The other day, I poured three candles that were one part soft soy wax to two parts unfiltered beeswax.
Candles poured into cylindrical glasses of roughly 55 mm interior diameter (ID) would burn as expected for awhile (perhaps an hour?) and then the flame would, judging by its appearance, become underfueled while standing in a pool of its own wax. In the case of the tall narrow candle (55 mm ID, 200 mm tall), the candle self-extinguished, leaving a rather tall wick standing out of the wax once it had cooled. At that point, the wax was approximately 37 mm from the top of the glass.
The shorter candle (55 mm ID, 87 mm height) exhibited the same behavior, but did not self-extinguish.
In both cases where candles became underfueled, the color of the flame became much redder, and the overall flame was much shorter, than in an all-paraffin control candle lit at the same time. Pouring out the currently molten wax immediately restored the bee/soy blend's flame to an appearance near to that of the paraffin, but the former underfueled condition soon reasserted itself.
The third of the three 2:1 bee:soy blended candles was in a glass with the shape of a conical frustum* with a 65 mm ID at its wider, upper base. Flame behaviour on this candle was much healthier, with the same wax blend, than in the other two candles in their cylindrical glasses. It burned very nearly a perfect cylindrical section out of its frustum mold with an OD of 40 mm, but this flame continued to burn healthily long after the short cylinder had started to labor, and the tall cylinder had self-extinguished. It was starting to labor when I extinguished both shorter candles myself last night.
The cylindrical glasses had formerly held all-paraffin candles, and the conical frustum had formerly held a soft soy scented candle (brand "Pacifica", scent "Avalon Juniper", purchased at Oakland Whole Foods Market because THERE WAS A RAVEN ON THE BOX; I am occasionally predictable). The previous candle had also had a problem with self-extinguishment, but not until it was nearly depleted.
The candle burns normally until it has melted enough of the blend to get underway. Then, it consumes the soy part of the blend (with its melting point of 49-52° C) first, but the flame is not hot enough to consume the beeswax portion of the blend (melting point 66° C). This would explain the flame slowly extinguishing in a thick pool of perfectly good melted wax. The wider ID of the frustum candle allows greater airflow, permitting a higher flame temperature and thus more complete combustion.
A candle with a lower soy content will burn with a hotter flame and combust more completely.
I returned to where I had purchased the previous batch of wax and obtained more. Unfortunately, there was no more pre-chipped unfiltered wax‡, necessitating the replacement of this with pelletized filtered beeswax. Approximately 765 g of filtered beeswax pellets and 85 g of soft soy were obtained for the next batch of experimental candles, all of which would use the previously used, and cleaned, cylindrical glasses I had had. Alas, no data is available on the compositions of the glasses, and there are small variations in size and shape. Still, all are roughly 200 mm tall and with an ID of approximately 55 mm. In this picture, the glasses are prepared for pouring, with a wick suspended in each between a wick grommet and a chopstick with a bit of Tacky Wax to affix the grommet to the bottom of the candle glass. I'm sorry it's a bit blurry:
I divided the waxes, by weight, as follows1:
From left to right:
A 284 g beeswax / 0 g soft soy (100% beeswax)
B 255 g beeswax / 28 g soft soy (90% beeswax / 10% soy)
C 227 g beeswax / 57 g soft soy (80% beeswax / 20% soy)
These were melted in my double boiler over 82° C (simmering) water2. A lack of a useful thermometer kept me from noting melting points, but as soon as the waxes were completely melted, I poured into the waiting glasses. I started with batch A (100% bee) and worked my way down in purity to avoid any soy contaminant in the all-bee batch.
Here they are just after the last one had been poured:
To avoid sinkholes, in each case, as the cooling wax formed a lid across the top of the candle, I opened the lid and pushed the wax down to join the main body of the candle. Properly, I should have reserved some wax from each pour and added it at this point instead, but time did not permit. As this only affects aesthetics until shortly after each candle has been lit, this was not deemed significant for this experiment.
Also, yes, I fixed that wobbly wick on the right, there.
After all candles had been poured and permitted to cool, they were lit. Well, we lit the 100% beeswax one as soon as we could, the mixes were lit a couple hours later, and the paraffin had already been going for twenty-four hours, but the blends are all fairly freshly lit:
These are standing in a lipped cookie tray in case one of the glasses breaks. As it is presumed that the glass holding the 100% beeswax candle will break, it is further protected by being placed in a glass container (in this case, a flower vase).
From L to R:
A - 100% beeswax
B - 90% beeswax, 10% soft soy
C - 80% beeswax, 10% soft soy
D - 66% beeswax, 33% soft soy
Z - 100% paraffin
The behavior cited in the original observation is not evident in Candle D (let alone B or C) because it does not appear until more than an hour after ignition.
In the time (72 min) since that picture was taken, D is starting to exhibit the expected, "sickly" behaviour. A, B, and C have healthy flames. A, D, and Z (the paraffin) are all performing as expected. In 6-8 house (i.e., "after I sleep"), I will see what comes of a longer run.
Speaking of sleep...I do that now. G'night!
* - A word I learned in High School Geometry that I never get to use, yet which I find strangely fun. Frustum frustum frustum! 8-)
† - Almost like Real Science™, only short on tools and a usefully-sized statistical universe. However, I plan to make up for this with HUMOR. Which one? Well, not bile, by preference.
‡ - Unfortunate in that it makes for an inconsistency among my beeswax sources and that they cost twice as much, fortunately in that the pellets melt cleanly and do not leave me earwax-looking goo composed, I am convinced, of bee bits. Ew.
1 - No, those aren't corn kernels topped with coconut shreds. Or breakfast cereal and cheese. It's all wax.
2 - Well, theoretically. It's an observed simmer. My probe thermometer (Polder brand thermistor) agreed that 212° F was boiling, but allowed that an obvious simmer was also boiling. I called it a dirty, filthy liar and gave the probe an icewater bath. When the temperature of the ice water (after temperature equilibrium had been reached) was reported as 40° F, the probe was discarded for being bloody useless--fortunately, I bought this brand in part because replacement probes are available. As a result of this, the part where my Science would look like SCIENCE! for being full of numbers? Not so much. Well, that and there's a lot of other numbers I have no way to get.3
3 - But science is not about numbers--the numbers do provide a consistent way to quantify data, yes, but science is a method. Observe. Hypothesize. Predict. Experiment--and back to Observe.