In order to look in more detail at the issues involved in practical biodiesel production we will take the example of a small-scale batch system making 100-litre batches. This sort of system might be appropriate for an individual, family or small community making biodiesel for their own use.
This reactor has been designed to use readily available materials to make a small-scale batch processor (see photo below).
Materials
Most of the processing takes place in the two tanks. It is generally said that stainless steel is the best material for biodiesel plant but it is expensive and hard to work with.
Some commercially available systems are made from plastic, normally HDPE (High Density Polyethylene). This is a cheaper option, and has the advantage that off-the-shelf tanks are available with conical bottoms. This allows for accurate and easy separation of different layers or ‘phases’.
However we have found that standard oil drums made from mild steel are suitable for a simple processor. They have the advantage of being readily available, very cheap, and easy to work with. They also stand up well to high temperatures, and give a degree of containment in case of fire. It is possible to weld conical bases onto them but in this design we have simply dented the bases, which works well enough.
Mild steel can be coated with chemical-resistant paint: some steel drums are already coated in a suitable paint. Oil drums are roughly 205 litres (45 UK gallons or 55 US gallons). Some biodieselers have used copper central heating cylinders as main reactors, because they already have useful fittings and a conical bottom when turned upside down.
Pipework is made from 22mm copper pipe with standard bronze compression fittings. In theory it might be better to use chemical resistant plastic hose but the fittings are expensive and hard to find. Bronze has very good chemical resistance, copper less so, but the liquids do not stay in the pipes for long enough to cause a problem. We use compression fittings not solder as the solder is not chemical-resistant. Joints are sealed with high-temperature silicon sealant or PTFE tape.
The heart of the system is a multifunction pump. It is used to transfer treated oil to the reactor, to mix the reaction, and to evacuate the finished product. This inexpensive water pump has a bronze impeller and a totally enclosed fan-cooled motor, which contributes to safety (no spark risk). A system of taps allows liquids to be diverted through particular pathways.
Heating is provided through standard domestic immersion heaters. These are attached to 2¼ inch flanges that can be brazed on to the tank. In our case we have used mechanical flanges, which are held on with a screw thread.
The electrical appliances are wired with heat-resistant flex into a control centre with illuminated switches. The unit runs off a standard 13 amp socket and is equipped with a RCD plug. All metal parts are connected and earthed.
Oil preparation
Good oil preparation is essential for making decent biodiesel. In this processor we have incorporated an oil preparation vessel into the design. Waste oil is poured through a sieve into the top of the oil preparation vessel. The immersion heater in the side of the tank is then switched on to bring the oil up to 50-55°C. With the lid on and an insulation jacket, this takes around an hour. It is also possible to heat via a heat exchanger if you have a readily-available source of heat. Some biodieselers run a generator on their own fuel and use the waste heat for this.
We then allow the oil to stand by opening the valve at the bottom of the vessel to draw off the water along with many of the solid impurities which made it through the sieve. We continue to draw off the nastiest oil from the bottom until the top of the oil reaches a pre-measured mark. At this point we know we have 100 litres of oil. This oil preparation vessel has an out-pipe positioned some way up the side of the vessel. The oil that leaves the tank is not drawn from the bottom but from this out-pipe. This means that residual solids at the bottom of the tank will not be drawn out with the oil (and gets round the issues associated with not having a conical base). Where oil passes out of the tank it goes through a simple in-line filter. This is a standard and inexpensive plumbing part, which has a washable stainless steel 400-micron filter. The fuel will be filtered again at the end of the process, but this pre-filtering avoids damage to the pump and unnecessary clogging of the final filter.
Methoxide preparation
While the oil is heating up we can prepare our methoxide. This is one of the most hazardous aspects of biodiesel production, as it involves manipulating significant quantities of hazardous chemicals. Therefore it is essential to do as much of the mixing process as possible in a sealed environment. In this processor the methanol is first pumped into the methoxide vessel from the methanol storage drum. We are using a methanol-resistant hand pump made from Ryton. This screws into the methanol drum. Note the gloves, glasses and overalls. The operator is also wearing a mask, which is no protection from methanol vapour but will be useful for the next part of the process. Using the funnel that you can see in the bottom left of the picture, the weighed lye is then added slowly to the methanol. This dissolves slowly over the time it takes the WVO to warm up.
There are a number of different strategies for methoxide preparation. Because we use KOH the lye dissolves fairly easily in the methanol. With NaOH it takes longer to dissolve. In both cases it is helpful if you can give the mixture a stir. Some processors incorporate mixing paddles operated by electric motors into their design. This ensures the methoxide is well mixed but adds to the cost and the possibility of sparks near methanol. One of the best approaches is to mix the methoxide in a separate 20l can that can be closed and then gently swirled by hand. It is also possible to make a large quantity of methoxide ahead of time when using KOH. With NaOH the methoxide goes off much faster. Some people keep two barrels, one containing methanol and the other methoxide prepared to a standard concentration. By mixing the two liquids they can create methoxide to match any titration.
The reaction
The reaction vessel is where the transesterification takes place. It needs to be well sealed to prevent the escape of methanol during the reaction. It must incorporate some way of stirring the WVO and methoxide together. In this design we use a pump to mix the reaction. Other designs often use large paddle mixers to stir. However using a pump gives a good mix and avoids the need for more electric motors.
In the past we have prepared the methoxide in the reaction vessel, used the pump to mix it, and then added the oil. In this system the methoxide is sucked into to the flow of oil. This improves the mixing as the methoxide gets churned up in the pump immediately after entering the oil stream. Some small-scale set-ups use a venturi fitting to suck up the methoxide. A venturi is a fitting which uses the flow of liquid to pull a vacuum. The advantage of this is that the methoxide container does not have to be above the point of mixing: it can be safely on the floor. Unfortunately venturi are hard to make and expensive to buy.
It’s best to insulate the reactor while the reaction takes place so the residual heat from the WVO preparation is not lost. You can use bubblewrap or loft insulation, though be careful as the plastics involved are often not up to the temperatures. It is also possible to heat the reactants with an immersion heater or heat exchanger, though in this system this is not necessary or indeed possible as using the pump and heater at the same time would use more power than an ordinary 13 amp electric outlet can supply.
Methanol recovery
Having reached this stage we normally leave the fuel to settle so we can remove the glycerine. However this reactor incorporates a condenser to recover the excess methanol at this point in the process.
Many homebrewers do not do methanol recovery. There are several reasons for this:
- it has a high calorific value, and is therefore a good fuel itself – improving range and power
- it thins the biodiesel, improving cold starting, and reducing the risk of coking injector heads
- it helps preserve the fuel in long-term storage, and stops bacterial growth
We think recovering methanol is important because it will otherwise end up in the glycerine and the wash-water as well as the fuel. We spilt wash water with methanol in on some grass 2 years ago and nothing has grown there since! There are other good reasons to recover methanol:
- so that the lower rate of duty is payable on your fuel
- to use it again; financially this is a good idea, as the methanol recovered will be more valuable than the energy used to remove it
- methanol can damage the sensors in new cars
In a 100-litre batch using 20% methanol we have about 8 litres of methanol to recover.
To operate the still, the reactor is left covered and the immersion heater in the tank is switched on. This brings the tank up to 80 degrees, which is above the boiling point of methanol. The tap leading from the top of the reactor to the condenser is opened. Methanol vapour enters the condenser, where it cools rapidly, turning the vapour back to methanol liquid. This drips down into the methoxide vessel. As the concentration of methanol in the mixture drops, the temperature has to be raised in the reaction vessel.
Recovering methanol at this stage has several drawbacks. You have to make sure the reactor is very well sealed or methanol vapour will escape. It takes a lot of energy to heat that much fuel and it’s hard to control the temperature precisely enough. There is also a risk as you remove the excess methanol that the reaction will start to go backwards, turning your precious fuel into fatty acids or triglycerides.
A better approach for the homebrewer may be to recover methanol from the glycerine only, after settling (see later section titled ‘What to do with the glycerine). This works because the majority of the methanol is in the glycerine. The small amount left in the fuel comes out in the wash water. There is a trade-off to be made here between fuel quality, environmental considerations, and effort. You could of course recover methanol from both the glycerine and the fuel separately but recovery from the fuel is an extra step involving energy and effort for little return. Another way of dealing with this problem is to use less excess methanol for the reaction. This means that the reaction will not proceed all the way to completion. Your product will include unreacted triglycerides and partly-reacted triglyceride in the form of mono- and di-glycerides. This is a lower quality fuel but will work well in many more fuel-tolerant vehicles. The advantage is that there is less excess methanol to worry about and the amount which ends up in the wash water may be trivial.
As you can probably tell, we do not feel that the definitive homebrew system for recovering methanol has been designed yet. Small scale biodiesel kits on the market tend to ignore the issue completely. We are continuing research on this; it’s probably the biggest technical challenge in small-scale biodiesel production.
Settling
Now the reaction is complete we can leave the products to settle out. Glycerine settles out best if the mixture is warm, so it’s good to keep the reaction vessel well insulated. You could also decant the mixture into another vessel for settling, freeing up the reactor for another reaction. In order to separate the glycerine accurately from the fuel a conical bottomed vessel is best. In this system we have a crudely dished oil drum. When we draw off the glycerine we inevitably draw off some fuel with it. This can be recovered from the glycerine when we come to deal with that later. In large biodiesel plants the settling process is achieved with a centrifugal separator that spins the heavier glycerine to the outside of a chamber and removes it. Often the process involves many separator stages. As these are very expensive bits of kit we’ve seen smaller scale plants that attempt to use separators designed for cream separation and even a fuel/water separator from a submarine! These experiments have not been a success so we are content to use gravity and time.
Water washing and de-watering
To get a good water wash we need to maximise contact between the water and fuel without excessively violent mixing. There are two main techniques used in biodiesel plants. The first is called mist washing. This involves spraying a very fine mist of water on top of the fuel. The tiny droplets dissolve the impurities in the fuel as they sink down through it. The water ends up as a layer underneath the fuel which can be drawn off the bottom of the vessel.
In this reactor we use bubble washing. First we add about 30 litres of water dribbled from a hose down the side of the vessel. Then we drop an aquarium air stone into the bottom of the vessel attached to a small air pump. The bubbles form in the water layer with a thin film of water around them. They then rise through the fuel, the thin film of water dissolving impurities in the biodiesel. At the top of the liquid the bubbles burst and the water travels back down the same as with the mist wash. The advantage of the bubble wash is that it can be left overnight, giving a very thorough wash. The air pump uses only a few watts of power.
Air pumps and air stones are easily available from pet shops. Some air stones dissolve in biodiesel: the ceramic white stones are more expensive but worth it. Some biodieselers make their own bubblers from grinding discs or perforated copper pipe.
Older recipes for making biodiesel will tell you to avoid bubble washing as it agitates the mixture too much and causes emulsions. They may also recommend the use of acid in the wash water to neutralise soaps. These precautions are generally only necessary if your original WVO was very high in FFAs and you have tried to neutralise these with a large amount of lye. In these cases it is better to use an acid/base method to esterify the FFAs. This is covered later in this section.
Some set-ups have a separate washing vessel. This is useful as it frees up the reactor. In this system we wash in the reactor for simplicity. The reactor vessel already has an immersion heater so we can use this to dewater the biodiesel if necessary. Some systems use zeolite to de-water. This is an inert substance that absorbs water, a bit like cat litter. You can put a bag of zeolite in the fuel to dewater it. The zeolite needs to be dried to regenerate it. This can be done in the sun or in an oven.
Another de-watering method is to bubble air through the fuel (bubble-drying, not to be confused with bubble-washing, but potentially using the same equipment). This may cause oxidation of the fuel however. On the biodiesel Internet forums the debates rage backwards and forwards and there is no ‘definitive’ technique for drying.
Fuel dispensing
As the finished fuel is still in the reactor we can use the pump to dispense it into storage vessels or directly into the vehicle. On the way out it passes through a 5 micron filter. In this case we are using an ordinary water filter. You can also use vehicle fuel filters. It helps to use a pump at this stage as 5 microns is quite fine and gravity filtering can take a very long time.