -> Na2O
  4. Na2O + H2O -> 2NaOH


The overall useful reaction is

  1. NaOH -> 1/2H2 + 1/2O2 + Na(l)

The half potentials quoted above are for aqueous reactions under standard conditions and so take into account the interaction energy of ions with water molecules rather than with the NaOH bath at the operating temperature of the Castner cell, or indeed the overpotentials at the electrodes, they are thus quoted for comparative purposes. One could alternatively use the formation energy of NaOH to derive the overall cell potential using DG = - nFE, which gives a decomposition potential of 2.23V, which is more representative of the actual reaction, however this does not give us the respective anode and cathode potentials.

The first half-cell reactions in each list are the actual desired processes leading to sodium evolution at the cathode. The actual voltage quoted for this is 4.5V (compared to 3.94V obtained from adding the half-cell potentials, or 2.23V from the decomposition potential). Comparing this voltage to that for the splitting of water (second cathode reaction) we see the latter occurs at a lower voltage of 2.03V under standard conditions, or 1.5V from the decomposition potential. Since water is always present in a fresh bath the initial stage of Castner cell operation is always the electrolysis of water - until the later is eliminated. Since this proceeds at a lower voltage use is made of this fact in the present cell to avoid production of sodium until all water has been electrolysed - as evidenced by a drop of current with the cell operating at the lower initial voltage. This almost entirely eliminates explosions, which occur if sodium is produced with significant excess water still present in the bath. This also serves to produce an initial protective cloud of hydrogen over the cathode. Once the all initial water has been electrolysed, the second cathode reaction subsides, the voltage is raised, and the first cathode reaction commences. The anode reaction continues as before, and since this generates water, the later is either electrolysed as before, or diffuses to the cathode and reacts with the sodium according to the third reaction. The effect is the same, almost one mole of water is electrolysed for every mole of sodium (very little water can escape a dry caustic soda bath at the anode) meaning the current efficiency of the cell is 50% maximum, based on the first cathode reaction. This can be seen explicitly from the overall reaction of the cell. This can be a handicap when compared to the Downs cell for the electrolysis of NaCl, where one and not two moles of electrons are needed to reduce one mole of sodium. This, as well as the fact that NaOH is overwhelmingly produced by electrolysis of common salt, means that the Downs process has replaced Castner almost everywhere.

The second anode reaction represents degradation of the cell anode. Nickel has a large potential in sodium hydroxide, and is chosen as the anode material for this reason.

Excess voltage, over that quoted above needs to be applied to the cell to overcome voltage drops due to ohmic losses and electrode overpotentials. The cathode represents a copper spiral 50mm long and 21mm diameter, the anode is also cylindrical, with a diameter of 70mm. Since the conductivity of molten NaOH is quoted at 2.1 Ohm-1cm-1, the resistance of the present cell is (1/2 Pi 2.1) * ln(70/21)/5 = 0.019 Ohm. To pass the standard cell current of 47 Amp a voltage drop of 1V is then required. This power goes towards heating the cell - its not really wasted since it replaces heat conduction losses and keeps the bath molten. This results in the convenient outcome, that after initial heating needed to melt the bath, subsequently the temperature is maintained by the electrolysis current, in commercial Casner cells, as wee as here.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig2.JPG">

Method

All reactions were carried out at an ambient temperature of 20C


Theory of construction

As mentioned in the introduction, the present construction is inspired by a pilot Castner cell presented in the German forum .de, to which the reader is referred. Problems identified with that design, which prevent reuse of equipment for more than one run and result in limited production, were

  1. fairly flimsy nature of apparatus - the Na collector just sits in the melt - if you heat it too high, it will move.
  2. The container walls are used as an anode, which is not how Castner operates, leading to wall erosion, as well as making it impossible to use nickel in the anode, by far the most sturdy material when it comes to dissolution under potential in molten NaOH
  3. The bottom of the vessel was ramed clay, exposed directly to NaOH at near melting temperature - although the original experiment tried to maintain a temperature differential so the bottom NaOH is solid, in practice its hard to do, and in any case at elevated temperatures the clay will rapidly dissolve (my experinments show molten NaOH attacks clay at the order of 1mm per 10 mins). The dissolution of the clay leads to substantial contamination of the bath, shortens its life, as well as leading to perforation at the bottom with time.
  4. The low amperage of the cell resulting from (3) means you have to wait all day to get even a few grams of Na.


A 3D drawing of the salient features of the cell and a table listing the appropriate materials and dimensions is given below. This is not a CAD scaled technical drawing, which would not be appropriate here. However adequate information is provided to enable the cell to be reproduced using the 3D drawing, the many photographs, and the dimensions given in the table.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add1.JPG">


Essential Components Length / Diameter / Thickess, in mm

  1. Cell Wall Carbon Steel Cylinder 160 / 80 / 5
  2. Collector Carbon Steel Cylinder 200 / 50 / 5
  3. Cell Bottom Carbon Steel Disk 90 / 5 with 24mm centre hole for cathode stem
  4. Cell Top Carbon Steel Disk 100 / 5 with 51mm centre hole for collector
  5. Stem Carbon Steel Cylinder 250 / 27 / 2
  6. Anode Nickel Cylinder 50 / 70 / 1
  7. Cathode Copper Wire/Pipe 550 / / 2 top is 18mm helix with 5 turns, bottom 290mm straight
  8. Heat Shield Stainless Steel Pot 130 / 130 / 0.7 with 92mm centre hole for cell wall
  9. Insulating Pot Stainless Steel Pot 230 / 220 / 0 with 28mm centre hole for cathode stem



Several areas of the original construction are used here unchanged

  1. The cathode was made of wound Cu tube of short cross section rather than solid. That is a great idea comapred to a solid Cu cathode- meaning much reduced heat loss through Cu notorious heat conductivity, a much more unifrom temperature of the bath, which is essential for any reasonable yield in this case.
  2. The Na collector was a slotted pipe into which a thin wire gauze (sold for toasters) was inserted, and held up extremely well.
  3. The cell was heated from the outside with insulation provided by fibreglass.


The present construction has the following features:

  1. A cylinder anode of 1mm Ni sheet of 6cm diameter by 5 cm long attached by adjustable SS bolts to a top circular plate separated from the collector and casing by ceramic insulators scavanged from a radiation heater. See figure 3.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig3.JPG">
  2. A NiCr heating element scavanged from the same heater wound around fibreglass matting and providing 300W of power. The fibregalss was held in a wound state around the outside of the cell by two stainless steel hose clips. Each of the hoseclips also acted as a terminal for the heating element. This provides a sturdy constrcution. See figure 4.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig4.JPG">
  3. An outer radiative-heat container of SS (an indian coffee jar from woolworths) of 15cm diameter and 15 cm long encased the cell with wound NiCr. See figures 5 and 6.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig5_6.JPG">
  4. Holes drilled in the radiative container with terminal posts attached to bolts drilled in the hose clamps allowed the passage of heating current at 240V. Note its independent of electrolysis current, and is only needed at startup. A hole also passed a thermocouple probe held by a screw clamp to the cell body. It was found that 365C was the ideal temperature for bath melting - while 320C was ideal for electrolysis. See figure 7.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig7.JPG">
  5. The cell measured 8cm diameter x 15.5cm and held 1 kg of caustic, which although full when cold, filled it half-way when molten - this reduced heat variation in the bath due to collector protrusion.
  6. A lided SS pot 25cm diameter 22cm long filled with glass wool formed the outer insulation. The pot was bolted to a long stand admitting a receptacle which was passed cathode. See figure 8.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig8.JPG">

    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add2.JPG">
  7. The cathode was 5cm long 15mm diameter wound copper tubing. Fixed at the bottom (cool end) to a threaded nut and sealed with exhaust cement. See figure 9.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig9.JPG">
  8. Temperatures measured were 100C for bottom of SS pot, 66C for top, 44C for lower end of cathode stem, 33C for cathode, 150C for anode terminals, 270C for top of Na collector, 320C for bath inside collector.
  9. The Na was collected by a perforated ladel. This let the bath pass thru as stated - has liquid Na got a high surface tension! A lot was wasted still clinging to the ladel - and you had to shake it really hard to get even half of it to drop into the parafin at 140C. See figure 10.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig10.JPG">
  10. The bath current was 47 - 50A. Experiment time was about an hour to produce the sodium shown.
  11. The main component of the power supply, figure 11, is the transformer sourced from a cheap welder with an adjustable current screw. Welders give 20-30V or so RMS open circuit, so I had to wind down the secondary. I took a tap at 4V 8V and total 12V. About 2.2 V of that is dropped in the diodes. These are 35A full wave bridges connected 4 in parallel to give a total capability of 140A. The PS is capable of about 70A before the transformer (which has an aluminium secondary coil) starts overheating. The heatsink for the diodes you see is actually a bit of overkill, they hardly warm to the touch in this arrangement at 70A. It would be better to mount the fan next to the transformer. The small red winding you see on the transformer is power for a digital 200mV LCD meter which shows the current you see. I got a 100A shunt from Jaycar across which the LCD meter is connected directly. It is switched with the switch you seeon the panel to show alternately the voltage/current. This is measured through a low pass filter (frequency about 10Hz) so you get to see the mean value, rather than the useless peak.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig11.JPG">
  12. Heater elements in radiative heaters are designed to operate at 800C plus, this does not melt the quartz tube in which they are encased, but it will melt the fibreglass. I ran the coil at 50% power. This is provided by a variable duty triac driver. After having run the cell many times I found the triac driver circuit is not really needed. That is because it turned out the optimum duty for the element, for exactly my type of cell, in 50% - and this can be achieved more easily with a diode half-wave rectifier in series with the winding. The diode cannot be a 1N4004, it must be a 5 amp variety at least.


Power Supply

The cell power supply must be capable of providing a single supply of rectified (though not necessarily smoothed) 3.3V and 4.5-4.8V RMS at 50A RMS, with the voltage in the second range variable. It must have an associated current and voltage meter. The minimum rating of the power supply must therefore be 240W to the load. Taking into account lead losses this is closer to 260-280W.

The simplest such power supply readily available to the amateur, is a PC supply with the 5V rail sensing made variable. However such supplies are normally rated 250-300W, and this was considered to leave too little margin to account for variability in cell operation due to instantenous high current demand associated with transients, such as infusion of small amounts of sodium into the anode compartment. Switching supplies are highly intolerant of such variability when operating near their maximum rating. Hence a bridge rectified step-down transformer configuration with switchable RMS current shunt / output voltage indicator was chosen. The power supply circuit diagram are shown below with an accompanying picture indicating layout.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add4.JPG">

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add3.JPG">



I have run the cell many times now, and can say its very reliable, just follow the procedure I outlined. The longest I ran the cell is 7 hours producing 63gms Na, figure 12.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig12.JPG">

The current adjusting screw, is essential, voltage control at high currents is very difficult (unless done on the primary). There is a screw on the transformer which adjusts the position of some shunt metal in the core, to reduce or increase the amount of magnetic flux from the primary passing thru the secondary. I have found that without such adjustment the procedure for getting sodium just doesnt work well. The cell easily overheats with all the attendant consequences.

The holes in the collector are not critical, make them as large as you can cosnistent with structural integrity (and of course containment of the Na at the top of the pipe). I made mine by milling 4 uniformly spaced straight channels 10mm wide by 60mm long. The collector is not floating, and is electrically connected to the cell body at the top, at the bottom theres about a 10mm gap from the bottom of the cell, which is covered with gauze, to allow good circulation.

Method of operation

If anyone repeats this PLEASE WEAR ENCLOSING EYE PROTECTORS. I never approach the cell with the NaOH molten without them. Molten NaOH is one of the best materials for destroying organic tissue. A spray bit landed on my forearm, about a 1mm sphere, and left exactly that cavity on my forearm.

If you dont follow the advice below, or your cell is constructed somewhat differently, there can be an EXPLOSION. This can eject a substantial portion of the bath straight up the Na collector.

If this hits your Cornea - you can kiss it goodbye. And a blind experimental chemist is not a very good one.


  1. The temperature of the cell must be raised GRADUALLY. In all it took about 2.5 hrs from room temperature till the contents inside the collector were liquid. AT no stage should electrolysis begin until this has occured. Heater elements in radiative heaters are designed to operate at 800C plus, this does not melt the quartz tube in which they are encased, but it will melt the fibreglass. I ran the coil at 50% power. This is provided by a variable duty triac driver. Once the set temperature is reached the triac is turned off. If it is required I can post a suitable circuit. You can basically set the heating, and then go do something else for 2.5hrs.
  2. The bath melts at about 350C-360C as displayed by the thermocouple. This is due to the latent heat of NaOH melting as well as an unavoidable T gradient between the thermocouple and the electrodes. Application of electrolysis current at this stage will lead to furious bubbling, and once the Na starts forming small EXPLOSIONS. This is normal, most descriptions of Castner mention this effect. However, in a large cell such an explosion is easily contained, doesnt eject bath, or ruin the cell. For a small cell, the explosion is of the same force, but is not well contained. It will lead to ejection of some bath. You can see the result of that in the spray present on the cover of the pot in the first picture I posted. I have found that explosions can be entirely avoided, so NO spatter from the cell occurs at all. The procedure is outlined below.
  3. Once the NaOH has melted turn of the heater. Apply electrolysis current at about 3.3V but no more. The temperature starts dropping, and the cell electrolysis residual water in the bath. It doesnt have enough voltage to electrolyse Na. As this happens the temperature starts to drop.
  4. When it reaches 320C raise the mean voltage to 4.6V (you can see the actual variation of voltage across the cell in the scope trace below), the current should approach 50A. Figure 13 shows the voltage waveform across the cell.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig13.JPG">
  5. If the temperature rises above 325 start dropping the current by screwing in the core shunt to keep it below 330. An alteration of 1 to 2 amps should be sufficient.
  6. If the temperature reaches 330C turn the electrolysis off, until its dropped to 325 (again to avoid explosions at higher temperatures due to the fact the the Na is more mobile in the less viscous NaOH melt).
  7. If you overcompensated the other way and the temperature drops below 315 turn on the heater.
  8. When starting a new run, assuming the cell is already full from the previous run, add some NaOH so bath level remains about 4cm above the cathode, using a funnel, into the NaOH colelctor pipe.
  9. Initially a brand-new ladel is a nuisance. The Na wets it perfectly and half of it stays in the ladel which is a waste. To avoid it dip the S/S ladel into the hot NaOH bath for a minute or so, then take it out wash, and dry at 280C. This treatment will cover it with a durable black-brown oxide cover. The Na will not stick to this.
    Figure 14 shows the sodium forming in the cell, and starting to cover the NaOH bath. This picture was taken with the collector cap raised - do not raise the cap on the collector unnecessarily, this risks both the splatter of sodium, and a fire due to escape of the protective hydrogen cloud.


<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig14.JPG">

Purification, yield, efficiency

The sodium laddled from the cell is in the form of globules, about 1c - 2cm across, figure 15, and generally has some adhering (green) bath paricles.
<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig15.JPG">
I have found the best method for coalesceing the globules is as follows. Once the Na has melted below the parafin surface the temperature needs to be brought up to 140C+ to decrease the surface tension of the Na. The small globules can be sucked up by a 5ml pipete on the picture, and injected into the larger globules. It is best if some parafin is sucked up first to reduce oxidation of the exposed Na surface at the top of the pipete. The action needs to be performed quickly to avoid the Na solodifying in the pipete. Once one big globule is obtained, large bits of crust, which due to the surface tension are ejected from the Na to the surface can be picked off with tweezers. When all the large bits are gone a fork can be used to trawl through the molten ingot several times, this will collect all the oxide and small crusts to the side where it can be easily picked off. The end result is an effectively pure Na ingot as seen in figure 16 (the slight brown precipitate on the surface, but not adhering to the sodium, is due to an earlier attempt at coagulation using isopropyl alcohol, which did not give good results). Purification by distillation is considered inappropriate for an amatuer set-up since the apparatus needs be completely evacuated.
<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig16.JPG">
Initially the cell was run for 150 minutes at an average current of 47A, resulting in 24.6 gms of crude Na. Assuming 1 gram of the weight is bath crud that gives a yield of 23.6 grams. Now 96000 Coulombs at 100% current efficiency give 23 grams of Na. In our case

47*150*60 = 432000 Coulombs.

The percentage current yield is then

(96000/432000)*(23.6/23) = 22.8%

The maximum theoretical current yield in a castner cell is 50% due to water from the anode diffusing to the cathode (where it generates H2, so necessary to ensure the Na doesnt catch fire). And 26% is a long-term yield quoted for industrial Castner's, so the result is not bad.

Another interesting number is the cost, which depends on the energy efficiency. As stated, and despite whats written in the literature, this 'small' Casnter cell does not need external heating during electrolysis. So the entire energy cost is the electric power. 47A at 6.4V for 150mins assuming 80% transformer efficiency (the diode efficiency has already been accounted for in the voltage drop) is 0.94 kW-Hr. Over where I live a Kw-Hr costs 17 cents. So the cost of Na with this method is AU$6.7/Kg Na. This is VERY good.

Of course I neglected the cost of the heating current in the prior-to-electrolysis phase. Its about 12cents. So if you produce say 100gms Na in a run, it would contribute about an extra $1.20, or $7.90 per Kg Na.

I have made some investigation of corrosion rates of different anodes in NaOH. The anode in this cell is the point most subject to attack because

  1. The positive anode voltage creates oxidizing conditions which tend to dissolve the anode material in the bath
  2. The O2 and H2O evolved at the anode tend to oxidize the anode material.

The greater concern here is contamination and shortening of bath life by the presence of anode cations in the bath, rather than the dissolution of the anode per-se, although, using nickel in the anode has allowed the anode material to be spot welded - something I would not have been able to do if it was subject to rapid dissolution. I have taken the cell apart after about 30hrs operation and 200gms of Na, and the nickel anode shows no signs of corrosion whatsover. I think it can be regarded as a permanent item of the cell. The copper cathode shows more corrosion, but not really significat. Copper can not be used in the anode, it dissolves and contaminates the bath rapidly. S/S I have used in mock-ups before. High Ni SS (which can be detected by the fact that they are non-magnetic) seem to stand up as anodes the best. Their dissolution rate just as a very rough estimate is less than about 100microns/hr at the current densities of this cell. The best material, without any doubt however, is the original Ni chosen by Castner.


Conclusion


I have run the cell several times and using the operating regime introduced the cell operates perfectly every time. Unlike what is generally assumed, the cost of sodium production at AU$7.90/kg sodium with an outlay of about $200 for cell and power supply fabrication, compares favourably with outside purchase and transport of the metal, where that is an available option.

[Edited on 24-3-2008 by len1]

[Edited on 24-3-2008 by The_Davster]

chemoleo - 16-1-2008 at 18:36

Absolutely fantastic!
The only thing that're missing are photogenic shots of the Na thus produced when undergoing various reactions with various environmental chemicals...:P

garage chemist - 16-1-2008 at 19:19

An extremely well-written and in-depth description of your sodium making process.
Your cell is really fantastic!

However, I am still not entirely able to figure out the construction of your cell.
Just one thing:

How is the bottom of your cell built up? As far as I can see, the cell simply sits on two ceramic (?) tiles and the cathode seems to be cemented into those. Or does your cell have a steel bottom?
How did you attach the cell itself to the ceramic tiles so that the melt doesnt flow out into the outer pot? I really don't understand this.
Can you provide pictures that show how the underside of the cell is built up?

len1 - 16-1-2008 at 19:33

Thanks garage chemist - and I also owe you a thanks for posting the link to the original german article - which inspired me to do this.

The cell does indeed sit on two ceramic plates - but these have a hole drilled to enable the cathode stem to pass. They are not the cell bottom - they are just thermal insulation and support for the cell. The cell itself has a welded carbon steel bottom - you can see the weld at the bottom of the cell in figure 4 - with an opening to which is welded the cathode stem - which protrudes through the ceramic plates, the outer insulating pot, the support, right to the bottom.

I have had several questions about construction so it is clear the pictures have not achieved their aim, it seems they are clear only if you have already seen the cell. So I intend to add a cross secional drawing as soon as I get a chance + will post more pictures. Len

[Edited on 17-1-2008 by len1]

The_Davster - 16-1-2008 at 19:48

I really love it!
I have been almost able to gather how you cell is constructed from this and the other thread, my only questions are on how the coil cathode is connected through the bottom, and how it is electrically isolated from the cell body, and with what material was used for isolation that is stable under these conditions. I know you mentioned that rammed clay was not.

Also would it be possible to post the schematic for the power supply? For those of us who know how to put that sort of thing together from schematics, but are not well versedin designing such things.

len1 - 17-1-2008 at 00:24

Thank you. Im sorry for making it hard to figure out how exactly the cell is put together. Its hard to tell what is clear and what is not after youve been working on it for a long time. I will post more pics and a cross-section soon. I will also change the text to address the questions Ive been getting.

To answer your question, I use the same method of insulation as Castner, the frozen cell bath. A 20cm long stem, a cast iron pipe 27mm outside diameter is welded in the centre to the base of the cell. The cathode, consisting of a coiled copper tube with a straight ending, as you see in the picture, is inserted through the middle of this, and the space inside the stem, between the copper and the stem walls is then filled with NaOH. This is liquid at the top of the stem, since its at the temperature of the bath, but the lower end protrudes through the insulation pot and the support, and its temperature was about 50C I think. So the bath freezes somewhere inside the stem. The copper cathode is cemented and filled to a sleeved nut (from plumbing supplies) that screws on to the end of the stem (which is also a plumbers pipe) - you can see it at the bottom of the cathode acting as support when the pic was taken. The cement is exhaust cement from automotive suppliers, mainly sodium silicate. Because that never contacts the molten bath it suffers no corrosion. I will post more pics. If there are any question please ask.

[Edited on 17-1-2008 by len1]

garage chemist - 17-1-2008 at 07:06

It seems like very few electrical insulators are able to withstand molten NaOH, right? Otherwise the cell would be easier to construct.

I had a thought: maybe Sorel cement can withstand molten NaOH?
Sorel cement is made from caustic magnesia (light-burned, reactive MgO) and magnesium chloride solution, it becomes hard as rock in a few hours upon mixing.
Since magnesium oxide is solely basic and does not form any "magnesates", it should be completely resistant against molten NaOH contrary to Al2O3 and SiO2 which are attacked. Think of the "basic brick lining" (magnesite bricks) in converters for steelmaking which is resistant against basic slag.
The only problem with this kind of cement is its limited resistance to heat.
One should probably add lots of dead-burned MgO to the cement to act as a filler against shrinking during heating.

I have both caustic and dead-burned MgO, I can give Sorel cement a try when I get around to it.

[Edited on 17-1-2008 by garage chemist]

a_bab - 3-2-2008 at 16:32

Congrats for your work. You are my hero, and possibly the first one who *finally* managed to get a reliable method for mass Na production (at an amateur level of course...)

I will try sometime this summer the reduction with C, for the sake of it.

len1 - 28-2-2008 at 00:55

Schematic diagram of cell, power supply, and more pictures of the cell can not now be posted as I intended, because the edit feature has been disabled by forum managers. Sorry

a_bab - 3-3-2008 at 02:35

A good idea is to make a pdf, or even a word file containing the info, that you can attach here.

len1 - 5-3-2008 at 19:44

I dont have a pdf converter. As far as the word file goes I could do that, but it means a lot of extra effort converting everything into word. And then the link to the file could not be included in the top post, which means it would not be the most current version. I dont see the point of the extra effort given the forum owners do not value this kind of thing anyway, they just want a discussion 'melee', rather than top quality up to date articles.

[Edited on 6-3-2008 by len1]

not_important - 5-3-2008 at 20:33

Open Office will write PDF files just fine, and costs nothing. PDFs are good for 'released' documents where most everything has been clarified and tweaked, because it's all in the same single file package and is easy to D/L and transfer about.

a_bab - 6-3-2008 at 05:08

You can be sure that everyone here really appreciate your efforts. As far as I know, nobody has done what you did in terms of home chemistry related to sodium production, or the research for the best chlorine making reaction with OTC stuff.

On the other hand, you seem to be very keen about the idea of posting the info in a thread; unfortunately this is not the appropiate way of doing it as the forum may go (database can get corrupted, pictures lost, etc). You can't update the post, can't attach files in old posts and other issues.

A pdf document on the other hand is probably the best way to go, as one can read it offline, and it "works" on almost any system (including pocket computers, PSP, you name it). There is a library you may didn't know about, where other members compiled their work, like the synthesis of benzene from purely-easy-to-find-OTC materials: http://www.sciencemadness.org/member_publications/index.html

This is the place where your work should eventualy go.

I can help you with the pdf creation if you want. PM for details.

len1 - 6-3-2008 at 17:29

@not_important & a_bab

Having the article in pdf rather than in a thread sounds all very nice and logical .. except - why have a prepublication section at all?

In actual fact a prepublication section IS in accodrance with all norms of copy editing, that is:

1) You write an article
2) received comments
3) edit article,
4) go back to step two until satisfied with results
5) write a pdf

Eliminating 3) (which the very short 24hrs editing time effectively does) breaks the whole process, making prepublication pointless - except for a one time crank of the above algorithm.

I add that I was asked to post into prepublication at the suggestion of others and to have the rules change as you go is not a lack of foresight on my part, and it annoys me because of time I spent on this.

There has been a solution, at least in the case of THIS thread. One moderator kindly offered to post into the original article the additions people asked for if I post them below, and I will now go ahead with that idea. Len

[Edited on 7-3-2008 by len1]

not_important - 6-3-2008 at 18:52

Bah, I said PDF once the noise and thrashing settled down, not necessarily right now. I just pointed out the advantage of a PDF when it felt like a good check point had been reached, and mentioned a way to make PDFs on the cheap.

len1 - 6-3-2008 at 19:52

I value your contribution. My annoyance was more aimed at the changes.

len1 - 23-3-2008 at 21:48

This article is part of a series of Illustrated Key Syntheses in Chemistry which I intend to post, with the purposes and guidelines outlined in the thread of the same name in the General Chemistry section.<br> I shall attampt to ensure that the first post in all syntheses contains all the relevant information. If something needs to be added later on, due to comments further down the thread for instance, I shall edit the first post.




Aim

It is the aim here to present an illustrated demonstration of a method for the amateur chemist to produce useful quantities of sodium metal for laboratory use. The method relies on electrolysis of molten NaOH, a process patented by Castner in 1891, which has been found to be the simplest method of sodium production for the amateur. The present method is a modification of the Castner process in both form and scale to enable construction by the amateur using only materials easily obtainable from hardware stores. It is nevertheless capable of producing up to several hundred grams per day. The present appartus is based on a simpler non-reusable cell presented on the German site .de which produced several grams sodium per day., here modified and increased in scale to make the equipment reusable, durable and allow easy replacement of consumable elements, such as the cathode and anode. One should note that many instances of alkali metal production using 'electrolysis in a can' have been previously presented. While these methods can be constructed in a matter of hours they are not capable of reliably producing quantities of more than fractions of a gram, the reason being that the electrolytic bath produces yields over a very narrow temperature range, 320-330C - for the sodium cell. Consequently in a simple cell initial production subsequently redisolves and there is no yield. The present equipment uses the minimum amount of sophistication necessary to achieve a consistent yield in a reusable cell.


Motivation


Sodium is an indispensible element in a serious amateur laboratory. Its primary use is in thoroughly drying organic solvents, such as Grignard reagents, in the formation of alkoxides used in the syntheses of ethers for instance, and in the synthesis of whole groups of inorganic compounds such as cyanides, azides, and metal amides. In the latter case, the disolution of metalic sodium in liquid ammonia produces an intense blue solution with electronic properties similar to metals. In industry sodium is used as a heat transfer agent in situations such as nuclear power plants due to its excelent thermal properties. Despite its plethora of uses, in the past there were no proven methods for the amateur to synthesise his own sodium except the subgram amounts available from electrolysis in a can. This presented a problem since in many locations sodium can not be purchased OTC, for reasons of liability and other considerations. Professional laboratories generally buy their sodium, and if it is required to be made, for reasons of purity for instance, decomposition of NaN3 at 300C is about the only laboratory preparation available. Since NaN3 is extremely poisnous and an explosive precursor, the availablity of sodium azide is generally poorer than that of sodium.

This leaves the amateur with industrial methods, many from the 19th century, for sodium production, methods which have generally not shown results due generally to the difficulty of adapting industrial processes to the laboratory. These include

  1. vacuum reduction of sodium salts with carbon at > 1000C, such as 4NaOH + 2C -> Na2CO3 + CO + 2H2 + 2Na
  2. vacuum reduction with metals such as lead, magnesium or zirconium at > 500C, such as 2NaCl + Pb -> PbCl2 + 2Na
  3. electrolysis of NaCl + CaCl2 at 600C (Na is soluble in the bath at the mp. of NaCl) with a consumable carbon anode
  4. variations of the above with mobile liquid cathodes, such as molten lead or mercury to deal with reactivity of Na at 600C
  5. electrolysis of NaOH (due to its low mp. of 318C) with iron cathode and S/S or nickel anode 2NaOH -> 2Na + H2 + O2 (some H2O also escapes)


The latter method due to its low temperature and evolution of hydrogen gas protective to the sodium, obviates the need for a vacuum and has been found suitable for modification to the amateur scale. The main problem which needs to be solved with this method is the extreme reactivity of molten sodium hydroxide, which degrades almost all materials except a few metals. The basis of this method, electrolysis, was the means used by Humphry Davy the discoverer of sodium who in 1807 electrolysed NaOH with platinum electrodes, but who was not able to produce more than subgram quantities. He in turn owes the discovery to Alexander Volta, who six years earlier invented the battery - the voltaic pile. This new force, was immediately applied by investigators all over Europe in an attempt to produce transformations in all kinds of materials, which previously defied change, and appeared elementary. One such material was lye. Davy used a large pile belonging to the Royal Society, and to make lye conduct dissolved it in water, but observed that water, rather than lye underwent transformation, and a new principal, of which there was a hint of formation at the cathode rapidly reacted with the water. As a result Davy undertook electrolysis of the molten salt - which he cleverly melted and electrolysed with the same current. For the next 80 years all sodium production was based on the high temperature methods above - which resulted in quick degradation of the apparatus, and expensive sodium. In 1891 the original method of Davy was adapted for industry by Hamilton Castner, in a process bearing his name. Castner observed that nickel was largely inert to the action of molten lye, and hence used this metal in th anode, where the positive potential in combination with the corrosive bath acted to dissolve most other materials. If one tries to conduct the electrolysis with an overhead cathode, the molten sodium adheres to it as it floats in the bath - and subsequent molten metal formed then attempts to form a bridge to the anode (see my experiments in the electrochemical sodium thread) due to the potential difference in solution making this the lowest energy configuration. If one attempts to eliminate this by using a non-conducting separator, the sodium will crawl under it, and commence forming in the anode compartment (resulting in explosions). Hence an underside cathode - not in contact with the floating accumulated sodium - is required, and this introduces isolation and hermiticity problems. Castner solved both these problems by using the bath material itself - a long underside stem encasing the cathode has a temperature gradient along it, such that the bath at the bottom is solidified. The Casner cell uses an iron gauze to separate the anode and cathode below the surface, this allows flow of ions, but not the sodium. Some sodium still manages to pass to the anode where it reacts with the oxygen formed there and produces small explosions. The standard Castner cell is a cylinder of about 60cm diameter and 80cm high, holding about 200kg of bath. Our cell has a diameter of 8cm and is 15cm high - a scaling down of about 250 - it holds 1.5kgs of bath, hence explosions are more serious for it - and I have found a procedure for almost eliminating them. The current density at the electrodes is almost the same as in the original.

The only chemical on which the present method relies to produce sodium is sodium hydroxide NaOH, which is freely available OTC. Even if difficulties arise in its purchase it can easily be manufactured from the electrolysis of salt NaCl, using a much simpler electrolysis cell than the present. It can also be easily made using such chemicals as Na2CO3 and Ca(OH)2 in the soda lime process. Since NaCl will always be available, the consequence of the present method is that it guarantees the same for sodium.

Note: A method described in a 1987 US patent 4725311 by metallgesellschaft aktiengesellschaft puportedly employing reduction with magnesium at 200C appeared simpler to execute in an amateur laboratory than any of the above. It has been tried in the unconventional sodium thread for the production of potassium following the procedure in example 1 of the patent closely. Despite repeating the experiment several times with different sources of magnesium absolutely no alkali metal was obtained, all observed indications of a reaction were consistent with the reaction of magnesium with water. This, combined with the obviously incorrect measurements of evolved H2 presented in the patent, suggest the results the patent reports are false.

Findings



  1. 30hrs operation produced 200gms of Na, figure 1, the nickel anode shows no signs of corrosion
  2. The cell operates at 47-49 amps, achieving a current yield of 23% - typical for small cells, compared to 50% max theoretical current yield
  3. Electrical energy cost about AU$7.90/kg Na produced.
  4. A simple purification procedure, which produces shiny pure sodium is presented.
  5. A method for starting and operating the cell is presented, which completely eliminates the small explosions normally accompanying Castner operation.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig1.JPG">
Theory


The construction of a typical Castner cell (not the scaled cell for amateur use presented here) us shown in figure 2.

The following are the major reactions occuring at the cathode

  1. Na+ + e- -> Na(l) (-2.71V in aqueous solution)
  2. 2H2O +2e- -> H2 + 2OH- (-0.83V in aqueous solution)
  3. H2O + Na(l) -> NaOH + 1/2H2
  4. Na(l) + NaOH -> Na2O + 1/2H2

The following are the major reactions occuring at the anode

  1. 2H2O -> O2 + 4H+ 4e- (1.23V in aqueous solution)
  2. Ni + 2OH- -> Ni(OH)2 + 2e-
  3. 2Na(l) + 1/2O2 -> Na2O
  4. Na2O + H2O -> 2NaOH


The overall useful reaction is

  1. NaOH -> 1/2H2 + 1/2O2 + Na(l)

The half potentials quoted above are for aqueous reactions under standard conditions and so take into account the interaction energy of ions with water molecules rather than with the NaOH bath at the operating temperature of the Castner cell, or indeed the overpotentials at the electrodes, they are thus quoted for comparative purposes. One could alternatively use the formation energy of NaOH to derive the overall cell potential using DG = - nFE, which gives a decomposition potential of 2.23V, which is more representative of the actual reaction, however this does not give us the respective anode and cathode potentials.

The first half-cell reactions in each list are the actual desired processes leading to sodium evolution at the cathode. The actual voltage quoted for this is 4.5V (compared to 3.94V obtained from adding the half-cell potentials, or 2.23V from the decomposition potential). Comparing this voltage to that for the splitting of water (second cathode reaction) we see the latter occurs at a lower voltage of 2.03V under standard conditions, or 1.5V from the decomposition potential. Since water is always present in a fresh bath the initial stage of Castner cell operation is always the electrolysis of water - until the later is eliminated. Since this proceeds at a lower voltage use is made of this fact in the present cell to avoid production of sodium until all water has been electrolysed - as evidenced by a drop of current with the cell operating at the lower initial voltage. This almost entirely eliminates explosions, which occur if sodium is produced with significant excess water still present in the bath. This also serves to produce an initial protective cloud of hydrogen over the cathode. Once the all initial water has been electrolysed, the second cathode reaction subsides, the voltage is raised, and the first cathode reaction commences. The anode reaction continues as before, and since this generates water, the later is either electrolysed as before, or diffuses to the cathode and reacts with the sodium according to the third reaction. The effect is the same, almost one mole of water is electrolysed for every mole of sodium (very little water can escape a dry caustic soda bath at the anode) meaning the current efficiency of the cell is 50% maximum, based on the first cathode reaction. This can be seen explicitly from the overall reaction of the cell. This can be a handicap when compared to the Downs cell for the electrolysis of NaCl, where one and not two moles of electrons are needed to reduce one mole of sodium. This, as well as the fact that NaOH is overwhelmingly produced by electrolysis of common salt, means that the Downs process has replaced Castner almost everywhere.

The second anode reaction represents degradation of the cell anode. Nickel has a large potential in sodium hydroxide, and is chosen as the anode material for this reason.

Excess voltage, over that quoted above needs to be applied to the cell to overcome voltage drops due to ohmic losses and electrode overpotentials. The cathode represents a copper spiral 50mm long and 21mm diameter, the anode is also cylindrical, with a diameter of 70mm. Since the conductivity of molten NaOH is quoted at 2.1 Ohm-1cm-1, the resistance of the present cell is (1/2 Pi 2.1) * ln(70/21)/5 = 0.019 Ohm. To pass the standard cell current of 47 Amp a voltage drop of 1V is then required. This power goes towards heating the cell - its not really wasted since it replaces heat conduction losses and keeps the bath molten. This results in the convenient outcome, that after initial heating needed to melt the bath, subsequently the temperature is maintained by the electrolysis current, in commercial Casner cells, as wee as here.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig2.JPG">

Method

All reactions were carried out at an ambient temperature of 20C


Theory of construction

As mentioned in the introduction, the present construction is inspired by a pilot Castner cell presented in the German forum .de, to which the reader is referred. Problems identified with that design, which prevent reuse of equipment for more than one run and result in limited production, were

  1. fairly flimsy nature of apparatus - the Na collector just sits in the melt - if you heat it too high, it will move.
  2. The container walls are used as an anode, which is not how Castner operates, leading to wall erosion, as well as making it impossible to use nickel in the anode, by far the most sturdy material when it comes to dissolution under potential in molten NaOH
  3. The bottom of the vessel was ramed clay, exposed directly to NaOH at near melting temperature - although the original experiment tried to maintain a temperature differential so the bottom NaOH is solid, in practice its hard to do, and in any case at elevated temperatures the clay will rapidly dissolve (my experinments show molten NaOH attacks clay at the order of 1mm per 10 mins). The dissolution of the clay leads to substantial contamination of the bath, shortens its life, as well as leading to perforation at the bottom with time.
  4. The low amperage of the cell resulting from (3) means you have to wait all day to get even a few grams of Na.


A 3D drawing of the salient features of the cell and a table listing the appropriate materials and dimensions is given below. This is not a CAD scaled technical drawing, which would not be appropriate here. However adequate information is provided to enable the cell to be reproduced using the 3D drawing, the many photographs, and the dimensions given in the table.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add1.JPG">


Essential Components Length / Diameter / Thickess, in mm

  1. Cell Wall Carbon Steel Cylinder 160 / 80 / 5
  2. Collector Carbon Steel Cylinder 200 / 50 / 5
  3. Cell Bottom Carbon Steel Disk 90 / 5 with 24mm centre hole for cathode stem
  4. Cell Top Carbon Steel Disk 100 / 5 with 51mm centre hole for collector
  5. Stem Carbon Steel Cylinder 250 / 27 / 2
  6. Anode Nickel Cylinder 50 / 70 / 1
  7. Cathode Copper Wire/Pipe 550 / / 2 top is 18mm helix with 5 turns, bottom 290mm straight
  8. Heat Shield Stainless Steel Pot 130 / 130 / 0.7 with 92mm centre hole for cell wall
  9. Insulating Pot Stainless Steel Pot 230 / 220 / 0 with 28mm centre hole for cathode stem



Several areas of the original construction are used here unchanged

  1. The cathode was made of wound Cu tube of short cross section rather than solid. That is a great idea comapred to a solid Cu cathode- meaning much reduced heat loss through Cu notorious heat conductivity, a much more unifrom temperature of the bath, which is essential for any reasonable yield in this case.
  2. The Na collector was a slotted pipe into which a thin wire gauze (sold for toasters) was inserted, and held up extremely well.
  3. The cell was heated from the outside with insulation provided by fibreglass.


The present construction has the following features:

  1. A cylinder anode of 1mm Ni sheet of 6cm diameter by 5 cm long attached by adjustable SS bolts to a top circular plate separated from the collector and casing by ceramic insulators scavanged from a radiation heater. See figure 3.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig3.JPG">
  2. A NiCr heating element scavanged from the same heater wound around fibreglass matting and providing 300W of power. The fibregalss was held in a wound state around the outside of the cell by two stainless steel hose clips. Each of the hoseclips also acted as a terminal for the heating element. This provides a sturdy constrcution. See figure 4.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig4.JPG">
  3. An outer radiative-heat container of SS (an indian coffee jar from woolworths) of 15cm diameter and 15 cm long encased the cell with wound NiCr. See figures 5 and 6.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig5_6.JPG">
  4. Holes drilled in the radiative container with terminal posts attached to bolts drilled in the hose clamps allowed the passage of heating current at 240V. Note its independent of electrolysis current, and is only needed at startup. A hole also passed a thermocouple probe held by a screw clamp to the cell body. It was found that 365C was the ideal temperature for bath melting - while 320C was ideal for electrolysis. See figure 7.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig7.JPG">
  5. The cell measured 8cm diameter x 15.5cm and held 1 kg of caustic, which although full when cold, filled it half-way when molten - this reduced heat variation in the bath due to collector protrusion.
  6. A lided SS pot 25cm diameter 22cm long filled with glass wool formed the outer insulation. The pot was bolted to a long stand admitting a receptacle which was passed cathode. See figure 8.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig8.JPG">

    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add2.JPG">
  7. The cathode was 5cm long 15mm diameter wound copper tubing. Fixed at the bottom (cool end) to a threaded nut and sealed with exhaust cement. See figure 9.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig9.JPG">
  8. Temperatures measured were 100C for bottom of SS pot, 66C for top, 44C for lower end of cathode stem, 33C for cathode, 150C for anode terminals, 270C for top of Na collector, 320C for bath inside collector.
  9. The Na was collected by a perforated ladel. This let the bath pass thru as stated - has liquid Na got a high surface tension! A lot was wasted still clinging to the ladel - and you had to shake it really hard to get even half of it to drop into the parafin at 140C. See figure 10.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig10.JPG">
  10. The bath current was 47 - 50A. Experiment time was about an hour to produce the sodium shown.
  11. The main component of the power supply, figure 11, is the transformer sourced from a cheap welder with an adjustable current screw. Welders give 20-30V or so RMS open circuit, so I had to wind down the secondary. I took a tap at 4V 8V and total 12V. About 2.2 V of that is dropped in the diodes. These are 35A full wave bridges connected 4 in parallel to give a total capability of 140A. The PS is capable of about 70A before the transformer (which has an aluminium secondary coil) starts overheating. The heatsink for the diodes you see is actually a bit of overkill, they hardly warm to the touch in this arrangement at 70A. It would be better to mount the fan next to the transformer. The small red winding you see on the transformer is power for a digital 200mV LCD meter which shows the current you see. I got a 100A shunt from Jaycar across which the LCD meter is connected directly. It is switched with the switch you seeon the panel to show alternately the voltage/current. This is measured through a low pass filter (frequency about 10Hz) so you get to see the mean value, rather than the useless peak.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig11.JPG">
  12. Heater elements in radiative heaters are designed to operate at 800C plus, this does not melt the quartz tube in which they are encased, but it will melt the fibreglass. I ran the coil at 50% power. This is provided by a variable duty triac driver. After having run the cell many times I found the triac driver circuit is not really needed. That is because it turned out the optimum duty for the element, for exactly my type of cell, in 50% - and this can be achieved more easily with a diode half-wave rectifier in series with the winding. The diode cannot be a 1N4004, it must be a 5 amp variety at least.


Power Supply

The cell power supply must be capable of providing a single supply of rectified (though not necessarily smoothed) 3.3V and 4.5-4.8V RMS at 50A RMS, with the voltage in the second range variable. It must have an associated current and voltage meter. The minimum rating of the power supply must therefore be 240W to the load. Taking into account lead losses this is closer to 260-280W.

The simplest such power supply readily available to the amateur, is a PC supply with the 5V rail sensing made variable. However such supplies are normally rated 250-300W, and this was considered to leave too little margin to account for variability in cell operation due to instantenous high current demand associated with transients, such as infusion of small amounts of sodium into the anode compartment. Switching supplies are highly intolerant of such variability when operating near their maximum rating. Hence a bridge rectified step-down transformer configuration with switchable RMS current shunt / output voltage indicator was chosen. The power supply circuit diagram are shown below with an accompanying picture indicating layout.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add4.JPG">

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig_add3.JPG">



I have run the cell many times now, and can say its very reliable, just follow the procedure I outlined. The longest I ran the cell is 7 hours producing 63gms Na, figure 12.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig12.JPG">

The current adjusting screw, is essential, voltage control at high currents is very difficult (unless done on the primary). There is a screw on the transformer which adjusts the position of some shunt metal in the core, to reduce or increase the amount of magnetic flux from the primary passing thru the secondary. I have found that without such adjustment the procedure for getting sodium just doesnt work well. The cell easily overheats with all the attendant consequences.

The holes in the collector are not critical, make them as large as you can cosnistent with structural integrity (and of course containment of the Na at the top of the pipe). I made mine by milling 4 uniformly spaced straight channels 10mm wide by 60mm long. The collector is not floating, and is electrically connected to the cell body at the top, at the bottom theres about a 10mm gap from the bottom of the cell, which is covered with gauze, to allow good circulation.

Method of operation

If anyone repeats this PLEASE WEAR ENCLOSING EYE PROTECTORS. I never approach the cell with the NaOH molten without them. Molten NaOH is one of the best materials for destroying organic tissue. A spray bit landed on my forearm, about a 1mm sphere, and left exactly that cavity on my forearm.

If you dont follow the advice below, or your cell is constructed somewhat differently, there can be an EXPLOSION. This can eject a substantial portion of the bath straight up the Na collector.

If this hits your Cornea - you can kiss it goodbye. And a blind experimental chemist is not a very good one.


  1. The temperature of the cell must be raised GRADUALLY. In all it took about 2.5 hrs from room temperature till the contents inside the collector were liquid. AT no stage should electrolysis begin until this has occured. Heater elements in radiative heaters are designed to operate at 800C plus, this does not melt the quartz tube in which they are encased, but it will melt the fibreglass. I ran the coil at 50% power. This is provided by a variable duty triac driver. Once the set temperature is reached the triac is turned off. If it is required I can post a suitable circuit. You can basically set the heating, and then go do something else for 2.5hrs.
  2. The bath melts at about 350C-360C as displayed by the thermocouple. This is due to the latent heat of NaOH melting as well as an unavoidable T gradient between the thermocouple and the electrodes. Application of electrolysis current at this stage will lead to furious bubbling, and once the Na starts forming small EXPLOSIONS. This is normal, most descriptions of Castner mention this effect. However, in a large cell such an explosion is easily contained, doesnt eject bath, or ruin the cell. For a small cell, the explosion is of the same force, but is not well contained. It will lead to ejection of some bath. You can see the result of that in the spray present on the cover of the pot in the first picture I posted. I have found that explosions can be entirely avoided, so NO spatter from the cell occurs at all. The procedure is outlined below.
  3. Once the NaOH has melted turn of the heater. Apply electrolysis current at about 3.3V but no more. The temperature starts dropping, and the cell electrolysis residual water in the bath. It doesnt have enough voltage to electrolyse Na. As this happens the temperature starts to drop.
  4. When it reaches 320C raise the mean voltage to 4.6V (you can see the actual variation of voltage across the cell in the scope trace below), the current should approach 50A. Figure 13 shows the voltage waveform across the cell.
    <IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig13.JPG">
  5. If the temperature rises above 325 start dropping the current by screwing in the core shunt to keep it below 330. An alteration of 1 to 2 amps should be sufficient.
  6. If the temperature reaches 330C turn the electrolysis off, until its dropped to 325 (again to avoid explosions at higher temperatures due to the fact the the Na is more mobile in the less viscous NaOH melt).
  7. If you overcompensated the other way and the temperature drops below 315 turn on the heater.
  8. When starting a new run, assuming the cell is already full from the previous run, add some NaOH so bath level remains about 4cm above the cathode, using a funnel, into the NaOH colelctor pipe.
  9. Initially a brand-new ladel is a nuisance. The Na wets it perfectly and half of it stays in the ladel which is a waste. To avoid it dip the S/S ladel into the hot NaOH bath for a minute or so, then take it out wash, and dry at 280C. This treatment will cover it with a durable black-brown oxide cover. The Na will not stick to this.
    Figure 14 shows the sodium forming in the cell, and starting to cover the NaOH bath. This picture was taken with the collector cap raised - do not raise the cap on the collector unnecessarily, this risks both the splatter of sodium, and a fire due to escape of the protective hydrogen cloud.


<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig14.JPG">

Purification, yield, efficiency

The sodium laddled from the cell is in the form of globules, about 1c - 2cm across, figure 15, and generally has some adhering (green) bath paricles.
<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig15.JPG">
I have found the best method for coalesceing the globules is as follows. Once the Na has melted below the parafin surface the temperature needs to be brought up to 140C+ to decrease the surface tension of the Na. The small globules can be sucked up by a 5ml pipete on the picture, and injected into the larger globules. It is best if some parafin is sucked up first to reduce oxidation of the exposed Na surface at the top of the pipete. The action needs to be performed quickly to avoid the Na solodifying in the pipete. Once one big globule is obtained, large bits of crust, which due to the surface tension are ejected from the Na to the surface can be picked off with tweezers. When all the large bits are gone a fork can be used to trawl through the molten ingot several times, this will collect all the oxide and small crusts to the side where it can be easily picked off. The end result is an effectively pure Na ingot as seen in figure 16 (the slight brown precipitate on the surface, but not adhering to the sodium, is due to an earlier attempt at coagulation using isopropyl alcohol, which did not give good results). Purification by distillation is considered inappropriate for an amatuer set-up since the apparatus needs be completely evacuated.
<IMG src="http://www.sciencemadness.org/scipics/Len1/Na_fig16.JPG">
Initially the cell was run for 150 minutes at an average current of 47A, resulting in 24.6 gms of crude Na. Assuming 1 gram of the weight is bath crud that gives a yield of 23.6 grams. Now 96000 Coulombs at 100% current efficiency give 23 grams of Na. In our case

47*150*60 = 432000 Coulombs.

The percentage current yield is then

(96000/432000)*(23.6/23) = 22.8%

The maximum theoretical current yield in a castner cell is 50% due to water from the anode diffusing to the cathode (where it generates H2, so necessary to ensure the Na doesnt catch fire). And 26% is a long-term yield quoted for industrial Castner's, so the result is not bad.

Another interesting number is the cost, which depends on the energy efficiency. As stated, and despite whats written in the literature, this 'small' Casnter cell does not need external heating during electrolysis. So the entire energy cost is the electric power. 47A at 6.4V for 150mins assuming 80% transformer efficiency (the diode efficiency has already been accounted for in the voltage drop) is 0.94 kW-Hr. Over where I live a Kw-Hr costs 17 cents. So the cost of Na with this method is AU$6.7/Kg Na. This is VERY good.

Of course I neglected the cost of the heating current in the prior-to-electrolysis phase. Its about 12cents. So if you produce say 100gms Na in a run, it would contribute about an extra $1.20, or $7.90 per Kg Na.

I have made some investigation of corrosion rates of different anodes in NaOH. The anode in this cell is the point most subject to attack because

  1. The positive anode voltage creates oxidizing conditions which tend to dissolve the anode material in the bath
  2. The O2 and H2O evolved at the anode tend to oxidize the anode material.

The greater concern here is contamination and shortening of bath life by the presence of anode cations in the bath, rather than the dissolution of the anode per-se, although, using nickel in the anode has allowed the anode material to be spot welded - something I would not have been able to do if it was subject to rapid dissolution. I have taken the cell apart after about 30hrs operation and 200gms of Na, and the nickel anode shows no signs of corrosion whatsover. I think it can be regarded as a permanent item of the cell. The copper cathode shows more corrosion, but not really significat. Copper can not be used in the anode, it dissolves and contaminates the bath rapidly. S/S I have used in mock-ups before. High Ni SS (which can be detected by the fact that they are non-magnetic) seem to stand up as anodes the best. Their dissolution rate just as a very rough estimate is less than about 100microns/hr at the current densities of this cell. The best material, without any doubt however, is the original Ni chosen by Castner.


Conclusion


I have run the cell several times and using the operating regime introduced the cell operates perfectly every time. Unlike what is generally assumed, the cost of sodium production at AU$7.90/kg sodium with an outlay of about $200 for cell and power supply fabrication, compares favourably with outside purchase and transport of the metal, where that is an available option.

[Edited on 24-3-2008 by len1]

len1 - 6-7-2008 at 05:04

The standard way to store sodium is under paraffin/oil. Even then sufficient oxygen and water penetrates the hydrocarbon, so that after several months the metal disappears under a thick white layer of oxide/hydroxide.

The standard way of dispensing the sodium out of such storage, is to cut the requisite amount of metal with a knife, dry it thoroughly off the oil by rubbing between tissues, and scrape-off the oxide coating under an inert atmosphere.

All-in-all a laborious procedure, which in addition often leads to contamination of the dispensed metal, as well as wastage in the off-cuts. When the sodium is produced in the laboratory - which necessarily means a small scale apparatus - the latter is particularly objectionable.

Heres a nice way I've devised to get round the problem.

The bulk sodium is melted at 140C under paraffin, at this temperature the oxide layer is easily scraped off by several trowles with a fork, leaving a shiny surface. Next a pipette, connected through a 1L 3-neck flask serving as gas reservoir to a water aspirator, is placed with its tip in the paraffin. The flask serves both as a safety device to isolate the sodium from the water aspirator, and as a bulk inert gas storage.

The pipette and flask is flushed with argon (or CO2), then the pipette is lowered so its tip lies below the surface of the melted sodium, and vacuum from the pump is slowly introduced until the sodium fills the pipette. A sufficient amount of sodium in the pipette lies below the paraffin surface to ensure the entire contents of the latter are liquid by thermal conduction through the sodium.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na-transfer1.JPG">

Next the pipette is disconnected from the vacuum and both its ends sealed with a glue gun. If very long term storage is required both end can be glass-sealed.

To dispense the clean sodium the pipette is placed in paraffn at 120C, after removing the lower seal, and sodium forced out (after flushing out any paraffin that entered the pipette) by positive pressure at the top end.

<IMG src="http://www.sciencemadness.org/scipics/Len1/Na-pipette1.JPG">

[Edited on 6-7-2008 by len1]

Panache - 6-7-2008 at 07:15

Brilliant storage solution, i'm going to store my sodium as thus immediately.

As an idea, instead of an aspirator to draw up the molten sodium you could put your mouth over the end of the pipette and suck.

No that was a joke but i do have a suggestion, using the suction from a dropping funnel/sep funnel would be very controllable and impossible to 'suck back'. Simply fill a dropping funnel, attach the suction line to the top of the funnel, let liquid run out until your sodium is drawn to the correct height, close the dropping funnel, seal blah blah.

The_Davster - 6-7-2008 at 07:21

I like it. Looks nicer than the sodium I glass sealed in some pipettes. Let me know if it turns blue at any point, as sodium in contact with glass when only a small amount of air was ever allowed in always turns blue for me.

len1 - 6-7-2008 at 19:36

Quote:

No that was a joke but i do have a suggestion, using the suction from a dropping funnel/sep funnel would be very controllable and impossible to 'suck back'. Simply fill a dropping funnel, attach the suction line to the top of the funnel, let liquid run out until your sodium is drawn to the correct height, close the dropping funnel, seal blah blah.


Thank you. I have actually tried exactly what you suggested first, but after trying it withe toluene as a hydrolic fluid in the separator funnel, decided on the water aspirator instead. I suppose its a matter of choice. Trouble was that the 'pump' has only a limited life before it has to be refilled with toluene. I was worried about the control Id get with with the water aspirator but it was actually OK after a bit of practice. The water was run quite slowly, and the tap opened gently. If overfill occured one could reduce the pressure by letting in some more argon.

I like the idea of a mouthful of molten sodium!:o

I have never seen sodium turn blue in glass before, even after 3 years storage, mine never had any blue tinge. The sodium stuck to the glass remains metalic and shiny. At the interface at the bottom of the pipette, where a small amount of air entered before being sealed the sodium, previously in contact with parafin, assumed a light purple colour. Does anyone know why???

You cant see this on the photo. Its quite a wrong tint - a tinge of red is present, god knows why. Ill have to learn to take better photos. The sodium actually looks much more impressive - especially to exhibit - as opposed to the white slimy mass it was before.

Fleaker - 6-7-2008 at 20:04

Len, my sodium also has that purple colour to it and I now have it stored as 1kg bricks in kerosene. I think it is a thin layer of oxide that develops for some reason gives the purple look to it. I'll take a photo next chance I get.

Unfortunately, I have too much of the stuff to store it as you do. What you have is perfect, but I'd need a monstrosity of a pipette to store it all. Perhaps I might buy some 4 inch glass tubing and make it into something resembling a pipette.

Panache - 6-7-2008 at 20:17

Quote:
Originally posted by Fleaker
Len, my sodium also has that purple colour to it and I now have it stored as 1kg bricks in kerosene. I think it is a thin layer of oxide that develops for some reason gives the purple look to it. I'll take a photo next chance I get.

Unfortunately, I have too much of the stuff to store it as you do. What you have is perfect, but I'd need a monstrosity of a pipette to store it all. Perhaps I might buy some 4 inch glass tubing and make it into something resembling a pipette.


Actually what would stop one from simply filling a decent branded reagent bottle, say 100mL Schott bottle, these have the narrower neck. They would be very convenient for throwing at religious people also. They would run in fear thinking fire and brimestone was being wrathed upon them by an angry god.

But seriously the caps are rated to 140C being an engineering polypropylene grade of high density. Hmmm.

not_important - 7-7-2008 at 06:41

Quote:
Originally posted by len1
...
You cant see this on the photo. Its quite a wrong tint - a tinge of red is present, god knows why. Ill have to learn to take better photos...


It can be difficult to get accurate colour rendering of some materials using digital cameras, especially under fluorescent or metal-halide lighting. Mant cameras have some sensitivity to wavelengths in the near UV and IR, even for the colour channels you'd not expect that on. Take a photo of a 390 or 800 nm LED and it may look rather pinkish, similarly photographing the spectrum of a thermal light source (sunlight or quartz-halogen) will show response part what you see, and odd colouring of those regions.

UV and IR filters can help, but it's sensor design and thus camera model dependent, so only experimentation will tell.


As for the blue tint formed, I'd not be surprised that it was an oxygen deficient film of sodium oxide.

len1 - 10-7-2008 at 06:26

Ive ironed out some small problems with the previous Na storage mechanism to my satisfaction, the pipette filling procedure is now perfect in my opinion. The problems were:

1) The Na did not fill the pipette completelly at the bottom (see picture), exposing a larger surface (than minimum) to oxidation, and more importantly allowing some paraffin to enter the pipette before being sealed - and so introducing contamination. The reason is unexpected and subtle - the sodium cools from the top when then pipette is withdrawn, compressing as it does so. This draws more sodium from the melt at the bottom. At a certain point the Na surface tension can not support the extra melt, a plug drops down and paraffin enters.

2) The paraffin glue over a period of several days allowed oxidation of sodium in the narrow channel to progress - clearly it is pervious to a degree. Its proper bonding to the glass cant be questions - these plugs are quite difficult to remove - so the O2/H2O must have been diffusingg in pores.

3) The sodium-mirror had some inclusions (see picture)

the solution after considerable trial is as follows

1) The Na is contained in a deep crucible with paraffin on top. The pipette inlet is lowered almost to the botton=m of the Na melt while passing slight Ar bubbles to prevent paraffin entering. After filling at 120C - which must not be much beyond the bulb as conduction into the narrow tube is poor - the Na is cooled to just below 98C, and while the Na i still soft pipette removed by twisting action - the paraffin at the tip is wiped off- theres no contamination.

2) The top entry is glass sealed in a flame - since this tube is long - each time Na is removed a 6mm bit can be broken off and then resealed. The bottom entry is glue sealed as before and the tip held in a small paraffin filled thimble with a pea of sacrificial sodium.

3) The pipette is washed with alcohol and preheated to 350C prior to sealing. This gives a silver mirror with no imperfections.

Picture coming.

I have now seen the blue -ink like - colour Davster was talking about after emptying several previously filled pipettes - interestingly it rapidly dissolves in the paraffin and disappears. I have never seen that colour working just with sodium, paraffin and air -> its produced by a reaction of Na with something in the glass and maybe O2 as well.

len1 - 12-7-2008 at 01:42

Here are the new pictures promissed in the last post (the colour balance is a bit blue this time - sodium refuses to be photographed in the right colour tone - have others encountered this problem?)

As an addendum to what I wrote before

1) Paraffin is an excellent medium for clearing oxidation off old sodium with minimum loss. After several hours all oxide flakes-off/disolved exposing a brilliant metal surface.

2) Coalescing of Na globules is almost impossible below about 120C, at about 140C its dead easy - the surface tension of the two are just right.

3) The lower tip of the pipette is sealed with paraffin glue and immersed in a PVC cup filled with liquid paraffin and containing a small sacrifical Na globule. The cup is sealed to the pipette by placing vertically in a stand and going round with a glue gun.

4) Na is dispensed by going round the seal with a heat gun and pulling the cup free. Next the paraffin glue plug is prised free, and the tip at the top of the pipette scratched about 3mm from end and broken by touching with hot rod. The Na is dispensed at 120C, and pipette resealed.

The Na should hopefully last pristine for many years with this storage.




[Edited on 29-7-2008 by len1]

len1 - 28-7-2008 at 18:58

Here is the sodium cell Mark II.

The changes from the previouds version were motivated by speed of setup - the present cell incorporates a number of simplifications over the previous design which enable it to be fanricated in one days work in a workshop equiped with no more than a drill press, a cut-off machine, and a welder.

The main simplification is in the method of heating - it is done using a standard 500W hotplate dispensing with the more complicated glass-fibre - heating coil - radiation shield arrangement. The stand is also simplified and consists basically of a cylindrical heat shield enclosing the cell and bottom reflector.

The anode is essentially the same - the main difference is that collector is pre-drilled rather than milled.

The cell wa stested and easily found able to achieve 400C - thus permitting production of K as welll as Na.

Are the pictures clear enough? Any suggestions of improvements appreciated.

NOTE: Im also looking for new synthesis to perform. If you would like to see am illustrated procedure demonstrated for a useful chemical reagent I would gladly consider writing it. Essentially Im looking for chemical reagents, which are useful, not instantly availale, whose procedure for manufacture is lacking in standard texts (in detail, purity analysis, kinetics etc.), is not drug related.


Fleaker - 29-7-2008 at 13:49

Once again Len, you have very much impressed me. Your construction skills are admirable. I can't wait until you make detailed plans available of this cell, once it is proven. I am very interested in building one such unit.
Exactly how large a batch can this cell run? I have many questions that you'll answer when you test it out!

As far as prepublication syntheses go, I have 2 that I am actively working on: synthesis of AlCl3 and perhaps other Lewis acids and also palladium on carbon, and other PGM catalysts and compounds. I have a few other, long term projects.

I will not get to white P until after I finish working on black P. I assume you'll have it cracked by then, and I greatly look forward to using your writeup as a guide when I attempt it sometime in October.

len1 - 16-4-2009 at 07:22

My notes.

Mark II has been upgraded

1) The anode on the long nickel tethers has been raised 3cm

2) A heater element (single coil insulated, bent to the outside radius of the cell) has been inserted at just above the mid point.

These changes are to allay a defect in the original design of mark II. With bottom heating the top of the cell was sufficiently cold, given the cell has to be operated at close to m.p of NaOH, to freeze most of the bath. Heat being generated internally in the electrolysis zone was too far below the NaOH surface to melt it. O2 generated below the solid top crust eventually pushed a sizeable portion of the bath to the top and denuded the bottom. Hence side heating is essential and the electrolysis zone has to be brought closer to the surface. This has now been found to have the desired effect. Raising the anode however has not eliminated the problem completely necessitating a 'defogging' cycle to be programmed every few hours (bath temp raised to 330C and maintened for 10mins) after collector emptied of Na. This seems to be an inherent feature of small cells (though I did not notice it in the Mark I)

3) The cell has been made sturdier by a welded tripod with inclined legs and a heavier piece of stainless steel for the anode ring.

4) Best of all a single GUI has been written to control the cell and display the time graphs of current, voltage, temperature, and moles. Now one only needs to ladel out the Na at regular intervals.

The control parameters are:

a) Temperature 20-365C in 2hrs - side element run at 500W

b) Bottom heater off at 250C

c) A drying cycle at 350C. When NaOH melts, if substantially wet, bubbles of H2O effuse for up to 30mins. When this ceases, an electrolysis current of 30A sets up by application of as little as 2.4V. Eventually the voltage rises to 3.1V to maintain a current of 18A. The drying cycle ends when hardly any bubbles of H2 are observed at the cathode - nearly calm surface. If this is not followed Na generated at this stage will lead to EXPLOSIONS when mixed with incompletely dry NaOH.

d) 4.6V applied at this stage - same as Mark I leads to a current of 53A rising to 56A. This is too high. Voltage can be decreased and cell stabilised by length of copper dropping 280mV @47A.

Achieving stability is quite a feat without current control. This is because the cell has an intrinsic positive feedback mechanism. Na will only be produced with any efficiency within a 10C temperature range of the mp of NaOH. Energy supplied by the current goes to

1) Produce the Na (and H2 and O2)

2) Heat the cell ohmically

The latter is normally enough for a well insulated cell such as this. However if the current should rise by a fluctuation - this increases ohmic dissipation in the cell and hence temperature in the electrolysis zone. Increased temperature decreases resistance of cell and so can lead to runaway - but this effect is small compared to the following: increased temp means the cell is less efficient - less Na is produced, more if it dissolves in the electrolyte and migrates to the anode where it is oxidized generatinh heat - this increases temperature further, making the cell even less efficient - and so runaway.

A power supply has positive effective resistance coefficient and so to a degree can counteract the negative one of the cell.

Each cell has a particular semi-stable equilibrium current, a balance between the heat this current generates, and the heat conducted away by the melt. Current can not be raised beyond that as extra ohmic heat this generates will not be conducted thermally away by the NaOH melt at the maximum10C differential allowed by the cell, will lead to unbalanced temperature rise in electrolysis zone, and eventually, no Na being produced. It is clear from this that this thershold current is mainly a function of the electrode geometry and size. Indeed so, the stable current for all cells with a 5cm 20mm diameter cathode, and 5m 70mm diameter anode has been found to be 47-49A.

The thermal parameters of the Mark II cell are somewhat differet due to better positioning and calibration of the thermocouple.

i) On temperature of side heater < 295C

ii) Off high temperature of electrolysis current 308C

iii) Off low temperature of electrolysis current 290C

iv) Temperature hysteresis 6C

v) Off temperature of bottom heater 250C

vi) Effective power of side heater 500W (35% at 1500W)

vii) Optimum current 47A

viii) Voltage at electrodes 4.0V




[Edited on 16-4-2009 by len1]

essential components?

thereelstory - 3-2-2010 at 18:19

ok, would it be safe to just make all the components as described with the listed dimensions for starters? gage

len2 - 7-2-2010 at 23:03

I would suggest doing it step wise. Make the outer part first. Wind with heater element, check that it heats up. Then the collector with holes in chamber, make sure it attaches tothe outer case. Then the anode, finally cathode - it takes about 1-2 weeks overall

plastics - 14-6-2010 at 12:31

Been a long time lurker on the forum. Two things interested me in particular: tube furnace by garage chemist and electrolytic production of Na by len1

Na has for me like many others held a particular fascination since schooldays. I even attempted electrolysis of NaOH on my mother's electric cooker using a 12v power supply, some carbon electrodes and a stainless steel egg cup. All I managed were some deep gouges on the cooker top where the molten NaOH splattered.

Somewhat older I revisited this, but without special equipment or welding skills I improvised. This time I used a gas cooker as a heat source. 1Kg of molten NaOH was contained in a stainless steel coffee pot. A stainless steel tube was fitted snugly into a piece of wood to act as the separator between anode and cathode and sat on top of the coffee container. A loop of nickel wire was situated on the outside of the inner tube by passing it through a hole in the wood to function as the cathode. The anode was a nickel wire spiral held inside the inner stainless steel tube by a retort clamp. Power supply was a 550W QTech PC PSU rewired as per woelens web page to provide 3.3, 5.0 and 12V outputs. Max current outputs were 18A, 30A and 16A respectively (these were obtained from the rating plate, my multimeter couldn't cope with these currents)

In use it took about 90mins to melt 1Kg NaOH prills. The melt fizzed for about 30 mins giving off an unpleasent mist (presumably water evaporating off with some NaOH). The 3.3v current was passed initially; this was accompanied by considerable effervescence which gradually died away after 90 mins (remainder of water been electrolytically cleaved). Small globules of Na formed at this stage and the voltage was increased to 5.0V Now appreciable globules of Na formed and rose to the surface eventually coalescing to form more substantial globules. Like len1 I found separating the molten globules from the melt was very hard. Molten Na is very 'sticky', and the NaOH quickly solidifies once removed from the bath. I eventually used a stainless steel teaspoon measure to scoop the globules up and then flick them into a container of liquid paraffin

End result 10g Na metal - not a lot but 10g more than I had before

Learning points

Put aside a whole day for one run

Only 'explosion' occurred when I re-used an NaOH melt. Think a small globule of Na had become trapped inside fused NaOH and reacted with some water when melted

Wood may sound a daft idea but it works to insulate cathode from stainless steel separator. It lasts 2 runs and then needs to be replaced as the heat and caustic vapour char it

Run the whole thing under an extracter fan - the mist doesn't do your chest any good

Temperature control is very difficult but I got a feel for the cell after a couple of runs

Wear goggles and protective gear. Molten NaOH hurts

Collecting and separating the molten Na is quite hard. It is very sticky. Even the individual globules were hard to fuse together when melted separately

IMG_44193.JPG - 51kB IMG_44182.JPG - 30kB

len1 - 15-6-2010 at 02:47

It is not possible to produce 10gm sodium from a single batch of electrolyte with this apparatus.

plastics - 15-6-2010 at 11:55

Quote: Originally posted by len1  
It is not possible to produce 10gm sodium from a single batch of electrolyte with this apparatus.


BS - got the evidence to prove it. On what basis is it not possible?

len1 - 15-6-2010 at 13:04

The theory of the NaOH cell is such, i have investigated it over many years. The one picture with the little globule of sodium you showed is about as much as youll get out of such a cell in a single go. If youve got some other evidence lets see it.

[Edited on 15-6-2010 by len1]

bbartlog - 15-6-2010 at 14:13

Hrm... not to step into the middle of an argument, but what is the presumed limiting factor that would prevent production of multiple grams of sodium here, if any were to be produced at all? Oxidation and solution of the nickel anode followed by deposition of nickel rather than sodium at the cathode? I can't think of anything other than poisoning of the cations that would limit production so severely.

plastics - 15-6-2010 at 14:21

Quote: Originally posted by len1  
The theory of the NaOH cell is such, i have investigated it over many years. The one picture with the little globule of sodium you showed is about as much as youll get out of such a cell in a single go. If youve got some other evidence lets see it.

[Edited on 15-6-2010 by len1]


Exactly how many pictures of little blobs or not so little blobs would you like me to post? Perhaps you would like a video to prove the globules are sodium and not lead

You know I thought I would post something to show that sophisticated equipment isn't always necessary and all I get is skeptical abuse. Well 10g may not be in your league but its 10g I didn't have at the beginning and that's good enough for me. If after all the years of study all you have acquired is a closed mind then I wouldn't give up the day job! If you're still speaking to me after that what were your experiences with KOH

IMG_45071.JPG - 42kB IMG_45092.JPG - 88kB IMG_44204.JPG - 55kB

watson.fawkes - 15-6-2010 at 14:56

Quote: Originally posted by bbartlog  
not to step into the middle of an argument
Oh, there's no argument here, at least not yet. Allow me to summarize the dialogue so far, in the form of a very short play.
Quote:
plastics / Man in Black

I have performed a feat.
len1 / Vizzini

Inconceivable!

len1 - 15-6-2010 at 16:47

Irrespective of all the stupidities you wrote about a closed mind etc. and the buffoons in this thread I will respond because there is some scientifically interesting input here. The amount of Na obtained is substantially more than I ever observed with with this kind of equipment - if this was indeed produced in a single run, and despite all the errors in the cell. Oh well, well done. Nonetheless soon you'll have to change the electrolyte.


[Edited on 16-6-2010 by len1]

Anders Hoveland - 15-6-2010 at 19:16

Actually sodium azide decomposes to metallic sodium and nitrogen when heated. Na3N is not very stable because Na will not form covalent bond with nitrogen, and true nitride ions have low thermal stability because the N atom is too small for 3 extra electrons repelling eachother. Lithium and magnesium, but not sodium, will burn in nitrogen. Aluminum will not burn either, but in that case it is because it is not reactive enough, Aluminum will burn hot in pure oxygen, however. Sodium azide can be made by heating sodium amide and nitrous acid solution, or NaNO2 and hydrazine. Danger: NaN3 in water easily partially hydrolyzes into HN3 gas which is very poisonous, like cyanide. Also AgNO3 solution and NH4OH left standing for a while or lightly warmed will precipitate some AgN3 explosive at the bottom. Adding thiosulfate to AgN3 will put the solid back into solution.

[Edited on 16-6-2010 by Anders Hoveland]

Lambda-Eyde - 16-6-2010 at 01:52

The decomposition of sodium azide can be quite energetic (read: explosive) when scaled up. Anyways, this is not the thread topic.

not_important - 16-6-2010 at 07:12

I too made small amounts of sodium with such simple apparatus, doing this with several friends back when I was a kid. Yields were highly variable, maybe averaging 5 grams a run using about 250 grams of NaOH and 25 grams of Na2CO3; several runs were a little too hot resulting in the aforementioned sodium fog in NaOH. The 1st time this happened we were puzzled, and then surprised as the cooling NaOH hydrated and starting sparking from the tiny globs of sodium trapped within it.

It works OK if you're after a few grams of sodium for a collection or whatever, for any serious amount of the metal I think a well made apparatus such as Len's is the way to go unless you lack the resources and skills needed to build it. While not as simple as making a pot of tea, his apparatus looks to be a much less troublesome way to produce moderate amounts of sodium, more practical for semi-continuous production runs.

len1 - 16-6-2010 at 08:23

I never managed to get the 5 or so grams here, with this kind kit, which is essentially two wires in a can (collector around anode can be done away with). The most i got when i started out with this kind of thing was sub 1 gram.

I suppose it shows if you dont have the inclination to make more complex gear it can be done. What time you save on making things you lose in operating the cell. And Youll have to chuck the electrolye much sooner. I now use a much simpler cell than in this thread, it gives about 12g/h and can run for days on same electrolyte

Mildronate - 16-6-2010 at 11:48

len1 maybe you can make PDF of this ilustrated guide?

len1 - 16-6-2010 at 15:14

Hi,

Yes eventually I will. It occurred to me that the guy who posted the pics earlier might have in reality got upset and gone away. If he is still out there this is just a clarification that he got it wrong. What he posted was the most interesting thing Ive seen on sodium in 3 years. I did not know you can make 5g with such an arrangement, now I do. Thanks for putting it right.

The other aspect is learning to take things properly. From his subsequent posts, its clear he took my initial statement as an attack. It wasnt. I was simply stating what I believed from several years experience, and a theory I had developped. What he seemed to be saying was not technically contrary to the theory - although it makes it hard to do, but completely contrary to my experience. I thought maybe we would get a clarification - which we did - but certainly not tears.

len1 - 16-6-2010 at 16:57

All your 985 posts are like this, trying to make trouble not one scientific contribution. I guess the name is apt ..

peach - 15-7-2010 at 15:49

That is absolutely outstanding work len, well done! ;)

metalresearcher - 17-10-2010 at 12:45

@len1, nearly three years after your post about your Castner's ceel replica I find this fantastic !
Do you still use it ?
I am now experimenting with making Na metal but more concentrating on the Na2CO3 + C chemical method as I am scared about the hot lye in the electrolysis.

Waffles SS - 22-12-2010 at 05:28

This is possible to get Lithium metal by this system from LiOH?sure temp should increase to 600c
*Molten Lithium float on melting LiOH.
*Molten Lithium has less dense than Molten Sodium.
Anybody know ready system(for buy) for doing that?
This is difficult for build this system

[Edited on 22-12-2010 by Waffles SS]

Waffles SS - 1-5-2011 at 21:52

I think it is interesting to get lithium metal from Lithium Nitrate by this method
Lithium nitrate melting point:255c

[Edited on 2-5-2011 by Waffles SS]

making sodium

questions - 4-5-2011 at 04:51

I loved your method of making sodium metal given below;

http://www.sciencemadness.org/talk/viewthread.php?tid=9797

I am thinking of using this method to make sodium metal and was wondering if any strong and noticeble fumes and odors were released durring the process?

[Edited on 5-4-2011 by Polverone]

len1 - 6-7-2011 at 23:11

A number of people have asked me questions about sodium, and also UV chlorination in the other thread.

I no longer recommend the method in this post. The Sodium chapter in my book gives a thorough explanation of a much simpler setup. UV chlorination is also explained in the CCl4 chapter, I havent time to repeat all that here.

If you want to see it its now on google books. Search for the title in the Book thread. The page restriction on google books can easily be got around by going to another computer.

len1 - 31-7-2011 at 02:41

Someone asked me for a link to the Sodium chapter I referred to. Here its is.

http://books.google.com.au/books?id=VqosZeMjNjEC&pg=PA14...

leu - 3-10-2013 at 05:44

The attached document describes the construction of a Castner cell for producing sodium :cool:

Attachment: ExperimentalElectrochemistryXI.pdf (718kB)
This file has been downloaded 1755 times

watson.fawkes - 3-10-2013 at 07:20

Quote: Originally posted by leu  
The attached document describes the construction of a Castner cell for producing sodium
This is a chapter from Experimental Electrochemistry, by Nevil Monroe Hopkins, 1905. Interesting that this was almost a full century after it was first done by Davy.

Zyklon-A - 7-5-2014 at 08:57

I might be building a Castner cell soon, and having not read this entire thread yet, I'm not sure if the insolating plug ( for the cathode) problem was completely solved.
In the first post, len1 used clay, which deteriorated after a while.
My idea is to use a light bulb screw-plug (or whatever it's called.) The idea would be to brake a bulb (preferably a dead one), wire the PSU to the wires that go into the bulb. Then solder cathode to the very bottom of the "screw-plug". So it screws in upside down, into the bottom of the cell.
The metal jacket terminal is directly connected (screwed) to the nickel cell casing which serves as the anode.
Some problems are:

•Light bulb terminals are made out of aluminum (at least the ones I've tested), which of course will react with the NaOH melt. Perhaps it can be coated in PTFE tape, and connected to the anode with solder too.
•I have not tested whether NaOH reacts with the black insolating ceramic yet - maybe it will.:(
•The wires going into the bulb are very thin, and will certainly over-heat at more than ~5 amps I'll bet. It's quite cramped inside the bottom of the bulb and very hard to solder a new thicker wire - with patience, I could do it I'm sure.

Other than that, it seems like it wouldn't be very hard to make this cell. I can't wait!
Here's a crappy illustration of the idea for those still confused.

Caster cell design - Resized.png - 8kB

plante1999 - 27-10-2014 at 06:30

Just an idea like that, you could lower the cathode socket a lot in a tube which would not be heated, this way, Solid NaOH could act as a plug, as such the clay would last almost indefinitely.

Metacelsus - 27-10-2014 at 08:19

PTFE will degrade in molten NaOH.

The solid NaOH plug is a good idea.

The Garbage Guardian - 14-12-2015 at 16:16

This is really late, but I have some questions that will hopefully be answered?:
1. Where did you obtain the larger metal parts like the cell wall and the collector? I checked local hardware stores, but they are all too small.
2. How were you able to tell if they and the stainless steel pots were galvanized/coated?
3. Is current for the cathode passed through the stub sticking out the bottom of the stem?

UncleJoe1985 - 14-9-2020 at 23:26

I was in the process of building my own cell after watching several Youtube videos, until I came across this excellent guide that makes me reconsider my cell design.

1. Is having the cathode at the bottom necessary? I know the design from 150 years ago did it that way to prevent bridging and that len1 is clearly in favor of that. The most successful cell I saw on YouTube also did it that way. I'm trying to make my cell from very thin titanium, which makes it very difficult to weld, so having a hole on the bottom seems a bad idea. Is it possible instead to passivate the top part of the anode so that it does not cause that awful bridging and short circuiting? What coating can withstand NaOH and not get reduced by the electron flow?


2. Is titanium a good cell body material compared to steel or nickel? According to this study, the erosion rate increases rapidly with temperature.
titanium_alkalai_corrosion_rate.png - 195kB

But even 1mm/year is pretty good - that's 44 kg at the 5g/hour rate of Nux's cell on Youtube. And this is for aqueous solutions, which is more damaging to the oxide layer than non-aqueous. I heard this from other people who constructed their cell using aluminum.

Here's a picture of my cell so far. The tube is a Ticon Industries perforated punch tube for mufflers. Also, I plan on using immersion heating with ceramic heater cartridges. They will be installed in nickel plated aluminum tubes.



cell_prototype.jpeg - 763kB