Introduction
Earlier work (Couture and Le Paillier-Malecot, 1975; Couture, 1977) has shown that some frozen dilute solutions of rare-earth salts exhibit optical absorption line spectra which reveal a splitting of the electronic energy levels of the rare-earth ions such as appears when such ions are located in a crystalline field.
High-pressure experiments at low temperatures (Jean-Louis and Couture, in press) have shown that this phenomenon appears only under conditions of temperature and pressure where the hexagonal Ih ice phase should be present (and not where phases II, III, IX, and Ic should appear).
We present here the results obtained from near-infrared absorption spectra of a frozen aqueous solution of ytterbium chloride YbC13 at atmospheric pressure and low temperatures. This study may shed light on the reason why the attempt to use frozen ytterbium chloride solutions as a substance for a proton-spin refrigerator was unsuccessful (Lichti and Stapleton, 1973).
Experimental Techniques
The ytterbium chloride solution was prepared by dissolving 86.29 mg of Yb203 in 10 cm3 of an aqueous solution of 1.2N hydrochloric acid; the result was an aqueous solution whose molarity in YbCl3, was 0.044 mol 1-1. The frozen solution was obtained by lowering the temperature at a rate of the order of 1 K min-1. The cell containing the frozen sample was then immersed in liquid nitrogen, hydrogen, or helium at temperatures of 77 K, 20 K, or 4.2 K. The experimental arrangement for absorption spectroscopy has been described by Couture and Le Paillier-Malecot (1975). The spectra have been recorded on Kodak IQ near-infrared-sensitive photographic plates.
Experimental Results
The absorption spectra of frozen dilute solutions of ytterbium chloride exhibit only two broad lines (Fig. 1). Table I gives the wave numbers σ, the line widths w, and the relative intensities I, of the lines at temperatures of 77 K, 20 K, and 4.2 K. The wave numbers are measured with an estimated error of ±5 cm-1 (1 cm-1 corresponds to a frequency of 30 GHz). Between 77 K and 4.2 K the spectra change very little with temperature.
Frozen dilute solution of ytterbium bromide YbBr3, ytterbium bromate Yb(Br03)3, or ytterbium nitrate Yb(NO3)3 at 77 K exhibit, in the same spectral region, only a very broad band whose width is about 300 cm-1, similar to the band presented by aqueous liquid solutions of ytterbium salts (Hoogschagen, 1946; Carnall and others, 1968).
Fig. 1. Absorption spectrum of frozen YbCl3 aqueous solution at T = 20 K, and (b) reference xenon emission spectrum.
Table 1. Parameters of Absorption Lines
Discussion
The ytterbium ion Yb3+ has a configuration 4f13; as one electron is lacking to fill the shell there is only one term 2F which gives two levels 2F7/2 and 2F5/2 by L-S coupling. For the free ion these two levels have degeneracies respectively of 8 and 6 and they are 10 214 cm-1 apart (Kaufman and Sugar, 1976). When ytterbium ions occupy definite sites in crystals each level is split into Stark components by the crystalline electric field and optical transitions may occur between those Stark levels. Figure 2 shows a possible schematic level diagram of the ytterbium ion Yb3+ in free space, in a cubic field, and in a field of lower symmetry (the degeneracies of the levels are indicated in parentheses).
The fact that in the absorption spectra that we have obtained there is no line of which the intensity varies with temperature in the range 4 K-77 K indicates from Boltzmann's law that there is no level in the vicinity of the lowest sub-level, at least in a region of 50 cm-1. As a consequence we conclude that the two lines arise from transitions starting from the lowest
sub-level. The separation of the two lines must then be interpreted as a Stark splitting of the excited level. The magnitude of that splitting (≈ 60 cm-1) is of the same order as the splittings occurring for rare-earth ions located in crystalline matrices.
The fact that only two lines are observed could be explained by a cubic or nearly cubic symmetry (cf. Fig. 2). But, as Dieke and Crosswhite (1956) have pointed out, such two-line spectra may also occur in a field of hexagonal symmetry; they thus explained the appearance of only two strong lines (at 10 282 and 10 297 cm-1) in the spectrum of an Yb doped YCl3 · 6H20 crystal for which the Zeeman effect had an approximately axial symmetry.
The width of the lines (Table I) varies very little with temperature and remains large even at 4.2 K. It is interesting to note that this width (≈ 50 cm-1) is characteristic of the spectra of frozen solutions, as it is about ten to one hundred times larger than the width of lines of rare-earth ions in crystals at low temperature and about ten times less than for liquid solutions.
The lack of success in performing proton-spin alignment in frozen aqueous solutions of ytterbium chloride (Lichti and Stapleton, 1973) even though there exists a Stark splitting, may be a consequence of the width of the levels, if the width of the fundamental Stark level is homogeneous and is of the same order of magnitude as that of the excited levels.
The results obtained from this work and related work done with other rare earths do not at present permit us to determine the exact nature of the phase responsible of the observed spectra, but we can limit the possibilities and infer what is most probable.
The number of lines for neodymium chloride solutions (Couture and Le Paillier-Malécot, 1975) indicate that there exists only one type of site for the rare-earth ions. Also the splitting of the excited 4F3/2 level of neodymium completely rules out the possibility of cubic symmetry at the site of Nd3+ ions, although it is consistent with any site symmetry lower than cubic.
Schematic energy levels of an ytterbium Yb3+ ion in various surroundings.
The absorption spectra of frozen ytterbium chloride aqueous solution strongly differs from that of ytterbium chloride hydrated crystal YbCl3 · 6H20 both in frequencies and widths of the lines. The same thing occurs with frozen solutions of other salts. The absorbing phase is then not one of the known crystalline hydrated chlorides.
Frozen aqueous solutions of neodymium chloride with and without HCl give the same spectra. The possibility of having a crystal containing hydrogen chloride has thus to be ruled out.
The Stark splittings observed in the spectra of frozen aqueous solutions of ytterbium chloride can then be accounted for by the following two possibilities. A vitreous solution of YbCl3 in water may present a quasi-crystalline order in the immediate vicinity of the rare-earth ion, with a quasi-symmetry lower than cubic; the widths of the lines could in this case be accounted for by inhomogeneities of the field. Hexagonal Ih ice may have definite sites occupied by ytterbium ions, thus forming a crystalline solid solution; this could explain easily the high-pressure observations ( Jean-Louis and Couture, in press), but the reason for the width of the lines should be found, possibly proton disorder in ice is responsible.
Conclusion
The absorption spectrum of frozen aqueous solutions of ytterbium chloride is essentially independent of temperature between 4 K and 77 K and shows characteristic line widths greater than the line widths of rare-earth ions in crystals at low temperature and with similar concentration, and less than the widths for liquid solutions.
A splitting of 60 cm-1 in the 2F7/2 2F 5/2 transition is observed, identified as a splitting of the excited 2F512 state and attributed to the Stark effect arising from a local quasi-crystalline or a crystalline environment of cubic or lower than cubic symmetry; however, related experiments on frozen aqueous solutions of neodymium chloride show that the symmetry of the surroundings of neodymium is lower than cubic.
Acknowledgements
We want to thank Dr P. Caro for kindly providing the ytterbium chloride solution and C. Milet-Vialle for her participation in the experiments. We also express our thanks to Dr H. Z. Cummins for a very fruitful discussion of this text.
Discussion
S. A. RICE: Is it not likely that when you freeze the solution of YbCl3 the Yb3+ ion is incorporated
along with its water of hydration ? It is known that the rate of exchange of water in the hydration shell of a trivalent ion is very small ; a half-life of a day is not unusual. If this conjecture is correct your experiment does no t probe the environment of ice Ih, but of a hydrated ion crammed into a nice Ih lattice, with distortion of the hydra tion sphere by the surrounding lattice. Perhaps the hydra tion sphere can conform to the lattice, perhaps not. It is likely that the line widths for electronic transitions will b e smaller than in a fluid solution (as in the Shpolski effect) , and perhaps a t sufficiently large pressure the original shell is distorted and even destroyed. It is also possible that your solid is heterogeneous and the Yb3+ is not incorporated substitutionally. Some check on the structure will be needed to check your suggested interpretation.
L. COUTURE: T he presence of water of hydra tion cannot be excluded. Those spectra show that Yb3+ ions are in definite sites, but not what sort of sites.
A. P. MACKENZIE: It seems to me that many, possibly all your findings are best explained on the basis of a freezing out of pure ice Ih from an initially dilute (c. 1% w/w ) solution, and the generation of a concentrated amorphous phase might reach 50 to 75% w/w YbCl3 and fail to yield a eutectic, persisting instead in a truly amorphous condition. Experience with the freezing of di- and trivalent cation salt solutions in a number of laboratories suggests the frequent reluctance of the salts to crystallize in practice. Such a physical mixture of ice and concentrated amorphous aqueous salt would satisfy your requirement that each Yb3+ ion occupy an essentially like site. Each Yb3+ ion could maintain some water of hydration. Clions would, similarly, persist in the same amorphous phase. What do you think of the possibility?
COUTURE: I do not exclude t his possibility. But why should the spectra disappear and reappear with the Ih ice phase? And phases with a high concentration in rare-earth ions do not show any fluorescence.
