When I blogged about DIY mineral water last year it was mainly a theoretical exercise since I didn’t have the required salts at hand. My experience was limited to adding some baking soda (sodium bicarbonate) to water before carbonation. Luckily Paul Hinrichs tested the calculator! In the meantime I have purchased the required salts and with several kilograms in total I’m probably well stocked for the next decade! Based on the output from the calculator, I mixed the salts required to clone San Pellegrino, added water and carbonated the mixture. And the good news is that it works! The water tastes great and I’ve been enjoying cloned mineral waters every day now for the last couple of weeks.

Some changes have been made to the mineral water calculator (Updated! – scroll down for download options) since I last posted:

• a simpler worksheet more suitable for printing has been added
• more mineral waters have been added to the database, covering TDS (total dissolved solids) levels all the way up to more than 4000 mg/L
• potassium bicarbonate, magnesium chloride and calcium nitrate are made optional and can be left out if desired (it’s still a little unclear to me to what extent these can be detected at the typical levels found in mineral waters)
• one can now chose between using either hydroxides or carbonates of calcium and magnesium, depending on availability (it should be noted however that some waters high in bicarbonate may require the use of the hydroxides – not quite sure about this though)

A spoon full of mineral salts is required for the preparation of 1 liter of San Pellegrino mineral water.

Instructions for how to prepare the mixture of salts
Start by chosing the mineral water you want to clone from the drop down list. My advice would be not to start with the waters having very high levels of total dissolved solids (TDS) (except Kessel and Vichy Saint-Yorre since sodium bicarbonate dissolves easily). Aim for a TDS in the range 200-1500 mg/L (the list of all mineral waters in the rightmost worksheet is viewable and sortable). At the lower end you may not detect much mineral taste at all. At the higher end the mineral taste becomes quite pronounced. You can click the check boxes to include/exclude some salts. If known enter the composition of your tap water (your local water company should be able to give you these figures). I suggest that you weigh out the salts for 10 or even 100 liters, otherwise the amounts of salts will be in the low milligram or microgram range, requiring expensive lab scales. Mix the salts well. It may be god to start by mixing the salts present in the lowest concentrations first to ensure a homogeneous mixture.

How to make a cloned mineral water
Weigh out the approximate amount of salt (prepared as described above) needed for the amount of water that your carbonation vessel holds. At this point it’s doesn’t need to be very accurate, so if you have weighed it once you can simply need to remember which spoon you used and the size of the heap. Note that the different mineral salts vary greatly in density, so you should calibrate the heap used for each mineral salt mixture. Add the salt to the carbonation vessel and fill it up to the mark with water. The water will now turn opaque and whitish as the salts are suspended in the water (see picture above). Carbonate carefully and, depending on whether the water is high in carbonation and/or bicarbonate, try to hold the carbonation pressure for a couple of seconds extra before letting the pressure out. This allows a little more carbon dioxide to dissolve. Screw on the cap immediately to prevent the carbon dioxide from escaping. In some cases it may be necessary to repeat the carbonation step after some hours. Once the salts have dissolved (i.e. the water becomes clear) you can enjoy your very own home-made mineral water!

Several of the mineral salts have are not soluble in tap water, hence the opaque look to the left. After carbonation however they dissolve rapidly.

So far I’ve made up the salt mixtures for San Pellegrino (total dissolved solids, TDS: 1109 mg/L) and Gerolsteiner (TDS: 2488 mg/L). The first works like a charm, even when all salts are added simultaneously. This is possibly due to the high amount of sulfates which seem to dissolve more easily. Gerolsteiner is more tricky, partly due to the high TDS and the low amount of sulfate. I made it using carbonates instead of hydroxides, hoping that this would require addition of less carbon dioxide to neutralize the base. But after two days and 2-3 extra additions of carbon dioxide the salts had still not dissolved completely and this puzzles me. I certainly need to repeat this experiment. Darcy O’Neil states in Fix the pumps that the order of addition does matter. I’m not quite sure if that really is the case as most of the salts have a very low water solubility to start with, and the carbonic acid is the reason they dissolve. But maybe there is something I’m overlooking here? It could be that Gerolsteiner is easier to do with hydroxides, but Paul Hinrichs also had some trouble getting all the salts to dissolve for Gerolsteiner.

Some of the salts may be tricky to obtain, but the synonyms and links to Amazon below may be of some help:

• CaSO₄·0.5H₂O = Plaster of Paris (check availability from Amazon)
• MgSO₄·7H₂O = Epsom salt (check availability from Amazon)
• CaCO₃ = Chalk (check availability from Amazon)
• NaHCO₃ = Baking soda
• NaCl = Table salt
• Mg(OH)₂ = Milk of magnesia (check availability from Amazon)
• Ca(OH)₂ = Slaked lime, pickling lime, CAL (check availability from Amazon)

Before you head of to Amazon or some other place to order salts I should probably add some words of warning: make sure that the source you find is suitable for consumption! Some technical qualities of mineral salts may not be intended for food use, for instance due to the presence of heavy metals or other contaminants.

Note that some of the salts are available with varying amounts of crystal water. If you use other salts than those specified (i.e. anhydrous salts or salts with more crystal water) the molecular weights in the spreadsheet need to be adjusted for this. I guess that if you are familiar with the concept of crystal water, you’ll easily figure out the correct molecular weight and how to update the calculator according to the specific salts you chose to use.

Screen shot of the simple version, best suited for printing (see below for download options):

Screen shot of the complete version (see below for download options):

Calculator download options
Version 5 (latest update)
Excel: mineral_water_calculator_v5.xlsx (44 kB)
Open office: mineral_water_calculator_v5.ods (44 kB)

Version 4 (the version originally provided with this blog post – contains errors)
mineral_water_calculator_v4.xlsx

Mineral waters included
Mineral waters included in the database that comes with the calculator: Acqua Panna, Antipodes, Apollinaris, Aquarel Birken, Artificial mineral water, Badoit, Borsec, Burton (beer brewing), Calistoga, Carola Rouge, Contrex, Dorna, Evian, Farris, Fiuggi, Gerolsteiner, Harghita, Hassia Sprudel, Henniez, Kessel, London (beer brewing), Mountain Valley Spring, Munich (beer brewing), Neuselters, Perrier, Pilsen (beer brewing), PurPur (coffee brewing), Rosbacher Klassich, Saint-Yorre, Salvus, San Benedetto, San Narciso, San Pellegrino, Selters, Tea brewing (max), Tea brewing (min), Tesanjski Dijamant, Ty Nant, Vittel, Volvic, Voss, Waiwera. And you can easily add data for other mineral waters. The websites mineralwaters.org, finewaters.com and Mineral water atlas of the world have data for several hundred waters available. And if you have a bottle of your favourite mineral water at hand you only need to check the label to find the required input for the calculator.

The products of the Maillard reaction provide tastes, smells and colors that are much desired and lend their charachteristics to a variety of foods. In this post I will focus on the factors that influence how fast the Maillard reaction proceeds. And more specifically I’ll give examples on how the Maillard reaction can be speeded up. This is not about fast food, nor is it about saving time. It’s more about controlling the browning reaction by speeding it up or slowing it down in order to get a desired end result.

The Maillard reaction is, to put it simple, a reaction between an amino acid and a sugar (there’s more on the chemistry at the end of the post). To speed it up you can do one or more of the following:

Chances are you have already utilized this in the kitchen without knowing. When eggs or milk are used for glazing, they act as a protein source for the Maillard reaction, giving a nice brown color. Milk also provides lactose which is a reducing sugar. You’ve probably also observed that temperature does influence browning. Water content is indirectly related to temperature – as long as there is water present, temperature will stay below 100 °C. But once the bread crust dries out the conditions are just right to get the Maillard reaction running.

The same principles are applied to microwaveable pies. The microwaves primarily interact with water and hence only bring the temperature up to the boiling point of water. In order to get sufficient Maillard productcs at these temperatures reducing sugars and amino acids are added to the crust (as exemplified in this patent where dextrose and whey solids are used). Not so surprisingly there is also a patent on how to avoid excessive browning in cookies which calls for addition of a polycarboxylic acid ester to lower pH and hence slow down the Maillard reaction.

Pretzels are an extreme example of how the Maillard reaction can be tweaked. Before baking the pretzels are brushed with lye, a dilute solution of sodium hydroxide, which is very basic. The high pH speeds up the bottleneck of the Maillard reaction (see end of post for details).

A pinch of baking soda can bring out a new taste dimension when browning onions

Another basic ingredient found in most kitchens is baking soda (sodium bicarbonate, NaHCO₃). It’s used as a leavning agent which requires addition of an acid to function. Since it is a weak base, it can be used to increase the pH and hence the speed of the Maillard reaction, for instance when browning onions. This basic task, which isn’t always so easy after all, benefits greatly from a pinch of baking soda (and surprisingly it seems that this hasn’t been done before!). To illustrate this I’ve made a time lapse video of chopped onions being fried with and without baking soda. The frying took 11 min, but things are speeded up about 10x.

Samples taken throughout the experiment are shown in the picture below. Even after 4 min there is a visible difference. After 11 min, the small addition of baking soda has yielded onions which taste remarkably sweet with strong caramel notes, compared to the control which tastes like fried onions.

Another example of how baking soda is used to speed up the Maillard reaction is dulce de leche, a popular sauce/caramel candy in Latin America. It’s made by slowly boiling sweetened milk. Baking soda is not a required ingredient, but is often included. The baking soda gives dulce de leche a darker color and also contributes to the flavor.

Photo by audinou from flickr.com.

It should perhaps be added that baking soda is frequently used in Chinese cooking, for instance in tempura batters and marinades. Once there, the baking soda will certainly speed up the Maillard reaction, but it also affects the texture of meat – I’ll have to return to that topic later.

To round of this post I will briefly touch upon one of the reasons why pH influences the Maillard reaction. The first step involves a reaction between a reducing sugar (depicted as R(C=O)H) and an amino acid (depicted as R’NH2) followed by loss of water to yield a Schiff base. The Schiff base rearranges to the Amadori product (not shown). Of these first steps the formation of the Schiff base is the bottleneck (rate limiting step). The reactivity of the amino acid is influenced by the pH. A simplified reasoning goes like this: At low pH the amino group is protonated, yielding it less nucleophilic. At higher pH, the nitrogen becomes more nucleophilic and at very high pH the amino group can even be deprotonated. It should be noted that the fate of the Amadori product is also in large determined by pH and hence pH will affect more than just the rate, but this is far beyond the scope of this blog post.