The purity test: is your water and ice letting your drinks down?

I received a press release recently for a bottled water called Isbre. It is sourced in Norway from an aquifer under a 5,000-year-old glacier at the end of the Hardanger Flord in Ulvik, where it is bottled on site. The Big Thing about Isbre is its purity, showing no more than four parts per million (ppm) of total dissolved solids, which is apparently the lowest figure ever for an unprocessed water.

Purity is an interesting concept in drinks. There is a natural tendency to associate it with safety and health, and producers of spirits, especially vodka (about which there is not much else to say) frequently use the p-word, either talking about the purity of the spring water they use to dilute their alcohol, or bragging about innovative filtration methods, often involving precious metals or stones. In fact if you look at the website of Technofilter, a leading manufacturer of filtration systems for the vodka industry, they make the point that, while filtration (traditionally through birch charcoal) was once necessary to remove impurities that imparted bad flavours or odours or even made the vodka unsafe, nowadays any producer can buy in incredibly pure alcohol and all distilleries will have the means to make distilled water on site. Modern vodka filtration is more about altering the flavour and perhaps fishing out unsightly particles that may have been introduced from machinery or from additives like sugar or honey that are sometimes introduced to create a perceived smoothness.

Naturally sparkling, Apollinaris pitched itself as having medicinal benefits, "the sworn enemy of gout,
rheumatism and indigestion", in this detail from an 1876 advert. Despite the high mineral content it also
speaks of "purity" and "softness". Note also that they recommend it as a mixer for brandy, gin and wine
(click to enlarge)

This 1903 ad is pushing Apollinaris (by now threatened
by mechanically aerated waters) as a mixer for Scotch

Once upon a time drinking water that came in bottles was generally referred to as “mineral water”—the point being that you drank it precisely because of the dissolved minerals in it, either for the taste or for imagined health benefits. (According to a table on the Isbre website Evian has 300 ppm total dissolved solids, Vittel 400, San Pellegrino 990 and Apollinaris, a German spring water with a long history, which I’ve not seen for sale in the UK, a whopping 1800.) Our modern interest in bottled water may have been prompted simply by fashion and marketing, so that we want to be seen clutching the hippest water brand, or perhaps by a rise in health-consciousness and a belief that tap water is not safe. (There have been scares in recent years about the levels of oestrogen in mains water, for example, or worries about fluoridation, although in fact UK tap water has to pass more stringent tests than bottled water does.)

I live in south-east London and my tap water is pretty tasty (not true of some parts of the UK I have visited). It is a hard water with, on average, 261 ppm of calcium carbonate, which comes from the chalky composition of local aquifers. (This can tend to leave lime scale deposits in your kettle or on the showerhead, but it makes the water quick to wash away soap and apparently forms a film on the inside of old lead pipes which prevents the metal from leaching into the water supply.) The UK Drinking Water Inspectorate regularly tests tap water and apparently London’s is the best in the country (mind you we are talking about a 99.98% pass rate compared to 99.94% in the worst-performing area, so there is not much in it). You can see a full chemical analysis here.) No fluoride is added but the total dissolved solids are about 350–400, compared to Isbre’s 4.

SW4 gin diluted half and half with tap water, Isbre and distilled water

All of which got me wondering about water in booze—not the water that is used to dilute spirits to bottling strength (over which I have no control), but water that we might add at the point of preparing a drink. Many people believe, for example, that adding a small amount of water to malt whisky helps unlock flavours and aromas. And even if you don’t add water, any drink that you serve on the rocks will be diluted as the ice melts.

I gather that in Japanese whisky bars it is de rigeur that if you add water to your whisky you use only water from the same spring where the whisky is made. Camper English of Alcademics has done some interesting experiments diluting different Scotch whiskies with spring water from different parts of Scotland. The general result was that water from a particular region brings out flavour characteristics associated with whisky made in that region.

But the test here today is about the benefits of purity in water. I decided to put Isbre* up alongside London tap water. And to push things to the other extreme, I also got hold of some commercially produced distilled water.**

The same test using Nikka Taketsuru Pure Malt whisky

Tasted neat, the three waters seem remarkably similar to me, though the purity of Isbre does give it a softer feel and a perceived sweetness to the finish. The hard tap water, particularly after it has been in the glass for a while, has a flatness, a hint of something sour or metallic. (I expected the distilled water to taste much the same as the Isbre, but to me it actually tasted less pleasant, with a slight bitterness to the finish.)***

I try diluting some SW4 gin (the high-strength 47% version), half and half with water. I don’t normally drink gin this way, though my father-in-law likes his gin half and with water and a dash of bitters. It’s quite a revelation, actually, because the botanicals in the gin are loud and proud, with the high, resinous juniper joined by orange sweetness and floral, woody notes, and a bitterness towards the end. To my surprise there was quite a noticeable difference between tap water and Isbre, with the former seeming to have a heaviness or roughness. Gin and Isbre, on the other hand, felt lighter and more ethereal on the tongue and, moreover, you had a sense of perceiving the botanicals more clearly and vividly. The distilled water gave the same effect, and I honestly couldn’t detect a difference between that and Isbre.

Ice cubes made from (left to right) tap water, Isbre and distilled water. As you
see, the purity of the water does not have an effect on the clarity of the ice

I try the same thing with Nikka Taketsuru Pure Malt Japanese whisky, again half water, half whisky, and get the same result. The tap water gives a hard, flat quality, compared to the sweet smoothness of Isbre. Distilled water is likewise smooth and light.

What about ice? If you pour yourself a whisky on the rocks and sip it slowly enough for the ice to melt, then you have diluted your drink quite a lot by the end. I always regret doing this, because ice made from my tap water leaves a gritty deposit in your drink when it melts. I make some ice from Isbre and try the same thing and, hey presto, no grit.

Nothing to see here: vodka with cubes of the three ices. They look identical until…

But if you think that ice made from pure water will be crystal clear, think again. The clarity of ice is affected far more by how fast it freezes and at what temperature (see this explanation). I make up trays of ice cubes using Isbre and distilled water and they look the same as ice cubes made from tap water.

I start off with vodka on the rocks, taking 25ml of Ketel One and adding two cubes each of the three types of ice. As you might expect the results are essentially the same as when we simply added water: at first the samples seem alike but as the ice melts the sample with tap water ice takes on a flatness, something a bit like soggy cardboard, with a slight sour note, compared to the other two, which seem to maintain the original flavour of the vodka better.

Once the ice has melted, the vodka with the tap water ice now has white mineral
deposits at the bottom of the glass. This was from just two cubes of ice. No such
deposits appeared with ice cubes from Isbre or distilled water

But in my household I suspect more ice gets used in the shaker (then discarded) than added directly to drinks. Would water purity still make a difference in this context? I rustle up three Martinis, using SW4 47% and Noilly Prat. I follow James Bond’s example and shake, rather than stir, on the grounds that this will presumably result in more dilution and an enhanced effect.

Three Martinis, shaken with ice made from (left to right) tap water, Isbre and distilled

I use a measure of gin and a teaspoon of vermouth for each sample, add six ice cubes and shake exactly 40 times. I eschew a garnish to help focus on the matter in hand. Not unexpectedly, the differences are less noticeable this time, though I still think that the Isbre has a softer, sweeter mouthfeel and a sense of greater depth and transparency in the flavours of the ingredients. The tap water ice seems to give a slight masking harshness and again even the distilled water seems less appealing, which I cannot explain. But these are subtle distinctions.

This whole process has been a revelation to me. There are many bottled waters out there, with their own flavour profiles that may or may not blend harmoniously with other ingredients, but as a simple starting point it is an eye-opener to use water that is much more neutral than what I am used to. While the contribution in ice used to shake cocktails is subtle to say the least, when it comes to ice over which drinks are served, and certainly where water is directly added, using purer water makes a noticeable difference. When Isbre comes to market in the UK (I can’t find it for sale anywhere yet) it will be about £1.50 a litre: for ice to serve in drinks (as opposed, perhaps, to shaking) I think this is certainly worth paying.

One thing I am keen to try now is absinthe—which is typically served with two to five times as much water as spirit…

* Of course Isbre isn’t the only bottled water one could try—in fact while searching my inbox I stumble across a press release sent to me a few years ago for a brand called Sno, made in Iceland from a glacier, this time 20,000 years old. They also bang on about purity, but Sno contains 68ppm of dissolved solids according to the press release (though on their current website this has dropped to 52), so for the purposes of our experiment Isbre is more useful.

** Curiously, almost all suppliers had a footnote that their product was not suitable for human consumption. Considering that the whole point of distilled water is that it contains nothing but water, this is odd. It could be that if you habitually drink nothing but distilled water you would miss out on vital minerals, but I suspect it is simply that any product that is intended for human consumption has to undergo various tests that the water distillers don’t want to pay for—their products are actually designed for laboratory use or cleaning delicate equipment.

*** Although Isbre’s website only talks about total dissolved solids, I did find another site which gave more of a breakdown. This suggests the only detectable minerals are nitrates (0.05 ppm) and silica (2 ppm) from the rocky aquifer. My distilled water, for the record, has <0.2 ppm nitrates, <0.1 ppm lead, <0.2 ppm ammonium and <10 parts per billion silicon and chloride. So the distilled water actually has more of all these contaminants than Isbre, with the exception of silicon—and I gather that 5–25 ppm silica is typical for natural water, so Isbre is still fairly low. This site also gives Isbre’s pH as 5.7, which is slightly acidic. I can only assume that the silica (silicon dioxide) is in solution in the form of silicic acid. Which, in case you were wondering, is considered to be good for you. However, I later got hold of the latest lab report from Isbre themselves: it doesn't give a figure for silica but I see that the pH is 6.6, which is pretty close to pH neutral. Nitrates in this sample are down to 0.015 ppm and there are around 0.5 ppm traces of calcium, sodium, sulphate and chloride and 0.1 ppm magnesium and potassium. Total dissolved solids are given as just 3.9 ppm.