Hibonite Blue: A New Class of Intense Inorganic Blue Colorants

Abstract

Commercially available spinel cobalt blue (CoAl₂O₄) utilizes a significant amount of carcinogenic Co²⁺, which makes its synthesis more hazardous and environmentally harmful. Considerable effort has been put into developing more environmentally benign and robust blue pigments to replace cobalt blue. A new class of blue pigments with tunable hue were prepared. The solid solution series, CaAl_12–2xCo_xTi_xO₁₉ (0 < x ≤ 1), crystallizes in a hexagonal mineral hibonite (CaM₁₂O₁₉) structure with five distinct crystallographic sites for M cations (M = Al, Co, and Ti). The origin of intense blue color is attributed to a synergistic effect of allowed d–d transitions involving the chromophore Co²⁺ in both tetrahedral and trigonal bipyramidal crystal fields. Compared with commercial cobalt blue, these tunable hibonite blues possess a reddish hue that intensifies the blue color as observed in Y(In,Mn)O₃ (YInMn) blues, with a significant reduction of Co²⁺ concentration from 33% to as low as 4% by mass. A significant advantage of hibonite blues over cobalt blue is the substantial reduction in carcinogenic cobalt content while enhancing the color properties at a reduced cost for raw materials.

Introduction

For thousands of years, civilizations around the world have sought inorganic materials that could be used to paint things blue, often with very limited success.^1,2 Discoveries of durable new inorganic blue colorants are rare occurrences in materials science. The first ancient blue material discovered was lapis lazuli, which was about 6000 years ago. It was more expensive than gold. Egyptians and Babylonians utilized naturally occurring lapis lazuli stones found in Afghanistan to make rudimentary blues. It was not until 1826 a chemical process for developing synthetic lapis lazuli (now called ultramarine blue, Na₇Al₆Si₆O₂₄S₃) was identified.² In 1706, Prussian blue (Fe₄[Fe(CN)₆]₃) was accidentally made by a German dye-maker Johann Jacob Diesbach.² Although cobalt blue containing mixtures had long been used in Chinese porcelain,³ the pure cobalt blue (CoAl₂O₄) pigment was independently synthesized by a French chemist in 1802.⁴ Nearly two centuries later, YInMn blue (YIn_1–xMn_xO₃) was serendipitously discovered when searching for magnetodielectric materials for electronics and sparked renewed worldwide interest in pigment discoveries.^2,5 Most of these compounds suffer from stability, cost, color, or toxicity issues. The origin of color in ultramarine blue is from a charge transfer within an unstable S₃^– cluster, which decomposes under mild acidic conditions and with time. Prussian blue is dark and dull because of a charge transfer between Fe²⁺ and Fe³⁺ and is unsuitable for many coloring applications. The YInMn blue family of pigments are durable and possess remarkable heat reflecting properties, which make them excellent candidates for energy-saving coatings, nonetheless they are not considered cost-effective for widespread general coating applications.^2,6

The intensity of the color, ease of synthesis, and wide applicability have resulted in CoAl₂O₄ becoming a dominant commercial blue pigment for the last 200 years. CoAl₂O₄ gives an intense bright blue color because of tetrahedral Co²⁺ spin-allowed d–d⁴A₂(⁴F)–⁴T₁(⁴P) transitions between 500 and 700 nm. CoAl₂O₄, which contains 33% Co²⁺ by mass, is expensive and environmentally unfriendly to produce. Finding replacement compounds that have the same durability, intensity, and hue of blue has proven even more difficult. The intensity of color generally scales with the chromophore composition; however, reducing cobalt content is infeasible in the CoAl₂O₄ spinel structure without significant loss in coloration. The search for materials with similar local environments for chromophores has led to the exploration of other crystal structures as color-producing hosts. The mineral hibonite, known as “blue angel”, has light blue coloration because of a charge transfer between Ti³⁺ and O^2–. Hibonite, with the general formula CaM₁₂O₁₉, has five distinct crystallographic M sites, including three different octahedra, a set of tetrahedra, and a set of trigonal bipyramids (TBPs), where the metal center, M, is in pseudotetrahedral coordination, offset from the trigonal plane (Figure 1). Hibonites crystallize into a hexagonal crystal structure, with spinel blocks identical to CoAl₂O₄ separated by an “R” block with TBPs, face-shared octahedral M sites, and Ca²⁺ sites.⁷ The tetrahedral and TBP sites are of interest because they are noncentrosymmetric and therefore may produce intense transitions as found in YIn_1–xMn_xO₃ and CoAl₂O₄.^5,8 A previous work has discovered many different colors in hibonite compounds; however, the resulting blue colors are less intense than CoAl₂O₄.⁹⁻¹⁴

Figure 1.

Schematic of the unit cell of hibonite. Symmetrically distinct sites are shown in different colors: light green (M1) and yellow (M5): edge-shared octahedra; red (M4): face-shared octahedra; light blue (M2): TBPs; dark blue (M3): tetrahedra.

We have investigated the effects of systematic Co²⁺/Ti⁴⁺ substitution into the hibonite structure and have found that Co²⁺ gives rise to an intense, tunable blue coloration with up to x = 1 Co²⁺ (CaAl_12–2xCo_xTi_xO₁₉, 8% Co²⁺ by mass). This coloration at such low Co²⁺ concentrations is characteristic because of the unique effect of the cosubstituted Ti⁴⁺ within the hibonite structure. This color is tunable through varying Co²⁺ concentration as well as cosubstitution with other transition metals, such as Ni²⁺, Mn³⁺, and Cr³⁺.

Results and Discussion

Addition of Co²⁺ and Ti⁴⁺ increases the unit cell parameters because of the increased size of Co²⁺ and Tⁱ⁴⁺ over Al³⁺ (Figure 2a).¹⁵ The deep blue hues seen in hibonites originate in the strong site preference of Co²⁺ and Ti⁴⁺ substitution, affirmed by structural analysis of neutron diffraction data of CaAl_10.6Co_0.7Ti_0.7O₁₉ and CaAl₁₀CoTiO₁₉ (Figure 2b). The face-shared octahedral sites are the primary occupied sites of Ti⁴⁺ for both compositions (Table 1). This is consistent with what was found in the CaAl_12–2xNi_xTi_xO₁₉ hibonites, with the primary difference being the increased site preference of Co²⁺ for the tetrahedral and TBP sites compared to Ni²⁺.⁹ Of the three octahedral coordination environments available in the structure (Figure 1), only the face-shared octahedral site is capable of a site distortion.¹¹ Because of the larger size and greater charge of Ti⁴⁺ over Al³⁺, a distortion must occur in order to satisfy valency. Therefore, Ti⁴⁺ distributes primarily into the face-shared octahedral environment where the local structure around the site can allow the metal to displace away from the shared face of the other octahedron. Neutron structure analysis reveals that 76 and 70% of Co²⁺ are in the tetrahedral site for CaAl_10.6Co_0.7Ti_0.7O₁₉ and CaAl₁₀CoTiO₁₉, respectively. A significant amount of Co²⁺ also resides in the TBP site, 24 and 17% of all Co²⁺ for CaAl_10.6Co_0.7Ti_0.7O₁₉ and CaAl₁₀CoTiO₁₉, respectively, with no Co²⁺ occupying the M1 edge-shared octahedral or M4 face-shared octahedral sites in either composition. Co²⁺ occupies the M5 edge-shared octahedral site in small amounts only in the CaAl₁₀CoTiO₁₉ composition.

Figure 2.

(a) Lattice parameter evolution with Co²⁺/Ti⁴⁺ substitution in CoAl_12–2xCo_xTi_xO₁₉ and (b) neutron Rietveld refinement of CaAl₁₀CoTiO₁₉.

Table 1. Summary of Rietveld Refinement Occupancy Results of CaAl_12–2xCo_xTi_xO₁₉.

	CaAl12O19a	CaAl10.6Co0.7Ti0.7O19	CaAl10CoTiO19
wRp (%)	4.72	4.57	4.60
Χ2	1.57	1.85	1.92
a (Å)	5.5592(1)	5.5970(1)	5.6106(2)
c (Å)	21.902(1)	22.042(1)	22.211(1)
M1 (Al)	1	1	1
M2 (Al)	0.5	0.369(6)	0.371(9)
M2 (Co)		0.084(1)	0.087(3)
M2 (Ti)		0.046(5)	0.042(6)
M3 (Al)	1	0.74(2)	0.65(4)
M3 (Co)		0.26(2)	0.35(4)
M4 (Al)	1	0.753(4)	0.605(7)
M4 (Ti)		0.247(4)	0.395(7)
M5 (Al)	1	0.981(2)	0.950(8)
M5 (Co)		0	0.024(4)
M5 (Ti)		0.019(2)	0.026(4)

^aFrom Li, et al.¹¹

The Co²⁺ ion displays a strong preference for the tetrahedral coordination environment, even though the crystal field stabilization energy for octahedral Co²⁺ is higher. This may be because the d⁷ electron configuration of Co²⁺ favors a weak Jahn–Teller distortion in octahedral coordination, which is unavailable in the hibonite structure.¹⁶ Neither the TBP nor the tetrahedral environments require a Jahn–Teller distortion for d⁷; however, the tetrahedral site has a much higher percent of Co²⁺ in both compositions refined. This is attributed to the displacement of the metal center of the TBP away from the trigonal plane. This displacement creates a pseudotetrahedral environment, which is smaller than the real tetrahedral site and thus less likely to be occupied by the larger Co²⁺ ion. The total TBP occupancy in both refined compositions is similar, despite the difference in the molar percent of cobalt out of all sites. This suggests that the TBP site becomes saturated by Co²⁺ at low concentrations of Co²⁺.

This site preference is much more pronounced in CaAl_12–2xCo_xTi_xO₁₉ than in the previously investigated CaAl_12–2xNi_xTi_xO₁₉ or Ca_1–xLa_xAl_12–xNi_xO₁₉.⁹ The high spin d⁸ electron configuration of Ni²⁺ may contribute to this reduced preference as octahedral crystal field stabilization energy is much larger than in Co²⁺.

The color of these materials has been evaluated by L*a*b* color measurements. L*a*b* uses three primary numbers to quantify the brightness (L*), green-red (a*), and blue-yellow (b*) values. Higher L* values are the result of brighter colors with higher light reflectivity. Positive a* values correspond to more red coloration while more negative b* values are more blue. Hibonite blue pigments give intense negative b* blue values rivaling that of CoAl₂O₄ and approaching that of YInMn blue (Table 2). We have found that, while CoAl₂O₄ gives no red hue like YInMn blue, Ca(Al,Co,Ti)₁₂O₁₉ substituted hibonite blues to give more reddish hues, approaching that of lapis lazuli stones and YInMn blue. These reddish hues are unobtainable with CoAl₂O₄ spinel, as reduction in Co²⁺ content by substitution in the spinel system only reduces the intensity of blue coloration and has no effect on hue. Introduction of other chromophores, such as Ni²⁺, into the Ca(Al,Co,Ti)₁₂O₁₉ hibonites reduces this redness and adds certain greenish hues to the hibonite blue. Material brightness can often be tuned through chromophore concentration; however, only recently, Y(In,Mn,Ti,Zn)O₃ has been shown to have a tunable hue.^17,18 This tunability is rare in blue pigments and shows that hibonite blue can be used as a versatile pigment with multiple hues to suit various needs. The CaAl₁₀CoTiO₁₉ composition was chosen for acid and base stability tests. Specimens of CaAl₁₀CoTiO₁₉ were soaked in either 50% HNO₃ or 1 M NaOH for 12 h, filtered, and examined via L*a*b* color measurements and X-ray diffraction (XRD). XRD patterns (Figure 3) show no change in structure and L*a*b* measurements reveal no change in color. This is consistent with other hibonites, which have high acid/base stability; therefore, we believe that Ca(Al,Co,Ti)₁₂O₁₉ hibonites are also stable in most pH environments.

Table 2. L*a*b* Color Coordinates of Ca(Al,Co,Ni,Ti)₁₂O₁₉ Samples Compared to Other Well-Known Blue Compounds.

composition	% mass Co	L*	a*	b*
CaAl11Co0.5Ti0.5O19	4.24	45.86	1.12	–38.34
CaAl10.8Co0.6Ti0.6O19	5.05	44.94	3.95	–42.24
CaAl10.6Co0.7Ti0.7O19	5.85	38.26	0.52	–39.47
CaAl10.4Co0.8Ti0.8O19	6.64	38.26	0.52	–39.85
CaAl10.2Co0.9Ti0.9O19	7.41	39.99	0.61	–38.51
CaAl10CoTiO19	8.18	29.91	1.07	–33.57
CaAl10Co0.8Ni0.2TiO19	6.54	42.52	–3.44	–36.87
CaAl10Co0.5Ni0.5TiO19	4.09	43.04	–4.39	–36.21
CaAl10Co0.2Ni0.8TiO19	1.64	50.08	–8.78	–31.37
CaAl10CoTiO19–HNO3c	8.18	30.19	0.92	–32.92
CaAl10CoTiO19–NaOHc	8.18	30.63	0.24	–31.28
CaAl10NiTiO19a		49.59	–14.72	–28.44
CoAl2O4b	33.31	43.51	–4.46	–44.39
lapis lazuli stone		30.06	7.51	–24.21
YIn0.8Mn0.2O3		34.47	9.08	–44.39
YIn0.9Mn0.1O3		40.00	11.90	–47.90

^aFrom Li, et al.⁹

^bSample from Shepherd Color Company.

^cData collected after soaking in 50% HNO₃ or 1 M NaOH.

Figure 3.

XRD patterns for as-synthesized CaAl₁₀CoTiO₁₉ and after 12 h of 50% HNO₃ soaking and 12 h of soaking in 1 M NaOH.

Neutron structure analysis reveals that 70% of Co²⁺ is in the tetrahedral site. The origin of the color of Ca(Al,Co,Ti)₁₂O₁₉ hibonite blue is therefore primarily due to a strong d–d transition of tetrahedral Co²⁺ between 500 and 675 nm (Figure 4a). The spectroscopy of Co²⁺ in tetrahedral coordination of oxides has been well studied, and the broadening of the ⁴A₂(⁴F)–⁴T₁(⁴P) Co²⁺ transition has been explained through spin–orbit interactions of the d⁷ ions.¹⁹ Typically, this gives rise to three noticeable transitions in the visible region, as is the case of CoAl₂O₄. However, we found that Ca(Al,Co,Ti)₁₂O₁₉ hibonite has four transitions within the same 500–675 nm region. This difference may be due to the distortion of the tetrahedral site in the hibonite that cannot exist in the cubic CoAl₂O₄ spinel.²⁰ This, however, does not explain the nearly identical intensity of the color of hibonite blue compared to CoAl₂O₄. Despite a decrease in Co²⁺ concentration from 33% by mass to as low as 4% by mass, hibonite blue maintains a dramatic blue coloration. Both the tetrahedral and TBP Co²⁺ may give rise to the intense absorption between 500 and 675 nm, as the TBP environment is more accurately described as a set of disordered face-shared tetrahedra, where the disordered site is offset from the face by approximately 0.28 Å. These pseudotetrahedra are smaller in volume than the tetrahedral site, giving the larger Co²⁺ ion preference for the tetrahedral site. The combined TBP and tetrahedral Co²⁺ then gives the spectrum seen in the Ca(Al,Co,Ti)₁₂O₁₉ hibonite. Indeed, substitution of Ni²⁺ for Co²⁺ reduces this absorption because of the decrease in site preference of Ni²⁺ for the tetrahedral and TBP environments as well as the difference in electron configuration.⁹ The brightness of hibonite blues (Figure 5) is affected by both the Co²⁺ and Ti⁴⁺ substitutions. The strong absorption in the UV region (250 nm) is the result of a metal–ligand charge transfer between Ti⁴⁺ and O^2–, which at higher Ti⁴⁺ concentrations tails into the visible region, darkening the color. This effect is also seen in other hibonites.

Figure 4.

(a) UV–vis absorbance and (b) NIR reflectance of CaAl_12–2xCo_xTi_xO₁₉ samples. Commercial Co blue (CoAl₂O₄) was measured for comparison.

Figure 5.

Images of selected CaAl_12–xCo_xTi_xO₁₉ compounds and other commercially available powders.

Lastly, we investigated the usefulness of hibonite blue as a near-infrared (NIR) reflective material. NIR reflectivity is a property necessary for compounds to be used as energy-saving heat reflection colorants (“cool pigments”) and is the subject of increasing interest.²¹ Current blue pigments containing cobalt, such as CoAl₂O₄, have very low NIR reflectivity. This is due to two d–d transitions for tetrahedral Co²⁺, ⁴A₂(⁴F)–⁴T₂(⁴F) at 1400 nm and ⁴A₂(⁴F)–⁴T₁(⁴F) at 1600 nm.²² The high concentration of Co²⁺ in CoAl₂O₄ gives a large absorption in the NIR region. Thus, most Co²⁺-containing compounds have very low NIR reflectivity, making them unsuitable for energy-saving coatings. The hibonite blues (e.g., CaAl₁₁Co_0.5Ti_0.5O₁₉) show an increase in NIR reflectivity by as much as 60% between 1200 and 1600 nm and as high as 125% increase between 800 and 1000 nm while still maintaining the intense blue color (Figure 4b).

Conclusions

We have found that hibonite blues are stable in both strongly acidic and basic environments, with no change in structure or color after treatment. This rivals the durability of both cobalt blue and YInMn blue. Furthermore, with increasing Co/Ti content, there is increasing ultraviolet absorbance because of the Ti–O charge transfer. UV radiation is a common degradant of polymer coatings. Additives, such as inorganic UV-absorbing pigments, are added to reduce this organic degradation. Both cobalt blue and YInMn blue were shown to be useful as functional colorants in this regard.⁶ Similarly, the UV absorption of hibonite blue makes it an excellent candidate for polymer-based coating applications.

Discoveries of blue pigments are still scarce as no criteria for investigations of novel systems exist. Despite the abundance of weakly blue cobalt-containing compounds, we could not have foreseen the intensity of color and redness coming from hibonite blue. Predicting the exact color of new materials remains the realm of intuitive guess-work rather than science.²³

Methods

Stoichiometric amounts of CaCO₃ (Sigma-Aldrich, 99.0%), Al₂O₃ (Cerac, 99.99%), CoCO₃ (Baker Analytical, 99%), and TiO₂ (Cerac, 99%) were ground in an agate mortar, pressed into pills, and heated between 1623 and 1773 K over 12 h. Intermittent grinding and reheating were implemented to ensure sample purity and homogeneity. Compositions for the formula, CaAl_12–2xCo_xTi_xO₁₉, were x = 0.2, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, and 1.0. Phase purity and unit cell parameters were determined using powder XRD on a bench-top Rigaku Miniflex II powder diffractometer with Cu Kα radiation and a graphite monochromator. Neutron powder diffraction data were obtained from the Center for Neutron Research at the National Institute of Standards and Technology to investigate chemical composition and site occupancies. Samples were loaded into vanadium canisters, and data were collected at room temperature using a Cu(311) monochromator yielding a 1.54 Å wavelength. Initial structures were refined using GSAS EXPGUI software. Diffuse reflectance spectroscopic data were collected in the UV–vis region on an in-house spectrometer with MgO as a reference. The reflectance data were converted to absorbance with the Kubelka–Munk equation. NIR reflectance measurements were performed using a Jasco V-670 spectrometer up to 2500 nm. Color coordinates in the CIELAB color space were measured using a Konica Minolta CM-700d spectrophotometer. Samples of CaAl₁₀CoTiO₁₉ were soaked in 50% HNO₃ and 1 M NaOH overnight and dried, and their color properties and phase purity were determined via XRD in order to determine acid and base resistance.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.9b03255.

• Description of L*a*b* color space, magnetic susceptibility data and results, and detailed Rietveld refinement results of CaAl_10.6Co_0.7Ti_0.7O₁₉ and CaAl₁₀CoTiO₁₉ (PDF)

Author Contributions

M.A.S. conceived the research interest. B.A.D. carried out the synthesis. B.A.D. and J.L. performed neutron refinements and performed optical property measurements. M.A.S., B.A.D., and J.L. discussed the structure characterization and optical properties. All authors contributed to the drafting of the manuscript.

This work was supported by NSF Grant DMR-1508527.

The authors declare no competing financial interest.