Calcium orthophosphates (CaPO4): occurrence and properties

MCPM

Monocalcium phosphate monohydrate [Ca(H ₂ PO ₄ ) ₂ ·H ₂ O; the IUPAC name is calcium dihydrogen orthophosphate monohydrate] is both the most acidic and water-soluble CaPO ₄ . Although acidic CaPO ₄ in general were known by 1795 as “super-phosphate of lime” (Fourcroy ¹⁸⁰⁴ ), their differentiation started in 1800s. Namely, by 1807, researchers first prepared a calcium phosphate, which could be attributed to MCPM (Aikin and Aikin ¹⁸⁰⁷ ).

MCPM crystallizes from aqueous solutions containing dissolved ions of H ₂ PO ₄ ⁻ and Ca ²⁺ at the Ca/P ratio ~0.5 and solution pH below ~2.0. Besides, MCPM might be precipitated from aqueous solutions containing organic solvents (Boonchom ²⁰⁰⁹ ; Kongteweelert et al. ²⁰¹³ ). At temperatures above ~100 °C, MCPM releases a molecule of water and transforms into MCPA but at temperatures >~500 °C MCPA further transforms into Ca(PO ₃ ) ₂ (Tynsuaadu ¹⁹⁹⁰ ). The results of the spectroscopic investigations of MCPM are available in literature (Xu et al. ¹⁹⁹⁸ ).

Due to high acidity and solubility, MCPM is never found in biological calcifications. Moreover, pure MCPM is not biocompatible with bones (Köster et al. ¹⁹⁷⁷ ). However, in medicine MCPM is used as a component of several self-hardening CaPO ₄ formulations (Huan and Chang ²⁰⁰⁹ ; Dorozhkin ^2013b ). In addition, MCPM is used as a nutrient, acidulant and mineral supplement for food, feed and some beverages (Budavari et al. ¹⁹⁹⁶ ; Stein et al. ²⁰⁰⁸ ). Coupled with NaHCO ₃ , MCPM is used as a leavening agent for both dry baking powders and bakery dough. MCPM might be added to salt-curing preserves, pickled and marinated foods. In addition, MCPM might be added to tooth pastes and chewing gums (Dorozhkin ^2013c ). Besides, MCPM might be added to ceramics and glasses, while agriculture is the main consumer of a technical grade MCPM, where it is used as a fertilizer, triple superphosphate (Nasri et al. ²⁰¹⁵ ).

MCPA (or MCP)

Monocalcium phosphate anhydrous [Ca(H ₂ PO ₄ ) ₂ ; the IUPAC name is calcium dihydrogen orthophosphate anhydrous] is the anhydrous form of MCPM. Although MCPM has been known since 1807 (Dorozhkin ^2012a , ^2013a ), MCPA was differentiated as “tetra-hydrogen calcium phosphate, H ₄ Ca(PO ₄ ) ₂ ” by 1879 (Roscoe and Schorlemmer ¹⁸⁷⁹ ). It crystallizes under the same conditions as MCPM but at temperatures above ~100 °C (e.g., from concentrated hot mother liquors during fertilizer production). In addition, MCPA might be prepared from MCPM by dehydration. Furthermore, it might be also prepared at ambient temperatures by crystallization in water-restricted or non-aqueous systems. Like MCPM, MCPA never appears in calcified tissues and is not biocompatible due to its acidity. There is no current application of MCPA in medicine. Due to the similarity with MCPM, in many cases, MCPA might be used instead of MCPM; however, highly hydroscopic properties of MCPA reduce its commercial applications (Becker ¹⁹⁸⁹ ; Budavari et al. ¹⁹⁹⁶ ).

DCPD

Dicalcium phosphate dihydrate [CaHPO ₄ ·2H ₂ O; the IUPAC name is calcium hydrogen orthophosphate dihydrate; the mineral brushite] has been known since, at least, 1804 (Fourcroy ¹⁸⁰⁴ ). As a mineral, brushite was first discovered in phosphatic guano from Avis Island (Caribbean) in 1865 (Moore ¹⁸⁶⁵ ) and named to honor an American mineralogist Prof. George Jarvis Brush (1831–1912), Yale University, New Haven, Connecticut, USA.

DCPD can be easily crystallized from aqueous solutions containing dissolved ions of HPO ₄ ²⁻ and Ca ²⁺ at the Ca/P ratio ~1 and solution pH within ~2.0 < pH < ~6.5 (Hamai et al. ²⁰¹³ ). Other preparation techniques such as neutralization of H ₃ PO ₄ and/or MCPM solutions by CaO, CaCO ₃ or more basic CaPO ₄ (α- or β-TCP, CDHA, HA, TTCP) are also known. Interestingly, that precipitation of DCPD by mixing a Ca(OH) ₂ suspension and a H ₃ PO ₄ solution in the equimolar quantities was found to occur in five stages, being HA the first precipitated phase (Ferreira et al. ²⁰⁰³ ; Oliveira et al. ²⁰⁰⁷ ). Besides, DCPD might be prepared in gels (Sivkumar et al. ¹⁹⁹⁷ ; Madhurambal et al. ²⁰⁰⁹ ). DCPD transforms into DCPA at temperatures above ~80 °C and this transformation is accompanied by ~11 % decrease in volume (MacDowell et al. ¹⁹⁷¹ ) and structural changes (Landin et al. ¹⁹⁹⁴ ). The value for Δ _r G ° for DCPD → DCPA transformation is −1.032 kJ/mol (Landin et al. ¹⁹⁹⁴ ). Briefly, DCPD crystals consist of CaPO ₄ chains arranged parallel to each other, while lattice water molecules are interlayered between them (Curry and Jones ¹⁹⁷¹ ). Liquid ordering at the {010} DCPD/water interface was determined (Arsic et al. ²⁰⁰⁴ ). In 2009, data on DCPD solubility were updated by solid titration technique (Pan and Darvell ^2009b ). The optical properties of DCPD were described (Lundager-Madsen ²⁰⁰⁸ ), while many additional data on DCPD including a good drawing of its atomic structure might be found in Ref. (Qiu and Orme ²⁰⁰⁸ ).

DCPD is of biological importance because it is often found in pathological calcifications (dental calculi, crystalluria, chondrocalcinosis and urinary stones) and some carious lesions (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; O’Neill ²⁰⁰⁷ ). It was proposed as an intermediate in both bone mineralization and dissolution of enamel in acids (dental erosion) (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; O’Neill ²⁰⁰⁷ ). In medicine, DCPD is used in self-setting CaPO ₄ formulations (Dorozhkin ^2013b ) and as an intermediate for tooth remineralization. DCPD is added to toothpaste both for caries protection (in this case, it is often coupled with F-containing compounds such as NaF and/or Na ₂ PO ₃ F) and as a gentle polishing agent (Dorozhkin ^2013c ). Other applications include a flame retardant (Mostashari et al. ²⁰⁰⁶ ), a slow release fertilizer, using in glass production, as well as calcium supplement in food, feed and cereals. In food industry, it serves as a texturizer, bakery improver and water retention additive. In diary industry, DCPD is used as a mineral supplement. In addition, plate-like crystals of DCPD might be used as a non-toxic, anticorrosive and passivating pigment for some ground coat paints (Budavari et al. ¹⁹⁹⁶ ).

DCPA (or DCP)

Dicalcium phosphate anhydrous (CaHPO ₄ ; the IUPAC name is calcium hydrogen orthophosphate anhydrate; the mineral monetite) is the anhydrous form of DCPD. Although DCPD has been known since, at least, 1804 (Dorozhkin ^2012a , ^2013a ), DCPA was differentiated as “mono-hydrogen CaPO ₄ , HCaPO ₄ ” by 1879 (Roscoe and Schorlemmer ¹⁸⁷⁹ ). As a mineral, monetite was first described in 1882 in rock-phosphate deposits from the Moneta (now Monito) Island (archipelago of Puerto Rico), which contains a notable occurrence (Shepard ¹⁸⁸² ).

Due to the absence of water inclusions, DCPA is less soluble than DCPD (Table  ¹ ). Like DCPD, DCPA can be crystallized from aqueous solutions containing Ca/P ratio ~1 at solution pH within ~2.0 < pH < ~6.5 but at temperatures >~90 °C. In addition, DCPA might be prepared by dehydration of DCPD. Furthermore, it might be also prepared at ambient temperatures in water-restricted or non-aqueous systems, such as gels (Sivkumar et al. ¹⁹⁹⁷ ), ethanol (Tas ²⁰⁰⁹ ), as well as in the oil-in-water and water-in-oil systems (Chen et al. ²⁰⁰⁵ ). DCPA is physically stable and resisted hydration even when dispersed in water for over 7 months in the temperature range of 4–50 °C (Miyazaki et al. ²⁰⁰⁹ ). A calcium-deficient DCPA was also prepared. It might be sintered at ~300 °C (Eshtiagh-Hosseini et al. ²⁰⁰⁸ ). Unlike DCPD, DCPA occurs in neither normal nor pathological calcifications. It is used in self-setting CaPO ₄ formulations (Dorozhkin ^2013b ). Besides, DCPA might be implanted as bioceramics (Tamimi et al. ²⁰¹⁰ ). Other applications include using as a polishing agent, a source of calcium and phosphate in nutritional supplements (e.g., in prepared breakfast cereals, enriched flour and noodle products), a tableting aid (Takami et al. ¹⁹⁹⁶ ) and a toothpaste component (Dorozhkin ^2013c ). In addition, it is used as a dough conditioner in food industry (Budavari et al. ¹⁹⁹⁶ ). DCPA of a technical grade of purity might be used as a fertilizer (Habashi ²⁰¹¹ ).

OCP

Octacalcium phosphate [Ca ₈ (HPO ₄ ) ₂ (PO ₄ ) ₄ ·5H ₂ O; the IUPAC name is tetracalcium hydrogen orthophosphate diorthophosphate pentahydrate, another name is octacalcium bis(hydrogenphosphate) tetrakis(phosphate) pentahydrate] is often found as an unstable transient intermediate during the precipitation of the thermodynamically more stable CaPO ₄ (e.g., CDHA) in aqueous solutions. To the best of my findings (Dorozhkin ^2012a , ^2013a ), OCP has been known since, at least, 1843, when Percy published a paper (Percy ¹⁸⁴³ ), in which he described formation of “a new hydrated phosphate of lime” with a chemical formula 2CaO + PO5 + 6HO, in which “1 equiv. water being basic and 5 constitutional”. However, according to Bjerrum (Bjerrum ¹⁹⁵⁸ ), a CaPO ₄ with the OCP composition was first described by Berzelius in 1836.

The preparation techniques of OCP are available in literature (LeGeros ¹⁹⁸⁵ ; Nakahira et al. ²⁰⁰¹ ; Arellano-Jiménez et al. ²⁰⁰⁹ ; Suzuki ^2010a ). Briefly, to prepare OCP, Ca- and PO ₄ -containing chemicals must be mixed to get the supersaturated aqueous solutions with the Ca/P ratio equal to 1.33. Typically, OCP crystals are smaller if compared to DCPD ones, extremely platy and almost invariably twinned. However, OCP might be non-stoichiometric and be either Ca-deficient (down to Ca/P = 1.26) or include excessive calcium (up to Ca/P = 1.48) in the structure (Suzuki ^2010a ). It has been proposed that the structure of a non-stoichiometric OCP contains an excess of hydrogen, resulting in a non-stoichiometric chemical formula Ca ₁₆ H _{4+ x} (PO ₄ ) ₁₂ (OH) _x ·(10 −  x )H ₂ O, which resembles the structure of HA even more closely than previously anticipated (Mathew et al. ¹⁹⁸⁸ ). Furthermore, a partially hydrolyzed form of OCP with Ca/P molar ratio of 1.37 might be prepared (Suzuki ^2010a ; Miyatake et al. ²⁰⁰⁹ ). At solution pH = 7.2 and temperature 60 °C, the full hydrolysis of OCP into CDHA occurs within ~6 h (Arellano-Jiménez et al. ²⁰⁰⁹ ). Ion-substituted OCP might be prepared as well (Matsunaga ^2008a ; Boanini et al. ^2010a ).

The triclinic structure of OCP displays apatitic layers (with atomic arrangements of calcium and orthophosphate ions similar to those of HA) separated by hydrated layers (with atomic arrangements of calcium and orthophosphate ions similar to those in DCPD) (LeGeros ¹⁹⁹¹ ; Elliott ¹⁹⁹⁴ ; Amjad ¹⁹⁹⁷ ; Suzuki ^2010a ). A similarity in crystal structure between OCP and HA (Brown ¹⁹⁶² ; Brown et al. ¹⁹⁶² ) is one reason that the epitaxial growth of these phases is observed. It is generally assumed that, in solutions, the hydrated layer of the (100) face is the layer most likely exposed to solution. The water content of OCP crystals is ~20 % that of DCPD and this is partly responsible for its lower solubility. The latest data on OCP structure (Davies et al. ²⁰¹² ) and solubility (Pan and Darvell ^2009c ) are available.

OCP is of a great biological importance because it is one of the stable components of human dental and urinary calculi (Chow and Eanes ²⁰⁰¹ ; Kakei et al. ²⁰⁰⁹ ). OCP was first proposed by W. E. Brown to participate as the initial phase in enamel mineral formation and bone formation through subsequent precipitation and stepwise hydrolysis of OCP (Brown ¹⁹⁶² , ¹⁹⁶⁶ ; Brown et al. ¹⁹⁶² ). It plays an important role in formation of apatitic biominerals in vivo (Suzuki ^2010b ). A “central OCP inclusion” (also known as “central dark line”) is seen by transmission electron microscopy in many biological apatites and in synthetically precipitated CDHA (Iijima et al. ¹⁹⁹⁶ ; Bodier-Houllé et al. ¹⁹⁹⁸ ; Rodríguez-Hernández et al. ²⁰⁰⁵ ). Although OCP has not been observed in vascular calcifications, it has been strongly suggested as a precursor phase to biological apatite found in natural and prosthetic heart valves (Tomazic et al. ¹⁹⁹⁴ ; Nancollas and Wu ²⁰⁰⁰ ). In surgery, OCP is used for implantation into bone defects (Suzuki et al. ²⁰⁰⁶ ; Kikawa et al. ²⁰⁰⁹ ; Murakami et al. ²⁰¹⁰ ; Suzuki ²⁰¹³ ). For the comprehensive information on OCP, the readers are referred to other reviews (Suzuki ^2010a ; Chow and Eanes ²⁰⁰¹ ; Suzuki ²⁰¹³ ).

β-TCP

β-tricalcium phosphate [β-Ca ₃ (PO ₄ ) ₂ ; the IUPAC name is tricalcium diorthophosphate beta, other names are CaPO ₄ tribasic beta or tricalcium bis(orthophosphate) beta] is one of the polymorphs of TCP. Although CaPO ₄ with the composition close to that of TCP, CDHA and HA were known in 1770s (Dorozhkin ^2012a , ^2013a ), α- and β-polymorphs of TCP were differentiated only by 1932 (Bredig et al. ¹⁹³² ; Trömel ¹⁹³² ).

β-TCP cannot be precipitated from aqueous solutions. It is a high-temperature phase, which can be prepared at temperatures above ~800 °C by thermal decomposition of CDHA or by solid-state interaction of acidic CaPO ₄ , e.g., DCPA, with a base, e.g., CaO. In all cases, the chemicals must be mixed in the proportions to get the Ca/P ratio equal to 1.50. However, β-TCP can also be prepared at relatively low temperatures (~150 °C) by precipitation in water-free mediums, such as ethylene glycol (Tao et al. ²⁰⁰⁸ , ²⁰⁰⁹ ). Apart from the chemical preparation routes, ion-substituted β-TCP can be prepared by calcining of bones (Hou et al. ²⁰⁰⁸ ; Santos et al. ²⁰¹² ): such type of CaPO ₄ is occasionally called “bone ash” (Lee et al. ²⁰¹³ ). At temperatures above ~1125 °C, β-TCP is transformed into a high-temperature phase α-TCP. Being the stable phase at room temperature, β-TCP is less soluble in water than α-TCP (Table  ¹ ). Both ion-substituted (Ito and LeGeros ²⁰⁰⁸ ; Kannan et al. ²⁰⁰⁹ , ²⁰¹⁰ ; Quillard et al. ²⁰¹² ) and organically modified (Karlinsey and Mackey ²⁰⁰⁹ ; Karlinsey et al. ^2010a , ^b ) forms of β-TCP can be synthesized, as well. The modern structural data on β-TCP are available in Refs. (Yashima et al. ²⁰⁰³ ; Yin et al. ²⁰⁰³ ; Liang et al. ²⁰¹⁰ ; Zhai and Wu ²⁰¹⁰ ), Raman spectra of β-TCP might be found if Ref. (Zhai et al. ²⁰¹⁵ ), while solubility data—in Ref. (Pan and Darvell ^2009d ). Furthermore, an ability of β-TCP to store an electrical charge by electrical polarization was studied and this material was found to have a suitable composition and structure for both ion conduction and charge storage (Wang et al. ²⁰¹⁰ ).

Pure β-TCP never occurs in biological calcifications. Only a Mg-substituted form [β-TCMP—β-tricalcium magnesium phosphate, β-(Ca,Mg) ₃ (PO ₄ ) ₂ , which is often called whitlockite to honor Mr. Herbert Percy Whitlock (1868–1948), an American mineralogist, the curator of the American Museum of Natural History, New York City, New York, USA (Frondel ¹⁹⁴¹ )] is found. Since β-TCMP is less soluble than β-TCP (Li et al. ^2009a ), it is formed instead of β-TCP in dental calculi and urinary stones, dentineal caries, salivary stones, arthritic cartilage, as well as in some soft-tissue deposits (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; O’Neill ²⁰⁰⁷ ; Kodaka et al. ¹⁹⁸⁸ ; Reid and Andersen ¹⁹⁹³ ; Scotchford and Ali ¹⁹⁹⁵ ; P’ng et al. ²⁰⁰⁸ ). However, it has not been observed in enamel, dentine or bone. In medicine, β-TCP is used in the self-setting CaPO ₄ formulations (Dorozhkin ^2013b ) and other types of bone grafts (Hou et al. ²⁰⁰⁸ ; Horch et al. ²⁰⁰⁶ ; Ogose et al. ²⁰⁰⁶ ; Kamitakahara et al. ²⁰⁰⁸ ; Liu et al. ²⁰⁰⁸ ; Epstein ²⁰⁰⁹ ; Liu and Lun ²⁰¹² ). Dental applications of β-TCP are also known. For example, β-TCP is added to some brands of toothpaste as a gentle polishing agent (Dorozhkin ^2013c ). Multivitamin complexes with CaPO ₄ are widely available in the market and β-TCP is used as the calcium phosphate there. In addition, β-TCP serves as a texturizer, bakery improver and anti-clumping agent for dry powdered food (flour, milk powder, dried cream, cocoa powder). Besides, β-TCP is added as a dietary or mineral supplement to food and feed (Güngörmüş et al. ²⁰¹⁰ ). Occasionally, β-TCP might be used as inert filler in pelleted drugs. Other applications comprise porcelains, pottery, enamel, using as a component for mordants and ackey, as well as a polymer stabilizer (Budavari et al. ¹⁹⁹⁶ ). β-TCP of a technical grade (as either calcined natural phosphorites or bone dust) is used as a slow release fertilizer for acidic soils (Becker ¹⁹⁸⁹ ).

To conclude, one should briefly mention on an existence of γ-TCP polymorph, naturally known as tuite, which was named after Prof. Guangzhi Tu (born in 1920), a founding director of the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou, China. Tuite appears to be a high-pressure polymorph of whitlockite with an empirical formula Ca2.51Na0.28Mg0.27Fe0.022+PO42.02 . Pure γ-TCP polymorph can be synthesized from β-TCP at pressures above ~4 GPa and temperatures above ~1000 °C (Murayama et al. ¹⁹⁸⁶ ; Xie et al. ²⁰⁰³ ; Zhai et al. ²⁰⁰⁹ , ²⁰¹⁴ ).

α-TCP

α-Tricalcium phosphate [α-Ca ₃ (PO ₄ ) ₂ ; the IUPAC name is tricalcium diorthophosphate alpha, other names are CaPO ₄ tribasic alpha or tricalcium bis(orthophosphate) alpha] is another polymorph of TCP, which was differentiated by 1932 (Bredig et al. ¹⁹³² ; Trömel ¹⁹³² ). α-TCP is also a high-temperature phase; therefore, it cannot be precipitated from aqueous solutions either. Thus, α-TCP is usually prepared by the same techniques as β-TCP (see the previous section) but, since the β-TCP → α-TCP transition temperature is ~1125 °C (Welch and Gutt ¹⁹⁶¹ ), calcining is performed at temperatures above ~1200 °C (Jokic et al. ²⁰⁰⁷ ). Consequently, α-TCP is often considered as a high-temperature polymorph of β-TCP. However, data are available that α-TCP might be prepared at lower temperatures. Namely, at the turn of the millennium, the previously forgotten data that the presence of silicates stabilized α-TCP at temperatures of 800–1000 °C (Nurse et al. ¹⁹⁵⁹ ) were rediscovered again. Such type of α-TCP is called “silica stabilized α-TCP” (Sayer et al. ²⁰⁰³ ; Reid et al. ²⁰⁰⁵ , ²⁰⁰⁶ ). Furthermore, sometimes, α-TCP might be prepared at even lower temperatures (~700 °C) by a thermal decomposition of low-temperature ACPs (Kanazawa et al. ¹⁹⁸² ).

Although α-TCP and β-TCP have exactly the same chemical composition, they differ by the crystal structure (Table  ⁴ ) and solubility (Table  ¹ ). In the absence of humidity, both polymorphs of TCP are stable at room temperatures; however, according to a density functional study, stability of β-TCP crystal lattice exceeds that of α-TCP (Yin et al. ²⁰⁰³ ). Therefore, of them, α-TCP is more reactive in aqueous systems, has a higher specific energy and in aqueous solutions it can be hydrolyzed to CDHA (TenHuisen and Brown ¹⁹⁹⁸ ; Durucan and Brown ²⁰⁰⁰ , ²⁰⁰² ). Milling was found to increase the α-TCP reactivity even more (Camiré et al. ²⁰⁰⁵ ). Although, α-TCP never occurs in biological calcifications, in medicine, it is used as a component of self-setting CaPO ₄ formulations (Budavari et al. ¹⁹⁹⁶ ). On the other hand, the chemically pure α-TCP has received not much interest in the biomedical field (Kamitakahara et al. ²⁰⁰⁸ ). The disadvantage for using α-TCP is its quick resorption rate (faster than formation of a new bone), which limits its application in this area. However, the silicon stabilized α-TCP (more precisely as a biphasic composite with HA) has been commercialized as a starting material to produce bioresorbable porous ceramic scaffolds to be used as artificial bone grafts (Sayer et al. ²⁰⁰³ ; Reid et al. ²⁰⁰⁵ , ²⁰⁰⁶ ). Upon implantation, α-TCP tends to convert to CDHA, which drastically reduces further degradation rate. Theoretical insights into bone grafting properties of the silicon-stabilized α-TCP might be found in Ref. (Yin and Stott ²⁰⁰⁵ ). The structure of α-TCP is well described in literature (Yin et al. ²⁰⁰³ ; Liang et al. ²⁰¹⁰ ), while the surface and adsorption properties are available in Ref. (Yin and Stott ²⁰⁰⁶ ). Similar to β-TCP, α-TCP of a technical grade might be used slow release fertilizer for acidic soils (Budavari et al. ¹⁹⁹⁶ ).

To conclude, one should briefly mention on an existence of α′-TCP polymorph, which was discovered in 1959 (Nurse et al. ¹⁹⁵⁹ ). However, this TCP polymorph lacks of any practical interest because it only exists at temperatures between ~1450 °C and its melting point (~1756 °C). It reverts to α-TCP polymorph by cooling below the transition temperature. Additional details on α-TCP are available in the topical review (Carrodeguas and de Aza ²⁰¹¹ ).

ACP

Amorphous calcium phosphates (ACPs) represent a special class of CaPO ₄ salts, having variable chemical but rather identical glass-like physical properties, in which there are neither translational nor orientational long-range orders of the atomic positions. To the best of my findings (Dorozhkin ^2012a , ^2013a ), ACP was first prepared in 1845 (Jones ¹⁸⁴⁵ ). Nevertheless, until recently (Dorozhkin ^2012c ), ACP has often been considered as an individual CaPO ₄ compound with a variable chemical composition, while, in reality, ACP is just an amorphous state of other CaPO ₄ . Therefore, in principle, all compounds mentioned in Table  ¹ might be somehow fabricated in an amorphous state but, currently, only few of them (e.g., an amorphous TCP) are known (Dorozhkin ^2012c ). Thus, strictly speaking, ACP should be excluded from Table  ¹ . Furthermore, since ACPs do not have the definite chemical composition, the IUPAC nomenclature is not applicable to describe them.

Depending on the production temperatures, all types of ACP are divided into two major groups: low-temperature ACPs (prepared in solutions, usually aqueous ones) and high-temperature ACPs (Dorozhkin ^2012c ). Low-temperature ACPs (described by the chemical formula Ca _x H _y (PO ₄ ) _z · n H ₂ O, n  = 3–4.5; 15–20 % H ₂ O) are often encountered as a transient precursor phase during precipitation of other CaPO ₄ in aqueous systems. Usually, an ACP is the first phase precipitated from supersaturated solutions (the higher supersaturation, the better) prepared by rapid mixing of solutions containing ions of calcium and orthophosphate (Elliott ¹⁹⁹⁴ ; Dorozhkin ^2012c ). Such ACP precipitates usually look like spherical particles with diameters in the range 200–1200 Å without a definite structure. Generally, the ACP particles are smaller if prepared under conditions of high supersaturation and/or high pH, while for a given pH, higher temperatures give larger particles (Blumenthal et al. ¹⁹⁷² ). The freshly precipitated ACPs contain 10–20 % by weight of tightly bound water, which is removed by vacuum drying at elevated temperature (Posner and Betts ¹⁹⁷⁵ ). The amorphization degree of ACPs increases with the concentration increasing of Ca- and PO ₄ -containing solutions, as well as at a high solution pH and a low crystallization temperature. A continuous gentle agitation of as precipitated ACPs in the mother solution, especially at elevated temperatures, results in a slow recrystallization and formation of better crystalline CaPO ₄ , such as CDHA (LeGeros ¹⁹⁹¹ ; Elliott ¹⁹⁹⁴ ). In addition, other production techniques of ACPs are known (Dorozhkin ^2012c ).

The lifetime of ACPs in aqueous solutions was reported to be a function of the presence of additive molecules and ions, pH, ionic strength and temperature. In addition, confinement was found to increase their lifetime (Wang et al. ^2014a ). Thus, ACPs may persist for appreciable periods and retain the amporphous state under some specific experimental conditions (Termine et al. ¹⁹⁷⁰ ). The chemical composition of ACPs strongly depends on the solution pH and the concentrations of mixing solutions. For example, ACPs with Ca/P ratios in the range of 1.18 (precipitated at solution pH = 6.6) to 1.53 (precipitated at solution pH = 11.7) (Elliott ¹⁹⁹⁴ , ¹⁹⁹⁸ ) and even to 2.5 (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; O’Neill ²⁰⁰⁷ ) were described. In deed and not in name, these data mean that various types of CaPO ₄ were prepared in an amorphous state (Dorozhkin ^2012c ). It should be noted that unsubstituted ACPs are unstable in aqueous solutions and even when stored dry they tend to transform into more crystalline CaPO ₄ , such as poorly crystalline CDHA. The presence of poly(ethylene glycol) (Li and Weng ²⁰⁰⁷ ), ions of pyrophosphate, carbonate and/or magnesium in solutions during the crystallization promotes formation of ACPs and slows down their further transformation, while the presence of fluoride has the opposite effect (LeGeros ¹⁹⁹¹ ; Elliott ¹⁹⁹⁴ ; Amjad ¹⁹⁹⁷ ; Daculsi et al. ¹⁹⁹⁷ ; Tadic et al. ²⁰⁰² ). In general, low-temperatures ACPs heated to ~550 °C (so that all volatiles have already escaped) remain amorphous, but further heating above ~650 °C causes their transformation into crystalline CaPO ₄ , such as α- or β-TCP, HA, mixtures thereof, depending on the Ca/P ratio of the ACP heated.

High-temperature ACPs might be prepared using high energy processing at elevated temperatures (Dorozhkin ^2012c ). This method is based on a rapid quenching of melted CaPO ₄ occurring, e.g., during plasma spraying of HA (Carayon and Lacout ²⁰⁰³ ; Keller and Dollase ²⁰⁰⁰ ; Kumar et al. ²⁰⁰⁴ ). A plasma jet, possessing very high temperatures (~5000 to ~20,000 °C), partly decomposes HA, which results in formation of a complicated mixture of products, some of which would be ACPs. Obviously, all types of high-temperature ACPs are definitively anhydrous contrary to the precipitated ACPs. Unfortunately, no adequate chemical formula is available to describe the high-temperature ACPs.

In general, as all amorphous compounds are characterized by a lack of long-range order, it is problematic to discuss the structure of ACPs (they are X-ray amorphous). Concerning a short-range order (SRO) in ACPs, it exists, just due to the nature of chemical bonds. Unfortunately, in many cases, the SRO in ACPs is uncertain either, because it depends on many variables, such as Ca/P ratio, preparation conditions, storage and admixtures. Infrared spectra of ACPs show broad featureless phosphate absorption bands. Electron microscopy of freshly precipitated ACPs usually shows featureless nearly spherical particles with diameters in the range of 20–200 nm. However, there is a questionable opinion that ACPs might have an apatitic structure but with a crystal size so small, that they are X-ray amorphous. This is supported by X-ray absorption spectroscopic data (EXAFS) on biogenic and synthetic samples (Harries et al. ¹⁹⁸⁶ , ¹⁹⁸⁷ ; Taylor et al. ¹⁹⁹⁸ ; Peters et al. ²⁰⁰⁰ ). On the other hand, it was proposed that the basic structural unit of the precipitated ACPs is a 9.5 Å diameter, roughly spherical cluster of ions with the composition of Ca ₉ (PO ₄ ) ₆ (Fig.  ⁷ ; Elliott ¹⁹⁹⁴ , ¹⁹⁹⁸ ; Posner et al. ¹⁹⁸⁰ ; Boskey ¹⁹⁹⁷ ). These clusters were found experimentally as first nuclei during the crystallization of CDHA and a model was developed to describe the crystallization of HA as a step-wise assembly of these units (Onuma and Ito ¹⁹⁹⁸ ) [see “ ^{HA (or HAp, or OHAp)} ”]. Biologically, ion-substituted ACPs (always containing ions of Na, Mg, carbonate and pyrophosphate) are found in soft-tissue pathological calcifications (e.g., heart valve calcifications of uremic patients) (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; O’Neill ²⁰⁰⁷ ).

Fig. 7

In medicine, ACPs are used in self-setting CaPO ₄ formulations (Dorozhkin ^2013b ). Bioactive composites of ACPs with polymers have properties suitable for use in dentistry (Dorozhkin ^2012c , ^2013c ) and surgery (Dorozhkin ^2012c ; Tadic and Epple ²⁰⁰² ). Due to a reasonable solubility and physiological pH of aqueous solutions, ACPs appeared to be consumable by some microorganisms and, due to this reason, it might be added as a mineral supplement to culture media. Non-biomedical applications of ACPs comprise their using as a component for mordants and ackey. In food industry, ACPs are used for syrup clearing. Occasionally, they might be used as inert filler in pelleted drugs. In addition, ACPs are used in glass and pottery production and as a raw material for production of some organic phosphates. To get further details on ACPs, the readers are referred to special reviews (Dorozhkin ^2012c ; Boskey ¹⁹⁹⁷ ; Combes and Rey ²⁰¹⁰ ).

CDHA (or Ca-def HA, or CDHAp)

Calcium-deficient hydroxyapatite [Ca _{10- x} (HPO ₄ ) _x (PO ₄ ) _{6- x} (OH) _{2- x} (0 <  x  < 1)] became known since the earliest experiments on establishing the chemical composition of bones performed in 1770s (Dorozhkin ^2012a , ^2013a ). However, the first appropriate term “subphosphate of lime” appeared by 1819 (Bache ¹⁸¹⁹ ). Other chemical formulae such as Ca _{10- x} (HPO ₄ ) _{2 x} (PO ₄ ) _{6−2 x} (OH) ₂ (0 <  x  < 2), Ca _{10− x − y} (HPO ₄ ) _x (PO ₄ ) _{6- x} (OH) _{2− x −2 y} (0 <  x  < 2 and y  <  x /2), Ca _{10− x} (HPO ₄ ) _x (PO ₄ ) _{6− x} (OH) _{2− x} (H ₂ O) _x (0 <  x  < 1), and Ca _{9− x} (HPO ₄ ) _{1+2 x} (PO ₄ ) _{5−2 x} (OH) were also proposed to describe its variable composition (Elliott ¹⁹⁹⁴ ). As seen from these formulae, Ca deficiency is always coupled with both OH deficiency and protonation of some PO ₄ groups with simultaneous formation of the ionic vacancies in the crystal structure (Wilson et al. ²⁰⁰⁵ ). In addition, CDHA often contains tightly bound water molecules, which might occupy some of these ionic vacancies. For example, there is an approach describing a lack of the hydroxide vacancies in CDHA: to perform the necessary charge compensation of the missing Ca ²⁺ ions, a portion of OH ⁻ anions is substituted by neutral water molecules (Zahn and Hochrein ²⁰⁰⁸ ). This water is removed by vacuum drying at elevated temperature. Concerning possible vacancies of orthophosphate ions, nothing is known about their presence in CDHA. It is just considered that a portion of PO ₄ ³⁻ ions is either protonated (as HPO ₄ ²⁻ ) or substituted by other ions (e.g., CO ₃ ²⁻ ) (Ivanova and Frank-Kamenetskaya ²⁰⁰¹ ). Since CDHA does not have any definite chemical composition, the IUPAC nomenclature is not applicable to describe it.

CDHA can be easily prepared by simultaneous addition of Ca- and PO ₄ -containing solutions in the proportions to get Ca/P ratio within 1.50–1.67 into boiling water followed by boiling the suspension for several hours (an aging stage). That is why, in literature, it might be called as “precipitated HA (PHA)” (Sinha et al. ²⁰⁰³ ; Mayer et al. ²⁰⁰³ ). Besides, it might be prepared by hydrolysis of α-TCP (TenHuisen and Brown ¹⁹⁹⁸ ; Durucan and Brown ²⁰⁰⁰ , ²⁰⁰² ). Other preparation techniques of CDHA are known as well (Vallet-Regí et al. ¹⁹⁹⁷ ; Siddharthan et al. ²⁰⁰⁴ ; Hutchens et al. ²⁰⁰⁶ ; Mochales et al. ²⁰¹¹ ). During aging, initially precipitated ACPs are restructured and transformed into CDHA. Therefore, there are many similarities in the structure, properties and application between the precipitated in alkaline solutions (pH > 8) ACPs and CDHA. Some data indicated on a presence of intermediate phases during further hydrolysis of CDHA to a more stable HA-like phase (Brès et al. ²⁰⁰⁵ ). In general, CDHA crystals are poorly crystalline and of submicron dimensions. They have a very large specific surface area, typically 25–100 m ² /g. On heating above ~700 °C, CDHA with Ca/P = 1.5 converts to β-TCP and that with 1.5 < Ca/P < 1.67 converts into a biphasic composite of HA and β-TCP (see “ ^{Biphasic, triphasic and multiphasic CaPO ₄ formulations} ”) (Dorozhkin ^2012d ). A solid-state transformation mechanism of CDHA into HA + β-TCP biocomposite was proposed (Dorozhkina and Dorozhkin ^2002a ; Dorozhkin ²⁰⁰³ ).

The variability in Ca/P molar ratio of CDHA has been explained through different models: surface adsorption, lattice substitution and intercrystalline mixtures of HA and OCP (Rodríguez-Lorenzo ²⁰⁰⁵ ). Due to a lack of stoichiometry, CDHA is usually doped by other ions (Rey et al. ²⁰⁰⁶ ). The doping extent depends on the counter-ions of the chemicals used for CDHA preparation. Direct determinations of the CDHA structures are still missing and the unit cell parameters remain uncertain. However, unlike that in ACPs (see “ ^ACP ”), a long-range order exists in CDHA. Namely, the following lattice parameters were reported for CDHA with Ca/P = 1.5: a  = 9.4418 (20) Å and c  = 6.8745 (17) Å (Mochales et al. ²⁰¹¹ ).

Systematic studies of defect constellations in CDHA are available in literature (Zahn and Hochrein ²⁰⁰⁸ ; Liou et al. ²⁰⁰⁴ ). As a first approximation, CDHA may be considered as HA with some ions missing (ionic vacancies) (Brown and Martin ¹⁹⁹⁹ ). The more amount of Ca is deficient, the more disorder, imperfections and vacancies are in the CDHA structure (Honghui et al. ²⁰⁰⁷ ). Furthermore, a direct correlation between the Ca deficiency and the mechanical properties of the crystals was found: calcium deficiency lead to an 80 % reduction in the hardness and elastic modulus and at least a 75 % reduction in toughness in plate-shaped HA crystals (Viswanath et al. ²⁰¹⁰ ). More recently, using first-principles calculations, a reduction in the elastic constants and moduli of CDHA due to vacancies was reported (Sun et al. ²⁰¹³ ; Bhat et al. ²⁰¹⁴ ). Theoretical investigations of the defect formation mechanism relevant to non-stoichiometry in CDHA are available elsewhere (Matsunaga ^2008b ).

Undoped CDHA (i.e., that containing ions of Ca ²⁺ , PO ₄ ³⁻ , HPO ₄ ²⁻ and OH ⁻ only) does not exist in biological systems. However, the ion-substituted CDHA: Na ⁺ , K ⁺ , Mg ²⁺ , Sr ²⁺ for Ca ²⁺ ; CO ₃ ²⁻ for PO ₄ ³⁻ or HPO ₄ ²⁻ ; F ⁻ , Cl ⁻ , CO ₃ ²⁻ for OH ⁻ , plus some water forms biological apatite—the main inorganic part of animal and human normal and pathological calcifications (LeGeros ¹⁹⁹¹ ; Rey et al. ²⁰⁰⁶ ; O’Neill ²⁰⁰⁷ ). Therefore, CDHA is a very promising compound for industrial manufacturing of artificial bone substitutes (Bourgeois et al. ²⁰⁰³ ), including drug delivery applications (Liu et al. ²⁰⁰⁵ ). Non-biomedical applications of CDHA are similar to those of ACP and HA. Interesting that CDHA was found to possess a catalytic activity to produce biogasoline (Tsuchida et al. ²⁰⁰⁸ ).

HA (or HAp, or OHAp)

Hydroxyapatite [Ca ₅ (PO ₄ ) ₃ (OH), but is usually written as Ca ₁₀ (PO ₄ ) ₆ (OH) ₂ to denote that the crystal unit cell comprises two molecules; the IUPAC name is pentacalcium hydroxide tris(orthophosphate)] is the second most stable and least soluble CaPO ₄ after FA. Apatites were recognized as calcium phosphates by, at least, 1789 (Dorozhkin ^2012a , ^2013a ). Here, it is worth noting that hydroxylapatite would be a more accurate abbreviation expansion of HA (perhaps, hydroxideapatite would be even better because it relates to calcium hydroxide) while by both the medical and material communities HA is usually expanded as hydroxyapatite .

Chemically pure HA crystallizes in the monoclinic space group P 2 ₁ /b (Elliott et al. ¹⁹⁷³ ). However, at temperatures above ~250 °C, there is a monoclinic to hexagonal phase transition in HA (space group P 6 ₃ /m) (Elliott ¹⁹⁹⁴ , ¹⁹⁹⁸ ; Mathew and Takagi ²⁰⁰¹ ; Rangavittal et al. ²⁰⁰⁰ ; Kim et al. ^2000a ). The structural difference between the two modifications represents the ordered, head-to-tail arrangement of OH groups located in the center of every other Ca ₂ triangle in the monoclinic low-temperature symmetry and the disordered arrangement of OH groups, where the head-to-tail and tail-to-head arrangements alternate throughout the channel in the hexagonal high-temperature symmetry. This induces strains that might be compensated by substitutions and/or ion vacancies. Some impurities, like partial substitution of hydroxide by fluoride or chloride, stabilize the hexagonal structure of HA at ambient temperature. Due to this reason, hexagonal HA is seldom the stoichiometric phase and very rare single crystals of natural HA always exhibit the hexagonal space group. The detailed description of the HA structure was first reported in 1964 (Kay et al. ¹⁹⁶⁴ ) and its interpretation in terms of aggregation of Ca ₉ (PO ₄ ) ₆ clusters, the so-called Posner’s clusters, has been widely used since publication of the article by Posner and Betts (Posner and Betts ¹⁹⁷⁵ ).

Due to the exceptional importance of HA for the human beings, its properties have been thoroughly investigated by many research groups and further studies are kept going. Namely, the crystal structure of HA is well described elsewhere (Elliott ¹⁹⁹⁴ ; White and Dong ²⁰⁰³ ; Mathew and Takagi ²⁰⁰¹ ), the detailed analysis of the electronic structure, bonding, charge transfer, optical and elastic properties is also available (Calderin et al. ²⁰⁰³ ; Rulis et al. ²⁰⁰⁴ ; Snyders et al. ²⁰⁰⁷ ; Ching et al. ²⁰⁰⁹ ), while the readers interested in Posner’s clusters are referred to still other papers (Treboux et al. ²⁰⁰⁰ ; Yin and Stott ²⁰⁰³ ; Kanzaki et al. ²⁰⁰¹ ). A shell model was developed to study the lattice dynamics of HA (Calderin et al. ²⁰⁰⁵ ), while a cluster growth model was created to illustrate its growth (Onuma and Ito ¹⁹⁹⁸ ). Polarization characteristics (Tanaka et al. ^2010a , ^b ), pyroelectric (Tofail et al. ²⁰⁰⁹ , ²⁰¹⁵ ) and piezoelectric (Tofail et al. ²⁰¹⁵ ; Bystrov ²⁰¹⁵ ) properties of HA as well as diffusion of protons inside the HA crystal lattice (Yashima et al. ²⁰¹⁴ ) were investigated. The thermodynamic properties of HA and other types of orthophosphate-based apatites were summarized (Drouet ²⁰¹⁵ ). First-principles calculations for the elastic properties of doped HA (Kawabata and Yamamoto ²⁰¹⁰ ) and vacancy formation in HA (Matsunaga and Kuwabara ²⁰⁰⁷ ) were performed. Other examples of the computer simulations of the structure and properties of HA are available elsewhere (de Leeuw ²⁰¹⁰ ; Corno et al. ²⁰¹⁰ ; Slepko and Demkov ²⁰¹¹ ; Aquilano et al. ²⁰¹⁴ , ²⁰¹⁵ ). Finally, an attempt was performed to explain an array of the curious characteristics of HA by referring to the hydroxyl ion channels extending in the direction of the c -axis, through a crystallographic column created by the overlapping calcium ion triangles (Uskoković ²⁰¹⁵ ).

Many techniques might be utilized for HA preparation; they can be divided into solid-state reactions and wet methods (Briak-Ben et al. ²⁰⁰⁸ ), which include precipitation, hydrothermal synthesis and hydrolysis of other CaPO ₄ . However, in all cases, Ca- and PO ₄ -containing chemicals must be mixed to get the Ca/P ratio strictly equal to 1.67. Nevertheless, even under the ideal stoichiometric conditions, the precipitates are generally non-stoichiometric, suggesting intermediate formation of precursor phases, such as ACP and CDHA. Usually, unsintered HA is poorly crystalline and often non-stoichiometric, resembling the aforementioned CDHA. However, well crystalline HA can be prepared from aqueous solutions at relatively high (10–11) pH and elevated (>90 °C) temperatures (Markovic et al. ²⁰⁰⁴ ). HA with the Ca/P ratio >1.67 (Ca-rich HA) might be prepared as well (Bonel et al. ¹⁹⁸⁸ ). The detailed information on HA synthesis is available elsewhere (Narasaraju and Phebe ¹⁹⁹⁶ ; Riman et al. ²⁰⁰² ; Koutsopoulos ²⁰⁰² ; Norton et al. ²⁰⁰⁶ ; Sadat-Shojai et al. ²⁰¹³ ). In addition, there are good reviews on HA solubility, crystal growth and intermediate phases of HA crystallization (Rakovan ²⁰⁰² ), as well as on HA dissolution (Dorozhkin ^2012e ).

Pure HA never occurs in biological systems. However, due to the chemical similarities to bone and teeth mineral (Table  ³ ), HA is widely used as coatings on orthopedic (e.g., hip joint prosthesis) and dental implants (Dorozhkin ^2013c ; Suchanek and Yoshimura ¹⁹⁹⁸ ; Sun et al. ²⁰⁰¹ ; Ong and Chan ¹⁹⁹⁹ ; Dey et al. ²⁰⁰⁹ ; Dorozhkin ^2015a ). HA bioceramics is very popular as well (Yuan et al. ²⁰⁰⁹ ; Engin and Tas ¹⁹⁹⁹ ; Mangano et al. ²⁰⁰⁸ ). Due to a great similarity to biological apatite, over a long time HA has been used in liquid chromatography of nucleic acids, proteins and other biological compounds (Bernardi ¹⁹⁶⁵ ; Brand et al. ²⁰⁰⁸ ; Hou et al. ²⁰¹¹ ; Hilbrig and Freitag ²⁰¹² ; Niimi et al. ²⁰¹⁴ ; Pinto et al. ²⁰¹⁴ ) and for drug delivery purposes (Uskoković and Uskoković ²⁰¹¹ ; Chen et al. ^2011a ; Li et al. ²⁰¹³ ; Long et al. ²⁰¹³ ; Feng et al. ^2013a ). Also, HA is added to some brands of toothpaste as a gentle polishing agent instead of calcium carbonate (Niwa et al. ²⁰⁰¹ ; Kim et al. ²⁰⁰⁶ ). Non-biomedical applications of HA include its using as an environmental-friendly filler for elastomers (Pietrasik et al. ²⁰⁰⁸ ), a low-temperature sorbent (Bailliez et al. ²⁰⁰⁴ ; Corami et al. ²⁰⁰⁸ ) and/or stabilizer (Wang et al. ^2014b ) of poisonous chemical elements, a high-temperature sorbent for carbon dioxide (Landi et al. ²⁰¹⁴ ), both a catalyst (Xu et al. ²⁰¹⁰ ; Rodrigues et al. ²⁰¹⁴ ) and a carrier for other catalysts (Domínguez et al. ²⁰⁰⁹ ; Sun et al. ²⁰⁰⁹ ; Vukomanović et al. ²⁰¹² ), a material for ultraviolet light protection (Holzmann et al. ²⁰⁰⁹ ) and sunscreen filter (Piccirillo et al. ²⁰¹⁴ ), as well as a component of various sensors (Nagai et al. ¹⁹⁸⁸ ; Petrucelli et al. ¹⁹⁹⁶ ; Tagaya et al. ²⁰¹⁰ ; Khairnar et al. ²⁰¹¹ ). Finally, highly flexible and nonflammable inorganic paper could be prepared from HA (Lu et al. ^2014a ).

FA (or FAp)

Fluorapatite [Ca ₅ (PO ₄ ) ₃ F, but is usually written as Ca ₁₀ (PO ₄ ) ₆ F ₂ to denote that the crystal unit cell comprises two molecules; the IUPAC name is pentacalcium fluoride tris(orthophosphate)] is the only ion-substituted CaPO ₄ , considered in this review. Since the presence of 2.5 % of fluorides in natural apatites was established by 1798 (Dobson ¹⁷⁹⁸ ), this date might be accepted as the earliest hearing of FA.

FA is the hardest (five according to the Mohs’ scale of mineral hardness), most stable and least soluble compound among all CaPO ₄ (Table  ¹ ). In addition, it is the most thermally stable CaPO ₄ with the melting point at ~1650 °C (Tõnsuaadu et al. ²⁰¹² ). Perhaps, such “extreme” properties of FA are related to the specific position of F ⁻ ions in the center of Ca(2) triangles of the crystal structure (Elliott ¹⁹⁹⁴ ; White and Dong ²⁰⁰³ ). Due to its properties, FA is the only CaPO ₄ that naturally forms large deposits suitable for the commercial use (McConnell ¹⁹⁷³ ; Becker ¹⁹⁸⁹ ; Rakovan and Pasteris ²⁰¹⁵ ; see also Fig.  ² ). Preparation techniques of the chemically pure FA are similar to the aforementioned ones for HA but the synthesis must be performed in presence of the necessary amount of F ⁻ ions (usually, NaF or NH ₄ F is added). Under some special crystallization conditions (e.g., in presence of gelatin or citric acid), FA might form unusual dumbbell-like fractal morphology that finally are closed to spheres (Fig.  ⁸ ; Busch et al. ¹⁹⁹⁹ ; Wu et al. ²⁰¹⁰ ). In addition, FA is the only CaPO ₄ , which melts without decomposition; therefore, big (up to 30 cm long and, for shorter lengths, up to 1.9 cm wide) single FA crystals might be grown from FA melts (Mazelsky et al. ^1968a ; Loutts and Chai ¹⁹⁹³ ). Similar to that for HA (see CDHA), an existence of CaF ₂ -deficient FA was also detected but for the crystals grown from the FA melt only (Mazelsky et al. ^1968a ; Warren ¹⁹⁷² ). In addition, FA with an excess of CaF ₂ was prepared (Mann and Turner ¹⁹⁷² ). A hierarchical structure for FA was proposed (Dorozhkin ^2007a ). The crystal structure of FA for the first time was studied in 1930 (Mehmel ¹⁹³⁰ ; Naray-Szabo ¹⁹³⁰ ) and is well described elsewhere (Elliott ¹⁹⁹⁴ ; White and Dong ²⁰⁰³ ; Mathew and Takagi ²⁰⁰¹ ). Computer simulations of the FA structure were performed as well (Li et al. ²⁰¹⁵ ). The detailed analysis of the electronic structure, bonding, charge transfer and optical properties of FA (Rulis et al. ²⁰⁰⁴ ), as well as its NMR study under pressure (Pavan et al. ²⁰¹² ) is available as well. In addition, there are reviews on FA solubility (Rakovan ²⁰⁰² ) and the dissolution mechanism (Dorozhkin ^2012e ).

Fig. 8

FA easily forms solid solutions with HA with any desired F/OH molar ratio. Such compounds are called fluorhydroxyapatites (FHA) (Nikcevic et al. ²⁰⁰⁴ ; Montazeri et al. ²⁰¹⁰ ; Zhu et al. ²⁰¹² ) or hydroxyfluorapatites (HFA) (Rodríguez-Lorenzo et al. ²⁰⁰³ ; Azami et al. ²⁰¹² ) and described with a chemical formula Ca ₁₀ (PO ₄ ) ₆ (OH) _{2− x} F _x , where 0 <  x  < 2. If the F/OH ratio is either uncertain or not important, the chemical formula of FHA and HFA is often written as Ca ₁₀ (PO ₄ ) ₆ (F,OH) ₂ . The lattice parameters, crystal structure, solubility and other properties of FHA and HFA lay in between of those for the chemically pure FA and HA. Namely, the substitution of F for OH results in a contraction in the a -axis with no significant change in the c -axis dimensions and greater resolution of the IR absorption spectra.

Similar to pure HA, pure FA never occurs in biological systems. Obviously, a lack of the necessary amount of toxic fluorides (the acute toxic dose of fluoride is ~5 mg/kg of body weight) in living organisms is the main reason of this fact (pure FA contains 3.7 % mass. F). Enameloid of shark teeth (Lowenstam and Weiner ¹⁹⁸⁹ ; Daculsi et al. ¹⁹⁹⁷ ; Prostak et al. ¹⁹⁹³ ; Dahm and Risnes ¹⁹⁹⁹ ; Carr et al. ²⁰⁰⁶ ; Enax et al. ²⁰¹⁴ ) and some exoskeletons of mollusks (Leveque et al. ²⁰⁰⁴ ) seem to be the only exclusions because they contain substantial amounts of fluoride, with is presented there as ion-substituted, non-stoichiometric FHA or HFA. Among all normal calcified tissues of humans, the highest concentration of fluorides is found in dentine and cementum, while the lowest—in dental enamel (Table  ³ ). Nevertheless, one should stress that the amount of fluorides on the very surface of dental enamel might be substantially increased using fluoride-containing toothpastes and mouthwashes (Schemehorn et al. ²⁰¹¹ ; Hattab ²⁰¹³ ). However, in no case, the total amount of fluorides is enough to form pure FA.

Contrary to the initial expectations (Heling et al. ¹⁹⁸¹ ), chemically pure FA is not used for grafting purposes. Presumably, this is due to the lowest solubility, good chemical stability of FA and toxicity of high amounts of fluorides. However, attempts to test FA-containing formulations (Gineste et al. ¹⁹⁹⁹ ; Agathopoulos et al. ²⁰⁰³ ; Yoon et al. ²⁰⁰⁵ ; Bogdanov et al. ²⁰⁰⁹ ; Nordquist et al. ²⁰¹¹ ), ion-substituted FA (Kheradmandfard et al. ²⁰¹² ; Sharifnabi et al. ²⁰¹⁴ ), FHA (Savarino et al. ²⁰⁰³ ; Vitkovič et al. ²⁰⁰⁹ ) and porous FA bioceramics (Chaari et al. ²⁰⁰⁹ ) are kept performing. The effect of fluoride contents in FHA on both osteoblast behavior (Qu and Wei ²⁰⁰⁶ ; Bhadang et al. ²⁰¹⁰ ) and leukemia cells proliferation (Theiszova et al. ²⁰⁰⁸ ) has been described. Non-biomedical applications of FA include luminescent light tubes (Davis et al. ¹⁹⁷¹ ) and thermometers (Fu et al. ²⁰¹⁵ ), laser materials (Mazelsky et al. ^1968b ; Ohlmann et al. ¹⁹⁶⁸ ) (in all these cases various dopants are necessary), as well as catalysts (An et al. ²⁰⁰⁹ ).

OA (or OAp, or OXA)

Oxyapatite [Ca ₁₀ (PO ₄ ) ₆ O; the IUPAC name is decacalcium oxide hexakis(phosphate), mineral voelckerite] is the least stable and, therefore, the least known CaPO ₄ , which, probably, does not exist at all. Nevertheless, a name “voelckerite” was introduced in 1912 by A.F. Rogers (1887–1957) for a hypothetical mineral with the chemical composition of 3Ca ₃ (PO ₄ ) ₂  + CaO (Rogers ¹⁹¹² , ¹⁹¹⁴ ), to honor an English agricultural chemist John Christopher Augustus Voelcker (1822–1884), who, in 1883, first showed an apparent halogen deficiency in some natural apatites (Voelcker ¹⁸⁸³ ). Therefore, 1883 might be accepted as the earliest hearing on OA.

To the best of my findings, phase pure OA has never been obtained at room temperatures; therefore, its properties are not well established. Furthermore, still there are serious doubts that pure OA can exist. Since hydroxyl ions in HA appear to be the most mobile ones and upon exposure to high temperatures are the first to leave the lattice, a mixture (or a solid solution?) of OA and HA (so-called “oxy-HA”, chemical formula: Ca ₁₀ (PO ₄ ) ₆ (OH) _2−2x O _x V _x , where V represents an OH ⁻ vacancy) might be prepared by a partial dehydroxylation of HA at temperatures exceeding ~900 °C (e.g., during plasma spray of HA) strictly in the absence of water vapor (Gross et al. ¹⁹⁹⁸ ; Hartmann et al. ²⁰⁰¹ ; Alberius-Henning et al. ²⁰⁰¹ ; Wang and Dorner-Reisel ²⁰⁰⁴ ; Liu and Shen ²⁰¹² ). It also might be crystallized in glass-ceramics (van’t Hoen et al. ²⁰⁰⁷ ). OA is very unstable and has no stability field in aqueous conditions (Duff ¹⁹⁷² ). Namely, data are available that oxy-HA containing less than 25 % HA (i.e., almost OA) during further dehydration decomposes to a mixture of α-TCP and TTCP. In addition, OA is very reactive and transforms to HA in contact with water vapor (HA reconstitution) (Gross et al. ¹⁹⁹⁸ ). The largest impediment to active research in OA is the non-availability of a user-friendly approach to measure the concentration of hydroxyl ions (Gross and Pluduma ²⁰¹² ).

Therefore, computer-modeling techniques have been employed to qualitatively and quantitatively investigate the dehydration of HA to OA (de Leeuw et al. ²⁰⁰⁷ ). OA has the hexagonal space group symmetry P 6¯ (174) of cesanite type (White and Dong ²⁰⁰³ ), while the space group symmetry for partially dehydrated HA was found to change from hexagonal P63/m to triclinic P 1¯ when more than ca. 35 % of the structurally bound water had been removed (Alberius-Henning et al. ²⁰⁰¹ ). On the c -axis, pure OA should have a divalent ion O ²⁻ coupled with a vacancy instead of two neighboring monovalent OH ⁻ ions.

Due to the aforementioned problems with OA preparation, it cannot be found in biological systems. In addition, no information on the biomedical applications of OA is available either. Plasma-sprayed coatings of CaPO ₄ , in which OA might be present as an admixture phase, seem to be the only exception (Dorozhkin ^2015a ).

To conclude, one should mention that various types of calcium peroxy-HA (oxygenated HA) could be prepared both in the presence of hydrogen peroxide (Rey et al. ¹⁹⁷⁸ ) and by HA heating in an oxygen-rich atmosphere (Zhao et al. ²⁰⁰⁰ ; Yu et al. ²⁰⁰⁷ ). Similar to that for OA, none of them was ever prepared as a phase pure individual compound and no examples of the biomedical applications of such types of CaPO ₄ were found.

TTCP (or TetCP)

Tetracalcium phosphate or tetracalcium diorthophosphate monoxide (Ca ₄ (PO ₄ ) ₂ O; the IUPAC name is tetracalcium oxide bis(orthophosphate); the mineral hilgenstockite) is the most basic CaPO ₄ , however, its solubility in water is higher than that of HA (Table  ¹ ). TTCP has been known since 1883 (Hilgenstock ¹⁸⁸³ ), while the mineral hilgenstockite was named to honor a German metallurgist Gustav Hilgenstock (1844–1913), who first discovered it in Thomas slag from blast furnaces (Hilgenstock ¹⁸⁸³ , ¹⁸⁸⁷ ). Its major industrial importance stems from the fact that it is formed by the reactions between orthophosphates and lime in the manufacture of iron, and through these reactions, TTCP has a significant role in controlling the properties of the metal.

TTCP cannot be precipitated from aqueous solutions. It can be prepared only under the anhydrous conditions by solid-state reactions at temperatures above ~1300 °C, e.g., by heating homogenized equimolar quantities of DCPA and CaCO ₃ in dry air, or in a flow of dry nitrogen (Elliott ¹⁹⁹⁴ , ¹⁹⁹⁸ ; Kai et al. ²⁰⁰⁹ ). These reactions should be carried out in a dry atmosphere, in vacuum or with rapid cooling (to prevent uptake of water and formation of HA). Easily DCPA might be replaced by ammonium orthophosphates (Romeo and Fanovich ²⁰⁰⁸ ; Jalota et al. ²⁰⁰⁵ ), while calcium carbonate might be replaced by calcium acetate (Jalota et al. ²⁰⁰⁵ ); however, in all cases, Ca/P ratio must be equal to 2.00. Furthermore, TTCP often appears as an unwanted by-product in plasma-sprayed HA coatings, where it is formed as a result of the thermal decomposition of HA to a mixture of high-temperature phases of α-TCP, TTCP and CaO (Moseke and Gbureck ²⁰¹⁰ ). Nevertheless, TTCP might be produced at 900 °C by calcining of a precipitated ACP with Ca/P = 2 in vacuum; the process is suggested to occur via a intermediate formation of a mixture of OA + CaO (Gross and Rozite ²⁰¹⁵ ).

TTCP is metastable: in both wet environment and aqueous solutions it slowly hydrolyses to HA and calcium hydroxide (Elliott ¹⁹⁹⁴ , ¹⁹⁹⁸ ; Martin and Brown ¹⁹⁹³ ). Consequently, TTCP is never found in biological calcifications. In medicine, TTCP is widely used for preparation of various self-setting CaPO ₄ formulations (Dorozhkin ^2013b ; Moseke and Gbureck ²⁰¹⁰ ); however, to the best of my knowledge, there is no commercial bone-substituting product consisting solely of TTCP. For the comprehensive information on TTCP, the readers are referred to a special review (Moseke and Gbureck ²⁰¹⁰ ), while the spectra (Jillavenatesa and Condrate ¹⁹⁹⁷ ) and solubility (Pan and Darvell ^2009a ) of TTCP are well described elsewhere.

To finalize the description of individual CaPO ₄ , one should mention on an interesting opinion, that all types of CaPO ₄ listed in Table  ¹ might be classified into three major structural types (Mathew and Takagi ²⁰⁰¹ ; Chow and Eanes ²⁰⁰¹ ). They comprise: (1) the apatite type, Ca ₁₀ (PO ₄ ) ₆ X ₂ , which includes HA, FA, OA, CDHA, OCP and TTCP; (2) the glaserite type, named after the mineral glaserite, K ₃ Na(SO ₄ ) ₂ , which includes all polymorphs of TCP and, perhaps, ACP; (3) the Ca-PO ₄ sheet-containing compounds, which include DCPD, DCPA, MCPM and MCPA. According to the authors, a closer examination of the structures revealed that all available CaPO ₄ could be included into distorted glaserite type structures, but with varying degrees of distortion (Mathew and Takagi ²⁰⁰¹ ; Chow and Eanes ²⁰⁰¹ ).

Biphasic, triphasic and multiphasic CaPO4 formulations

CaPO ₄ might form biphasic, triphasic and multiphasic (polyphasic) formulations, in which the individual components cannot be separated from each other. Presumably, the individual phases of such compositions are homogeneously “mixed” at a far submicron level (<0.1 μm) and strongly integrated with each other. Nevertheless, the presence of all individual phases is easily seen by X-ray diffraction technique (Dorozhkin ^2012d ).

The usual way to prepare multiphasic formulations consists of sintering non-stoichiometric compounds, such as ACP and CDHA, at temperatures above ~700 °C. Furthermore, a thermal decomposition of the stoichiometric CaPO ₄ at temperatures above ~1300 °C might be used as well; however, this approach often results in formation of complicated mixtures of various products including admixtures of CaO, calcium pyrophosphates, etc. (Dorozhkin ^2012d ; Nakano et al. ^2002b ). Namely, transformation of HA into polyphasic CaPO ₄ by annealing in a vacuum occurs as this: the outer part of HA is transformed into α-TCP and TTCP, while the α-TCP phase of the surface further transforms into CaO. Besides, in the boundary phase, HA is transformed into TTCP (Nakano et al. ^2002b ).

Historically, Nery and Lynch with co-workers first used the term biphasic calcium phosphate (BCP) in 1986 to describe a bioceramic that consisted of a mixture of HA and β-TCP (Ellinger et al. ¹⁹⁸⁶ ). Based on the results of X-ray diffraction analysis, these authors found that the “tricalcium phosphate” preparation material used in their early publication (Nery et al. ¹⁹⁷⁵ ) was in fact a mixture of ~20 % HA and ~80 % β-TCP. Currently, only biphasic and triphasic CaPO ₄ formulations are known; perhaps, more complicated formulations will be manufactured in future. Furthermore, nowadays, multiphasic and/or polyphasic compositions consisting of high-temperature phases of CaPO ₄ , such as α-TCP, β-TCP, HA and, perhaps, high-temperature ACP, OA and TTCP, are known only. No precise information on multiphasic compositions, containing MCPM, MCPA, DCPD, DCPA, low-temperature ACP, OCP and CDHA has been found in literature (Dorozhkin ^2012d ). Perhaps, such formulations will be produced in future.

All BCP formulations might be subdivided into two major groups: those consisting of CaPO ₄ having either the same (e.g., α-TCP and β-TCP) or different (e.g., β-TCP and HA) molar Ca/P ratios. Among all known BCP formulations, BCP consisting of HA and β-TCP is both the most known and the best investigated (Dorozhkin ^2012d ). In 1986, LeGeros in USA and Daculsi in France initiated the basic studies on preparation of this type of BCP and its in vitro properties. This material is soluble and gradually dissolves in the body, seeding new bone formation as it releases calcium and orthophosphate ions into the biological medium. Presently, commercial BCP products of different or similar HA/β-TCP ratios are manufactured in many parts of the world as bone graft or bone substitute materials for orthopedic and dental applications under various trademarks and several manufacturers (Dorozhkin ^2012d ). A similar combination of α-TCP with HA forms BCP as well (Reid et al. ²⁰⁰⁵ ; Pan et al. ²⁰⁰⁶ ; Kui ²⁰⁰⁷ ; Li et al. ^2009b ).

Recently, the concept of BCP has been extended by preparation and characterization of biphasic TCP (BTCP), consisting of α-TCP and β-TCP phases (Wang et al. ^2006a ; Li et al. ²⁰⁰⁷ , ^2008a ). It is usually prepared by heating ACP precursors (Wang et al. ^2006a ; Li et al. ²⁰⁰⁷ , ^2008a ), in which the α-TCP/β-TCP ratio can be controlled by aging time and pH value during synthesis of the amorphous precursor (Li et al. ²⁰⁰⁷ ). Furthermore, triphasic formulations, consisting of HA, α-TCP and β-TCP (Vani et al. ²⁰⁰⁹ ) or HA, α-TCP and TTCP (Nakano et al. ^2002b ) have been prepared (Dorozhkin ^2012d ).

It is important to recognize, that the major biomedical properties (such as bioactivity, bioresorbability, osteoconductivity and osteoinductivity) of the multiphasic formulations might be adjusted by changing the ratios among the phases. When compared to both α- and β-TCP, HA is a more stable phase under the physiological conditions, as it has a lower solubility (Table  ¹ ) and, thus, slower resorption kinetics. Therefore, due to a higher biodegradability of the α- or β-TCP component, the reactivity of BCP increases with the TCP/HA ratio increasing. Thus, in vivo bioresorbability of BCP can be adjusted through the phase composition. Similar conclusions are also valid for both the biphasic TCP (in which α-TCP is a more soluble phase) and the triphasic (HA, α-TCP and β-TCP) formulations. Further details on this subject might be found in a topical review (Dorozhkin ^2012d ).

Ion-substituted CaPO4

Last, one should very briefly mention on existence of carbonated HA (Lafon et al. ²⁰⁰⁸ ; Tonegawa et al. ²⁰¹⁰ ; Yahia and Jemal ²⁰¹⁰ ; Silvester et al. ²⁰¹⁴ ; Fleet ²⁰¹⁵ ), chlorapatite (Kannan et al. ²⁰⁰⁷ ; García-Tuñón et al. ²⁰¹² ; Zhao et al. ²⁰¹³ ), as well as on a great number of CaPO ₄ with various ionic substitutions (CaPO ₄ with dopants) (Pan and Fleet ²⁰⁰² ; Rey et al. ²⁰⁰⁶ ; Kannan et al. ²⁰⁰⁸ ; Boanini et al. ^2010b ; Shepherd et al. ²⁰¹² ; Adzila et al. ²⁰¹² ; Grigg et al. ²⁰¹⁴ ; Šupová ²⁰¹⁵ ). In principle, any ion in CaPO ₄ might be substituted by other ion(s). Usually, the ion-substituted CaPO ₄ are of a non-stoichiometric nature with just a partial ionic substitution and there are too many of them to be described here. Currently, this is a hot investigation topic; therefore, the readers are referred to the special literature (Pan and Fleet ²⁰⁰² ; Rey et al. ²⁰⁰⁶ ; Kannan et al. ²⁰⁰⁸ ; Boanini et al. ^2010b ; Shepherd et al. ²⁰¹² ; Adzila et al. ²⁰¹² ; Grigg et al. ²⁰¹⁴ ; Šupová ²⁰¹⁵ ). In addition, there is a very good review, in which the structures of more than 75 chemically different apatites have been discussed (White and Dong ²⁰⁰³ ).

To finalize this brief topic of ion-substituted CaPO ₄ , it is important to note that chemical elements not found in natural bones can be intentionally incorporated into CaPO ₄ biomaterials to get special properties. For example, addition of Ag ⁺ (Pushpakanth et al. ²⁰⁰⁸ ; Zhang et al. ^2010b ; Kim et al. ¹⁹⁹⁸ ; Kolmas et al. ²⁰¹⁴ ), Zn ²⁺ (Kim et al. ¹⁹⁹⁸ ; Kolmas et al. ²⁰¹⁴ ; Stanić et al. ²⁰¹⁰ ) and Cu ²⁺ (Kim et al. ¹⁹⁹⁸ ; Kolmas et al. ²⁰¹⁴ ; Stanić et al. ²⁰¹⁰ ; Li et al. ²⁰¹⁰ ) was used for imparting antimicrobial effect, while radioactive isotopes of ⁹⁰ Y (Thomas et al. ²⁰⁰⁸ ), ¹⁵³ Sm (Chinol et al. ¹⁹⁹³ ; Argüelles et al. ¹⁹⁹⁹ ; O’Duffy et al. ¹⁹⁹⁹ ) and ¹⁸⁶ Re (Chinol et al. ¹⁹⁹³ ) were incorporated into HA bioceramics and injected into knee joints to treat rheumatoid joint synovitis (Thomas et al. ²⁰⁰⁸ ; Chinol et al. ¹⁹⁹³ ; O’Duffy et al. ¹⁹⁹⁹ ). More to the point, apatites were found to incorporate individual molecules, such as water, oxygen and carbon dioxide (Rey et al. ²⁰⁰⁶ ).

Finally, to conclude the description of the known CaPO ₄ , one should mention that in spite of the well-defined crystallographic data (Table  ⁴ ) and, therefore, the well-defined shapes of the single crystals, various types of CaPO ₄ can be prepared with the controllable sizes from nano- to macro-scale (up to centimeter size) and the diverse shapes including zero- (particles), one- (rods, fibers, wires and whiskers), two- (sheets, disks, plates, belts, ribbons and flakes) and three-dimensional morphologies (Sadat-Shojai et al. ²⁰¹³ ; Lin et al. ²⁰¹⁴ ). The latter might be of versatile morphologies and shapes (Fig.  ⁸ is an example), including porous and hollow structures.

Bone

Bone, also called osseous tissue (Latin: os ), is a type of hard endoskeletal connective tissue found in many vertebrate animals. All bones of a single animal are, collectively, known as the skeleton. True bones are present in bony fish (osteichthyes) and all tetrapods. Bones support body structures, protect internal organs and, in conjunction with muscles, facilitate movement (Loveridge ¹⁹⁹⁹ ). In addition, bones are also involved with blood cell formation, calcium metabolism and act for mineral storage. For a material scientist, bone is a dynamic, highly vascularized solid tissue that is formed from a complicated biocomposite containing both inorganic (Table  ³ ) and bioorganic (chiefly, collagen) compounds, in which nanodimensional crystals of the inorganic phases are dispersed in the meshes of the bioorganic ones (Palmer et al. ²⁰⁰⁸ ; Nightingale and Lewis ¹⁹⁷¹ ; Currey ²⁰⁰² ; Rho et al. ¹⁹⁹⁸ ; Tzaphlidou ²⁰⁰⁸ ). More than 20 types of collagen have been reported in the human body, among which type I collagen is the most abundant protein and provides much of the structural integrity for connective tissue, particularly in bones, tendons and ligaments. Furthermore, for a physiologist, bone is a living organ populated by living cells (mainly, osteoblasts, osteoclasts and osteocytes); however, that is another story. The inorganic to bioorganic ratio is approximately 75–25 % by dry weight and about 65–35 % by volume. This ratio not only differs among animals, among bones in the same animal and over time in the same animal but also it exerts a major control on the material properties of bone, such as its toughness, ultimate strength and stiffness. In general, load-bearing ability of bones depends on not only architectural properties, such as cortical thickness and bone diameter, but also intrinsic, size-independent, material properties such as porosity, level of mineralization, crystal size and properties derived from the organic phase of bone (Davison et al. ²⁰⁰⁶ ). A higher mineral to collagen ratio typically yields stronger, but more brittle, bones (Turner and Burr ¹⁹⁹³ ; Currey ²⁰⁰⁴ ; Currey et al. ²⁰⁰⁴ ). For example, bones from the leg of a cow have a relatively high concentration of CaPO ₄ (for support), whereas bones from the antler of a deer have a relatively high concentration of collagen (for flexibility) (Pan and Darvell ^2009a ). It is interesting to note that bones exhibit several physical properties such as piezoelectricity (Anderson and Eriksson ¹⁹⁷⁰ ) and pyroelectricity (Lang ¹⁹⁶⁶ ).

Stability of the mineral composition of bones has a very long history: CaPO ₄ were found in dinosaur fossils (Elorza et al. ¹⁹⁹⁹ ; Eagle et al. ²⁰¹⁰ ; Haynes ¹⁹⁶⁸ ; Rensberger and Watabe ²⁰⁰⁰ ; Kolodny et al. ¹⁹⁹⁶ ; Trueman and Tuross ²⁰⁰² ). Therefore, organisms have had a great deal of time to exploit the feedback between composition and structure in apatite, on the one hand, and benefit from its biological functionality, on the other. Bones of the modern animals are a relatively hard and lightweight porous composite material, formed mostly of biological apatite (i.e., poorly crystalline CDHA with ionic substitutions). It has relatively high compressive strength but poor tensile strength (Currey and Brear ¹⁹⁹⁰ ). While bones are essentially brittle, they have a degree of significant plasticity contributed by their bioorganic components.

The distribution of the inorganic and bioorganic phases depends on a highly complex process that takes place during bone formation. Each of these components may be assembled in different proportions creating two different architectural structures depending on the bone type and function. They are characterized by different structural features that strongly correlate with the mechanical performance of the tissue. These two types of bones are: the cortical bones (or compact bones) which are dense and the cancellous bones (also known as trabecular or spongious bones) which are less dense and less stiff than the compact ones. Usually, bones are composed of a relatively dense outer layer of cortical bone covering an internal mesh-like structure (average porosity of 75–95 %) of cancellous bone, the density of which is about 0.2 g/cm ³ but it may vary at different points (Fig.  ¹⁰ ). Cortical bones make up a large portion of the skeletal mass; they have a high density (~1.80 g/cm ³ ) and a low surface area. Cancellous bones have an open meshwork or honeycomb-like structure. They have a relatively high surface area but form a smaller portion of the skeleton. Bones are a porous material with the pore sizes ranging from 1 to 100 μm in normal cortical bones and 200–400 μm in trabecular ones. 55–70 % of the pores in trabecular bones are interconnected. The porosity reduces the strength of bones but also reduces their weight (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; Lowenstam and Weiner ¹⁹⁸⁹ ; O’Neill ²⁰⁰⁷ ; Skinner ²⁰⁰⁵ ; Daculsi et al. ¹⁹⁹⁷ ; Weiner and Wagner ¹⁹⁹⁸ ; Currey ²⁰⁰² ; Rho et al. ¹⁹⁹⁸ ; Rey et al. ²⁰⁰⁹ ; Lerebours et al. ²⁰¹⁵ ).

Fig. 10

Bones can be either woven or lamellar. The fibers of woven bones are randomly aligned and as the result have a low strength. In contrast, lamellar bones have parallel fibers and are much stronger. Woven bones are put down rapidly during growth or repair (Watt ¹⁹²⁵ ) but as growth continues, they are often replaced by lamellar bones. The replacement process is called “secondary bone formation” and described in details elsewhere (Olszta et al. ²⁰⁰⁷ ). In addition, bones might be long, short, flat, sutural, sesamoid and irregular (Fig.  ¹¹ ). The sizes and shapes of bones reflect their function. Namely, broad and flat bones, such as scapulae, anchor large muscle masses, flat skull bones protect the brain, ribs protect the lungs, pelvis protects other internal organs, short tubular bones in the digits of hands and feet provide specific grasping functions, hollow and thick-walled tubular bones, such as femur or radius, support weight and long bones enable locomotion (Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ). Long bones are tubular in structure (e.g., the tibia). The central shaft of a long bone is called the diaphysis and has a medullar cavity filled with bone marrow (Fig.  ¹⁰ ). Surrounding the medullar cavity is a thin layer of cancellous bone that also contains marrow. The extremities of the bone are called the epiphyses and are mostly cancellous bone covered by a relatively thin layer of compact bone. Short bones (e.g., finger bones) have a similar structure to long bones, except that they have no medullar cavity. Flat bones (e.g., the skull and ribs) consist of two layers of compact bone with a zone of cancellous bone sandwiched between them. Irregular bones (e.g., vertebrae) do not conform to any of the previous forms. Thus, bones are shaped in such a manner that strength is provided only where it is needed. All bones contain living cells embedded in a mineralized organic matrix that makes up the main bone material (Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ; Boskey and Roy ²⁰⁰⁸ ). The structure of bones is most easily understood by differentiating between seven levels of organization because bones exhibit a strongly hierarchical structure (Palmer et al. ²⁰⁰⁸ ; Weiner and Wagner ¹⁹⁹⁸ ; Nightingale and Lewis ¹⁹⁷¹ ; Currey ²⁰⁰² , ²⁰⁰⁵ ; Rho et al. ¹⁹⁹⁸ ; Rensberger and Watabe ²⁰⁰⁰ ; Cui et al. ²⁰⁰⁷ ; Fratzl and Weinkamer ²⁰⁰⁷ ; Boskey and Coleman ²⁰¹⁰ ; Meyers et al. ²⁰⁰⁸ ; Reznikov et al. ^2014a , ^b ). One should stress that, taking into account the presence of ordered and disordered materials, nine hierarchical levels for lamellar bones have been differentiated (Fig.  ¹² ; Reznikov et al. ^2014b ).

Fig. 11

Fig. 12

The mechanical properties of bones reconcile high stiffness and high elasticity in a manner that is not yet possible with synthetic materials (Meyers et al. ²⁰⁰⁸ ). Cortical bone specimens have been found to have tensile strength in the range of 79–151 MPa in longitudinal direction and 51–56 MPa in transversal direction. Bone’s elasticity is also important for its function giving the ability to the skeleton to withstand impact. Estimates of modulus of elasticity of bone samples are of the order of 17–20 GPa in longitudinal direction and of 6–13 GPa in the transversal direction (Athanasiou et al. ²⁰⁰⁰ ). The elastic properties of bone were successfully modeled at the level of mineralized collagen fibrils via step-by-step homogenization from the staggered arrangement of collagen molecules up to an array of parallel mineralized fibrils (Nikolov and Raabe ²⁰⁰⁸ ). The investigations revealed that bone deformation was not homogeneous but distributed between a tensile deformation of the fibrils and a shearing in the interfibrillar matrix between them (Gupta et al. ²⁰⁰⁵ ; Peterlik et al. ²⁰⁰⁶ ). Readers, who are interested in further details, are addressed to a good review on the effects of the microscopic and nano-scale structure on bone fragility (Ruppel et al. ²⁰⁰⁸ ).

The smallest level of the bone hierarchy consists of the molecular components: water, biological apatite, collagen and other proteins (Fig.  ¹³ ; Duer ²⁰¹⁵ ). The second smallest hierarchical level is formed by mineralization of collagen fibrils, which are of 80–100 nm thickness and a length of a few to tens of microns (Fig.  ¹² ). Thus, biocomposites of biological apatite and molecules of type I collagen are formed (Pasteris et al. ²⁰⁰⁸ ; Weiner and Wagner ¹⁹⁹⁸ ; Nightingale and Lewis ¹⁹⁷¹ ; Tzaphlidou ²⁰⁰⁸ ; Fratzl et al. ²⁰⁰⁴ ). Some evidences for direct physical bonding between the collagen fibers and apatite crystals in bone were found (Marino and Becker ¹⁹⁶⁷ ). Atomic force microscopy was used to measure the crystals of biological apatite in mature cow bones (Eppell et al. ²⁰⁰¹ ). The crystals are always platelet like (elongated along the crystallographic c -axis) and very thin (Wopenka and Pasteris ²⁰⁰⁵ ; Clark and Iball ¹⁹⁵⁴ ; Rubin et al. ²⁰⁰³ ; Su et al. ²⁰⁰³ ), with remarkably uniform thicknesses (determined in transmission electron microscopy) of 2–4 nm (i.e., just a few unit cells thick—see Table  ³ ). They exist in bones not as discrete aggregates but rather as a continuous phase, which is indirectly evidenced by a very good strength of bones. This results in a very large surface area facing extracellular fluids, which is critically important for the rapid exchange of ions with these fluids. The nano-sized crystals of biological apatite are inserted in a nearly parallel way into the collagen fibrils, while the latter are formed by self-assembly of collagen triple helices (Weiner and Wagner ¹⁹⁹⁸ ; Nightingale and Lewis ¹⁹⁷¹ ; Cui et al. ²⁰⁰⁷ ; Landis et al. ¹⁹⁹⁶ ; Rosen et al. ²⁰⁰² ) using the self-organization mechanism (Sato ²⁰⁰⁶ ; Hartgerink et al. ²⁰⁰¹ ). In addition, experimental data from electron diffraction studies revealed that that the mineral plates of biological apatite are not quite as ordered as previously assumed (Olszta et al. ²⁰⁰⁷ ). This imperfect arrangement of nearly parallel crystals has been supported by SAXS and transmission electron microscopy studies (Burger et al. ²⁰⁰⁸ ). According to the latest data, water, which always present in bone mineral, appears to glue the mineral platelets together in a way akin to how water sticks two glass plates together—it provides firm binding between the plates, but at the same time allows slippage between them—flexibility for a stack of mineral platelets held together in this way (Duer ²⁰¹⁵ ).

Fig. 13

The lowest level of hierarchical organization of bone has successfully been simulated by CDHA precipitation on peptide-amphiphile nanodimensional fibers (Hartgerink et al. ²⁰⁰¹ ). However, apatite platelets nucleating on the surface of peptide tubules are not similar to the nanostructure of bone and they are only an example of surface induced nucleation (and not accurately characterized either), while the nanostructure of bone consists of intra-fibrillar platelets intercalated within the collagen fibrils. However, the interface between collagen and biological apatite is still poorly understood; for the available details, the readers are referred to a review devoted to the structure and mechanical quality of the collagen/mineral nanodimensional biocomposite of bones (Fratzl et al. ²⁰⁰⁴ ). There is still no clear idea why the crystals of biological apatite are platelet shaped even though dahllite has hexagonal crystal symmetry (Weiner and Wagner ¹⁹⁹⁸ ; Currey ²⁰⁰² ; Rho et al. ¹⁹⁹⁸ ; Rey et al. ²⁰⁰⁹ ). One possible reason is that they grow via an OCP transition phase, in which crystals are plate-shaped (Weiner and Wagner ¹⁹⁹⁸ ). Another explanation involves the presence of citrates, which strongly bound to (10Ī0) surface of biological apatite because of space matching (Hu et al. ²⁰¹⁰ ; Xie and Nancollas ²⁰¹⁰ ). Therefore, the crystal growth in the [10Ī0] direction becomes inhibited, while the citrate effect on other crystal surfaces of biological apatite appears to be very small owing to poor space matching. Thus, after crystal growth, the (10Ī0) crystal face becomes predominant resulting in plate-like morphology of biological apatite (Xie and Nancollas ²⁰¹⁰ ).

The processes of bone formation (ossification) and growth are very complicated ones and it is difficult to describe them without making a deep invasion into biology. It has been studied for decades (Watt ¹⁹²⁵ ) but still there are missing points. From the standpoint of chemists, formation of bones could be considered as calcification (i.e., deposition or precipitation of CaPO ₄ ) within the bioorganic matrix of connective tissues (mainly cartilage), resulting in formation of biocomposites (Palmer et al. ²⁰⁰⁸ ; Olszta et al. ²⁰⁰⁷ ). Cartilage is composed of collagen fibers, cells (chondrocytes and their precursor forms known as chondroblasts) and extracellular matrix (proteoglycans, which are a special class of heavily glycosylated glycoproteins). Very briefly, the initial stage of ossification involves synthesis and extracellular assembly of the collagen matrix framework of fibrils. At the second stage, chondrocytes calcify the matrix before undergoing the programmed cell death (apoptosis). At this point, blood vessels penetrate this calcified matrix, bringing in osteoblasts (they are mononuclear cells primarily responsible for bone formation), which use the calcified cartilage matrix as a template to build bone, and ions of calcium and orthophosphate to be deposited in the ossifying tissue (Hall ²⁰¹⁵ ). Therefore, the role of collagen in the nucleation, growth, structure and orientation of apatite crystals appears to be predominant (Wang et al. ²⁰¹² ). The biomineralization process is controlled to some extent by cells and the organic matrices made by those cells facilitate the deposition of crystals (Boskey and Roy ²⁰⁰⁸ ). As bone crystals grow, there is greater association with proteins, such as osteocalcin, that regulate remodeling (George and Veis ²⁰⁰⁸ ). Thus, in vivo formation of hard tissues always occurs by mineral reinforcement of the previously formed network of soft tissues (Palmer et al. ²⁰⁰⁸ ; Olszta et al. ²⁰⁰⁷ ; Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ; Cui et al. ²⁰⁰⁷ ).

Since the bioorganic matrix has control over the size and orientation of CaPO ₄ crystals, the latter grow with a specific crystalline orientation—the c -axes of the crystals are roughly parallel to the long axes of the collagen fibrils within which they are deposited (Palmer et al. ²⁰⁰⁸ ; Grynpas et al. ¹⁹⁹³ ; Weiner and Wagner ¹⁹⁹⁸ ; Currey ²⁰⁰² ; Rho et al. ¹⁹⁹⁸ ; Tzaphlidou ²⁰⁰⁸ ; Olszta et al. ²⁰⁰⁷ ). Earlier, it was believed that this process occurred via epitaxial growth mechanism (Marino and Becker ¹⁹⁷⁰ ). The same was suggested for dentine and enamel (Jodaikin et al. ¹⁹⁸⁸ ; Fincham et al. ¹⁹⁹⁵ ) (see “ ^Teeth ”), as well as for more primitive living organisms. For example, in the shell of the fossil marine animal Lingula brachiopod unguis that consists of a biological apatite, the crystal c -axes are oriented parallel to the β-chitin fibrils (Leveque et al. ²⁰⁰⁴ ; Williams et al. ¹⁹⁹⁸ ; Rohanizadeh and LeGeros ²⁰⁰⁷ ; Neary et al. ²⁰¹¹ ). Therefore, the orientation of biological apatite crystals parallel to the long axes of the organic framework could be a general feature of the CaPO ₄ biomineralization. However, the degree of the crystal orientation appears to be a useful parameter to evaluate an in vivo stress distribution, a nano-scale microstructure and a related mechanical function, a regenerative process of the healed bone, as well as to diagnose bone diseases such as osteoarthritis (Nakano et al. ²⁰⁰⁷ , ²⁰⁰⁸ ). It is interesting to note that contrary to what might be expected in accordance with possible processes of dissolution, formation and remineralization of hard tissues, no changes in phase composition of mineral part, crystal sizes (length, width and thickness) and arrangement of crystals on collagen fibers were detected in abnormal (osteoporotic) human bones compared to the normal ones (Suvorova et al. ²⁰⁰⁷ ).

Some animals, such as newts, are able to regenerate amputated limbs. This is, of course, of a high interest for the regenerative medicine. Therefore, bone regeneration in the forelimbs of mature newts was studied by noninvasive X-ray microtomography to image regenerating limbs from 37 to 85 days. The missing limb skeletal elements were restored in a proximal-to-distal direction, which reiterated the developmental patterning program. However, in contrast to this proximal–distal sequence, the portion of the humerus distal to the amputation site was found to fail to ossify in synchrony with the regenerating radius and ulna. This finding suggests that the replacement of cartilage with mineralized bone close to the amputation site is delayed with respect to other regenerating skeletal elements (Stock et al. ²⁰⁰³ ).

Unlike other mineralized tissues, bone continuously undergoes a remodeling process, as it is resorbed by specialized cells called osteoclasts and formed by another type of cells called osteoblasts (so-called “bone lining cells”) in a delicate equilibrium to ensure that there are major net changes in neither bone mass nor its mechanical strength (Palmer et al. ²⁰⁰⁸ ; Olszta et al. ²⁰⁰⁷ ; Boskey and Roy ²⁰⁰⁸ ; Teitelbaum ²⁰⁰⁰ ; Rodan and Martin ²⁰⁰⁰ ). The purpose of remodeling is the release of calcium and the repair of micro-damaged bones from everyday stress. Osteoblasts are mononuclear cells primarily responsible for bone formation. They contain alkaline phosphatase, which enzymatically produces orthophosphate anions needed for the mineralization. In addition, there is one more type of the cells called osteocytes that originate from osteoblasts, which have migrated into, become trapped and surrounded by bone matrix, which they themselves produce (Palmer et al. ²⁰⁰⁸ ; Currey ²⁰⁰² ; Olszta et al. ²⁰⁰⁷ ; Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ; Boskey and Roy ²⁰⁰⁸ ; Weiner et al. ²⁰⁰⁵ ).

If osteoblasts are bone-forming cells, osteoclasts are multinuclear, macrophage-like cells, which can be described as bone-destroying cells because they mature and migrate to discrete bone surfaces (Boskey and Roy ²⁰⁰⁸ ; Teitelbaum ²⁰⁰⁰ ; Rodan and Martin ²⁰⁰⁰ ). Upon arrival, active enzymes, such as acid phosphatase, are secreted to produce lactic acid that causes dissolution of biological apatite. This process, called bone resorption, allows stored calcium to be released into systemic circulation and is an important process in regulating calcium balance (Teitelbaum ²⁰⁰⁰ ; Rodan and Martin ²⁰⁰⁰ ). The iteration of remodeling events at the cellular level is influential on shaping and sculpting the skeleton both during growth and afterwards. That is why, mature bones always consist of a very complex mesh of bone patches, each of which has both a slightly different structure and a different age (Palmer et al. ²⁰⁰⁸ ; Grynpas et al. ¹⁹⁹³ ; Elliott ²⁰⁰² ; Weiner and Wagner ¹⁹⁹⁸ ; Currey ²⁰⁰² ; Rho et al. ¹⁹⁹⁸ ; Olszta et al. ²⁰⁰⁷ ). The interested readers are suggested to read a review on the interaction between biomaterials and osteoclasts (Schilling et al. ²⁰⁰⁶ ).

Still there is no general agreement on the chemical mechanism of bone formation. It is clear that the inorganic part of bone consists of biological apatite, i.e., CDHA with ionic substitutions but without the detectable amounts of hydroxide (Rey et al. ^1995a ; Loong et al. ²⁰⁰⁰ ; Seo et al. ²⁰⁰⁷ ). However, the results of solid-state nuclear magnetic resonance on fresh-frozen and ground whole bones of several mammalian species revealed that the bone crystal OH ⁻ was readily detectable; a rough estimate yielded an OH ⁻ content of human cortical bone of about 20 % of the amount expected in stoichiometric HA (Cho et al. ²⁰⁰³ ). Various in vitro experiments on precipitation of CDHA and HA revealed that none of these compounds is directly precipitated from supersaturated aqueous solutions containing calcium and orthophosphate ions: some intermediate phases (precursors) are always involved (LeGeros ¹⁹⁹¹ , ²⁰⁰¹ ; Elliott ¹⁹⁹⁴ ; O’Neill ²⁰⁰⁷ ; Iijima et al. ¹⁹⁹⁶ ; Bodier-Houllé et al. ¹⁹⁹⁸ ; Rodríguez-Hernández et al. ²⁰⁰⁵ ; Tomazic et al. ¹⁹⁹⁴ ; Nancollas and Wu ²⁰⁰⁰ ; Dorozhkin ^2012c ). Depending on both the solution pH and crystallization conditions, three types of CaPO ₄ compounds (DCPD, ACP and OCP) have been discussed as possible precursors of CDHA precipitation in vitro. Due to this reason, the same CaPO ₄ are suggested as possible precursors of biological apatite formation in vivo.

The transient nature of the precursor phase of bone, if it exists at all, makes it very difficult to detect, especially in vivo (Grynpas and Omelon ²⁰⁰⁷ ). However, in 1966 Brown proposed that OCP was the initial precipitate that then acted as a template upon which biological apatite nucleates (Brown ¹⁹⁶⁶ ). This idea was extended in his further investigations (Tung and Brown ¹⁹⁸³ , ¹⁹⁸⁵ ; Brown et al. ¹⁹⁸⁷ ; Siew et al. ¹⁹⁹² ). The principal support for this concept derived from the following: (1) the close structural similarity of OCP and HA (Brown ¹⁹⁶² ; Brown et al. ¹⁹⁶² ); (2) formation of interlayered single crystals of OCP and HA (pseudomorphs of OCP); (3) the easier precipitation of OCP compared with HA; (4) the apparent plate- or lath-like habit of biological apatites that does not conform to hexagonal symmetry, but looks like a pseudomorph of triclinic OCP; (5) the presence of HPO ₄ ²⁻ in bone mineral, particularly in newly formed bones (Elliott ²⁰⁰² ). Some evidences supporting this idea were found using high-resolution transmission electron microscopy: computer-simulated lattice images of the “central dark line” in mineralized tissues revealed that it consisted of OCP (Brown ¹⁹⁶⁶ ; Bodier-Houllé et al. ¹⁹⁹⁸ ; Rodríguez-Hernández et al. ²⁰⁰⁵ ). In addition, Raman spectroscopic indications for an OCP precursor phase were found during intra-membranous bone formation (Crane et al. ²⁰⁰⁶ ). Other evidences of OCP to HA transformation, including a mechanistic model for the central dark line formation, might be found in literature (Tseng et al. ²⁰⁰⁶ ).

Simultaneously with Brown, the research group led by Posner proposed that ACP was the initially precipitated phase of bone and dentine mineral formation in vivo, thus explaining the non-stoichiometric Ca/P ratio in bones and teeth (Eanes et al. ¹⁹⁶⁵ ; Termine and Posner ^1966a , ^b ). This conclusion was drawn from the following facts: (1) when CaPO ₄ are prepared by rapid precipitation from aqueous solutions containing ions of calcium and orthophosphate at pH > 8.5, the initial solid phase is amorphous; (2) mature bone mineral is composed of a mixture of ion-substituted ACP and poorly crystallized ion-substituted CDHA; (3) early bone mineral has a lower crystallinity than mature bone and the observed improvement in crystallinity with the age of the bone mineral is a result of a progressive reduction in the ACP content (Elliott ²⁰⁰² ; Eanes et al. ¹⁹⁶⁵ ; Termine and Posner ^1966a , ^b ; Harper and Posner ¹⁹⁶⁶ ; Posner ¹⁹⁶⁹ , ¹⁹⁷³ , ¹⁹⁷⁸ ; Boskey and Posner ¹⁹⁷⁶ ; Glimcher et al. ¹⁹⁸¹ ). Interestingly, thermodynamic data prove that the transition of freshly precipitated ACP into CDHA involves intermediate formation of OCP (Meyer and Eanes ^1978a , ^b ). However, the discovery of a stable amorphous calcium carbonate in sea urchin spines (Politi et al. ²⁰⁰⁴ ) reawakened the suggestion that a transient amorphous phase might also exist in bones (Olszta et al. ²⁰⁰⁷ ; Weiner et al. ²⁰⁰⁵ , ²⁰⁰⁹ ; Weiner ²⁰⁰⁶ ; Pekounov and Petrov ²⁰⁰⁸ ; Gower ²⁰⁰⁸ ). Afterwards, evidences of an abundant ACP phase in the continuously forming fin bones of zebrafish were found (Mahamid et al. ²⁰⁰⁸ , ²⁰¹⁰ ). The new bone mineral was found to be delivered and deposited as packages of nanodimensional spheres of ACP, which further transformed into platelets of crystalline apatite within the collagen matrix (Mahamid et al. ²⁰¹⁰ ).

Furthermore, to investigate how apatite crystals form inside collagen fibrils, researchers carried out a time-resolved study starting from the earliest stages of mineral formation (Nudelman et al. ²⁰¹⁰ ). After 24 h of mineralization, CaPO ₄ particles were found outside the fibril, associated with the overlap region, in close proximity to the gap zone. Cryogenic energy-dispersive X-ray spectroscopy confirmed that these precipitates were composed of CaPO ₄ , while a low-dose selected-area electron diffraction technique showed a diffuse band characteristic of ACP. After 48 h, CDHA crystals started to develop within a bed of ACP and after 72 h, elongated electron-dense crystals were abundant within the fibril, in many cases still embedded within a less dense matrix. A low-dose selected-area electron diffraction technique demonstrated that the mineral phase consisted of both ACP and oriented apatite, the latter identical to bone apatite (Nudelman et al. ²⁰¹⁰ ). This process is schematically shown in Fig.  ¹⁴ (Cölfen ²⁰¹⁰ ). Further details on the bone formation mechanisms are available in literature (Olszta et al. ²⁰⁰⁷ ; Hall ²⁰¹⁵ ; Driessens et al. ²⁰¹² ; Boonrungsiman et al. ²⁰¹² ), where the interested readers are referred.

Fig. 14

In spite of the century-plus long studies (Dorozhkin ^2012a , ^2013a ), the maturation mechanism of bone minerals remains to be not well established, mainly due to the difficulties involved in the nanostructural analyses (Olszta et al. ²⁰⁰⁷ ; Sahar et al. ²⁰⁰⁵ ). Still, indirect evidences for the in vivo bone mineral maturation are available only. For example, X-ray diffraction patterns of bones from animals of different age show that the reflections become sharper with age increasing (Meneghini et al. ²⁰⁰³ ; Bilezikian et al. ²⁰⁰⁸ ). This effect is more pronounced in the crystallographic a -axis [(310) reflections] as compared to the c -axis [(002) reflections] (Burnell et al. ¹⁹⁸⁰ ; Weiner and Traub ¹⁹⁸⁶ ). The most comprehensive report describing how normal human bone mineral changes in composition and crystal size as a function of age was based on X-ray diffraction analyses by Hanschin and Stern (Hanschin and Stern ¹⁹⁹⁵ ), who examined 117 homogenized iliac crest biopsies from patients aged 0–95 years. They found that the bone mineral crystal size and perfection increased during the first 25–30 years and then decreased thereafter, slightly increasing in the oldest individuals. The same 117 homogenized biopsy samples were analyzed by wavelength-dispersive X-ray fluorescence to quantify the carbonate substitution in biological apatite as a function of age. Although the changes observed in carbonate substitution were relatively slight (at most 10 %), there was a general increase from 0 to 90 years that is distinct from the absence of a change in crystallinity after age 30 in these samples (Boskey and Coleman ²⁰¹⁰ ). In addition, other changes, like an increase of Ca ²⁺ content and a decrease of HPO ₄ ²⁻ , occur in bone mineral with age (Verdelis et al. ²⁰⁰⁷ ; Yerramshetty et al. ²⁰⁰⁶ ; Kuhn et al. ²⁰⁰⁸ ; Donnelly et al. ²⁰⁰⁹ ; Li and Pasteris ²⁰¹⁴ ; Lowenstam ¹⁹⁸¹ ; Lowenstam and Weiner ¹⁹⁸³ ; Plate et al. ¹⁹⁹⁶ ; Stratmann et al. ¹⁹⁹⁶ ; Jahnen-Dechent et al. ¹⁹⁹⁷ , ²⁰⁰¹ ; Schinke et al. ¹⁹⁹⁹ ; Savelle and Habu ²⁰⁰⁴ ; Chen et al. ^2011b ; Delloye et al. ²⁰⁰⁷ ; Araújo et al. ²⁰⁰⁹ ; Koussoulakou et al. ²⁰⁰⁹ ; Jones ²⁰⁰¹ ; Avery ²⁰⁰¹ ; Nanci ²⁰¹² ; Xue et al. ²⁰⁰⁸ ; Gaft et al. ¹⁹⁹⁶ ; Huang et al. ²⁰⁰⁷ ; Yin et al. ²⁰¹³ ; Ho et al. ^2009a ). Both the crystal sizes and carbonate content were found to increase during aging in rats and cows (LeGeros et al. ¹⁹⁸⁷ ; Rey et al. ^1995b ). The increase in carbonate content with age was also reported in other studies (Yerramshetty et al. ²⁰⁰⁶ ; Kuhn et al. ²⁰⁰⁸ ; Donnelly et al. ²⁰⁰⁹ ; Li and Pasteris ²⁰¹⁴ ). Simultaneously with carbonate content increasing, the content of sodium increased as well (Li and Pasteris ²⁰¹⁴ ). From a chemical point of view, these changes indicate to a slow transformation of poorly crystallized non-apatitic CaPO ₄ into a better-crystallized ion-substituted carbonate-containing CDHA. In addition, during aging, edge areas of bones were found to become less porous, whereas the concentration of organics in the edges was reduced (Li and Pasteris ²⁰¹⁴ ). While there are still many gaps in our knowledge, the researchers seem to be comfortable in stating that in all but the youngest bone and dentine, the only phase present is a highly disordered, highly substituted biological apatite.

In general, the biomineralization process (therefore, bone formation) can happen in two basic ways: either the mineral phase is developed from the ambient environment as it would from a supersaturated solution of the requisite ions, but just requires the living system to nucleate and localize mineral deposition, or the mineral phase is developed under the direct regulatory control of the organism, so that the mineral deposits are not only localized but may be directed to form unique crystal habits not normally developed by a saturated solution of the requisite ions. In a very famous paper (Lowenstam ¹⁹⁸¹ ) and two extended elaborations (Lowenstam and Weiner ¹⁹⁸³ , ¹⁹⁸⁹ ), the first type of biomineralization was called “biologically induced” mineralization and the second “(organic) matrix-mediated” biomineralization. In some papers, the former process is called “passive” and the latter one—“active” biomineralization (Lowenstam and Weiner ¹⁹⁸⁹ ). Briefly, an “active process” means an assembly of nano-sized crystals of biological apatite into bones due to an activity of the suitable cells (e.g., osteoblasts), i.e., within a matrix vesicle. Such structures have been discovered by transmission electron microscopy for bone and teeth formation (Plate et al. ¹⁹⁹⁶ ; Stratmann et al. ¹⁹⁹⁶ ). A “passive process” does not require involvement of cells and means mineralization from supersaturated solutions with respect to the precipitation of biological apatite. In the latter case, thermodynamically, biomineralization might occur at any suitable nucleus. The collagen fibrils have a specific structure with a 67 nm periodicity and 35–40 nm gaps or holes between the ends of the collagen molecules where bone mineral is incorporated in the mineralized fibril (Weiner and Wagner ¹⁹⁹⁸ ; Weiner et al. ¹⁹⁹⁹ ; Tzaphlidou ²⁰⁰⁸ ; Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ). Such a nucleation within these holes would lead to discrete crystals with a size related to the nucleating cavity in the collagen fibril (Fig.  ¹² ). It was proposed that a temporary absence of the specific inhibitors might regulate the process of bone formation (Jahnen-Dechent et al. ¹⁹⁹⁷ , ²⁰⁰¹ ; Schinke et al. ¹⁹⁹⁹ ).

To conclude the bone subject, let me briefly mention on the practical application of bones. In the Stone Age, bones were used to manufacture art, weapons, needles, catchers, amulets, pendants, headdresses, etc. In the coastal regions, big bones of whales were used to construct houses (Savelle and Habu ²⁰⁰⁴ ). Nowadays, cut and polished bones from a variety of animals are sometimes used as a starting material for jewelry and other crafts. Ground cattle bones are used as a fertilizer (Chen et al. ^2011b ). Furthermore, in medicine, bones are used for bone graft substitutes, such allografts from cadavers (Delloye et al. ²⁰⁰⁷ ) and xenografts (Araújo et al. ²⁰⁰⁹ ). However, since recently, bone-derived biomaterials are attempting to avoid due to serious concerns about bovine spongiform encephalopathy (“mad cow disease”) and other possible diseases that might be transmitted even by heated bones.

Teeth

Teeth (singular: tooth) are dense mineralized hard tissues found in the jaws, mouth and/or pharynx of many vertebrates (Koussoulakou et al. ²⁰⁰⁹ ). They have various structures to allow them to fulfill their different purposes. The primary function of teeth is to tear and chew food, while for carnivores it is also a weapon. Therefore, teeth have to withstand a range of physical and chemical processes, including compressive forces (up to ~700 N), abrasion and chemical attack due to acidic foods or products of bacterial metabolism (Jones ²⁰⁰¹ ). The roots of teeth are covered by gums. From the surface teeth are covered by enamel of up to ~2 mm thick at the cutting edges of the teeth, which helps to prevent cavities on the teeth. The biggest teeth of some gigantic animals (elephants, hippopotamuses, walruses, mammoths, narwhals, etc.) are known as tusks or ivory.

Similar to the various types of bones, there are various types of teeth. The shape of the teeth is related to the animal’s food, as well as its evolutionary descent. For example, plants are hard to digest, so herbivores have many molars for chewing. Carnivores need canines to kill and tear and since meat is easy to digest, they can swallow without the need for molars to chew the food well. Thus, the following types of teeth are known: molars (used for grinding up food), carnassials (used for slicing food), premolars (small molars), canines (used for tearing apart food) and incisors (used for cutting food). While humans only have two sets of teeth, some animals have many more: for example, sharks grow a new set of teeth every 2 weeks. Some other animals grow just one set during the life, while teeth of rodents grow and wear away continually through the animal gnawing, maintaining constant length (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ).

Similar to bones, the inorganic part of teeth also consist of biological apatite (Xue et al. ²⁰⁰⁸ ). The stability of the mineral composition of teeth also has a very long history: namely, CaPO ₄ were found in fossil fish teeth (Gaft et al. ¹⁹⁹⁶ ). Similarly, investigations of biological apatite from fossil human and animal teeth revealed its similarity to the modem biological apatite (Huang et al. ²⁰⁰⁷ ). The same was found for modern and fossil (mammoth) elephant ivories (Yin et al. ²⁰¹³ ).

The structure of teeth (Fig.  ¹⁵ top) appears to be even more complicated than that of bones. Unlike bones, teeth consist of at least two different CaPO ₄ materials: enamel, which is a rigid, inert and acellular outer layer, and dentine, which is a bone-like Mg-rich hard tissue that forms the bulk of vertebrate teeth. In addition, there is cementum, which is a thin layer of a bone-like calcified tissue that covers dentine at the roots of teeth and anchors them to the jaw (Ho et al. ^2009a , ^b ; Bosshardt and Selvig ²⁰⁰⁰ ; Yamamoto et al. ²⁰¹⁰ ). Finally, there is the core called pulp (commonly called “the nerve”)—it is a remnant of the embryologic organ for tooth development and contains nerves and blood vessels necessary for tooth function. All these tissues form a highly organized and complex structure (Fig.  ¹⁵ top) with ideal functional and structural capabilities, which assist in sustaining mastication-induced mechanical loading and preventing their mechanical failures during function (Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ; Avery ²⁰⁰¹ ; Nanci ²⁰¹² ). It is interesting to note that the structure of teeth is common to the most species. Namely, fish, reptiles and mammals share the same architecture of a harder external layer and a tougher core.

Fig. 15

Both dentine and cementum are mineralized connective tissues with an organic matrix of collagenous proteins, while the inorganic component of them consists of biological apatite. As shown in Table  ³ , dentine, cementum and bone are quite similar and for general purposes of material scientists they are regarded as being essentially the same material (Elliott ²⁰⁰² ; Weiner and Wagner ¹⁹⁹⁸ ; Currey ²⁰⁰² ; Rho et al. ¹⁹⁹⁸ ; Fratzl et al. ²⁰⁰⁴ ; Rubin et al. ²⁰⁰³ ; Su et al. ²⁰⁰³ ; Pellegrino and Blitz ¹⁹⁷² ; LeGeros et al. ¹⁹⁸⁷ ; Jones ²⁰⁰¹ ). However, strictly speaking, there are some differences among them. For example, the hardness of live dentine is less than that of enamel but is greater than that of bone or cementum (Ho et al. ^2009a ). When pulp of the tooth dies or is removed by a dentist, the properties of dentine change: it becomes brittle, liable to fracture and looses a reparative capability. In addition, unlike bones, both dentine and cementum lack vascularization. In spite of these minor differences, let us consider that the majority of the statements made in the previous section for bones are also valid for dentine and cementum.

Dental enamel is the outermost layer of teeth. It is white and translucent and its true color might be observed at the cutting edges of the teeth only. Enamel is highly mineralized and acellular, so it is not a living tissue. Nevertheless, it is sufficiently porous for diffusion and chemical reactions to occur within its structure, particularly acidic dissolution (dental caries) and remineralization from saliva (possible healing of caries lesions). Therefore, it acts as a selectively permeable membrane, allowing water and certain ions to pass via osmosis (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ).

Enamel is the hardest substance in the body and forms a solid, tough and wear-resistant surface for malaxation (Chai et al. ²⁰⁰⁹ ). In the mature state, it contains up to 98 % of inorganic phase (Table  ³ ). Other components of enamel include remnants of the organic matrix and loosely bound water molecules. Interestingly, from the mechanical point of view, dental enamel was found to be a “metallic-like” deformable biocomposite, because enamel was found to have a similar stress–strain response to that of cast alloy and gold alloy, all of which showed work-hardening effect (He and Swain ²⁰⁰⁷ ). The crystals of biological apatite of enamel are much larger as evidenced by higher crystallinity (reflecting greater crystal size and perfection) demonstrated in their X-ray diffraction patterns than those of bone and dentine. Besides, enamel apatite has fewer ionic substitutions than bone or dentine mineral and more closely approximates the stoichiometric HA (Boskey ²⁰⁰⁷ ). The organic phase of enamel does not contain collagen. Instead, enamel has two unique classes of proteins called amelogenins and enamelins. While the role of these proteins is not fully understood yet, it is believed that both classes of proteins aid in the enamel development by serving as a framework support (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ; Margolis et al. ²⁰⁰⁶ ). The large amount of minerals in enamel accounts not only for its strength but also for its brittleness. Dentine, which is less mineralized and less brittle, compensates for enamel and is necessary as a support (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ). Shark enameloid is an intermediate form bridging enamel and dentine. It has enamel-like crystals of fluoridated biological apatite associated with collagen fibrils (Lowenstam and Weiner ¹⁹⁸⁹ ; Rey et al. ²⁰⁰⁶ ; Prostak et al. ¹⁹⁹³ ; Dahm and Risnes ¹⁹⁹⁹ ; Carr et al. ²⁰⁰⁶ ; Enax et al. ²⁰¹⁴ ). Due to the presence of fluorides, biological apatite of shark enameloid shows both higher crystal sizes and a more regular hexagonal symmetry if compared to non-fluoridated biological apatite of bones and teeth (Daculsi et al. ¹⁹⁹⁷ ). Similar correlation between the presence of fluorides and crystal dimensions was found for enamel (Vieira et al. ²⁰⁰³ ).

Like that for bones, seven levels of structural hierarchy have been also discovered in human enamel; moreover, the analysis of the enamel and bone hierarchical structure suggests similarities of the scale distribution at each level (Palmer et al. ²⁰⁰⁸ ; Athanasiou et al. ²⁰⁰⁰ ; Cui and Ge ²⁰⁰⁷ ). On the mesoscale level, there are three main structural components: a rod, an interrod and aprismatic enamel. Among them, the enamel rod (formerly called an enamel prism) is the basic unit of enamel. It is a tightly packed mass of biological apatite in an organized pattern. Each rod traverses uninterrupted through the thickness of enamel. They number 5–12 million rods per crown. The rods increase in diameter (4 up to 8 microns) as they flare outward from the dentine-enamel junction (DEJ). Needle-like enamel rods might be tens of microns long (up to 100 µm) but sometimes only 50 nm wide and 30 nm thick (Fig.  ¹⁵ bottom; Avery ²⁰⁰¹ ; Nanci ²⁰¹² ; Jandt ²⁰⁰⁶ ; Chen et al. ²⁰⁰⁴ ; Rönnholm ¹⁹⁶² ; Nylen et al. ¹⁹⁶³ ; Miake et al. ¹⁹⁹³ ; Daculsi et al. ¹⁹⁸⁴ ; Jodaikin et al. ¹⁹⁸⁴ ; Bres and Hutchison ²⁰⁰² ). They are quite different from the much smaller crystals of dentine and bone (Table  ³ ), but all of them consist of biological apatite (Schroeder and Frank ¹⁹⁸⁵ ; Brès et al. ¹⁹⁹⁰ ). In cross section, an enamel rod is best compared to a keyhole, with the top, or head, oriented toward the crown of the tooth and the bottom, or tail, oriented toward the root of the tooth.

The arrangement of the crystals of biological apatite within each enamel rod is highly complex. Enamel crystals in the head of the enamel rod are oriented parallel to the long axis of the rod. When found in the tail of the enamel rod, the crystals’ orientation diverges slightly from the long axis (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ). The arrangement of the enamel rods is understood more clearly than their internal structure. Enamel rods are found in rows along the tooth (Fig.  ¹⁵ bottom) and, within each row, the long axis of the enamel rod is generally perpendicular to the underlying dentine (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ; Jandt ²⁰⁰⁶ ; Chen et al. ²⁰⁰⁴ ; Rönnholm ¹⁹⁶² ; Nylen et al. ¹⁹⁶³ ; Miake et al. ¹⁹⁹³ ). An AFM study indicated that apatite crystals in enamel exhibited regular sub-domains or subunits with distinct chemical properties related to topographical features and gave rise to patterned behavior in terms of the crystal surface itself and the manner in which it responded to low pH (Robinson et al. ²⁰⁰⁴ ). In addition, the results of the investigations of human single apatite crystals of both enamel and dentine revealed an absence of the mirror plane perpendicular to the c -axis leading to the P 6 ₃ space group (Mugnaioli et al. ²⁰¹⁴ ) instead of the P 6 ₃ /m space group typical for HA and FA (Table  ⁴ ).

The second structural component of the enamel matrix is the interrod (or interprismatic) enamel, which surrounds and packs between the rods. The difference between the rod and the interrod is the orientation of apatite crystals; the rod contains aligned crystallites, whereas the mineral in the interrod is less ordered. These structures coalesce to form the tough tissue of enamel, which can withstand high forces and resist damage by crack deflection. The third structure, aprismatic enamel, refers to the structures containing apatite crystals that show no meso-scale or macro-scale alignment (Palmer et al. ²⁰⁰⁸ ).

The in vivo formation and development of teeth appears to be even more complicated when compared with the aforedescribed process of bone formation. It is a very complex biological process, by which teeth are formed from embryonic cells, grow and erupt into the mouth (Boskey and Roy ²⁰⁰⁸ ). For human teeth enamel, dentine and cementum must all be developed during the appropriate stages of fetal development. Primary (baby) teeth start to form between the sixth and eighth weeks in utero, while the permanent teeth begin to form in the twentieth week in utero (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ). Experimental data confirmed the necessity of CaPO ₄ in the diet of pregnant and nursing mother to prevent early childhood dental caries (Warf and Watson ²⁰⁰⁸ ).

As teeth consist of at least two materials with different properties (enamel and dentine), the tooth bud (sometimes called “the tooth germ”—that is an aggregation of cells that eventually forms a tooth) is organized into three parts: the enamel organ, the dental papilla and the dental follicle. The enamel organ is composed of at least four other groups of cells [for the biological details see Refs. (Avery ²⁰⁰¹ ; Nanci ²⁰¹² )]. Altogether, these groups of cells give rise to ameloblasts, which secret enamel matrix proteins. The protein gel adjacent to ameloblasts is supersaturated with CaPO ₄ , which leads to the precipitation of biological apatite. Similarly, the dental papilla contains cells that develop into odontoblasts, which are dentine-forming cells. The dental follicle gives rise to three important entities: cementoblasts, osteoblasts and fibroblasts. Cementoblasts form the cementum of a tooth (Bosshardt and Selvig ²⁰⁰⁰ ). Osteoblasts give rise to the alveolar bone around the roots of teeth (see bone formation above). Fibroblasts develop the periodontal ligaments that connect teeth to the alveolar bone through cementum (Boskey ²⁰⁰⁷ ; Glimcher ²⁰⁰⁶ ; Boskey and Roy ²⁰⁰⁸ ; Hall ²⁰¹⁵ ; Avery ²⁰⁰¹ ; Nanci ²⁰¹² ).

The first detectable crystals in enamel formation are flat thin ribbons (Rönnholm ¹⁹⁶² ; Nylen et al. ¹⁹⁶³ ; Miake et al. ¹⁹⁹³ ) that were reported to be OCP (Brown and Chow ¹⁹⁷⁶ ; Simmer and Fincham ¹⁹⁹⁵ ; Diekwisch et al. ¹⁹⁹⁵ ; Aoba ¹⁹⁹⁶ ), β-(Ca,Mg) ₃ (PO ₄ ) ₂ (Diekwisch et al. ¹⁹⁹⁵ ), DCPD (Rey et al. ^1995a ; Bonar et al. ¹⁹⁹¹ ) or ACP (Beniash et al. ²⁰⁰⁹ ). The formation process of enamel is different from that for bone or dentine: amelogenin being hydrophobic self-assembles into nano-sized spheres that guide the growth of the ribbon-like dental enamel crystals. During maturation of enamel, the mineral content increases from initially ~45 wt.% up to ~98–99 wt.% (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ; Bonar et al. ¹⁹⁹¹ ). The enamel crystal rods widen and thicken by additional growth (Rey et al. ^1995b ; Bonar et al. ¹⁹⁹¹ ; Smith ¹⁹⁹⁸ ) with a simultaneous increase of the Ca/P molar ratio (Smith ¹⁹⁹⁸ ) and a decrease in carbonate content (Sydney-Zax et al. ¹⁹⁹¹ ; Rey et al. ¹⁹⁹¹ ; Takagi et al. ¹⁹⁹⁸ ), finally resulting in the most highly mineralized and hardest substance produced by vertebrates. It is interesting to note that in the radular teeth of chitons, ACP was found to be the first-formed CaPO ₄ mineral, which over a period of weeks was transformed to dahllite (Lowenstam and Weiner ¹⁹⁸⁵ ).

The crystal faces expressed in enamel are always (100) face and at the ends presumably (001) (Selvig ¹⁹⁷⁰ , ¹⁹⁷³ ), which are the ones usually found in HA. The centers of enamel crystals contain a linear structure known as the “central dark line” (this line was also observed in bone and dentine), which consists of OCP (Iijima et al. ¹⁹⁹⁶ ; Bodier-Houllé et al. ¹⁹⁹⁸ ; Rodríguez-Hernández et al. ²⁰⁰⁵ ; Tseng et al. ²⁰⁰⁶ ). However, data are available that the “central dark line” appears to be the calcium richest part of the human tooth enamel crystallites (Gasga et al. ²⁰⁰⁸ ). As described above for bones, X-ray diffraction studies revealed that the biological apatite of younger teeth was less crystalline than that of more mature ones, while the maximum crystallinity was found for teeth belonged to people of 31–40 years old (Pankaew et al. ²⁰¹² ). Therefore, maturation of teeth also means a slow transformation (re-crystallization?) of biological CaPO ₄ from ion-substituted ACP to a better-crystallized ion-substituted CDHA. However, crystallinity of teeth belonged to older people (41–50 and 51–60 years old) was found to decrease with age. Simultaneously, thermogravimetric studies revealed bigger weight loss for teeth of the older people (Pankaew et al. ²⁰¹² ), which could be due to a greater amount of carbonates.

The development of individual enamel and dentine crystals was studied by high-resolution transmission electron microscopy. Both processes appear to be roughly comparable and were described in a four-step process. The first two steps include the initial nucleation and formation of nano-sized particles of biological apatite. They are followed by ribbon-like crystal formation, which until recently was considered as the first step of biological crystal formation (Cuisinier et al. ¹⁹⁹³ ; Houllé et al. ¹⁹⁹⁷ ). These complicated processes, starting with the heterogeneous nucleation of inorganic CaPO ₄ on an organic extracellular matrix, are controlled in both tissues by the organic matrix and are under cellular control (Bodier-Houllé et al. ¹⁹⁹⁸ ; Mann ¹⁹⁹³ ). To complicate the process even further, regular and discrete domains of various charges or charge densities on the surface of apatite crystals derived from the maturation stage of enamel development were discovered by a combination of atomic and chemical force microscopy (Kirkham et al. ²⁰⁰⁰ ). Binding of organic molecules (e.g., amelogenin (Kirkham et al. ²⁰⁰⁰ )) at physiological solution pH appears to occur on the charged surface domains of apatite. The recent visions on dental tissue research are available elsewhere (Smith and Tafforeau ²⁰⁰⁸ ).

As teeth consist of several materials, there are mutual junctions among them. For example, a dentine–enamel junction (DEJ) is a thin [2.0 ± 1.1 μm (Habelitz et al. ²⁰⁰¹ )] but gradual interface with characteristics transiting from those of dentine to those of enamel. Namely, the collagen fibers in dentine were found to enter into the enamel side of DEJ and terminate in a region in which crystals of biological apatite begin to show enamel characteristics (Chan et al. ²⁰¹¹ ). In addition, the average contents of organic matrix and carbonate ions were found to increase in the order: enamel < DEJ < dentin. Furthermore, HPO ₄ ²⁻ ions were not detected in the DEJ, while in dentin their content was higher than in enamel (Kolmas et al. ²⁰¹⁰ ). Comparable results were obtained in another study (Desoutter et al. ²⁰¹⁴ ). Besides, both chemical and molecular structure of DEJ appeared to be dependent on the intratooth location (Xu et al. ²⁰⁰⁹ ). Genetically, it is a remnant of the onset of enamel formation because enamel grows outwards from this junction (Nanci ²⁰¹² ). Mechanically, DEJ plays an important role in preventing crack propagation from enamel into dentine (Imbeni et al. ²⁰⁰⁵ ). Hierarchically, DEJ has a three-level structure: 25–100 μm scallops with their convexities directed toward the dentin and concavities toward the enamel, 2–5 μm microscallops and a smaller scale structure (Marshall et al. ²⁰⁰³ ). Therefore, high-resolution elastic modulus mapping has indicated that the DEJ is a band with a graded mechanical property rather than a discrete interface (Sui et al. ²⁰¹⁴ ). The major steps of enamel crystal growth at the junction have been described above but the mechanism of the junction formation is still debatable. Some authors claim that enamel crystals grow epitaxially on the pre-existing dentine crystals because of a high continuity between enamel and dentine crystals (Arsenault and Robinson ¹⁹⁸⁹ ; Hayashi ¹⁹⁹² , ¹⁹⁹³ ). Others have shown that enamel crystals are formed at a given distance from the dentine surface (Simmer and Fincham ¹⁹⁹⁵ ; Diekwisch et al. ¹⁹⁹⁵ ; Aoba ¹⁹⁹⁶ ; Bodier-Houllé et al. ²⁰⁰⁰ ) and could either reach dentine crystals by a subsequent growth (Takano et al. ¹⁹⁹⁶ ) or remain distant (Bodier-Houllé et al. ²⁰⁰⁰ ; Dong and Warshawsky ¹⁹⁹⁶ ). In addition, there are both a cementum–enamel junction (CEJ) (Wang et al. ^2006b ), which is quite similar to DEJ, and a cementum–dentine junction (CDJ) (Ho et al. ²⁰⁰⁴ , ²⁰⁰⁵ ; Jang et al. ²⁰¹⁴ ).

Enamel formation, or amelogenesis, is a highly regulated process involving precise genetic control as well as protein–protein interactions, protein–mineral interactions and interactions involving the cell membrane. Much is still unknown about the interactions among proteins present in enamel matrix and the final crystalline phase of biological apatite (Palmer et al. ²⁰⁰⁸ ; Paine et al. ²⁰⁰¹ ). At some point before a tooth erupts into the mouth, the ameloblasts are broken down. Consequently, unlike bones, enamel has no way to regenerate itself using the process of “active mineralization” (see bone formation) because there is no biological process that repairs degraded or damaged enamel (Avery ²⁰⁰¹ ; Nanci ²⁰¹² ). In addition, certain bacteria in the mouth feed on the remains of foods, especially sugars. They produce lactic acid, which dissolves the biological apatite of enamel in a process known as enamel demineralization that takes place below the critical pH of about 5.5. Similar process called enamel erosion occurs when a person consumes acid-containing (citric, lactic, phosphoric, etc.) soft drinks (Jandt ²⁰⁰⁶ ; Barbour and Rees ²⁰⁰⁴ ; Low and Alhuthali ²⁰⁰⁸ ; White et al. ²⁰¹⁰ ). Evidences exist that there is a preferential loss of carbonates and Mg during acidic dissolution of mineral in dental caries. Luckily, saliva gradually neutralizes the acids that cause pH on teeth surface to rise above the critical pH. This might cause partial enamel remineralization, i.e., a return of the dissolved CaPO ₄ to the enamel surface. Until recently, it was generally agreed that if there was sufficient time between the intake of foods (generally, 2–3 h) plus a damage was very limited, teeth could repair themselves by the “passive mineralization” process (LeGeros ¹⁹⁹⁹ ). Data on increased remineralization of dental enamel by CaPO ₄ -containing compounds (Cochrane et al. ²⁰⁰⁸ ; Langhorst et al. ²⁰⁰⁹ ; Weir et al. ²⁰¹² ) are in support of this hypothesis.

However, studies performed using atomic force microscopy nano-indentation technique revealed that previously demineralized samples of dental enamel further exposed to remineralizing solutions did show a dense crystalline layer of CaPO ₄ formed on their surface. Unfortunately, the re-precipitated deposits of CaPO ₄ always consisted of loosely packed crystals and did not protect the underlying enamel from a subsequent acid attack. Furthermore, these surface deposits were completely removed by either a toothbrush or a short exposure to an erosive acidic solution (Jandt ²⁰⁰⁶ ; Lippert et al. ^2004a , ^b , ^c ). In this context, it should be emphasized that the term “remineralization”, which is often misused in the literature, should imply the process of mineral growth that goes hand in hand with a strengthening effect of the weakened enamel surface. Since no strengthening of an exposure to remineralizing solutions was observed, it might be considered that no “passive mineralization” was found (in spite of the real evidence of the re-precipitated surface deposits of CaPO ₄ ) (Jandt ²⁰⁰⁶ ; Lippert et al. ^2004b , ^c ).

An interesting hypothesis that nano-sized apatite crystallites occur in the oral cavity during extensive physiological wear of the hierarchical structured enamel surface due to dental abrasion and attrition has been published (Hannig and Hannig ²⁰¹⁰ ). These nano-scaled apatite enamel crystallites might promote remineralization at the tooth surface. However, this idea should be verified experimentally. Thus, according to the current knowledge, the enamel self-repairing ability by a passive remineralization appears to be doubtful, while an active remineralization is impossible. Nevertheless, investigations in this field keep going (Karlinsey and Mackey ²⁰⁰⁹ ; Karlinsey et al. ^2010a , ^b ; Busch ²⁰⁰⁴ ; Onuma et al. ²⁰⁰⁵ ; He and Feng ²⁰⁰⁷ ; Li et al. ^2008b , ^2014a ; Wang et al. ²⁰⁰⁹ ; Roveri et al. ^2009a , ^b ; Peters et al. ²⁰¹⁰ ; Orsini et al. ²⁰¹⁰ ; Uysal et al. ²⁰¹⁰ ; Niu et al. ²⁰¹⁴ ). For example, ACP-containing orthodontic biocomposite resins might reduce the enamel decalcification found in patients with poor oral hygiene (Uysal et al. ²⁰¹⁰ ). Further details on CaPO ₄ application in dentistry are available in a topical review (Dorozhkin ^2013c ).

A content of fluoride added to either toothpaste or mouthwash lowers the solubility of CaPO ₄ (by formation of FHA on the surface) and therefore improves the acid resistance of dental enamel (Schemehorn et al. ²⁰¹¹ ; Hattab ²⁰¹³ ; Driessens ¹⁹⁷³ ; Moreno et al. ¹⁹⁷⁴ ; McClendon ¹⁹⁶⁶ ). Furthermore, fluorides also reduce production of acids by bacteria in the mouth by reducing their ability to metabolize sugars. However, dental treatment by fluorides must be used with care because an improper treatment results in formation of CaF ₂ globules deposited on the enamel surface (Wang et al. ^2008a ).

To conclude the teeth subject, let me briefly mention on the practical application of teeth. Due to relatively small dimensions of normal teeth, only tusks and ivory of giant animals are widely used. For example, both the Greek and Roman civilizations used large quantities of ivory to make high value works of art, precious religious objects and decorative boxes for costly objects. Ivory was often used to form the whites of the eyes of statues. Prior to introduction of plastics, it was used for billiard balls, piano keys, buttons and ornamental items. The examples of modern carved ivory objects are small statuary, netsukes, jewelry, flatware handles and furniture inlays.

Antlers

Deer antlers (Fig.  ¹⁶ top) are unique biological structures since their growth rate is without parallel in vertebrates and because they are the only bony appendages in mammals capable of complete regeneration. This allows for basic research in bone biology without the interference of surgical procedures and their adverse effects in animals where samples are obtained. In addition, antlers also allow for the gathering of a large amount of samples from different populations to assess nutritional and ecological effects on bone composition and structure (Yue et al. ²⁰⁰⁵ ; Zhao et al. ²⁰⁰⁶ ; Landete-Castilleijos et al. ^2007a ). They are costly sexual secondary characters of male deer and constitute 1–5 % of the body weight (Huxley ¹⁹³¹ ). Recent studies suggest that antler regeneration is a stem cell-based process and that these stem cells are located in the pedicle periosteum (Kierdorf et al. ²⁰⁰⁹ ; Kierdorf and Kierdorf ²⁰¹¹ ).

Fig. 16

Antlers are not true horns; they are a simple extension of bone. Both antlers and bones utilize the same basic building blocks, namely the type I collagen in the protein phase and the carbonated CDHA in the mineral phase, which was verified by amino acid analysis, X-ray diffraction and TEM observations (Chen et al. ²⁰⁰⁹ ). Antlers are large and complex horn-like appendages of deer consisting of bony outgrowths from the head with no covering of keratin as found in true horns. Usually, they begin growing in March and reach maturity in August. In winter, antlers fall off; this is known as shedding. Similar to bones, antlers contain pores and can withstand applied stresses of over 300 MPa (Landete-Castilleijos et al. ^2007b ; Evans et al. ²⁰⁰⁵ ; Akhtar et al. ²⁰⁰⁸ ; Currey et al. ²⁰⁰⁹ ), which is even higher than that of bones (Table  ³ ). Therefore, antlers are occasionally considered an almost unbreakable bone (Currey et al. ²⁰⁰⁴ ). However, there are several distinct differences between them. Namely, antlers and skeletal bones have different functions. The primary functions of antlers are social display, defense against predators and combat between male species. Skeletal bones contain bone marrow, whereas antlers have no marrow. There exists a transition zone between cortical and cancellous bones in antlers, whereas there is no such transition zone in skeletal bones. The cancellous bone is well aligned and uniformly distributed through the entire antler. In bovine femur, the cancellous bone is mainly located in the femur head and its density decreases progressively toward the central region of the femur, which correlates to the external loading conditions. In addition, antlers have a lower mineral content and consequently lowest elastic modulus among mineralized calcified tissues with a mineral content of ~50 wt% (or ~30 vol. %). Interesting that antlers may act as large hearing aids; namely, moose with antlers have far more sensitive hearing than moose without them (Bubenik and Bubenik ²⁰⁰⁸ ).

Since antlers are accessible, shed after mating season and cast every year, they appear to be a good model to study bone biology (Price et al. ²⁰⁰⁵ ; Landete-Castilleijos et al. ^2007c ; Li et al. ^2014b ). Each antler grows from an attachment point on the skull called a pedicle. While an antler is growing, it is covered with highly vascular skin called velvet, which supplies oxygen and nutrients to the growing bone. Antlers bud and branch as they grow. Once they have achieved the proper dimensions, the velvet starts to dry out, cracks and breaks off, while the antler’s bone dies. Therefore, fully developed antlers consist of dead bone only (Kierdorf and Kierdorf ²⁰⁰⁰ , ²⁰⁰⁵ ; Pathak et al. ²⁰⁰¹ ; Yuxia et al. ²⁰⁰² ; Li et al. ²⁰⁰⁵ ). However, the formation and mineralization process of antlers is still not fully understood. To clarify this, researchers used oxytetracycline injections to label different stages of bone formation in antlers of 14 red deer between days 28 and 156 of antler growth. The results revealed that initially a trabecular scaffold of woven bone was formed which largely replaced a pre-existing scaffold of mineralized cartilage. Lamellar bone was then deposited and from about day 70 onwards, primary osteons filled in the longitudinal tubes lined by the scaffold in a proximal to distal sequence (Gomez et al. ²⁰¹³ ). In addition, it was found that food processing cannot supply the mineral needs required for antler growth and thus, male deer must temporarily resorb CaPO ₄ minerals from their own skeleton for antler growth (Meister ¹⁹⁵⁶ ; Muir et al. ¹⁹⁸⁷ ; Baxter et al. ¹⁹⁹⁹ ). Detailed studies revealed that daily food intake provided between 25 and 40 % of calcium needed for antler mineralization, which resulted in a temporary skeleton demineralization (Muir et al. ¹⁹⁸⁷ ; Baxter et al. ¹⁹⁹⁹ ).

One should note that people seldom come across the antlers in the woods. Rabbits and rodents such as mice and chipmunks eat antlers (and bones of wild animals after they die) for calcium. Rodents and rabbits also gnaw bones and antlers to sharpen their incisors. Due to an extremely high growth rate, which can achieve 2–4 cm per day, combined with a very fast biomineralization, these unique appendages might be a well-suited animal model for studying the disturbances of bone formation induced by additives (e.g., by excess of fluoride) (Kierdorf and Kierdorf ²⁰⁰⁵ ). Antler size and external characteristics were found to be influenced by nutrition, climatic variability and other factors. Thus, since antlers are periodically replaced, the analysis of naturally cast antlers offers the opportunity for a continuous and a noninvasive monitoring of the environmental pollution by these additives (Kierdorf and Kierdorf ²⁰⁰⁵ ).

To conclude this part, let me briefly mention on the practical application of antlers. Associated with aristocracy, antlers have adorned European castles and hunting lodges for centuries. Today, furnishings and accessories made from antlers are featured in fine homes throughout the world and are a reflection of grace and elegance (Fig.  ¹⁶ bottom). Concerning biomedical applications, the initial attempts to evaluate a potential use of deer antlers as xenogenic bone grafts have been already performed (Baciut et al. ²⁰⁰⁷ ; Hasan et al. ²⁰¹² ; Zhang et al. ²⁰¹² , ²⁰¹³ ).