Attempting a more D2-like Candy Syrup

In my last post I discussed whether homebrewers could reproduce a syrup with the complexity -- and fermentability -- of D2. After a great deal of research, I decided to give it a go.

Dark Candi D2

Dark Candi claims that D2 is produced solely by repeatedly heating and cooling beet sugar, with no acid or amino acid additions, which would make it fairly difficult to make regular old beet sugar produce maillard reactions or rapid caramelization. Dark Candi labels D2 as highly fermentable, which makes their claims very accurate. If you caramelize sugars or brown them through maillard reactions, they are no longer fermentable. There are other Belgian syrups, made for breweries, that seem similar to D2 and provide some additional information on the make up of these sugars. They tend to be almost entirely maltose (the sugar produced by hydrolysis -- mashing) with a small amount of maltriose (a similar sugar) and fructose.

This suggests one of two things. Either D2 is created by adding enzymes to sugar beets to "mash" them and create maltose, which can then be refined down to a cleaner product, or D2 is a product of the sugar beet refinement process that is less refined than the pure white sugars. Belgium has several unrefined beet sugar products (such as beet molasses and sugar beet syrup) that are commercially available that may be the basis of D2. I'm not sure that maltose can be produced by any other process than hydrolysis so I suspect there is a complex mash and refinement process involved. Either way, it has to be less refined than the white crystal end product because you can't produce that kind of color without caramelization or maillard reactions.

One other unique piece that I suspect is that D2 is made, at least in part, by boiling sugar in a vacuum. Beet sugar is normally refined by boiling in a vacuum. It makes sense that D2 would be made the same way, since it would allow for creating different products at different temperatures that, when boiled under normal conditions, may help produce that chocolate-like flavor D2 has that most homemade syrups lack.

So those two elements being true, it is highly unlikely homebrewers will be able to produce a candy syrup like D2 because we lack access to unrefined beet sugar products (at least most of us in the US) and I don't know of many people with the ability to cook in a vacuum. However, there is hope yet. I wanted to combine my suspicions with elements that would be available to the typical homebrewer to see what I could get.

Attempted Recipe

So maltose is a combination of two glucose molecules, which made me want to use corn sugar (which is individual molecules of glucose) as the base, rather than sucrose (which is both fructose and glucose). To make it less refined and add color -- so less caramelization would be necessary -- I am adding regular cane molasses from the grocery store. I will not be attempting to create maillard reactions by adding DAP or some other ammonia source because D2 claims not to do this (and the molasses may provide some amino acids necessary to create maillard reactions). I am also adding a lot of water early on to keep lower temperatures and possibly create more precursors to flavors at higher temperatures.

This recipe creates 4 ounces of sugar in about 6-8 ounces of syrup.

3 oz corn sugar
1 oz cane molasses
Filtered water

Process

I combined 2 cups of water, 3oz of corn sugar and 1oz of light molasses (non black-strap molasses) over medium heat. I stirred every couple of minutes for the first 15 minutes. Began to boil after about seven minutes. Cooled saucepan in water for 5 minutes with an additional half cup of water. Began to sputter and spray liquid at the end of the 15 minutes. The consistency is still very watery. Taste is slightly rummy and very sweet.

Returned to heat for another 25 minutes. The long cooling process has already darkened the syrup from where it was when I started. It has a slight rummy flavor developing but right now it is mostly sugary-sweet with lighter fruit flavors. At the end of the 25 minute period it appears to have boiled off most of the second water addition. The flavor is slightly more intense. It is starting to develop some darker fruit flavors and a hint of raisin, maybe chocolate. It's hard to tell since it is still very watery. I added another half cup of water and cooled the saucepan in water for another 5 minutes.

Returned to heat for 45 minutes. It has cooked down to a thick syrup with a good dark, rummy flavor. I cooled it back down with 1/4 cup of water and a water bath for another 5 minutes. It still needs to cook down some more for the dark flavors to emerge well.

Repeated process but had to add water in 1/4 cup after 15 minutes and 1/2 cup after another 10.

Poured into pre-heated mason jar. Added 1/2 cup water to saucepan and turned up heat to boil to remove residual sugars. Boiled water down for several minutes until there was only a thin layer, so as not to add unnecessary water content.

Outcome

The resulting flavor is interesting. It is full of dark fruits, raisins, rum and some chocolate. I took it off the heat before it started to burn so it did not develop the toffee-like flavors that table sugar starts to make as it gets the really dark color as it starts to burn. The color is very, very dark, like D2 but without the burnt flavors and aromas that normally accompany producing sugar at that color. This makes me believe that I am right that D2 is either mashed, a less-refined sugar beet product, or maybe a combination of several products in the process of refining sugar beets.

I would be interested to see what would happen if you mashed sugar beets with some 6 row and boiled down the resulting wort. Also, some Asian food stores carry a maltose syrup, like honey, that is made from rice that might produce a more D2-like product. I also intend on trying to find this syrup and trying it out. Maybe mix it with a little molasses.

Ultimately I think my product in this test is closer to D2 but definitely not quite there. I would say it is 80-90% there, which is pretty good and much cheaper.