Sunday, September 18, 2022

Refining Oil, Making Varnish


Testing to see if the varnish is done.

I originally meant the recipes here to be the second half of my Mortise & Tenon article, but it was way too long and didn't really work. So I'm including them here. I recommend reading the article for context--a lot is explained there, like why turpentine is preferable to mineral spirits, and why you should only use Diamond G brand.

There are two things that need to be understood before trying to make varnish. First, it’s dangerous. High temperatures are required, the materials are flammable, and the fumes can be hazardous. Never cook inside, and never cook over an open flame. I recommend setting up an electric hotplate on sawhorses, at a safe distance from the house, the shop, the car, or the cat. Wear long sleeves, gloves, safety glasses, and a mask or respirator. Have a fire extinguisher close by. Above all, think about what could go wrong, and plan accordingly.

Second, remember that cooking varnish is like baking bread: Sticking the loaf in the oven is the easy part. Success or failure is determined by all the preparatory steps like feeding the starter, kneading, and rising the dough. Varnish is similar: Preparing the oil and resin is most of the ballgame; if you do those steps correctly, the rest is a piece of cake. Or bread.

The first step for any varnish is to prepare the oil. If you don’t want to do any preparation, you can use an alkali refined oil like "varnish maker's oil" from Wood Finishing Enterprises. I don’t use this oil—my results have always been better with oil I refine myself, and the resulting varnish will dry faster (unless you add metal driers), but if you don’t want the hassle of refining oil, skip this section.


 Refining Oil

The following is a pretty simple method of washing oil. It takes a relatively long time, but requires very little effort. For a comprehensive look at oil washing, see Tad Spurgeon’s book Living Craft, available as a download in the previous post


 Stuff you’ll need:

  • Cold pressed linseed oil. Avoid most oil sold as flax oil in health food stores, and avoid all bargain basement oils available online. The Swedish cold pressed oils (Ottosson, Allbäck) are very good. I especially like Ottosson, but it has to be raw; do not use boiled linseed oil. 
  • Water. Tap water is usually fine for this method, especially if it’s relatively soft, but if it’s hard it may be too aggressive. If you have difficulty getting the oil and mucilage layers to separate, try spring or distilled water.
  • Canning or kosher salt. Look at the label and make sure it’s just salt, with no extra ingredients like anti-caking additives. Diamond Crystal brand is good and cheap.
  • Screw top jars (Ball jars work great) and a bulb baster.

Washing: Fill a jar three-quarters full with two parts water, one part oil, and a cup of salt for every 16 ounces of oil. Screw the lid on tight and shake vigorously for one to two minutes, then let the jar sit undisturbed, on a sunny windowsill if possible. Perform the shaking procedure three times per day.

After the first shaking, the oil and water will separate within a few minutes. But after several rounds of shaking, a mucilage layer will form in the middle. With more rounds of shaking, the mixture will separate more slowly, and the mucilage layer will grow larger. This may take a few days, or it may take a week: It depends on the oil and the water. If you reach a point where the mucilage takes a long time to separate, change the water. If not, change the water after a week.


Oil, mucilage, and water.


To change the water, use a bulb baster to transfer most of the oil to a clean jar. To remove the last couple ounces (you can’t do it with the baster), gently pour water into the jar until it’s completely full. Wait a few minutes for the remaining oil to separate,, then use a spoon to remove as much of the remaining oil as you can. Try to avoid picking up mucilage, but don’t worry if you get a little–you don’t need to be really careful until the final wash.

An alternative method of removing the oil is to simply freeze the jar, then pour off the oil. There are pros and cons to each method: Freezing takes longer and not all the mucilage may freeze, but it’s much less fuss. Jars may crack after repeated freezing cycles, so you need to inspect them carefully.

After you’ve removed the oil, repeat the whole process. The second wash may require fewer rounds of shaking to reach the point when the water needs to be changed. Then do a third and fourth wash, but do not use salt for these last two washes. After the final wash, transfer the oil to a clean jar, being careful to avoid picking up any mucilage. About a third of the oil will be lost during the washing process.

Optional Pre-wash: Before the first wash, you can do a pre-wash with water and vinegar. The vinegar helps the mucilage separate more easily in subsequent washes. To 16 ounces of oil, add 3 ounces water and 1 ounce apple cider vinegar. Shake together for two minutes, allow a few hours or overnight to separate, then remove the oil. 

Safety note: Always dispose of rags soaked with linseed oil safely--there is a small but real chance they could catch fire as they dry. I like to spread them out, weighted down with pebbles, until they dry thoroughly.

Troubleshooting: If the mucilage is separating very slowly, try gently spinning the jar for about thirty seconds. Repeat every few hours. A stubborn emulsion can sometimes take several days to fully separate. If it’s not separating at all, put the jar in the freezer overnight (make sure it’s not completely full, and don’t screw the lid on too tight). Take it out the next day and let it thaw completely; this will break the emulsion. After the emulsion breaks, change the water; if you’ve been using tap water, try distilled or spring water instead.

Clearing: After the last wash the oil is clean, but still turbid due to residual water. The safest and easiest way to clear the oil is to cover the jar with cheesecloth, and let it sit in front of a sunny window for a few days. When the oil is completely transparent, it is ready to use–just decant it to another container, leaving any residual water behind. 


You can also clear the oil with low heat, but be aware that this is potentially quite dangerous. You must heat the oil slowly to the temperature of boiling water (212°F/100°C); do not exceed this temperature by more than a few degrees. If you heat the oil too fast or too hot, the boiling water, trapped on the bottom, may erupt in jets, causing serious burns. You will need to monitor the temperature carefully, and if there is much water, it can take a long time to evaporate. Even when done correctly, this method is messy–the water makes the oil spatter, covering everything within a few feet with a fine, greasy spray. I much prefer the first method.

When the oil is cleared, you will have a raw, refined oil that greatly exceeds in quality anything you can buy. It can be used anywhere you’d normally use raw linseed oil, but it will dry considerably faster. However, there is still one more step before we can use it to make varnish.

Breaking: This step eliminates impurities that cannot be washed out, and can coagulate when the resin and oil are combined. 


Stuff you’ll need: 

  • Electric hot plate. For this step, almost any cheap hot plate will work. 
  • Thermometer. Infrared thermometers are cheap, can read very hot temperatures, and don’t require immersion, which comes in handy when melting resins. Immersion thermometers are more accurate, but good ones are expensive and can’t read the high temperatures needed for some resins. Get both if you can afford it; if not, get an infrared.
  • Stirring rod (glass or stainless steel is best). 
  • Saucepan. For oil, and for rosin varnishes, a cheap, stainless steel 1 quart pan will work well. 

Heat the oil (remember, outside, on an electric hotplate) to 450°- 475°F (230° - 250°C), and hold for thirty minutes. If you did not use heat to clear the oil, watch carefully as the temperature reaches 212°F/100°C, and if you see bubbles, hold the temperature until the bubbles disappear. Monitor the temperature carefully; if you go much above 500°F, linseed oil starts to break down, so avoid getting it too hot. You may actually see the “break,” which will appear as an opaque cloud that suddenly dissipates; but you may not–it happens pretty quickly. Safety note: Don't walk away while the oil is heating. The temperature can climb pretty quickly, and if it climbs high enough, the whole thing could catch fire.  

After heating, the oil will dry faster, and it will also appear much darker than before. However, if you put a drop on a piece of glass and spread it out, it will be completely clear. If you apply it to wood, it will look no darker than before it was heated. The color of oil is fugitive; how it looks in a jar has almost no relationship to the color of a thin film.


 Limed Rosin Varnish

This varnish is a great way to get started–it’s not difficult to make, and the equipment needs are modest. Historically, anywhere from 2% - 8% lime was used. The more you add, the harder the resin becomes; however, you can add so much that the resin will not unite with oil. Furthermore, brands of rosins differ in terms of how much lime they will absorb. The recipe below adds roughly  3% - 4.5% rosin; you can modify the percentage after gaining some experience.


Stuff you’ll need: 

  • A scale for weighing ingredients. It should read in 0.1 ounces and/or grams.
  • A jelly roll pan, or something similar, to pour hot resin into. 
  • A small piece of glass or marble (steel or even wood will work in a pinch).
  • Rosin (Diamond G rosin is excellent) 
  • Powdered lime (calcium hydroxide) 
  • Turpentine (use only Diamond G brand), or mineral spirits (if you must use mineral spirits, Real Milk Paint brand is best, by far). 

Heat four ounces of rosin over moderate heat until it’s completely melted. Don’t stir until the rosin is liquid. Increase the heat and bring the rosin to 450°F/230°C. Add 0.12 - 0.2 ounces of lime (by volume, this is roughly 3-4 level teaspoons). Add the lime one teaspoon at a time while stirring. As you add lime, gradually raise the temperature to 525°F/275°C. Wait until each teaspoon is absorbed before adding another. It can be helpful to cover the pan and let the rosin simmer for a while, but don't cover the pan with an airtight seal: All resins give off volatile vapors that need to be let out.

When all the lime is absorbed, pour the hot rosin into the jelly roll pan. After it has cooled, use the putty knife to scrape it out (it should come out easily), and weigh it. If you started with four ounces, you should have approximately three left. Proceed to the next step, or bag and store the rosin for future use.

Measure out your oil, using whatever oil:rosin ratio you choose. For a brilliant, glossy varnish, try 1:1 or even 1:2 (that’s one part oil to two parts rosin). For furniture, a more typical ratio is 3:2 or 2:1. Remember, more oil means more durable, flexible, and satin; more resin means harder, glossier, but more likely to chip. I usually use a 2:1 ratio, so if I had three ounces of resin, I would use six ounces of oil (that’s six ounces by weight, not volume).

Reheat the limed rosin to approximately 450°F/230°Ç. Gradually add the oil, stirring as you add it. Let the varnish cook for at least 15 minutes, then place a drop on a piece of glass (or whatever you have). If it looks cloudy rather than clear, it is not ready. After a few seconds, dip your fingertip into the drop, and lift it away. Do this a number of times. If the varnish is ready, it will pull away from the surface, forming a thin filament (see photo at the top of the post). This is called the “string” test. The strings indicate that the oil and resin have combined, and will not separate upon cooling. If the varnish feels greasy, it’s not ready. Keep checking every five to ten minutes. When it starts to feel sticky, it’s close to done–in fact, a varnish that is merely sticky without forming strings can be quite acceptable. Ideally though, you will keep cooking until you get strings of at least an inch or two. Old varnish texts mention strings of six, eight, even ten inches, but when you are a beginner, it’s wise to declare victory if you get any string at all. How long this takes depends on how much oil is in the mixture; a 1:1 varnish will be done sooner than a 2:1 varnish.

When you’re satisfied that the oil/resin mixture is sufficiently cooked, turn off the heat and get ready to add turpentine. Some old varnish texts suggest adding turpentine when the oil/resin mixture is slightly above turpentine’s boiling point (309°F/154°C). However, this is quite dangerous, and I do not recommend it. To reduce risk, wait until the temperature drops below the boiling point. Safety note: Some people may prefer to use mineral spirits. If you do, wait until the temperature drops below 250°F/120°C.

Determining how much turpentine to add is complicated. When I make a 2:1 varnish, I add approximately 40 % of the weight of the solids, e.g. if I start with three ounces of resin and six ounces of oil, I’ll add 3.6  ounces of turpentine (by weight). However, the amount depends on so many factors that it’s impossible to generalize. A 2:1 varnish will require less turpentine than a 1:1 varnish, but it depends on the consistency you want. A sensible approach is to add a conservative amount at first. Stir it in thoroughly, let it sit for a couple minutes, then put a spoonful of the varnish into a small measuring cup. Let the spoonful cool to room temperature, then swirl it around, stir it, and determine if it’s thin enough; if not, add more turpentine.

With no siccatives (driers), this varnish should dry in about 24 hours. Sunlight speeds drying time , and violin makers often use UV boxes that mimic sunlight. I don’t have much experience with siccatives, but Don Williams suggests that you decide what’s the longest drying time you can live with, and add only enough siccative to achieve that result. That’s because siccatives are not only toxic; they also change the nature of the varnish film, and can cause all sorts of problems if you add too much. If you want to experiment, I suggest getting a small bottle of high quality Japan drier from an art supply store, not the big metal can from the hardware store, which contains enough to last several lifetimes. 

When the varnish has cooled to 140°F/60°C, pour it through a paint filter into a clean jar. I recommend letting it sit for at least a week before using.

Here's a video of a guy (not me) making this varnish. He does a fine job. He mixes his lime with a little water, which is unnecessary, but doesn't hurt anything either.



Rosin, Sandarac, and Mastic Varnish

Achille Livache, author of one of the classic texts on varnish making, gives a recipe that uses these three resins. It’s a great introduction to “running” resins and combining them into a complex whole. I use a lot of this varnish in my shop. The sandarac adds hardness, and the mastic is a plasticizer that makes the varnish less brittle.

Stuff you’ll need:

  • Sandarac and mastic resins.
  • Limed rosin from the previous recipe.
  • A second hotplate. This one should be in the 1300-1500 Watt range. Two-burner hot plates are also available, and not very expensive.
  • A second saucepan. This one should have thick sides and bottom. Tall and narrow is good; wide and shallow is not. I use a Corning Vision Ware 1 quart saucepan, which can be found second hand at reasonable prices. It’s very heavy, temperature-resistant glass–don’t use thin, ordinary glass or ceramic. Cast iron works, but can darken varnish over the long term, even if you season the pot regularly. Heavy stainless steel is great, but expensive.

The first step is to "run" or “crack” the sandarac. Like amber, copal, and some other resins, sandarac must be melted completely and cooked until various impurities boil off; otherwise, it will not combine with oil. The technical term for this is “destructive distillation.” Safety note: This process generates a lot of nasty fumes, so wear a mask or respirator. The sandarac will expand a lot in the pan, so don’t fill the saucepan more than a quarter to a third full. I recommend cooking four to six ounces in a one quart saucepan. Less than four ounces, and you're liable to burn the resin.

Heat the sandarac at a moderate temperature until it begins to soften, then increase the heat to high. You can partially cover the pan to hold in heat, but watch very carefully to make sure it doesn’t overflow. Remove the lid periodically to let the fumes out. Don’t stir, but when the resin puffs up like a soufflé, puncture the surface to let the gasses escape. Do this carefully, so that hot liquid doesn’t squirt out and burn you! Resin will adhere to the sides, but eventually the saucepan will get hot enough that the resin will run to the bottom of the pot. 


Sandarac starting to bubble.


You may find it helpful to check the temperature occasionally; the target temperature is 575°F/300°C. But visual clues are more important. When all the resin is liquified and the surface is gently simmering, with only small bubbles and minimal froth, the sandarac is ready. Pour it out, cool it, and set it aside for weighing. You should lose about a third of the resin in the running process. 


This sandarac has been sufficiently "run."


The proportion of the resins in this recipe is flexible (Livache didn’t list any). For a 2:1 varnish, I use 2.5 oz. limed rosin, 2 oz. sandarac, 0.5 oz. mastic, 10 oz. linseed oil, and approximately 6 oz. turpentine (again, all measures are by weight). But feel free to scale or adjust the proportions.

Combine the rosin and sandarac in the pot and heat until you have a thin liquid that runs easily off your stirring rod. Separately, heat the oil to roughly 300°F/150°C. Pour a little bit of oil into the resin and stir. If the mixture congeals, keep stirring and don’t add oil until it’s liquid again. Then gradually add the rest of the oil. Check the temperature: You want to cook the varnish at 450°-500°F (230°-260°C). As in the previous recipe, cook to a string, turn off the heat, and add most of the turpentine when the temperature drops below 300°F/150°C. Add the mastic right after the turpentine. Stir for a few minutes, then check for consistency and add more turpentine if needed. When all is right, cool the varnish to 140°F/60°C, then filter and bottle as before. Wait at least a week before using.

Troubleshooting: If the varnish congeals when you add oil, or if you discover lumps of undissolved resin when you pour out the varnish, you probably did not run the sandarac sufficiently. The higher the proportion of sandarac in the mix, the more it needs to be run. 

 If you make this concoction successfully, you’ll have mastered the essential skills of varnish making. Running copal or amber is a little more difficult than running sandarac, but the principles are the same. If you want to go further, check out Livache's book, or some of the additional sources mentioned in my M&T article. Experiment, have fun, and be safe.



  1. Thanks for sharing. As a chemist, I'd be inclined to use a separatory funnel do remove the water directly from the bottom but most folks won't have that. If one had a gravy strainer that may be another simple way to do. That mucilage layer isn't unique to just this oil. I've done thousands of organic solvent with salt solution extractions and almost all of them had a mucilage layer. All the tricks you mention are the same ones I've used over my career to help them separate.

    1. I have never studied organic chemistry. What is going on with the salt water and oil? What is the composition of the muscilage ? This seems like it may be a way to make my own faster drying linseed oil.
      Thanks Dan

  2. Thanks for the comment. For this post, I wanted to keep the equipment and techniques as simple as possible, since most folks don't have experience with it. The funnels are expensive, but you're right that a more efficient way of separating would be nice. I've got a couple gallon jars with metal lids, and I was thinking about trying to install some kind of valve in one of them (and then inverting the jar while it separates, obviously).