Irradiation Damage and Semiconducting Properties of \(CdF_2:Eu\)

Thermoluminescence data have been used to unravel the fluorescence spectrum of Eu3+ in the insulating CdF 2 host. Optical absorption and EPR measurements of semiconducting CdF 2: Eu are the bases of a model of this semiconductor.


I. INTRODUCTION
Cadmium fluorite has the well-known fluorite structure.The series of MeF2 crystals (Me= Cd, Ca, Sr, Ba) are easily doped with rare earths.The rareearth impurity, usually in the trivalent state, substitutes for the divalent cation.The charge compensation needed is usually provided by an interstitial Fatom.Kingsley and Prener 1 discovered that baking the CdF2 doped with some rare earths in Cd vapor results in transforming the colorless and insulating crystal into a colored semiconductor.During the process of baking of MeF 2 in the metal vapor, interstitial fluorine atoms are believed to diffuse to the surface, forming MeF2 with the metal vapor, and for each one molecule two electrons are liberated into the lattice.In the case of Me= Ca, Sr, Ba 2 the rare earth is converted to the divalent state.In CdF2, however, even at low temperatures,3 there is no evidence of kinds of Eu H sites.The converted sample exhibits two absorption bands in the near-infrared region, one is peaked at 2650 cm-I (0.33 eV) and the second at 1100 cm-1 (0.136 eV).The EPR measurements show that the electron is indeed trapped at the europium ion which results in a "normal" EPR spectrum of Eu2+.In addition, a broad signal is present at a certain angle 5 which indicates the existence of exchange interaction between the conducting electron and the paramagnetic Eu2+.This effect disappears at 77°K or lower temperature.On the basis of these experimental results we propose a model to explain the semiconducting properties of CdF 2 : Eu.

A. Irradiation Damage and Fluorescence of Insulating
CdF 2 :Eu most of the rare earth in the divalent state.The Unlike CaF2, SrF2, and BaF2, where the europium difference can most probably be attributed to the ion is embedded in the host mainly as Eu2+, in insulatgreater electron affinity of the Cd2+ compared to ing CdF2, only the trivalent state is found in our Ca 2 +, Sr2+, and Ba2+. 4 The electrons are located at crystals.The EPR spectrum of our samples at room cadmium sites and form a conduction band.It has temperature and liquid nitrogen temperature shows been shown by EPR and optical studies that 5 some of only a small amount of Mn2+ [see Fig. 1 (a)].X-ray the rare earths represent a "shallow trap" for the irradiation of the CdF2:Eu results in partial conversion conducting electron in the sense that the electron is of Eu H to Eu2+ and a trapped hole in the lattice free to move around the trivalent rare earth on the [see Fig. 1 (b)].Angular variation of the spectrum of cadmium sites without converting it to the divalent a crystal oriented in (100) plane shows that the local state.
symmetry is cubic or very close to cubic.The g factor In CdF2 :Eu, however, it has been shown 6 that the (g= 1.988) and crystal field parameters are identical converted crystals exhibit Eu2+ optical properties.to those reported by Glaser and Glist.7 The nature of The resistivity of the converted sample is of the order the hole which accompanies the Eu2+ ion has not of a few hundred ohm• centimeters compared to a been established.It is possible that VK 8 centers similar few ohm• centimeters for the shallow traps such as to the VK centers in CaF2, SrF2, and BaF2 exist. 9d, Sm, Tb, Dy, Ho, Tm.Resistivity measurements Since the g factor of these hole centers is also close to 2, show the energy gap of the europium-doped semi-but their intensity is much weaker than the intensity conducting CdF2 is 0.33 eV.6  of the paramagnetic divalent europium, it is not The purpose of the present work is to confirm that possible to confirm this assumption.The total angular Eu H in CdF2 represents a deep trap for the electron.momentum of ~ for Eu2+ and t for the VK centers is We have studied the insulating crystals by means of the reason for the difference in the intensity since the irradiation damage (thermoluminescence and EPR) spin magnetization is proportional to S (S + 1) .and fluorescence; the semiconducting crystal has been Heating the irradiated sample from 77°K up to studied in terms of optical absorption and EPR.270 0 K results in gradual disappearance of the EPR The fluorescence spectrum of Eu H shows two kinds spectrum of Eu2+ according to Fig. 2  previously described in detail for CaF2 host lO doped with various rare earths, is particularly useful since the spectroscopy of the Eu3+ ion is revealed.In Figs.
3-5 the high resolution spectrum reveals that the europium ion is in symmetry sites lower than axial.The spectroscopy of the Eu3+ ion in various crystal field environments is well known. 11 -13The thermoluminescence of Eu3+ below 1700K is very different from the thermoluminescence above 170 0 K (see Figs. 3, 4 and Table I).It is easy to identify these thermoluminescence spectra with the aDo to 7Fl transition (see Refs. 5-7).Since oDD is a single level and cannot be split, the group of three lines belongs to the splitting of 7Fl in the crystal field.The Hamiltonian of this level with spin S= 1 is given by: The constants A and ."describe the crystal field splitting.Formally the Hamiltonian has the same form as the one conventionally used to describe the pure nuclear quadrupole interaction,14 ."describes the discrepancy from axial symmetry.For ." = 0 there are two degenerate energy levels corresponding to the states I ± 1) and single state I 0).This is the case of axial symmetry.For .,,~O the three levels are split.The coefficient."changes from 0 to 1 in the Hamiltonian (1), where the z direction is defined as the direction of the largest component of the second derivative of the electric potential which interacts with the quadrupole moment of the electronic wavefunction of  Eu3+ in the TFI group of levels:

3
(3) where EI and E2 are the energy gaps according to Fig. 6.
The fluorescence data shown in Fig. 5 of CaF2: Eu are different from Kingsley and Prener l3 (see Table XIII).They reported data in which it is clear that the symmetry of the Eu3+ ion is very close to axial symmetry.The reason for the discrepancy is probably that the charge compensation, which depends on the crystal growing conditions, is quite different.We are able to detect neatly the following groups of lines: f>Do-7FI, f>DO-7F2, f>Do-7F3, and r.DO-7F4 (see Figs. 6 and 7).The results are tabulated in Table II and shown in Fig. 7.The two spectra of thermoluminescence can be seen in the fluorescence spectrum [Fig.6(b)].
The advantages of thermoluminescence in spectroscopy of ions in the crystal field are obvious.For comparison we also examine the fluorescence spectrum of double doped CdF2 (with Eu and Na): see Table II.In this case we obtained only the f>Do-7 FI transition which was very intense.This spectrum also consists of three lines and indicates symmetry sites lower than axial.The parameters of A and TJ [see Eq.

B. Optical Absorption and EPR of Semiconducting
CdF 2 :Eu If CdF 2 :Eu is baked in Cd vapor the insulating CdF 2 (r,,-,10 9 g'cm) is converted to a semiconductor (r .. ....,SOOg•cm).Since Eu3+ is a much deeper trap than the shallow traps such as Cd, Sm, Y, Vb, the resistivity is two or three orders of magnitude larger for this semiconductor.Figure 8 shows the nearinfrared absorption spectrum of converted CdF2: Eu.Two broad absorption peaks are found peaked at 2650 cm-I (0.33 eV) and 1100 cm-I (0.136 eV).These peaks do not exist in the unconverted sample.The absorption in the visible region reveals the ::J 5868 existence of Eu2+.The 337-mJ.! peak is identical with previously reported results. 6he converted CdF2 :Eu is studied by EPR technique.The spectrum at 77°K confirms the existence of Eu 2 +.The g factor, the crystal field parameters and the hyperfine structure are identical to those of the x-irradiated sample.We studied the EPR spectrum at room temperature, 77, and 4.2°K.The intensity of Eu2+ spectrum at 77°K is 4.5 times greater than the spectrum at room temperature.In both cases the

5968A
intensity of the spectrum of Eu2+ is compared to the intensity of Mn2+ at low power to avoid saturation effects.At room temperature in addition to the normal Eu2+ spectrum a broad line is observed.This broad line is observed only when the magnetic field is at a certain angle, so th~t the seven Zeeman split energy gaps of the 8S7/2 ground state are approximately equalY' As in the case of Gd3+ in CdF 2 , when converted, this eight resonance is interpreted as an average line due to the exchange interaction between the con-  --------------------- duction electron and Eu2-!-.This broad signal disappears when the sample is cooled down to 77°K and the intensity of Eu2-!-is increased.No difference is observed between the ElI2-!-spectrum at 4.2 and n°K.
The width of the "eighth resonance" line at room temperature close to the "magic angle" (when the energy levels are equally spaced) is changing according to the following table:  The crystal is rotated in the (100) plane relative to the magnetic field.
It is difficult in the case of europium to estimate the exchange interaction because of the anisotropic hyperfine interaction with the two nuclear isotopes (mEu and 153Eu).

III. EXPERIMENTAL TECHNIQUE
The CdF 2 single crystals used in this work were grown in graphite crucibles by use of Brodgener-Stockbarger technique.16 -19 The CdF2 powder was obtained from the General Electric Chemical Products Plant.The pure powder was first purified by repeated growth of single crystals.After each growth the top end was cut off and the remainder was ground up to be used as initial powder for the next growth.After purification the dopant was added to the pure powder and thus regrown.As grown, the samples were highly insulating.A Pyrex tube was cleaned and baked under vacuum at 500°C for 2 h.The doped sample and a piece of cadmium metal were placed in the tube so FIG. 6. Scheme of the crystal field split of 7 FJ level.

-----r---r-------eo
that they were physically separated.The tube was sealed and placed in a furnace.The crystal was heated at about 600°C for half an hour and removed from the furnace while hot.Electron spin resonance data were taken with Varian V-4500 35-Gc/sec spectrometer.A special attachment has been built 2 to allow both x-ray irradiation and collection of EPR data, keeping the sample at 77°K. Figure 9 shows the attachment which consists of a quartz Dewar with two suprasil quartz windows, 0.020 in.thick each.X-ray exposures are made with a molybden x-ray tube powered by a General Electric supply operated at SO kV and 20 mAo The bottom part of the 35-Gc/sec cavity and the cold finger are supported by springs.The sample is at the level of the window during x-ray exposure.After the irradiation, the cavity and the wave guide with the iris is screwed up into the bottom part of the cavity.They are pushed down so the sample sits on the middle of the 8-in.pole of the Harvey-Wells Nuclear Corp. magnet.The sample temperature is measured by a copper-constantant thermocouple (made from 30gauge wire) with reference junction in icewater.The thermocouple junction is pressed to the bottom part of the cavity, as close to the sample as possible.The static magnetic field is measured by a free precession NMR signal with spin-echo apparatus.
The absorption measurements in the visible region are done in a Cary 14 spectrometer and the nearinfrared region in a Cary-White 90 spectrometer.A liquid nitrogen Dewar with KBr windows is used for this purpose.
The high-resolution spectrum of thermoluminescence as well as the fluorescence spectrum is taken with Jarrel-Ash 1-m spectrometer (model 78-420).The excitation of the fluorescence spectrum is done by llV mercury lamp (General Electric Co., Hendersonville, N. C.), model B-H6 (900 W).The dimensions of the sample used in the fluorescence and thermoluminescence are: lOX 10X6 mm; for EPR experiments-6X 1.6X 1.6 mm; the thickness of the sample used in optical absorption was about 1 mm.

IV. DISCUSSION
Irradiation damage (x-ray irradiation at nOK) is almost not manifested in pure CdF2 or CdF2 doped with "shallow traps" rare earth.When Eu ions are not present the thermoluminescence intensity is about three orders of magnitude less and is probably not connected with the specific rare earth.The thermoluminescen(:e of Eu H in CdF 2 is quite strong and it has the feature that is identical to the luminescence spectrum excited by uv source.This proves the assumption that Eu H is a deep electron trap in the CdF2 host.The shallow electron traps (rare earth) repel the electron mainly on the 5s shell of the cadmium.The EPR of Eu H due to the x-ray irradiation at 77°K (Fig. 1) in cubic or almost cubic sites proves that the electron excited from the valence fluorine band is excited into the cadmium conduction band, which lies at least 6 eV higher than the valence band, and trapped into the Eu H , thus converting it into Eu H .As shown in Fig. 2, the electron (or the hole center) is mobile and thermally activated at 170oK.As a result a recombination occurs 9 ,IO and one is able to observe the spectroscopy of Eu H .The experimental results show that there are two kinds of low symmetry sites for Eu H .It is logical to assume that interstitial fluorine,S or cadmium vacancies cause this low symmetry ; Interstitial fluorine in (210) position from the europium (elh symmetry) will split the three levels of 7 F I .
It is clear from Figs. 3, 4, and 7 that if the sample is more doped, the intensity of one kind of spectrum is increased relative to the intensity of the second kind.This suggests that this second kind of spectrum is related to a constant amount of some defect (Cd H vacancy, for instance), which does not depend on the amount of doping.It is striking that the Eu H is in cubic symmetry but the thermoluminescence spectrum is a low symmetry one.A possible explanation is that the level 85 7 /2 belongs to an 5-site ion where the crystal field splitting affects it only in higher orders, whereas the amount of splitting for the 7 FI is much greater (of the order of 100 cm-I ).We wish to compare this result with the zero field splitting of Gd H in CaF2 in the ground (857/2) and an excited state ( 6P 7/2).From the cubic EPR spectrum the zero field splitting is 0.149 cm-I ; however, the splitting of 6P 7 / 2 level from fluorescence measurement of Mokovski o is 56.6 cm-I .
On the basis of optical absorption and EPR data, we wish to propose the following model for the semiconducting CdF 2 :Eu; this model is consistent with the experimental results obtained for the insulating CdF 2 :Eu.The ground state of the conduction electron is the Eu H , which merely means that the low-lying energy level represents the europium impurity site (trap).The first excited state which is 0.136 e V above the ground state is 12 times degenerate; it represents the 12 cadmium nearest-neighbor sites at (110) type of direction from the europium impurity.At nOK the ground state (Eu H ) is four times more populated (Boltzmann factor) than at room temperature.This property is exhibited in the increase of the intensity of EPR of the Eu H at nOK relative to the intensity at room temperature.The increase in intensity is relative to the Mn H signal.The conduction band consists of more distant cadmium sites relative to the rare earth in 0.33 eV above the ground state.The rate of the reaction of aquation of the azidopentaaquochromium (III) ion has been measured at two different concentrations of hydrogen ion in the presence of a number of salts.The values of kl and ko of the equation were calculated and both show positive salt effects and obey the Olson-Simonson rule.Calculations performed using the Mayer theory in the approximation "DHLL+B2" have shown that at high dilution both these effects can be rationalized using appropriate values of the distances of closest approach.The possibility of using this form of the Mayer theory for extrapolation purposes is suggested.The use of different distances of closest approach for different pairs of ions does not lead to any inconsistency.Kinetic salt effects on reactions involving ions of the same sign cannot, in general, be satisfactorily interpreted on the basis of the simple Br¢nsted-Debye theory.lFor reactions between anions the rate does not depend upon the ionic strength, i, but upon the nature and concentration of the cations, whereas for reactions between cations it is the concentration of the anions which is importan t. 2 This can not be ra tionalized in terms of the Br¢nsted-Debye theory, even with the assumption of the formation of reactive ion pairs. 3Moreover, salt effects are sometimes observed also for unimolecular reactions or for reactions between an ion and a neutral molecule,4 and these cannot be interpreted on the basis of the simple Br¢nsted-Debye theory but require the assumption of a specific ionic interaction. 5Quantitative studies on this subject are rather scarce. 6Moreover, the instances of salt effects in reactions between cations which are found in the literature are usually complicated by the formation of complexes of the reactants with the added salts and are not suitable for a study in terms of an electrostatic theory.Therefore we considered it useful to measure the salt effects in the aquation of the azidopentaaquochromium ion which offers interesting opportunities.In fact this reaction has been studied by Swaddle and King 7 who found that the rate in a wide range of H+ concentration and at an ionic strength 1= 1.0M, could be expressed by the equation where [CrJ is the concentration of the complex.The first rate constant, k l , corresponds to a reaction between two cations, whereas the second, ko, corresponds to a unimolecular reaction, or, possibly, to a reaction between a cation and a neutral solvent molecule.Under suitable conditions it is possible to neglect the contributions of the third and fourth terms.The Mayer theory,S even in a very crude form, appears to be able

Kinetic Salt Effects on the Aquation Reaction of the Azidopentaaquochromium (III) Ion and Predictions of the Mayer Theory
FIG. 1.(a) EPR spectrum around g=2 of the irradiated CdF2 :Eu(O.02%).The magnetic field is oriented along the (100) direction.The seven lines of Eu+ + are superposed to the natural impurity of Mn+ + in the crystal.(b) Nonirradiated crystal: only Mn+ + signal is present.

2 FIG. 2 .
FIG. 2. (a) The solid line represents the total intensity of the thermoluminescence as a function of the temperature for CdF2:Eu(O.02%);the dotted line represents the thermoluminescence of CdF2 :Eu(O.07%)as a function of the temperature.(b) The EPR intensity of x-irradiated CdF2 :Eu(O.02%) is plotted vs temperature.
FIG. 3. Thermoluminescence spectra of CdF2 :Eu(O.02%)as a function of the temperature; trace a as the lowest temperature range; b, c, and dare spectra at intermediate temperatures; e the highest temperature range.

TABLE I .
Thermoluminescence

TABLE II .
Fluorescence data of CdF,:Eu at 77°K.