Making Belgian Candi Sugar

The Belgian Candi Sugars
Left: Attempt #1, Right: Attempt #2

With the Krampus Kristmas Ale transferred to the secondary & spiced my thoughts have turned to the next big beer - probably not to be brewed for a few months (some stouts & porters are in my new future) - but I like to think ahead.

One style I've always enjoyed, but rarely brewed, are the strong Belgian (trappist-style) ales. With lots of fruity esters and spicy phenolics, these beers are flavourful & well balanced. Their high-alcohol & dry finish makes them easy to drink, with the dryness deceptively hiding their strength.

The key to making these beers strong (7-12% alcohol) with a dry finish is the use of Belgian candi sugar - beet sugar which is heated & otherwise treated to create a candi (or syrup). Depending on how it is prepared, this candi can be anything from a light-amber with little flavour through to near-black candis with flavours of coffee, chocolate and dark fruits (e.g. plums & dates). It is the cost (often over $16/kg), and difficulty finding this product in Canada which has largely held back my ability to brew & explore Belgian strong ales.

So imagine my joy when finding out that it may be possible to produce Belgian candi sugar at home. Armed with one youtube video & a short write-up I embarked on my first experiment. After nearly two hours of heating, with little colour formation, I started reading further. It turns out I - and many other brewers out there - are doing it wrong!

Below the fold is my attempt to fix my mistake, and perhaps establish a method others can employ...

EDIT: I have solved some of the crystalizaiton issues people have been reporting when making candi sugar. Details can be found in this blog post.

---

What Makes this Stuff Special?

A lot has been written elsewhere about what makes this sugar special, so I'll give the coles notes version. During the preparation of this candi, beet or cane sugar (both being nearly-pure sucrose) is heated under conditions which do two things:

1. "Invert" the sugar - a fancy way of saying it breaks the sucrose (which is made of two simpler sugars - glucose & fructose - attached together) into individual molecules of fructose and glucose, and
2. Reacts these "inverted" sugars with amino acids to form the various colour and flavour compounds that are desired. This occurs through Maillard Reactions; the same reactions which develop colour during long boils, decoction mashing, and during malting.

---

A Warning

Danger: 125-135C sugar
+ water = pain!

I would hope this goes without saying, but this is potentially dangerous. My instructions here are based on my own experience, which means that they are far from safe or perfect.  If you follow these instructions, you do so at your own risk.

Why is it dangerous - aside from the obvious (working with a hot sticky substance that is near-impossible to get off your skin before it transfers all of its heat into your skin), we are also using small additions of water to control the temperature of the sugar mix.  The high heat of the sugar mix - between 25C and 60C over the boiling point of water - converts these water additions near-instantly to steam; leading to globs of molten sugar being propelled rocket-like through the air - most often landing on some piece of exposed skin.

So if you're going to do this, please be careful. Like bacon, making this in the nude would be a very bad idea.  Wear shoes (I'm speaking from experience here), and some sort of eye protection would not be untoward. Heat-resistant gloves (oven mitts are not great, as they kill dexterity) are a must.  This is something to do after the kiddies are safe in bed...and is not something to do after enjoying a few pints of homebrew.

---

How Are Brewers "Doing It Wrong"?

The process most brewers have attempted has been largely stolen from conventional candy-making methods. While this produces a good candy, it doesn't achieve the desired end point. The process is, in a nut-shell:

2 hours of 125-135C = light amber.
WTFerment is up with that?

1. Mix table sugar and water, using the minimal amount of water needed to dissolve the sugar (an astoundingly low 250ml per 1kg of sugar (roughly 1 cup per 2 lbs). Heat on the stove - a small amount of acid (a tablespoon of lemon juice, or 1/4 tsp of cream of tartar) is added once the sugar dissolves, to aid in the inversion process. At this time, many also add yeast nutrient to provide amino acids (proteins) required for the Maillard reactions.
2. Using a candy thermometer, heat to 125C (260F), hold at this temperature by adding small amounts of water (a few tablespoons) when the temperature hits 135C (275F). Hold at 125-135C for 20 minutes to complete inversion. You now have a light candi sugar!
3. Optional: Continue heating as in 2, to achieve a darker colour & some flavour. Be careful to not over-darken, as you are primarily charring the sugar (not conducting Maillard reactions). A medium-dark amber is about as far as it is safe to go.
4. Heat to 150C (300F) to create a 'hard-crack'; a fancy way of saying, to drive out enough water to allow the candi to form a hard/brittle product. If you skip this, you get something with the consistency of nougat.

So what exactly is wrong with this process? The answer is not very obvious - its the acid. The acid lowers the pH of the sugar, which brings the Maillard reactions to a near stop. In my first attempt I added well over double the amount of cream of tartar I was supposed to - meaning after nearly two hours of heating I was left with a light/medium-amber product, with a mild caramel-like flavour (a sign it was beginning to burn).

So what we have been making is less like a Belgian candi sugar than it is a caramelized English-style invert sugar. Good for stouts & bitters, but not really the product we're looking for when making a Belgian-style ale.

---

How Do we Fix This?

Too much acid = nothing but
mild caramelizaiton

The short answer is "we don't know - yet", as the making of this product is a closely held secret by Belgian sugar manufactures. But Ryan, author of the Ryan Brews blog, is on the case! I would recommend reading Ryan's post (follow the previous link) for all the details. Be sure to read the comments - a lot more information, plus the results of some additional trials, are outlined in the comments.

Long story short, we need to create an environment where:

1. Inversion can take place. Luckily, this is easy; heating sugar in water for 30 minutes at 125-135C (260-275F) will do it - no acid needed (sugar can act as its own acid).
2. Create a basic environment, preferably with something that won't create off-flavours. Easily done with a concentrated solution of pickling lime (aka slacking lime or calcium hydroxide) added after the inversion process. Food-grade lye (sodium hydroxide) could be used as well - but with caution! Lye is highly corrosive!
3. Provide a source of amino acids (proteins). Many (including Ryan) use yeast nutrient for this, but I have different plans...

---

My Second Plan (the first being a failure):

My plan here is simple: in the processing of beet sugar into candi sugar, the beet's proteins act as an amino acid source for the Maillard reactions. I lack sugar beet proteins, but I do have a ready source of another plants protein - namely barley, in the form of dry malt extract (DME). DME is fairly high in protein (4% minimum) and as such should provide an ample amount of protein. To alkalinize the mix, I will add a small amount of pickling lime (I don't have any, instead I'm using lye [sodium hydroxide]), and a two-stage heating process will be used to selectively drive the Mailliard and caramelization reactions I desire.

My plan is:

1. For every 500g (~1lb) of sugar, add 1 tablespoon DME
2. Dissolve DME/sugar mixture in water, at 250ml water per 1kg of sugar/DME mix.
3. Heat to 125-135C for 20 minutes.
4. Add a small amount of lime or lye (~2 teaspoons/kg of lime or 1 teaspoon/kg lye), pre-dissolved in water. This is added slowly, as to not drop the temperature below 125C.
5. Continue heating at 125-135C (maintaining temperature with the additions of small amounts of cold water). This will drive Maillard reactions, but only allow for minimal caramelization (aka 'burning').
6. If needed, more lime/lye can be added later on - caramelization reactions decrease pH, and thus can impair Maillard reactions. Heating at 125-135C is continued until a deep-red colour is achieved. This should produce a lot of the fruity & nutty flavours we expect - but we'll be lacking the roasted flavours.
7. To achieve the desired roasted flavours, and further darken the candi, the sugar mixture will be allowed to heat to between 150C and 165C (300-330F). At this temperature a lot of caramelization should occur - rapidly darkening the sugar, providing roasted flavours and the hard-beak all-in-one.
8. The sugar mixture is then poured into a silicone tray and allowed to cool.

---

The Results:

Maillard reactions in action - left: candi immediately before
lye addition, right: candi 30 seconds after lye addition.

For consistency I decided to make the same amount of candi as in my first attempt - 225g. This was measured out and 0.5 tablespoons of DME mixed in (twice what I had planned). The dry mix was then added to a pot, along with 100ml of water, and heated gently until dissolved. Once dissolved, the heat was increased to about 1/2 the stove's maximum, and the temperature held at 125-135C for 20 minutes. The sugar mix was super-foamy (see picture to the left), probably due to the high protein content. Any time the temperature got too hot, a tablespoon or so of water was added to cool the solution.

After the 20 minutes were up, I added 1ml of 1 molar NaOH (lye), and continued to hold the temperature at 125-135C. The initiation of Maillard reactions was almost instant (picture to left), with a red/brown colour change occurring within seconds of adding the lye. At the 5 minute mark it was already darker than the stuff prepared in my first attempt! As it cooked the sugar emitted wonderful fruity/nutty aromas, with a hint of malt. The build-up of colour was going slow, so after the 20 minute mark I added another 1 ml of lye - brining me up to the planned amount (described above). I continued heating until the 40 minute mark, where I guessed the Maillard reactions had progressed far enough to move onto the next step.

Colour development during the cook
(click for full-sized). Time 0 = 20min
of inversion & time of lye addition

At this point I raised the temperature further, holding at 150-165C. At this point the mixture began emitting roasted/caramelization aromas, and began to darken. The temperature of this mixture was hard to control - it heated quickly and small water additions cooled it far more than the 125-135C solution. 15 minutes into the hotter portion I decided to call it quits as I didn't want a mess of burnt sugar - in retrospect I didn't need to worry and could have continued heating for longer, and aimed for a much darker finishing product. It was at this point that I poured the mixture into a silicone bread pan and allowed it to cool for several hours.

You may notice that the candi became cloudy at the end time points - this was because sugar had begun to condense and crystallize on the sides of the pot; I started scraping this into the pot, leading to crystallized sugar in the mix. While it ruins the appearance, this isn't something to fear - its still good sugar, which will dissolve in hot wort and be fermented by yeast.

Every 5 minutes, starting just before the lye addition, I dribbled a small sample onto a piece of parchment paper (see picture to right). These acted as both a colour control, and as tasting standards.

Time	Flavour
0 minutes *	Sweet, no other notable flavours
5 minutes	My wife ate it
10 minutes	My wife ate it
15 minutes	My wife ate it
20 minutes *	Mild fruit character, some caramel
25 minutes	My wife ate it
30 minutes	My wife ate it
35 minutes	Stronger fruit character, caramel still mild
40 minutes +	My wife ate it
45 minutes	My wife ate it
50 minutes	My wife ate it
55 minutes	Tastes great - nutty flavour plus some plum; roastines is mild, caramel is modest

* = 1 ml lye addition

+ = temperature increased to 150-165C

---

Conclusions:

The Belgian Candi Sugars
Left: Attempt #1, Right: Attempt #2

The obvious conclusion is my wife liked the candi, but the other conclusion is this worked and provided a desired result - namely, stone-fruit & nutty flavours. I probably could have cooked it longer to get a stronger caramel/roast character.

One idea I may try next time is to hold the temperature mid-way between the Maillard & caramelization temperatures (i.e. 135-145C) for most of the cook; this should accelerate the process, while still keeping caramelization under control. To finish, a heat-up to 150-165C could be added to provide more caramelization & roasted flavours. The good news is that my concern about excessive caramelization was overblown, so these higher temperatures are not to be scared of. It also appears that a much darker colour is possible, without getting excessive burnt flavours.

Between the first and second batches of sugar, I now have about nearly enough candi for my planned batch of Belgian strong ale - one block of light amber, one block of medium/dark amber. Next time I'm going to try a full half-kilo batch, aiming for a much darker product, to provide the final bit of sugar needed.  The plan is simple:

1. Invert the sugar/DME mix (1 tablespoon DME per 1 kg [2.2lbs] sugar + 250ml [1 cup] water) for 20 minutes at 125-135C (260-275F)
2. Add lye (~teaspoon per 1kg [2.2lbs] of sugar) & raise temperature to 135-145C (275-290F).  Hold at this temperature until the desired colour & flavour is achieved*
3. Heat to 150C (300F) to create a hard-break; if more roasted flavours are desired, sugar solution can be held at 150-165C (300-330F).
4. Pour into a silicone pan

*Flavour & colour can be assessed by dropping a small amount of the hot sugar solution into a glass of water.  This will cool it quickly; and once retrieved, the sugar blob can be used to assess colour & flavour.