Structural variability in M²⁺ 2-hydroxyphosphonoacetate moderate proton conductors

General information

A racemic mixture of R,S-2-hydroxyphosphonoacetic acid (50% w/w stock solution in water), H₃HPAA, was from Biolabs, UK. In-house, deionized (DI) water was used for all syntheses. Stock solutions of HCl, HNO₃ and NaOH were used for pH adjustments. A parallel synthesis procedure was applied to obtain solids by hydrothermal reactions. The autoclave block is made of aluminum and contains six reaction chambers in a 3×2 array. Teflon reactors have an inner diameter of 19 mm and a depth of 18 mm, with a total volume of about 5 mL. A thin sheet of Teflon covers the reaction vessels, which are then sealed inside a specially designed aluminum autoclave. Elemental analyses (C, H, N) were measured on a Perkin–Elmer 240 analyzer. Thermogravimetric analysis (TGA) data were recorded on an SDT-Q600 analyzer from TA instruments. The temperature varied from RT to 1000°C at a heating rate of 10°C·min⁻¹. Measurements were carried out on samples in open platinum crucibles under air flow.

Materials syntheses

1D compounds M²⁺(OOCCH(OH)PO₃H)·2H₂O (M²⁺=Mg, Zn); M-H₁HPAA-1D were prepared according to literature [14]. MgCl₂·6H₂O (4.39 mmol) or ZnNO₃·6H₂O was added to a solution formed by 1.0 mL of the ligand stock solution (4.39 mmol) and DI water (25 mL) under vigorous stirring. Solution pH was adjusted to 2.7 with 5.0 M or 1.0 M NaOH stock solutions. The clear, slightly yellow solutions were stored at RT. Microcrystalline materials precipitated after 3 days. The precipitates were isolated by filtration (without washing) and air-dried. Elemental analysis for Mg-H₁HPAA-1D: calculated (wt%): C 11.21, H 3.30; found (wt%): C 10.55, H 3.54, Yield ~70% based on the metal. Elemental analysis for Zn-H₁HPAA-1D: calculated (wt%): C 9.40, H 2.76; found (wt%): C 9.17, H 2.82, Yield ~70% based on the metal.

Mg(OOCCH(OH)PO₃H)·2H₂O, Mg-H₁HPAA-2D, was prepared according to literature [14]. MgCl₂·6H₂O (4.39 mmol) or ZnNO₃·6H₂O was added to a solution formed by 1.0 mL of the ligand stock solution (4.39 mmol) and DI water (25 mL) under vigorous stirring. Solution pH was adjusted to 2.7 with 5.0 M or 1.0 M NaOH stock solutions. The clear, slightly yellow solutions were stored at RT. After 1 day the precipitates were isolated by filtration (without washing) and air-dried. Elemental analysis for Mg-H₁HPAA-2D: calculated (wt%): C 11.21, H 3.30; found (wt%): C 11.14, H 3.43, Yield ~84% based on the metal.

Zn(OOCCH(OH)PO₃H)·2H₂O, Zn-H₁HPAA-2D. A modified, previously reported synthesis was followed [19]. Zn(OOCCH(OH)PO₃H)(H₂O)₂ was prepared by dissolving Zn(OOCCH₃)₂(H₂O)₂ (11.17 mmol) in a solution formed by 2.5 mL of the ligand stock solution (11.17 mmol) and DI water (25 mL) under vigorous stirring. Solution pH was adjusted to 2.0 with 5 M KOH stock solution. The solution was transferred into a flask and heated under reflux to 100°C for 72 h. The precipitate was isolated by filtration and air-dried. Elemental analysis for Zn-H₁HPAA-2D: calculated (wt%): C 9.40, H 2.76; found (wt%): C 9.00, H 2.86, Yield ~71% based on the metal.

KZn₆(OOCCH(OH)PO₃)₄(OH)·5H₂O, KZn₆-HPAA-3D. A KOH solution was used for all pH adjustments. A quantity of Zn(H₂CCOOH)₂·2H₂O (0.4902 g, 2.23 mmol) was added to a solution containing 0.5 mL of the ligand stock solution (2.23 mmol) and 1.5 mL of DI water under stirring. The pH of the resulting solution was 0.74 and it was adjusted to 1, 2.5 or 4.5 with a 4 M stock solution of KOH. A total filling volume of ~3 mL per reactor was used. The reaction was conducted at 180°C for 3 days. The resulting solid was filtered off, washed twice with DI water and acetone, and dried at 50°C. Elemental analysis for KZn₆-HPAA-3D: calculated (wt%) for Zn₆KP₄O_31.5C₈H_25.5: C 8.35, H 1.66; found (wt%): C 8.664, H 1.375, Yield ~85% based on the metal.

NH₄Zn(OOCCH(OH)PO₃), NH₄Zn-HPAA-3D. A modified literature synthesis was followed [20]. NH₄Zn-HPAA-3D was prepared by dissolving ZnSO₄·7H₂O (10 mmol) in a solution formed by 2.25 mL of the ligand stock solution (10 mmol) and DI water (20 mL) and NH₄Cl (125 mmol) in DI water (20 mL) under vigorous stirring. Solution pH was adjusted to 2.8 with solid NaOH. The solution was transferred into a flask and heated under reflux to 100°C for 48 h. The precipitate was isolated by filtration and air-dried. Elemental analysis for NH₄Zn-HPAA-3D: calculated (wt%): C 10.16, H 2.56; N 5.92, found (wt%): C 10.48, H 2.77; N 5.70, Yield ~65% based on the metal.

Post-synthesis modifications with NH₃ and HCl

Adsorption of NH₃vapors: Based on the derivatization procedure followed in literature [21]. In separate experiments, an amount of 600 mg of Mg-H₁HPAA-1D, or Zn-H₁HPAA-2D or KZn₆-HPAA-3D was placed in contact with vapors from an aqueous solution of 14 wt% of NH₃ for various contact times (17 h for Mg-H₁HPAA-1D; 3.5 h for Zn-H₁HPAA-2D and 1 h for KZn₆-HPAA-3D) in a closed container and then dried for 1 h.

Adsorption of HCl: KZn₆-HPAA-3D (0.3801 g, 0.33 mmol) was exposed to HCl vapors generated from the reaction of an excess of 95% H₂SO₄ with NaCl for 30 min. After the adsorption reaction the solid was washed with 96% ethanol, filtered and air-dried. Elemental analysis for: Mg(HO₃PCHOHCO₂)(H₂O)_2.6(NH₃)_0,2, Mg-H₁HPAA-1D_NH₃, calculated (wt%): C 11.22, H 3.30; N 1.29, found (wt%): C 11.14, H 3.43; N 1.29; Zn(HO₃PCHOHCO₂)(H₂O)₂(NH₃), Zn-H₁HPAA-2D_NH₃, calculated (wt%): C 8.82, H 5.14; N 3.70, found (wt%): C 8.61, H 5.48; N 3.35.

Structural characterization

Laboratory X-ray powder diffraction (XRPD) patterns were collected on a PANanalytical X’Pert Pro diffractometer equipped with an X’Celerator detector. XRPD patterns corresponding to the single phases were auto-indexed using the DICVOL06 program [22] and the space groups were derived from the observed systematic extinctions. The crystal structure of KZn₆-HPAA-3D was successfully solved following an ab initio methodology. For this purpose, a high resolution synchrotron powder data set was collected on ID31 powder diffractometer of ESRF, European Synchrotron Radiation Facility, (Grenoble, France) using a wavelength λ=0.2998 Å selected with a double-crystal Si (111) monochromator and calibrated with Si NIST (a=5.43094 Å). The Debye-Scherrer configuration was used with the sample loaded in a rotating borosilicate glass capillary of diameter of 1.0 mm. The overall measuring time was ~100 min to have very good statistics over the angular range 1.7–16° (in 2θ). The data from the multi-analyzer Si(111) stage were normalized and summed into 0.003° step size with local software. The integrated intensities extracted with the program Ajust [23] were introduced in the direct methods program XLENS [24]. The starting framework model, containing the totally of atoms in the asymmetric part of the unit cell except those corresponding to the solvent, was derived from the interpretation of the electron density map computed with the set of refined phases with the highest combined figure of merit. The crystal structure was refined by the Rietveld method [25] using the GSAS package [26] and the EXPGUI graphic interface [27]. The following soft constraints were imposed in order to preserve chemically reasonable geometries for the phosphonate, alkyl chain and amine groups. The soft constraints were as follows: PO₃C₁ tetrahedron, P–O [1.53(1) Å], P–C₁ [1.80(1) Å], O···O [2.55(2) Å], O···C₁ [2.73(2) Å]; C₁OH–C₂OO group, C₁–C₂ [1.50(1) Å], C₂–O_carb [1.23(1) Å], C₁–OH [1.40(1) Å], P···OH [2.68(2) Å], C₂–OH [2.40(2) Å], O_carb···O_carb [2.21(2) Å], C₁···Ocarb [2.36(2) Å]. No attempts to locate the H atoms were carried out due to the limited quality of the XRPD data. Two isotropic atomic displacement parameters were fixed for all atoms. Selected structural data are reported in Table 1 and the final Rietveld plot is given in the Supporting Information (SI) as Figure S1.

Thermodiffractometric studies at different relative humidity values were collected on a D8 ADVANCE (Bruker AXS) diffractometer equipped with a Johansson Ge(111) primary monochromator, which gives a strictly monochromatic Mo radiation (λ=0.7093 Å), and an Anton Paar MHC-trans chamber. The X-ray tube was operating at 50 kV and 50 mA. The energy-dispersive linear detector LYNXEYE XE 500 μm, optimized for high energy radiation, was used with the maximum opening angle. Data were collected at the temperature range 25 and 80°C and 95% RH using a heating rate of 1°C/min. Samples were measured between 2 and 21° (2θ) with a step size of 0.017° and counting time of 154 s/step. Samples were held at each temperature for 10 min before recording any pattern, giving sufficient time for any transformation to take place.

Conductivity characterization

Impedance measurements were carried out on cylindrical pellets (diameter ~5 mm and thickness ~1 mm) obtained by pressing 40 mg of sample at 250 MPa, for 1 min. The pellets were pressed between porous C electrodes (Sigracet, GDL 10 BB, no Pt). The sample cells were placed inside a temperature- and humidity-controlled chamber (Espec SH-222) and connected to a HP4284A impedance analyzer. AC impedance data were collected over the frequency range from 20 Hz to 1 MHz with an applied voltage of 1 V. To equilibrate water content, pellets were first preheated (0.2°C/min) from 25 to 80°C and RH 95%. Impedance spectra were recorded on cooling using a stabilization time of 3 h for each temperature (80, 70, 60, 50, 40, 30, 25°C). Water condensation on sample was avoided by reducing first the relative humidity before decreasing temperature. All measurements were electronically controlled by the winDETA package of programs [28].

Proton conductivity studies

The members of this class of coordination networks exhibit high structural versatility, acidic sites and, usually, both non-coordinating and metal-bound water molecules. All these structural features may be the basis for creating a variety of extended H-bonding networks to tune proton conductivity. Although water-mediated is the common mechanism of proton conductivity [5], [6], [9], [34], it may be enhanced by incorporating other species, such as NH₄⁺/NH₃. Furthermore, post-synthesis modifications, such as ion exchange, where possible, and/or protonation through gas-solid reactions could be also, in principle, processes that could be envisaged for improving proton conduction properties [21], [35], [36], [37]. In the present study, impedance spectroscopy measurements were carried out in the range between room temperature and 80°C at 95% RH in order to correlate proton conductivities with structural and/or morphological changes for all synthesized solids as well as for some derivatives obtained by post-synthesis treatments by exposure to NH₃ or/and HCl gases.

Impedance spectra at different temperatures and 95% RH are shown in Figure S8 for Mg-H₁HPAA derivatives as representative example. The total pellet resistence was calculate from the intercept of the spike or, alternatively, from the arc (low frequency end) on the Z’ axis. The overall pellet conductivities for M²⁺-H₁HPAA (M=Mg²⁺ and Zn²⁺) compounds, in Arrhenius format, are displayed in Fig. 7.

Fig. 7:

Arrhenius plots for: (a) Mg-H₁HPAA derivatives [▼ Mg-H₁HPAA-2D as synthesized; ■ Mg-H₁HPAA-1D as-synthesized; ● Mg-H₁HPAA-2D after conversion; ▲ Mg-H₁HPAA-2D_NH₃] and (b) Zn-H₁HPAA derivatives [◆ KZn₆-HPAA-3D; ∇ Zn-H₁HPAA-2D as-synthesized; ○ Zn-H₁HPAA-2D_NH₃; ▸ NH₄Zn-HPAA-3D; ★ Zn-H₁HPAA-2D after exposure of KZn₆-HPAA-3D to HCl(g)]. Activation energies are given for each compound.

As commented above, crystalline 1D phases can be synthesized for Mg²⁺, Zn²⁺ from simple crystallization at room temperature. The 1D phase is not stable in the entire temperature range studied and transforms to a 2D phase that is stable, at least, up to 80°C and 95% RH. The 2D solids (Zn²⁺ and Mg²⁺ derivatives), either as synthesized or derived from the 1D phase, show moderate proton conductivities range between 2.1×10⁻⁵ S/cm and 6.7×10⁻⁵ S/cm at 80°C and 95% RH. The transition 1D→2D, shown only for Mg-H₁HPAA-1D, is detected in the Arrhenius plot (Fig. 7a) as an abrupt change in the slope and, hence, in conductivity values between 80 and 60°C (■ points). Once the 1D→2D phase transformation has taken place, no further change in slope occurs down 60°C. The converted 2D material exhibits a behavior practically similar to that corresponding to the as-synthetized Mg-H₁HPAA-2D solid (▼). A second cycle of measurements for the converted Mg-H₁HPAA-2D phase (●), previously checked by XRD (Figure S9), confirms a monotonic behavior in proton conductivity, similar to the as-synthesized 2D solid, over the complete range of temperature studied, the small increase observed being attributed to slight changes in morphology. This result also indicates that even if morphological changes occur, the proton conductivity is essentially an intrinsic feature of the structure of these solids. Zn-H₁HPAA-1D follows a quite similar behavior (not shown).

All the aforementioned compounds adsorb NH₃ vapors together with water from an aqueous solution. This, generally causes an increase in proton conductivity, however, an explanation of this effect is not simple. Mg-H₁HPAA-1D adsorbs a small amount of ammonia to give a compound with the stoichiometry Mg(HO₃PCH(OH)CO₂(H₂O)_2.6(NH₃)_0.2, Mg-H₁HPAA-1D_NH₃. The XRD pattern of this solid, Figure S10, shows splitting of diffraction peaks observed at lower angles, which may be attributed to a higher separation of the chains upon NH₃ adsorption in a relatively ordered fashion. This solid transforms to Mg-H₁HPAA-2D at high relative humidity, but still retains a small fraction of NH₃. This process causes a further enhancement of the proton conductivity (Fig. 7a, ▲ points), up to 2.6×10⁻⁴ S/cm, accompanied by partial amorphization of the solid, as observed by XRD (Figure S10). Since no structural changes in the crystalline 2D phase could be noted, NH₃-induced amorphization may be related to a particle size reduction. However, Zn-H₁HPAA-2D differs from Mg-H₁HPAA-2D in the capacity to adsorb NH₃. Thus, upon 3.5 h of exposure, the resulting solid, Zn-H₁HPAA-2D_NH₃, has the composition Zn(HO₃PCHOHCO₂)(H₂O)₂(NH₃). As a result, proton conductivity increases slightly with respect to the as-synthesized 2D material (Fig. 7b, ○ points). No intercalation of NH₃ into the crystalline Zn-H₁HPAA-2D phase was detected by XRD but a partial solid amorphization, possibly due to particle size reduction as it was also observed for Mg-H₁HPAA-2D_NH₃. The activation energy values (Fig. 7) for these compounds fall into a range characteristic of a Grotthuss-type proton transfer mechanism, i.e. below 0.5 eV.

The 3D solids were also investigated for their proton conductivity, in order to evaluate the influence of the network architecture. These 3D frameworks present cavities containing H₂O and/or NH₄⁺ guest molecules that may, in principle, contribute to establish hydrogen transport pathways and, hence to proton conductivity in different ways. The Arrhenius plots for the as-synthesized and derivatized solids are displayed in Fig. 7b. Solid KZn₆-HPAA-3D (◆) exhibits a very low proton conductivity, 1.6×10⁻⁶ S·cm⁻¹ at 80°C and 95% RH, due to the presence of basic groups rather than acidic groups and mobile lattice water, as happens with lower dimensionality solids. Hydrated K⁺ ions are not thought to contribute to this residual conductivity as these ions occupy fixed positions in the framework playing a structural role. Given the low proton conductivity of KZn₆-HPAA-3D, the solid was put in contact with NH₃ vapors in order to increase its proton conductivity. However, no changes were found for this solid, due to the low amount of adsorbed ammonia in this relatively basic framework. Exposure of this solid to dry HCl gas led again to the phase Zn-H₁HPAA-2D as main phase, which is considered the most thermodynamically stable phase under high humidity conditions (Fig. 8). This process implies partial removal of Zn²⁺ and total depletion of K⁺ ions due to ligand protonation, while the remaining Zn²⁺ ions adopt exclusively an octahedral coordination. The proton conductivity is further improved in the resulting 2D solid HCl-treated compound (★), 4.5×10⁻⁴ S·cm⁻¹ at 80°C and 95% RH, as compared to the as-synthesized layered material, Zn-H₁HPAA-2D, which may be attributed to extrinsic effects caused by morphological changes upon acid treatment [38], [39].

Fig. 8:

X-ray powder diffraction patterns for KZn₆-HPAA-3D before and after exposure to HCl(g) (KZn₆-HPAA-3D_HCl).

Since NH₄Zn-HPAA-3D exhibits weak H-bonding interactions (Figure S11) an increase of the proton conductivity should be expected. As can be seen in Fig. 7b, this compound shows proton conductivity values 1.4×10⁻⁴ S·cm⁻¹ at 80°C and 95% RH, two orders of magnitude higher than those of the potassium derivative. These may be rationalized with the presence of more effective proton transfer pathways in this framework, in comparison to those present in KZn₆-HPAA-3D. In these solids, a Grotthuss-type proton transfer mechanism is also observed.