An outstanding second-harmonic generation material BiB₂O₄F: exploiting the electron-withdrawing ability of fluorine

Abstract

It is well known that the incorporation of fluorine usually blue-shifts the absorption edge of a crystal due to its large electronegativity. Herein, we validate the new functionality of fluorine that leads to the strong second harmonic generation (SHG) effect in a unique nonlinear optical (NLO) bismuth borate, BiB₂O₄F, where F coordinates to both B³⁺ and Bi³⁺ cations. The F-incorporation not only forms a more asymmetric unit [BO₃F] compared to the almost optical isotropic [BO₄], but also withdraws the non-bonding electrons of Bi³⁺, increasing its local polarization; thus the total SHG efficiency is increased significantly. Indeed, experimental measurements and first-principles calculations both support our anticipation that BiB₂O₄F is a phase-matchable NLO material with an exceptionally large SHG response. In order to further consolidate the functionality of fluorine, a comparison study with the structurally similar compound BiB₂O₄(OH) was performed.

---

Introduction

Metal borates are well-known candidates for ultraviolet (UV) nonlinear optical (NLO) materials because of their tremendous structural flexibility and outstanding intrinsic physical characteristics such as high thermal stability and good transparency in the UV region.^1–3 Usually, a noncentrosymmetric structure consisting of asymmetric or polar structural units exhibits a large second harmonic generation (SHG) response.^4,5 For example, the large SHG effects of β-BaB₂O₄ and LiB₃O₅ are mainly contributed by the planar polyanions [B₃O₆]³⁻ and [B₃O₇]⁵⁻, respectively.^6,7 Meanwhile, the metal-centered MO_x polyhedra could also make a substantial or even major contribution to SHG efficiency. Here, M could be cations with second-order Jahn–Teller distortion such as d⁰, d¹⁰ cations showing a polar displacement in an octahedral coordination^8–11 or p-block cations with stereochemical lone pair electrons (i.e. Pb²⁺ and Bi³⁺).^12–14 Apparently, a combination of the above-mentioned functional units can lead to a synergistic effect, achieving excellent SHG properties, like Cd₄BiO(BO₃)₃, which shows an SHG response about six times that of KH₂PO₄ (6 × KDP).¹¹ Pb₂B₅O₉I is another representative showing a promising SHG performance arising from all three components, the Pb²⁺ cation, the I⁻ anion, and the polyborate anion.¹⁵

It is known that the introduction of halogen atoms into metal borates is beneficial to the transparency in the UV-region,¹⁶ such as KBe₂BO₃F (KBBF).¹⁷ Generally, F can be either coordinated to metal cations (forming M–F bonds) or boron atoms (forming B–F bonds, e.g. in the [BO₃F]⁴⁻ group). The former is quite usual, like KBBF,¹⁸ Pb₃B₆O₁₁F₂,¹⁹ and Cd₅(BO₃)₃F.²⁰ However, borates containing B–F bonds are rare; BaBOF₃ was the first reported example,²¹ followed by Na₃B₃O₃F₆,²² Li₂B₆O₉F₂,²³ Li₂B₃O₄F₃ ²⁴ and LiB₆O₉F.²⁵ Because the fluorine atom is more electronegative than the oxygen atom, it can be expected that [BO₃F] is more asymmetric than the [BO₄] unit. In fact, most of the above mentioned fluorooxoborates show observable SHG effects.^18–20,26

Is there another way to further enhance the SHG response rather than merely superimposing polarizations from individual functional units? Our idea is to take full advantage of the electron-withdrawing ability of the F atom, not only creating an asymmetric structural unit ([BO₃F]) but also enhancing the local polarization of Bi³⁺ cations. BiB₂O₄F is a unique fluorooxoborate in which F coordinates to both boron and bismuth.²⁷ It crystallizes in the space group P3₂ as shown in Fig. 1a, however, it was previously reported to be a moderate SHG material (about 1.8 × KDP), which is apparently lower than our expectation. In order to clarify this discrepancy, herein we present a re-investigation of the SHG properties of BiB₂O₄F by both theoretical and experimental methods. Our first-principles calculations suggested a very strong SHG effect of BiB₂O₄F (>10 × KDP), which was further validated by the re-measured optical properties. Furthermore, the comparison between the SHG properties of BiB₂O₄F and its structurally similar compound BiB₂O₄(OH) helps us to understand the function of F in NLO materials.

Fig. 1 Projection of the structures of (a) BiB₂O₄F and (b) BiB₂O₄(OH) along the [100] and [010] directions respectively. The pink arrows imply the orientation of lone pair electrons on Bi³⁺ cations. The unit cells are indicated by the black boxes.

Experimental

Synthesis

Typically, to obtain BiB₂O₄F crystals, a mixture of 0.3638 g Bi(NO₃)₃·5H₂O, 0.9275 g H₃BO₃ and 0.0630 g NaF with a Bi/B/F ratio of 1/20/2 was charged into the autoclave with an additional 0–0.5 mL of water. The autoclave was sealed and heated at 200 °C for 3 days. After the reaction, the product was washed with distilled water, and colorless crystals of BiB₂O₄F with a yield of ∼90% (based on Bi) were obtained. For the synthesis of BiB₂O₄(OH) (yield: ∼90%), a mixture of 0.3495 g Bi₂O₃ and 1.855 g H₃BO₃ with a Bi/B ratio of 1/20 was ground in an agate mortar, and then transferred into a 25 mL Teflon autoclave, which was sealed and heated at 220 °C for 3 days after adding 2 mL of water.

The use of a bismuth source, NaF and an appropriate amount of water strongly affected the final product. For example, the utilisation of Bi(NO₃)₃·5H₂O as the bismuth source, together with an appropriate amount of NaF, is necessary for the synthesis of BiB₂O₄F. BiB₂O₄(OH), on the other hand, can only be synthesized by using Bi₂O₃ as the bismuth source, and the purity of the product was sensitive to the content of water. If the added water was between 1 and 3 mL, pure BiB₂O₄(OH) could be obtained; if it was less than 1 mL, Bi[B₄O₆(OH)₂]OH appeared as an impurity; if it was more than 3 mL, a mixture of BiB₂O₄(OH) and Bi₃B₅O₁₂ formed. Here the purity of as-made BiB₂O₄F and BiB₂O₄(OH) was verified by powder X-ray diffraction (see details in the ESI†). Only phase-pure samples were used for further characterization.

Structure analysis and characterization

Single-crystal X-ray diffraction (XRD) data were collected on a Super Nova-CCD using graphite-monochromated Mo Kα radiation (λ = 0.71073 Å) at 104 K. Powder XRD data were collected to confirm the phase purity on a PANalytical X'pert diffractometer with a PIXcel 1D detector (Cu Kα, 40 kV and 40 mA). Thermogravimetric analysis (TG) was performed on a Mettler-Toledo TGA/DSC instrument under a N₂ flow. The UV-Vis diffuse reflectance spectra were recorded at room temperature using a powder sample with BaSO₄ as a standard (100% reflectance) on a Shimadzu UV-3100 spectrophotometer. Reflectance spectra were converted to absorbance spectra using the Kubelka–Munk function. The electronic structures and SHG coefficients for BiB₂O₄F and BiB₂O₄(OH) were calculated using the first-principles plane-wave pseudopotential method implemented in the CASTEP program.^28,29 Details of the computational method are given in the ESI.†

Powder SHG signals were measured using the experimental method adapted from Kurtz and Perry.³⁰ Samples of BiB₂O₄F and BiB₂O₄(OH) were ground and sieved into different particle sizes. As a reference, a crystalline α-BiB₃O₆ (BIBO) sample was ground into the range of 105–125 μm. The powder sample was pressed between quartz slides and secured with a tape in a 1 mm thick aluminum holder that had a Φ8 mm hole. The holder was then irradiated with a pulsed infrared beam (970 mV, 10 ns, 1 Hz, 18 mW cm⁻²) from a Q-switched YAG:Nd³⁺ laser of 1064 nm wavelength. A cutoff filter was used to minimize the background, and an interference filter (530 ± 10 nm) was used to select the second harmonic for detection with a photomultiplier tube attached to an oscilloscope. This procedure was then repeated using BIBO (as a standard). No index-matching fluid was used in any of the experiments.

Results and discussion

Structural and electronic characteristics

In the literature, the crystal structure of BiB₂O₄F was convincingly solved by the powder XRD technique in the space group P3₂ (a = 6.7147(1) Å and c = 6.4688(1) Å).²⁷ Our hydrothermal synthesis leads to aggregated particles in tens of micrometers and the EDX analysis on a selected area confirms the existence of fluorine in the compound (see Fig. 2). As shown in Fig. 1a, helical borate chains in BiB₂O₄F, which is constructed by corner-shared borate tetrahedra (2BO₄ + BO₃F), are interconnected by Bi³⁺ to form a three-dimensional condensed structure.

Fig. 2 (a) SEM image and (b) EDX results for BiB₂O₄F. The observed atomic ratio was presented, which is consistent with the calculated value within the experimental error.

In the same article reporting on BiB₂O₄F,²⁷ there was another compound with the formula “Bi₃[B₆O₁₃(OH)]”, possessing a highly related structure to BiB₂O₄F. That was also solved by the powder XRD technique but with very low symmetry, i.e. triclinic. In ref. 27, the authors claimed that the triclinic structure was a tentative (probably inaccurate) model, and during the structural refinements with powder XRD data, constraints were applied to obtain reasonable structure parameters. An exact structure determination requires single crystal X-ray diffraction. Moreover, the formula “Bi₃[B₆O₁₃(OH)]” was also doubtful, because the TG weight losses observed in ref. 27 and here in our work are 2.5 and 2.77 wt% (Fig. S1 in the ESI†), respectively, which is more consistent with the formula BiB₂O₄(OH) (cald 2.9 wt%) rather than Bi₃[B₆O₁₃(OH)] (cald 1.0 wt%).

We made an effort towards the syntheses of “Bi₃[B₆O₁₃(OH)]” and obtained small single crystals, thus a single crystal XRD become possible. More than a dozen single crystals were selected for XRD data collection, and in most cases, the R_int values were quite high (>0.20), using either a triclinic or trigonal symmetry during the manual data reduction. We speculate that the very large absorption of Bi³⁺ to X-ray is the possible cause, especially in the case that the spatial density of Bi³⁺ in BiB₂O₄(OH) is very high.

Fortunately, we eventually obtained a suitable dataset by cracking the large crystal into pieces, and selecting a small fraction for data collection at 104 K even though only Bi³⁺ could be refined anisotropically, while other light atoms (boron and oxygen) were refined isotropically. The obtained atomic coordinates were chemically reasonable with regular bond distances. The final crystal structure was refined in the polar space group P3₁ by full-matrix least-squares refinement.³¹ Detailed crystallographic information is listed in Table S1.† The atomic coordinates and selected bond lengths are listed in Table S2 and S3,† respectively. The CIF file and details of the structure are also provided in the ESI.† However, there are still small residual electron densities around Bi³⁺, which is therefore not surprising in such a case. Overall, the currently P3₁, and the previously reported P1, structures are basically the same, except for a small difference in formula and the coordination of some of the boron. In fact, the higher symmetry in P3₁ leads to much less structural parameters, which gave a green light for DFT calculations.

For BiB₂O₄(OH), there are 8 crystallographically independent non-hydrogen atoms, including 1 Bi, 2 B, and 5 O (shown in Fig. 1b). All atoms locate at general positions. Coordinates for H were determined based on structural analysis and bond valence sum (BVS) calculations. Both boron atoms are coordinated by four oxygen atoms in regular tetrahedral environments. The structure of BiB₂O₄(OH) is very similar to BiB₂O₄F; the main difference is the replacement of F⁻ by OH⁻, as shown in Fig. 1. Clarifying the structural similarity between BiB₂O₄F and BiB₂O₄(OH) would facilitate our further discussion about the structure–property relationship (shown later).

The bandgap energies for BiB₂O₄F and BiB₂O₄(OH) were estimated to be 4.43 and 4.28 eV, respectively, by the maximum peak of the first order differential curve for UV-Vis-NIR absorption spectra (see Fig. 3). No obvious absorption in the range of 350–1000 nm indicates the transparency for both samples in the entire visible region.

Fig. 3 UV-Vis-NIR absorption spectra for BiB₂O₄F and BiB₂O₄(OH).

The calculated electronic band structure of BiB₂O₄F along the symmetry lines in the Brillouin zone of the unit cell is plotted in Fig. 4a. This structure has an indirect bandgap (3.36 eV) which is smaller than the experimental value (4.43 eV) due to the discontinuity of XC energy. Fig. 4b gives the total and partial density of states (DOS and PDOS) projected on the constituent atoms in BiB₂O₄F. Several electronic characteristics considering the respective groups can be summarized as follows: (i) the energy region below −10 eV is mainly composed of the isolated inner orbitals of Bi (5d), O (2s), and F (2s), which are strongly localized deeply in the valence band (VB) and have a negligible influence on the neighboring atoms; (ii) the upper VB (from −10 eV) shows large hybridization between B (2p) and O/F (2p) orbitals, but the top of the VB is mainly composed of O (2p) orbitals, which is similar to the situation in BIBO;¹² (iii) the bottom of the conduction band (CB) is mostly contributed by the orbitals of Bi (6p) and O (2p). It should be emphasized that the states on both sides of the band gap are mainly composed of the orbitals from [BiO₅].

Fig. 4 Electronic structure in BiB₂O₄F (a) electronic band structure and (b) DOS/PDOS.

Similar to BiB₂O₄F, the electronic band structure, DOS and PDOS of BiB₂O₄(OH) are shown in Fig. 5. The theoretical energy bandgap is calculated to be 3.41 eV (indirect gap). The top of the VB is mainly composed of O (2p) orbitals and the bottom of the CB is mainly contributed by Bi (6p) and O (2p) orbitals, which is similar to the case of BiB₂O₄F. Additionally, the H 1s orbitals have hybridization with the electronic states on the O atoms due to the formation of hydroxyl groups, but make a rather small contribution to both VB and CB. Therefore, one can easily find that the [BiO₅] group is also the major functional group in BiB₂O₄(OH).

Fig. 5 Electronic structure in BiB₂O₄(OH) (a) electronic band structure and (b) DOS/PDOS.

Optical properties relevant to SHG output

The calculated linear refractive indices and birefringence for BiB₂O₄F and BiB₂O₄(OH) at 1064 nm are listed in Table 1. The birefringence values are slightly small, but still satisfy the phase-matching requirement for the SHG output in the IR and visible spectral regions. Since the space group is P3₂/P3₁, there are four independent non-zero SHG coefficients due to the Kleinman symmetry relationship, as shown in Table 1. The calculated SHG coefficients of BiB₂O₄F are much higher than that of BiB₂O₄(OH), and the values of d₃₃ are significantly larger than other d_ij coefficients, which are −11.43 and −5.23 pm V⁻¹ for BiB₂O₄F and BiB₂O₄(OH) respectively. Moreover, it is interesting that these values for BiB₂O₄F are even larger than the d_ij values of BIBO (expt. d₂₂ = 2.53 pm V⁻¹, d₁₆ = 2.8 pm V⁻¹, d₁₄ = 2.4 pm V⁻¹, d₂₃ = −1.3 pm V⁻¹),³² which is known as a promising SHG material in borates. Consequently, the powder SHG coefficients for BiB₂O₄F and BiB₂O₄(OH) are estimated to be 5.82 and 3.62 pm V⁻¹ according to the Kurtz–Perry method,³⁰ which are ∼1.77 and 1.10 times of BIBO (d_powder = 3.28 pm V⁻¹, ∼8 × KDP).³²

Table 1 Calculated refractive index, birefringence, and the SHG/PSHG effect in BiB₂O₄F and BiB₂O₄(OH), and the experimental PSHG effect as a comparison

	Calculated refractive index n and birefringence Δn at 1064 nm			Calculated SHG coefficients	PSHG effect dpowder
	n o	n e	Δn	d ij (pm V^−1)	Cal.	Exp.
a Powder SHG effect in BIBO is about 8 × KDP (about 3.28 pm V^−1), from ref. 32.
BiB2O4F	2.0699	2.0879	0.018	d 11 = −d12 = 4.49;	5.82	4.72
				d 16 = −d22 = 2.28	1.77 × BIBO^a	1.44 × BIBO^a
				d 15 = d24 = 1.44
				d 33 = −11.60
BiB2O4(OH)	2.1570	2.1403	0.017	d 11 = −d12 = 2.82;	3.62	2.48
				d 16 = −d22 = 3.10	1.10 × BIBO^a	0.76 × BIBO^a
				d 15 = d24 = 0.16
				d 33 = −5.23

---

Apparently, the theoretical calculations suggest excellent NLO performances for both BiB₂O₄F and BiB₂O₄(OH). In order to verify the calculations, the powder SHG effects of BiB₂O₄F and BiB₂O₄(OH) were re-measured by the Kurtz and Perry method, despite the fact that they were previously reported to be moderate in the literature.²⁷ It should be noted that a low stability under laser irradiation was observed during the SHG measurements for both compounds. For instance, once the power density of the laser emitter increased to ≥20 mW cm⁻², the powder burned from white to gray color, accompanied by a sharp decline of the SHG signal. It is well known that the output of the SHG signal increases with the enhancement of the input laser energy. Here, the final experimental data were collected by fixing the power density to be 18 mW cm⁻².

Fig. 6 shows the intensities of the SHG signal of powder BiB₂O₄F and BiB₂O₄(OH) samples as a function of the particle size. It is clear that the SHG intensities increase within the range below 100 μm and then becomes almost independent of the particle size, exhibiting a typical phase-matching behavior, which is consistent with the theoretical anticipation. The further comparison with the BIBO sample (105–125 μm, as the reference) reveals that the SHG signals of BiB₂O₄F and BiB₂O₄(OH) are about 2.07 and 0.57 times of BIBO, respectively. Therefore, the experimental d_powder are estimated to be 4.72 and 2.48 pm V⁻¹, respectively, which are smaller than our calculated values (5.82 and 3.62 pm V⁻¹).

Fig. 6 Dependence of the measured powder SHG signal intensity with the particle size of BiB₂O₄F and BiB₂O₄(OH) (α-BiB₃O₆ as a reference).

It is reasonable that the currently measured d_powder are lower than the theoretical values, because the sieved particles for optical measurements are in fact aggregations of smaller crystallites of BiB₂O₄F (see the SEM images as shown in Fig. 2a). We spent months on the hydrothermal synthesis to obtain large single crystals, however all attempts failed. Nevertheless, the very strong SHG signals, especially for BiB₂O₄F, already validate our original anticipation that they are good NLO materials with much stronger SHG effects than the previously reported values.

Here, we should explain why there is such a large difference in SHG efficiency between our results and that in ref. 27. The previous study mainly focused on the structural chemistry of three bismuth borates, and the observable SHG output was only supporting evidence of the noncentrosymmetric structures. The authors of ref. 27 interpreted data to infer that the SHG response might come from the one-dimensional borate chains, like in the case of Pb₆B₁₁O₁₈(OH)₉, which is, however, not correct (please refer to the discussions below). As mentioned above, both BiB₂O₄F and BiB₂O₄(OH) have a low damage threshold and probably the burning phenomenon was not noticed at that time. We believe this is the major reason why a significantly low SHG effect was reported in ref. 27.

In addition, in ref. 27, they report the SHG effect based on a powder sample. We did our best to synthesize crystals. Although we did not obtain perfect and large single crystals for BiB₂O₄F, the average size of the crystals is tens of micrometers (see Fig. 2a). Larger crystals would give stronger SHG signals. This might be the second reason why our experiments gave a higher d_powder.

More importantly, we move on to further analyze the structure–property relationship and try to figure out the significance of the incorporation of fluorine, which may have a general instructive effect on finding new NLO materials.

Structure–property relationship

The geometrical sum of the microscopic second-order susceptibility tensors of active units determines the overall SHG coefficient of a crystal. Hence an efficient spatial arrangement of the microscopic groups will generate a relative large macroscopic SHG coefficient. In BIBO, for example, the Bi³⁺ cations bond to the neighbor O²⁻ anions, forming the (BiO₄)⁵⁻ groups with a pyramidal shape,¹² so the Bi 6s lone pair electrons “point” toward the “empty” space along the polar b-axis, indicating a strong anisotropism of (BiO₄)⁵⁻ (Fig. 7c). Moreover, the orientations of all (BiO₄)⁵⁻ groups are identical, so their microscopic second-order susceptibility tensors are simply superposed. This leads to the very large overall SHG coefficient for BIBO and the predominant contribution is from (BiO₄)⁵⁻ (giving almost 90% of the total value).¹²

Fig. 7 Coordination environments of Bi³⁺ cations in (a) BiB₂O₄F, (b) BiB₂O₄(OH), (c) α-BiB₃O₆, and Pb²⁺ cation in (d) Pb₃B₆O₁₁F₂.

Both BiB₂O₄F and BiB₂O₄(OH) have a unique polar axis along the c-direction. Fig. 7a and 7b give the coordination environment of Bi³⁺ in BiB₂O₄F and BiB₂O₄(OH), respectively, which suggest that the 6s lone pair electrons are basically aligned along the c-axis. The microscopic second-order susceptibility tensors from BiO₅ groups are canceled out within the ab-plane but have a superposed effect along the polar direction. The interesting question is why these two closely related structures generate largely different NLO responses, i.e. the SHG output of BiB₂O₄F is much stronger than that of BiB₂O₄(OH).

Accordingly, the respective contributions of each ionic group to the overall optical effects were investigated by adopting the real-space atom-cutting method.⁶ There are two crystallographically independent B atoms in BiB₂O₄F; one presents as [BO₄], and the other is coordinated by three oxygen and one F as [BO₃F] (Fig. 1). The bismuth atom is linked to five oxygen atoms by strong Bi–O bonds with the lengths of 2.24–2.54 Å (Fig. 7a). Therefore, the respective contribution of the three anionic groups, i.e., [BO₄], [BO₃F] and [BiO₅], to the birefringence and SHG coefficients were analyzed, as shown in Table 2. The sums of the contributions from various anionic groups to the optical properties are always larger than the original total values because the oxygen ions were calculated twice in the atom-cutting procedures. It can be seen that the [BiO₅] and [BO₃F] groups are the main contributors to the birefringence, and the contributions are comparable.

Table 2 Calculated results and real-space atom-cutting analysis for the nonlinear optical properties in BiB₂O₄F at the incident light wavelength of 1064 nm

SHG coefficients (pm V^−1)	Original	Only BO4	Only BO3F	Only BiO5
d 11	4.49	0.42	1.48	4.59
d 16	2.28	0.31	0.81	1.68
d 15	1.44	0.42	−0.07	1.13
d 33	−11.60	−0.65	−2.18	−10.91
d powder	5.82	0.57 (10%)	1.50 (26%)	5.59 (96%)
Δn	0.018	0.008	0.014	0.016

---

For the contribution of SHG coefficients, one may easily find that the [BiO₅] groups make the dominant contribution (giving ∼96% of the total value) to BiB₂O₄F, which is similar to the situation of BIBO. Since BiB₂O₄F and BiB₂O₄(OH) have spatial arrangements similar to those of [BiO₅] and borate groups (see Fig. 1), the substitution of OH⁻ by F⁻ should be responsible for the stronger SHG output of BiB₂O₄F. First, the contribution to the total NLO response of [BO₃F] groups (∼26%) is larger than that of [BO₄] groups (∼10%) in BiB₂O₄F, because the [BO₃F] tetrahedron exhibits a larger spatial anisotropy due to the partial substitution of O atoms by F atoms. More importantly, the presence of F also influences the surrounding environment of Bi³⁺. As shown in Fig. 7a, two F atoms were found above Bi³⁺ with distances of 2.86 and 2.98 Å. Fluorine is the most electronegative atom, therefore, the non-bonding electrons (6s² electrons) of Bi³⁺ can be withdrawn towards two F atoms (along the c-axis) thus leading to a higher local polarization, which is beneficial to enhancing the total SHG response.

There also exists another interesting noncentrosymmetric lead borate, Pb₃B₆O₁₁F₂,¹⁹ which exhibits a SHG response arising from the superposing effects of the borate groups and the distorted [PbO_xF₂] (x = 4–6) polyhedra. The presence of F also affects the stereotactic chemistry of Pb²⁺. However, the F–Pb–F angles are ∼160° (Fig. 7d), and the majority of their electron-withdrawing effects on the lone pair electrons of Pb²⁺ are neutralized, so only a moderate SHG response was observed. Then, looking back into the case of BiB₂O₄F, these two F atoms locate at the same side of Bi (the F–Bi–F angle is ∼70°), indicating that their electron-withdrawing effects are synergistic. Moreover, this leads to a superior performance of BiB₂O₄F compared with BiB₂O₄(OH).

BIBO is the star compound in borate NLO materials,^12,32 and was used as the reference for SHG signals in our study. From the structural viewpoint, the strong SHG effect of BIBO originates from the perfect spatial arrangement of [BiO₄] groups.¹² In our study, the powder SHG effect of BiB₂O₄F is about two times of that of BIBO, according to theoretical and experimental results. As discussed above, the incorporation of fluorine into BiB₂O₄F leads to a synergistic increase of local polarizations of [BiO₅] and [BO₃F], and hence, a very strong NLO response. In addition, the spatial density of the [BiO₅] groups in BiB₂O₄F is about 1.33 times more than that of the [BiO₄] groups in BIBO.

Conclusion

In the literature, there are numerous metal borates showing excellent UV NLO properties. Most studies are devoted to those materials with appropriate arrangements of polar structure units, which has already become the fundamental principle for NLO materials. Here in our work, we rationalize the additional function of fluorine, which is the most electronegative atom, in further enhancing the SHG efficiency by withdrawing lone pair electrons of Bi³⁺. In view of this, we re-investigated the SHG properties of a unique bismuth borate, BiB₂O₄F, where F coordinates to both B and Bi atoms. DFT calculations and experimental results both point to a very strong d_powder, even higher than the star compound BIBO. Our detailed analyses on the structure–property relationship, especially the comparison study with its relative compound BiB₂O₄(OH), validate our judgement on the functionality of F atoms, which may have a general instructive effect on finding new NLO materials.

Acknowledgements

This work was supported by the Nature Science Foundation of China (grants 21101175, 21171178, 91222106, 91022036 and 11174297) and National Basic Research Project of China (grants 2010CB630701 and 2011CB922204). We acknowledge Prof. Z. G. Hu (Technical Institute of Physics and Chemistry, Chinese Academy of Science) for the assistance on the SHG measurements.

Notes and references

1. P. Becker, Adv. Mater., 1998, 10, 979–992.
2. C. Chen, Z. Lin and Z. Wang, Appl. Phys. B, 2005, 80, 1–25.
3. F. Li, X. L. Hou and S. L. Pan, Chem. Mater., 2009, 21, 2846–2850.
4. P. Shiv Halasyamani and K. R. Poeppelmeier, Chem. Mater., 1998, 10, 2753–2769.
5. C. F. Sun, C. L. Hu and J. G. Mao, Chem. Commun., 2012, 48, 4220–4222.
6. J. Lin, M. H. Lee, Z. P. Liu, C. T. Chen and C. J. Pickard, Phys. Rev. B: Condens. Matter, 1999, 60, 13380–13389.
7. Z. S. Lin, J. Lin, Z. Z. Wang and C. T. Chen, Phys. Rev. B: Condens. Matter, 2000, 62, 1757–1764.
8. H. S. Ra, K. M. Ok and P. Shiv Halasyamani, J. Am. Chem. Soc., 2003, 125, 7764–7765.
9. E. O. Chi, K. M. Ok, Y. Porter and P. Shiv Halasyamani, Chem. Mater., 2006, 18, 2070–2074.
10. Y. Inaguma, M. Yoshida and T. Katsumata, J. Am. Chem. Soc., 2008, 130, 6704–6705.
11. W. L. Zhang, W. D. Cheng, H. Zhang, L. Geng, C. S. Lin and Z. Z. He, J. Am. Chem. Soc., 2010, 132, 1508–1509.
12. Z. S. Lin, Z. Z. Wang, C. T. Chen and M. H. Lee, J. Appl. Phys., 2001, 90, 5585–5590.
13. S. H. Kim, J. Yeon and P. Shiv Halasyamani, Chem. Mater., 2009, 21, 5335–5342.
14. F. Kong, S. P. Huang, Z. M. Sun, J. G. Mao and W. D. Cheng, J. Am. Chem. Soc., 2006, 128, 7750–7751.
15. Y. Z. Huang, L. M. Wu, X. T. Wu, L. H. Li, L. Chen and Y. F. Zhang, J. Am. Chem. Soc., 2010, 132, 12788–12789.
16. H. P. Wu, S. L. Pan, K. R. Poeppelmeier, H. Y. Li, D. Z. Jia, Z. H. Chen, X. Y. Fan, Y. Yang, J. M. Rondinelli and H. S. Luo, J. Am. Chem. Soc., 2011, 133, 7786–7790.
17. B. C. Wu, D. Y. Tang, N. Ye and C. T. Chen, Opt. Mater., 1996, 5, 105–109.
18. Z. S. Lin, Z. Z. Wang, C. T. Chen, S. K. Chen and M. H. Lee, Chem. Phys. Lett., 2003, 367, 523–527.
19. H. Y. Li, H. P. Wu, X. Su, H. W. Yu, S. L. Pan, Z. H. Yang, Y. Lu, J. Han and K. R. Poeppelmeier, J. Mater. Chem. C, 2014, 2, 1704–1710.
20. G. H. Zou, L. Y. Zhang and N. Ye, CrystEngComm, 2013, 15, 2422–2427.
21. D. M. Chackraburtty, Acta Crystallogr., 1957, 10, 199–200.
22. G. Cakmak, T. Pilz and M. Jansen, Z. Anorg. Allg. Chem., 2012, 638, 1411–1415.
23. T. Pilz and M. Jansen, Z. Anorg. Allg. Chem., 2011, 637, 2148–2152.
24. T. Pilz, H. Nuss and M. Jansen, J. Solid State Chem., 2012, 186, 104–108.
25. G. Cakmak, J. Nuss and M. Jansen, Z. Anorg. Allg. Chem., 2009, 635, 631–636.
26. B. Andriyevsky, T. Pilz, J. Yeon, P. Shiv Halasyamani, K. Doll and M. Jansen, J. Phys. Chem. Solids, 2013, 74, 616–623.
27. L. Y. Li, G. B. Li, Y. X. Wang, F. H. Liao and J. H. Lin, Chem. Mater., 2005, 17, 4174–4180.
28. M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias and J. D. Joannopoulos, Rev. Mod. Phys., 1992, 64, 1045–1097.
29. S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. J. Probert, K. Refson and M. C. Payne, Z. Kristallogr., 2005, 220, 567–570.
30. S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968, 39, 3798–3813.
31. G. M. Sheldrick, SHELXS 97, Program for the Solution of Crystal Structures, and SHELXL 97, Program for the Refinement of Crystal Structures, University of Göttingen, Göttingen, Germany, 1997.
32. H. Hellwig, J. Liebertz and L. Bohatý, Solid State Commun., 1999, 109, 249–251.

---

Footnote

† Electronic supplementary information (ESI) available: Computational methods, tables of crystallographic data, TG curve and SEM photo for BiB₂O₄(OH), powder XRD patterns for both BiB₂O₄(OH) and BiB₂O₄F. CCDC 1014612. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4qi00192c

---