SilicaMineral

• Bone mineral densityPossibly helps

  This is the best-supported use, which is not the same as a settled one. The Framingham Offspring cohort and the Aberdeen Prospective Osteoporosis Screening Study both found higher dietary silicon linked to higher hip and spine density [13][10].
  
  That association held in men, in pre-menopausal women, and in post-menopausal women on hormone replacement — but not in post-menopausal women generally [13][10].
  
  A randomized placebo-controlled trial in osteopenic and osteoporotic women found a significant femoral density increase at 6 mg Si/day of ch-OSA, added to 1 g/day calcium and 800 IU vitamin D3 [10][10].
  
  Bone-formation markers trended upward as the ch-OSA dose rose [10]. Small older trials of monomethyl trisilanol reported increased trabecular bone volume and femoral and lumbar density, in one case above what sodium fluoride or etidronate achieved [10].
  
  RxList and WebMD rate silicon possibly effective for bone strength when it comes from food [14][15]. Their reading of the mechanism: silicon seems to help build bone, not to slow the breakdown side, which is why older post-menopausal women see no benefit [14][15]. ^[10]

• Where the framing splitsPossibly helps

  Government and academic sources — the IOM report, the epidemiological cohorts — describe silicon's bone role as suggestive and associative rather than causal [1][13].
  
  Retailer pages state it far more definitively. We are following the hedged version, and flagging that the confident version exists. ^[1]

• Hair, skin, and nailsPossibly helps

  Two randomized, double-blind, placebo-controlled trials carry nearly all of this claim, and both came from the same manufacturer-funded group [11][12].
  
  The 2005 trial gave 50 women with photodamaged skin 10 mg Si/day as ch-OSA or placebo for 20 weeks, reporting improved skin roughness and elasticity plus less hair and nail brittleness [11].
  
  The 2007 trial gave 48 women with fine hair the same dose for 9 months [12]. Hair elastic gradient and break load declined significantly less than on placebo, and cross-sectional area — thickness — rose in step with urinary silicon [12].
  
  Reviewers note both studies were run or funded by the ingredient's maker and did not meet the highest design and reporting standards [11].
  
  One consequence matters when you shop: those results belong to that specific branded ch-OSA ingredient. They do not carry over to colloidal silica or horsetail products without their own trials [11]. ^[11]

• Whether silicon is essential at allPossibly helps

  No silicon deprivation study has ever been run in humans, and no deficiency syndrome has been characterized in people with normal kidney function on an ordinary diet [6][1].
  
  Everything pointing toward essentiality comes from animals. Carlisle's classic chick experiments found retarded growth, narrower skull frontal areas, thinner and paler legs and beaks, and bones that fractured more easily [17][18].
  
  The biochemistry was the interesting part: collagen content in bone dropped sharply while mineral content was largely unaffected, placing silicon's role in the organic matrix rather than in mineralisation [17][6]. The skull effect happened independently of vitamin D status [18].
  
  The rat literature disagrees with itself, and we are not resolving it. Early studies reported dramatic skull, eye-socket, and incisor-enamel defects [6][1].
  
  Later, more controlled work by Seaborn and Nielsen could not reproduce them, finding instead modest drops in bone mineral density, bone mineral content, and collagen synthesis, plus increased collagen breakdown [6][1].
  
  With no functional human biomarker and no characterized syndrome, the IOM does not list silicon as an essential nutrient [1][2]. ^[6]

• Proposed mechanismPossibly helps

  Silicon is proposed to support prolyl hydroxylase, an enzyme collagen synthesis depends on, and to influence transcription of the type I collagen gene [1][6].
  
  Cell-culture and animal work shows silicon or orthosilicic acid raising osteoblast proliferation, type I collagen synthesis, alkaline phosphatase activity, and osteocalcin [16][6].
  
  None of that amounts to an established physiological function in humans, which the IOM report says plainly [1]. ^[1]

• Reducing aluminium absorptionPossibly helps

  A real research signal, and one where the chemical form decides everything. Human studies found oligomeric (polymerized) silica reduced aluminium absorption from the gut, while monomeric orthosilicic acid did not [16].
  
  Animal work found dietary silicon lowered aluminium accumulation in tissue [6]. Researchers have explored this against the still-debated aluminium hypothesis of Alzheimer's disease [16][6].
  
  This is research-stage evidence about how the body handles aluminium, not an established clinical use. ^[16]

• Heart disease, Alzheimer's, sprains, and digestive complaintsPossibly helps

  WebMD and RxList rate the evidence for all four as insufficient to rate [14][15].
  
  That is a blank, not a negative finding. ^[14]

• Choline-stabilized orthosilicic acid (ch-OSA, sold as Biosil)The form behind essentially every positive human trial on this page

  Elemental content

  Orthosilicic acid stabilized with choline; trials used 6–10 mg elemental Si/day [10][11]

  Absorption

  16% of the dose appeared in urine in a controlled crossover comparison [19].
  
  That is below plain dilute orthosilicic acid in the same study, yet reviewers describe ch-OSA as reliably bioavailable and biologically active [6].

  Best for

  The form behind essentially every positive human trial on this page [10][11][12]

• MonomethylsilanetriolBest measured uptake of the forms compared, with little modern trial evidence attached

  Elemental content

  A soluble organosilicon compound rather than a plant extract

  Absorption

  The highest absorber tested — 63% of the dose recovered in urine [19][5]

  Best for

  Best measured uptake of the forms compared, with little modern trial evidence attached [19]

• Dilute orthosilicic acid solutionShowing what monomeric silicon can do when it stays unpolymerized

  Elemental content

  Monomeric Si(OH)4 already in solution, at low concentration [19]

  Absorption

  43% of the dose, second only to monomethylsilanetriol [19]

  Best for

  Showing what monomeric silicon can do when it stays unpolymerized [19]

• Colloidal silica and magnesium trisilicateWidely sold, barely absorbed. Reviewers note market products range from under 1% to over 50% bioavailability

  Elemental content

  Polymerized silica rather than the absorbable monomer [6][19]

  Absorption

  The worst performers by a wide margin — 1% for colloidal silica and 4% for magnesium trisilicate [19].
  
  Only the small uncharged monomer crosses the gut lining readily, which is the whole explanation [6].

  Best for

  Widely sold, barely absorbed. Reviewers note market products range from under 1% to over 50% bioavailability [6]

• Horsetail extract (Silica+)A cautionary example of a label claim tested and found wanting

  Elemental content

  12 mg Si per tablet, marketed with a manufacturer claim of 85% bioavailability [6]

  Absorption

  Independent testing found it significantly less bioavailable than ch-OSA [6].
  
  The 85% figure did not survive that comparison.

  Best for

  A cautionary example of a label claim tested and found wanting [6]

• Liquid drops versus microencapsulated powderSettling the liquid-versus-powder argument — it is close to irrelevant

  Elemental content

  Three commercial orthosilicic-acid products, all dosed at 21.6 mg Si [5]

  Absorption

  A 2021 crossover trial in 5 healthy men found statistically equivalent urinary excretion of 27% to 35% over 6 hours [5].
  
  Liquid and powder presentations behaved the same once dissolved, so format is not the variable that matters [5].

  Best for

  Settling the liquid-versus-powder argument — it is close to irrelevant [5]

• Phytolithic silica from whole foodsOrdinary eating. Beer and mineral water carry silicon already dissolved as orthosilicic acid, with at least 50% of beer…

  Elemental content

  Silica bound into plant cell walls, historically assumed poorly available [6]

  Absorption

  Recent work puts mean absorption from solid foods at about 41% [6].
  
  That implies digestion breaks phytolithic silica down into soluble species such as orthosilicic acid [6].

  Best for

  Ordinary eating. Beer and mineral water carry silicon already dissolved as orthosilicic acid, with at least 50% of beer's silicon absorbed [6]

• Silicate kidney stones

  This is the clearest documented supplement risk. Silicate stones — as opposed to the far more common calcium oxalate kind — are extremely rare, but case reports keep tying them to long-term silicate intake [20][21][22].
  
  A 2009 report describes a woman who formed silicate calculi while taking three over-the-counter supplements that all carried silicon dioxide as an excipient [20].
  
  Her stones resolved when she stopped the products and returned when she restarted them [20]. That on-off pattern is about as close to causal as a case report gets.
  
  A more recent report describes first-time silica-and-calcium-carbonate stones in a 47-year-old man on multiple OTC vitamins, noting only two prior cases had linked silica stones to supplement use [22][24].
  
  The counterweight, which every source states: silicate stones sit well under 1% of all kidney stones [22]. Read the case reports as a rare risk made plausible, not a common one. ^[20]

• Inhaled crystalline silica is a separate hazard entirely

  Worth stating flatly, because the two get conflated constantly. Chronic inhalation of fine crystalline silica dust — mining, construction, quarrying, sandblasting — causes silicosis, and IARC classifies it as a cause of lung cancer [31][29].
  
  That is an occupational dust exposure. It is not a finding about eating food or swallowing a supplement containing amorphous silica [31][29].
  
  It runs the other way too: the safety findings for oral amorphous silica say nothing about breathing crystalline dust. ^[31]

• Impaired kidney function

  The kidneys are what keep blood silicon low, so losing that clearance changes the picture sharply.
  
  Plasma silicon in uraemic patients was measured at roughly 24 times the healthy level — about 3.8 mg/L against 0.16 mg/L [6]. ^[6]

• Food amounts versus concentrated doses

  WebMD and RxList describe silicon taken by mouth in ordinary food amounts as likely safe [14][15].
  
  Concentrated supplement doses are a different question, and one the reliable human safety data does not answer well [14][15]. ^[14]

• Pregnancy and breastfeeding

  WebMD rates food amounts likely safe in pregnancy and breastfeeding, but says there is not enough reliable information about concentrated supplement doses, and advises sticking to food [14].
  
  A UK retailer's health-hub summary also lists pregnant women among groups advised to avoid silica supplements [27]. No dedicated human pregnancy trial exists. ^[14]

• After stomach removal

  WebMD notes blood silicon levels can run lower after surgical removal of the stomach [14]. ^[14]

• Levothyroxine — nothing found, which is not the same as nothing there

  A systematic review of 63 studies on levothyroxine interactions with foods and supplements flagged coffee, soy, fibre, calcium, and iron, all reducing absorption [25].
  
  Silica and silicon dioxide were not among them [25]. That is an absence of evidence rather than a demonstrated absence of interaction — the review may simply not have looked. ^[25]

Plants carry far more silicon than animal foods, because they take it up from soil as a structural material [6].

The monocots — cereals and grasses — are the accumulators, holding roughly 10 to 20 times more silicon than dicots such as legumes [6].

Using the UK's provisional food-silicon database, the categories above 5 mg/100g are grain-heavy. Breakfast cereals average 7.79 mg/100g across a wide 1.34 to 23.36 range, with oat bran topping the grain products because silicon concentrates in the husk [6].

Dried fruit averages 10.54 mg/100g, the richest fruit category at 6.09 to 16.61 [6].

Vegetables average 1.79 mg/100g and run as high as 8.73, led by Kenyan, green, and runner beans, spinach, and coriander [6]. Legumes such as lentils and soya average 1.46 mg/100g [6].

Fresh fruit is thinner at 1.34 mg/100g, highest in bananas, pineapples, and mangoes [6]. Bananas come with a catch — under 2% of their silicon is actually absorbed [6].

Nuts and seeds average 0.78 mg/100g. Dairy at 0.31 and meat at 0.1 to 1.89 mg/100g are consistently low [6][6].

Rice hulls are the extreme case at about 110 mg Si/g, which is why milling matters [6]. Refining that strips hulls, husks, and bran lowers silicon, though flour, bread, rice, and pasta stay moderate sources [6].

Beer is the outlier among drinks, averaging 1.92 mg/100g and ranging to 3.94 [6]. Malting and mashing convert the plant silica in barley and hops into soluble orthosilicic acid, which is why it absorbs so well [6].

In one Finnish study beer supplied 44% of adult men's silicon intake, while cereals supplied 68% of children's [6][6].

Wine sits at 1.35 mg/100g and mineral or spring water at 0.55, above tap water [6]. Fizzy drinks and spirits are low.

A few very high-silicon mineral waters, Fiji and Spritzer among them, have been reported at 30–40 mg/L [6].

One absorption note that is not about heat. No source here measured silicon loss from cooking, but a controlled feeding study found a high-fibre diet dropped absorbed silicon to 35% of intake against 58% on a low-fibre diet [26].

The evidence here is thin enough that we will name its single source. The only concrete guidance found is the dosing protocol of one randomized placebo-controlled bone trial [10].

That protocol had participants drink the liquid ch-OSA immediately after preparing it, preferably 30 minutes before a meal or 2 hours after one [10].

Separating from meals matches general practice for water-soluble mineral solutions. But the trial never compared with-food against without-food, so this is the routine that worked, not a timing finding [10].

No source found any evidence that morning versus evening changes absorption or effect. WebMD and RxList simply point to the product label, given how little reliable dosing information exists at all [14][15].