Working With HF and Rare Earths: What the Literature Doesn't Tell You
The pKa of hydrofluoric acid is commonly listed as 3.17 at 25°C in undergraduate textbooks. That number becomes almost meaningless the moment you start working with rare earth elements in actual practice. I spent about six years running extraction and precipitation experiments with lanthanide fluorides before I stopped trusting standard pKa tables for anything beyond rough estimates. Here is the problem: HF doesn't just dissociate in water. It forms ion pairs, polymeric species, and strong complexes with metal cations. When you add a trivalent rare earth ion to an HF solution, you are no longer dealing with a simple monoprotic acid equilibrium. The free fluoride concentration drops dramatically because LnF2+, LnF2+, and higher complexes form. This shifts the apparent acidity of the solution in ways that standard pKa calculations completely miss. I encountered this directly during a purification run where I was trying to separate neodymium from praseodymium using selective fluoride precipitation. The protocol called for adjusting the solution to pH 2.0 with HF and then adding more fluoride to precipitate one element while keeping the other in solution. The precipitation didn't follow the predicted selectivity at all. I ended up co-precipitating nearly everything.
The workaround was straightforward once I understood what was happening. Instead of targeting a bulk pH, I calculated the free fluoride concentration using conditional stability constants that accounted for HF speciation andLn-F complexation. I used the NIST Critical Stability Constants database values for each lanthanide fluoride complex. The key equation is that the free [F-] is determined by the equilibrium between HF dissociation and metal-fluoride complex formation, not by the measured pH alone. A pH meter reading in an HF-rare earth system is essentially measuring the activity of H+ ions in a solution where fluoride is being consumed by complexation. The apparent pKa shifts with ionic strength and metal concentration. One counter-intuitive point that people miss: in concentrated HF solutions above about 1 M, the effective acidity actually decreases as you add more HF. This is because HF begins to self-associate into polymeric species like H2F- and H3F2-, and the activity coefficient of the proton changes non-linearly. For rare earth work at high fluoride concentrations, the Hammett acidity function H0 becomes a more useful measure than pH. I stopped using pH meters for anything above 0.5 M HF and switched to indicator-based H0 measurements. It took me longer to get reliable readings but the data was actually consistent. Another thing worth noting is temperature dependence. The pKa of HF changes by roughly -0.02 units per degree Celsius near room temperature, but in the presence of rare earth ions the shift is larger and non-linear. If you are running extractions at elevated temperatures, which many rare earth processes do, the standard 3.17 value is off by enough to matter for process control. I observed a shift of about 0.15 pKa units when going from 25°C to 60°C in a neodymium-HF system, and that shift alone accounted for a 20% difference in precipitation efficiency.
The practical takeaway is this: when working with HF and rare earth elements, you need to model the full speciation. Use a program like PHREEQC, Visual MINTEQ, or build your own equilibrium solver. Feed it the stability constants for each Ln-F complex, the HF dissociation constant at your actual temperature and ionic strength, and the total concentrations of all components. The output will give you free fluoride, free metal ion, and the actual pH or H0 of the solution. Relying on textbook pKa values without this background will give you results that look wrong until you figure out why. I also learned the hard way that glass containers are not suitable for storing HF-rare earth solutions. The hydrofluoric acid etches glass and releases silicates into the solution, which then compete with the rare earth for fluoride binding. I once lost an entire batch of purified lanthanum fluoride precipitate to silica contamination because I stored the mother liquor in a glass beaker overnight. Switched to PTFE vessels after that. The lesson extends to any analytical work where you need to know the exact free fluoride concentration. If you need stability constant data to build your speciation model, the review by Martell and Smith in the NIST database is the most comprehensive source I have found. It covers lanthanide fluoride complexes from La through Lu with formation constants measured at controlled ionic strengths. The values vary between different research groups by up to an order of magnitude for some of the heavier lanthanides, so pick a consistent source and stick with it throughout your calculations.
Get the Full Details
