Sulfate Exchange of the Nitrate-Type Layered Hydroxide Nanosheets of Ln2(OH)5NO3·nH2O for Better Dispersed and Multi-color Luminescent Ln2O3 Nanophosphors (Ln = Y0.98RE0.02, RE = Pr, Sm, Eu, Tb, Dy, Ho, Er, and Tm)

Through restricting thickness growth by performing coprecipitation at the freezing temperature of ~4 °C, solid-solution nanosheets (up to 5-nm thick) of the Ln2(OH)5NO3·nH2O layered hydroxide (Ln = Y0.98RE0.02; RE = Pr, Sm, Eu, Tb, Dy, Ho, Er, and Tm, respectively) were directly synthesized without performing conventional exfoliation. In situ exchange of the interlayer NO3− with SO42− produced a sulfate derivative [Ln2(OH)5(SO4)0.5·nH2O] of the same layered structure and two-dimensional crystallite morphology but substantially contracted d002 basal spacing (from ~0.886 to 0.841 nm). The sulfate derivative was systematically compared against its nitrate parent in terms of crystal structure and phase/morphology evolution upon heating. It is shown that the interlayer SO42−, owing to its bonding with the hydroxide main layer, significantly raises the decomposition temperature from ~600 to 1000 °C to yield remarkably better dispersed oxide nanopowders via a monoclinic Ln2O2SO4 intermediate. The resultant (Y0.98RE0.02)2O3 nanophosphors were studied for their photoluminescence to show that the emission color, depending on RE3+, spans a wide range in the Commission Internationale de l’Eclairage (CIE) chromaticity diagram, from blue to deep red via green, yellow, orange, and orange red.

Background Y 2 O 3 is a widely used host lattice in the phosphor field, owing to its excellent structure stability, chemical durability, and particularly its ability to accept a substantial amount of various trivalent rare-earth activators for a broad range of optical functionalities. Due to their technical importance in the lighting and display areas, Y 2 O 3 -based phosphors are being widely investigated to correlate their luminescent performance with the characteristics of the phosphor powder [1,2]. Controllable synthesis has always been an active area of phosphor study, and the well-adopted processing technologies may include flux-assisted solid state reaction [3,4], solution synthesis, combustion [5], spray pyrolysis [6][7][8], and gas-phase condensation [9,10].
The hydroxide main layer of LRH is a close-packed low-energy plane, and thus, its two-dimensional growth needs lower activation energy than the thickness growth along the [001] direction. We recently demonstrated that, through suppressing thickness growth by lowering the synthesis temperature to~4°C, NO 3 − -LRH nanosheets of only~4-nm thick can be directly crystallized, without exfoliation, for a wide spectrum of single Ln (Ln = Pr-Er, and Y) [43]. With this technique, similarly thin nanosheets were produced in this work for the LRH solid solutions of Y/RE (RE = Pr, Sm, Eu, Tb, Dy, Ho, Er, and Tm) in good batch quantity (0.03 mol of LRH or~10 g). The effects of SO 4 2− exchange for interlayer NO 3 − on crystal structure and thermal behavior of the nanosheets and also characteristics and luminescent properties of the derived (Y 0.98 RE 0.02 ) 2 O 3 nanophosphors were studied in detail.

Freezing Temperature Crystallization of LRH Solid-Solution Nanosheets
The starting rare-earth sources are Pr 6 ). The other reagents of ammonium hydroxide solution (25 %), nitric acid (63 wt.%), and ammonium sulfate are of analytical grade and were purchased from Shenyang Chemical Reagent Factory (Shenyang, China). Nitrate solution of the rare earth was prepared by dissolving the oxide with a proper amount of nitric acid, followed by evaporation to dryness at 95°C to remove superfluous HNO 3 and a final dilution to 1.0 mol/L.
In a typical synthesis, a diluted ammonium hydroxide solution (1.0 mol/L) was slowly dripped (~2.0 mL/ min) into 300 mL of a 0.2 mol/L nitrate solution of Ln 3+ (Ln = Y 0.98 RE 0.02 ) kept at~4°C until pH~8 to produce Ln 2 (OH) 5 NO 3 ·nH 2 O nanosheets (30 mmol of Ln 2 (OH) 5 NO 3 ·nH 2 O per batch) [43]. For anion exchange with SO 4 2 −100 mL aqueous solution containing 15 mmol of (NH 4 ) 2 SO 4 (one SO 4 2− would replace two NO 3 − ) was added into the nanosheets suspension after 20 min of magnetic stirring (in situ anion exchange). The final product was collected via centrifugation after reaction for 1 h, followed by sequential washing with distilled water three times and ethanol one time and then air drying at 50°C for 15 h. Calcination of the dried nanosheets was performed in flowing oxygen gas (200 mL/min), using a heating rate of 5°C/min at the ramp stage and a holding time of 4 h. The final phosphor powders are all calcined at 1100°C, but with the Pr-and Tb-containing samples being subjected to an additional reduction in flowing H 2 (200 mL/min) at 1100°C for 1 h.

Characterization Techniques
Chemical analysis of the products was performed for Ln via inductively coupled plasma (ICP) spectroscopy (Model IRIS Advantage, Jarrell-Ash Japan, Kyoto), for NO 3 − via spectrophotometry (Ubest-35, Japan Spectroscopic Co., Ltd., Tokyo), and for S via combustion-infrared absorptiometry (Model CS-444LS, LECO, St. Joseph, MI). The detection limits of these analyses are all 0.01 wt.%. Phase identification was made via X-ray diffractometry (XRD; Model PW3040/60, Panalytical B.V., Almelo, the Netherlands) operated at 40 kV/40 mA, using nickel-filtered Cu-Kα radiation (λ = 0.15406 nm) and a scanning speed of 1.0°2θ/min. Fourier transform infrared spectroscopy (FTIR; Model Spectrum RXI, Perkin-Elmer, Shelton, CT) was performed by the standard KBr method. Powder morphology was analyzed by transmission electron microscopy under an acceleration voltage of 200 kV (TEM; Model JEM-2000FX, JEOL, Tokyo) and field emission scanning electron microscopy (FE-SEM; Model S-5000, Hitachi, Tokyo) under 10 kV. Thermogravimetry (TG; Model 8120, Rigaku, Tokyo) of the nanosheets was conducted in flowing air (100 mL/min) with a constant heating rate of 10°C/min. Specific surface area of the oxide phosphor was obtained with an automatic analyzer (Model TriStar II 3020, Micromeritics Instrument Corp., Norcross, GA) using the Brunauer-Emmett-Teller (BET) method via nitrogen adsorption at 77 K. Particle size/size distribution analysis was made with a laser-diffraction particle sizer (Model LA-920, Horiba Scientific, Kyoto), after ultrasonically dispersing the oxide powder in ethanol. Photoluminescence was analyzed at room temperature using an FP-6500 fluorospectrophotometer (JASCO, Tokyo) equipped with a 60-mm-diameter integrating sphere (Model ISF-513, JASCO) and a 150-W Xe-lamp for excitation. Measurements were conducted under identical conditions for all the samples, using a scan speed of 100 nm/min and slit widths of 5 nm for both excitation and emission. Fluorescence lifetime of the luminescence was analyzed with the FP-6500 equipment for Sm 3+ , Eu 3+ , and Tb 3+ , and with a DeltaFlex lifetime fluorescence spectrometer (Horiba Scientific) for the fast decay of Pr 3+ , Dy 3+ , Ho 3+ , Er 3+ , and Tm 3+ . L(Y 0.98 Eu 0.02 )H and its SO 4 2− derivative. The NO 3 − -type exhibits the characteristic 00l and non-00l diffractions of Ln 2 (OH) 5 NO 3 ·nH 2 O layered compounds [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. The strong and sharp 220 diffraction, arising from the ab plane (the hydroxide main layer), suggests that the host layers of L(Y 0.98 Eu 0.02 )H are well crystallized. Compared with the thick LRH crystals synthesized under high temperature [21-31, 33, 34], the 002 diffraction has a substantially lower intensity relative to the 220 one, implying that the primary crystallites are much less developed along the c-axis or rather thin, as also confirmed later by TEM analysis. SO 4 2− exchange of NO 3 − shortens the interlayer distance, as perceived from the obvious shifting of the 00l diffraction to a higher angle, and the basal spacing (d 002 ) calculated from the center of the 002 peak is~0.886 nm for NO 3 − -L(Y 0.98 Eu 0.02 )H and 0.841 nm for the SO 4 2− derivative. The values are close to those found for NO 3 − -LYH and its exchange product, respectively [43]. The sulfate derivative still exhibits quite strong 220 diffraction, indicating that sulfate exchange did not appreciably damage the hydroxide main layers. Shifting of the 220 peak from 2θ~28.86°to 29.02°and decreased intensity of the 400 diffraction by the anion exchange, however, suggests that the intercalated SO 4 2− is interacting with the hydroxide layers to deteriorate crystallinity of the sample.  [21][22][23][24], leads to lattice distortion and thus the slight peak-shifting [43]. The hydrogen bonding is also responsible for the observed interlayer contraction, since it would draw closer the adjacent positively charged hydroxide layers [43]. − is virtually complete. The results of the chemical analysis conform to those of FTIR spectroscopy (Fig. 1b). It is clearly seen that the intense NO 3 − absorption at~1385 cm −1 (ν 3 vibration, as of free anion) vanished, and meanwhile the ν 3 (~1105 cm −1 ) and ν 1 (~982 cm −1 ) absorptions, being characteristic of SO 4 2− , appeared from the exchange product. The non-splitting feature of ν 3 suggests that SO 4 2− is not directly coordinated to the metal center in the hydroxide layer while the emergence of ν 1 implies that the SO 4 2− tetrahedron is distorted owing to the effects of hydrogen bonding [43][44][45]. It is also owing to the effects of hydrogen bonding that the stretching vibrations of hydroxyls (~3565 cm −1 ) and the O-H radicals in hydration water (~3370 cm −1 ) are both substantially enhanced [44,45]. The twin absorption bands at~1520 and 1375 cm −1 indicate contamination of the product by CO 3 2− , mostly from dissolved atmospheric CO 2 during synthesis. All the SO 4 2− -LLnHs made in this work show almost identical interlayer distances owing to the limited content of RE, but the 220 diffraction successively shifts towards a higher angle along with the decreasing ionic size of the RE 3+ dopant (Fig. 2c). The 220 spacing (d 220 ) is shown in Fig. 2d as a function of RE 3+ size (for eightfold coordination) [46]. As the 220 diffraction reflects metalto-metal distance in the hydroxide layer [21][22][23][24], the d 220 value thus monotonically decreases towards a smaller RE 3+ as expected. The results also provide direct evidence of solid-solution formation. Figure 2 shows the results of electron microscopy for the two types of L(Y 0.98 Eu 0.02 )Hs. FE-SEM observation found that the NO 3 − -L(Y 0.98 Eu 0.02 )H is composed of 3D flower-like assemblies of nanoflakes having lateral dimensions up to~300 nm (Fig. 2a, coated with 10-nmthick tungsten for electrical conduction), while TEM analysis found entangled nanosheets of up to~5-nm thick (the inset). Calculated from the d 002 basal spacing of~0.886 nm, each single nanosheet would have only~5-6 stacking repetitions along the c-axis. Selected area electron diffraction (SAED) yielded a well-arranged spot-like pattern (the inset) for the hydroxide layer, indicating that the individual nanosheets are primary of single crystalline and are well crystallized. Anion exchange with SO 4 2− did not incur any appreciable morphology change to either the overall flower-like assemblies or the individual nanosheets (Fig. 2b), in compliance with our previous observations on NO 3 − -LYH [43].

Decomposition and Phase/Morphology Evolution of the Nanosheets upon Heating
Thermal behaviors of the nanosheets were analyzed via TG, and the results are compared in Fig. 3  − type clearly decomposes via three well-defined stages as previously observed for thick LRH crystals [25][26][27][28][29][32][33][34], with the first one being dehydration to form Ln 2 (OH) 5 NO 3 (up to~175°C), the second one being dehydroxylation to yield an intermediate mass with nominal composition of (Y 0.98 Eu 0.02 ) 2 O 2 (OH)NO 3 (up to~325°C), and the last one being further dehydroxylation and denitration to form oxide (up to~545°C). Though the NO 3 − -LLnH obtained in this work are nanometer thin, it shows a thermal behavior almost identical to that reported for thick crystals [25][26][27][28][29][32][33][34]. The sulfate derivative similarly decomposes via three steps, but only up to the significantly higher temperature of~1135°C and with a rather sluggish step in the~175-1005°C range (stage II).
To better understand the thermal decomposition of SO 4 2− -L(Y 0.98 Eu 0.02 )H, which has not been addressed prior to us, FTIR analysis was performed on the products obtained at various selected temperatures for 4 h (Fig. 4). It is clearly seen that the powders calcined up to 900°C are characterized by strong SO 4 2− absorptions and successively weaker ones of hydroxyls/water. The results may thus imply that the sluggish weight loss observed on the TG curve from~175 to 1005°C is mainly owing to successive dehydroxylation rather than desulfuration. The weight loss (~14.4 %) calculated for dehydroxylation of (Y 0.98 Eu 0.02 ) 2 (OH) 5 (SO 4 ) 0.5 to the nominal  -L(Y 0.98 Eu 0.02 )H is mainly due to the intramolecular hydrogen bonding between SO 4 2− and OH − groups. Raising the calcination temperature from 900 to 1000°C simultaneously eliminates the strong SO 4 2− and the already rather weak hydroxyl absorptions, suggesting that the sudden weight loss observed on the TG curve from~1005 to 1135°C is dominantly resulted from desulfuration. Again, the weight loss calculated for this thermal event (~14.9 %) is close to the value revealed by TG (~13.5 %). The SO 4 2− anions exhibit significantly split ν 3 and ν 4 vibrations for the 800 and 900°C products, implying that the tetrahedrons of SO 4 2− are substantially distorted by direct coordination to Ln 3+ [44,45,[47][48][49].
Phase evolution of the SO 4 2− -L(Y 0.98 Eu 0.02 )H upon heating was studied via XRD analysis of the products calcined at different temperatures, and the results are displayed in Fig. 5. It is seen that dehydration of the  layered compound at 200°C leads to an amorphous mass, which persists up to~600°C despite the already occurrence of dehydroxylation (Figs. 3 and 4). Further removal of hydroxyls at 800°C produced a phase mixture of poorly crystallized cubic Ln 2 O 3 (Ln = Y 0.98 Eu 0.02 , JCPDS: 00-043-1036) and monoclinic Ln 2 O 2 SO 4 (JCPDS: 00-053-0168), whose diffraction intensities both remarkably improve for the 900°C product. As the original layered compound has the approximate composition of Ln 2 (OH) 5 (SO 4 ) 0.5 ·nH 2 O, it can thus be said that the 900°C product is approximately composed of 1/2 mol of Ln 2 O 2 SO 4 and 1/2 mol of Ln 2 O 3 . Since the SO 4 2− in Ln 2 O 2 SO 4 is bidentately coordinated to Ln 3+ [47][48][49], the significant splitting of the ν 3 and ν 4 IR bands was thus observed once the oxy-sulfate compound was formed (Fig. 4). The Ln 2 O 2 SO 4 component desulfurates at the higher temperature of 1000°C, and thus, only cubic structured Ln 2 O 3 was found. The results of XRD comply well with those of FTIR (Fig. 4). That is, the ν 3 and ν 4 vibrations of SO 4 2− present as single bands for the 200-600°C products, as split bands for the 800 and 900°C products, and vanish for the 1000°C product. Similar phase evolution analysis of the NO 3 − -L(Y 0.98 Eu 0.02 )H found that cubic (Y 0.98 Eu 0.02 ) 2 O 3 crystallizes at~600°C via an amorphous state at lower temperatures. The results are not shown here since they are essentially identical to those previously reported by us for thick NO 3 − -LRH crystals [33]. Figure 6 compares XRD patterns of the oxides calcined from SO 4 2− -L(Y 0.98 RE 0.02 )H at 1100°C for 4 h. Only enlarged view of the 2θ = 25-35°region was given to show the effects of RE 3+ dopant. It is clearly seen that both the (222) and (400) diffractions steadily shift to higher angles along with decreasing ionic radius of RE 3+ , indicating the formation of solid solution. The lattice parameter (a, in angstrom) calculated from the strongest (222) diffraction indeed becomes successively smaller towards a smaller RE 3+ (Fig. 6). The crystallite size assayed from the (222) diffraction by applying Scherrer equation is~35 nm for all the oxides, irrespective of the dopant type. The cell parameters determined herein are all larger than the 10.547 Å reported for pure Y 2 O 3 (JCPDS: 00-043-1036, Y 3+ close to Ho 3+ in radius), possibly owing to the limited crystallite size of the present powders.
Morphology evolution of the nanosheets during calcination was studied with NO  Figure 7 exhibits typical FE-SEM morphologies for the powders calcined at some representative temperatures. It is seen that the 800°C product from NO 3 − -L(Y 0.98 Eu 0.02 )H well retained the overall morphology of its precursor, despite that it has been a phase-pure oxide, and the flower-like assemblies (domains) and the individual nanosheets within the domains are clearly observable. Calcination at 900°C led to substantial collapse of the nanosheets into nanoparticles within each domain, owing to the thermal stress arising from crystallite growth, and the domain boundary is still identifiable. Significant crystallite growth was observed at 1100°C, together with densification of some of the domains via inter-particle sintering to form dense aggregates. The final powder was found to have a specific surface area of~8.3 m 2 /g, corresponding to an average  to 1100°C causes complete collapse of the domains to yield a substantially better dispersed oxide powder, and the evolution of SO x gas was believed to promote disintegration of the domains. Accordingly, the final oxide has a much higher specific surface area of~17.5 m 2 /g (average particle size~67 nm).
The amount of residual sulfur in the oxide phosphors calcined at 1100°C was assayed via ICP elemental analysis to be up to 0.18 wt.% in our previous work [50]. Figure 8 shows the particle size/size distribution of the two kinds of (Y 0.98 Eu 0.02 ) 2 O 3 powders calcined at 1100°C.  It is clearly seen that the powder from NO 3 − -L(Y 0.98 Eu 0.02 )H exhibits a bimodal size distribution owing to the presence of hard aggregates (Fig. 7). The finer portion (~56.5 vol.%) has an average particle (cluster) size of~320 ± 43 nm while the coarser part has a value of~6.21 ± 0.53 μm. A unimodal size distribution was observed for the (Y 0.98 Eu 0.02 ) 2 O 3 powder from SO 4 2− -L(Y 0.98 Eu 0.02 )H, and the average particle size was analyzed to be~219 ± 94 nm. The above results are in agreement with morphology observations (Fig. 7), and further confirm that SO 4 2− exchange of the interlayer NO 3 − is beneficial to the derivation of finer and better dispersed oxide powders.
Photoluminescent Properties of the (Y 0.98 RE 0.02 ) 2 O 3 Nanophosphors Figure 9 shows photoluminescence excitation/emission spectra for the (Y 0.98 RE 0.02 ) 2 O 3 nanophosphors calcined at 1100°C, with the excitation and emission wavelengths used for the measurements indicated in each part of the figure. The origins of these main bands [19,[51][52][53] are summarized in Table 1, together with the chromaticity coordinates of emission and the fluorescence lifetime. The origins of the other PLE/PL bands in each part of Fig. 9 are well documented and can be found in the literature [19,[51][52][53]. It is seen from the Commission Internationale de l'Eclairage (CIE) chromaticity diagram that the phosphors synthesized in this work span a wide range of emission colors, from blue (Tm 3+ ) to deep red (Pr 3+ ) via green (Tb 3+ , Ho 3+ , and Er 3+ ), yellow (Dy 3+ ), orange (Sm 3+ ), and orange red (Eu 3+ ).

Conclusions
It is shown in this work that coprecipitation at the freezing temperature of~4°C can directly produce, without exfoliation, solid-solution nanosheets of the nitrate-type layered hydroxides of Ln 2 (OH) 5  via in situ anion exchange was achieved to produce the sulfate derivative of SO 4 2− -LLnH. Detailed characterizations of both the types of layered materials and their calcination products via the combined techniques of XRD, FTIR, DTA/TG, FE-SEM/TEM, BET, particle sizing, and photoluminescence spectroscopy have led to the following main conclusions: (1) anion exchange did not bring about any appreciable change to the layered structure and the two-dimensional crystallite morphology, but induces a basal-spacing contraction from~0.886 to 0.841 nm, (2) the interlayer SO 4 2− significantly raises the decomposition temperature of the nanosheets from~600 to 1000°C to yield oxide via a monoclinic-structured Ln 2  -LLnH are much better dispersed and finer than those from NO 3 -LLnH, and exhibit emission colors, depending on RE 3+ , covering a wide range in the CIE chromaticity diagram, from blue to deep red via green, yellow, orange, and orange red.