摘要:

J. MATER. CHEM., 1991, 1(3), 415-421 Computer-simulation Study of Alkali-metal Insertion into a-lJ30, Richard G. J. Ball*"nb and Peter G. Dickens" a Inorganic Chemistry Laboratory, South Parks Road, Oxford OX1 3QR, UK Materials and Chemistry Division, AEA Technology, Harwell Laboratory, Oxfordshire OX1 1 ORA, UK Atomistic simulation techniques have been used to study the insertion of lithium and sodium into a-U,O,. Calculations for isolated guest ions predict that lithium will occupy five-co-ordinate trigonal bipyramidal sites whereas sodium will occupy nine-co-ordinate sites. The modelling of the stoichiometric phases MU,08 (M=Li, Na) reinforces these predictions. Calculations of ion migration are presented; lithium is demonstrated to be fairly mobile within the lattice whereas the diffusion of sodium is much more difficult. Keywords: Computer simulation; Uranium oxide; Intercalation; Defects; Interatomic potential 1.Introduction Over the last 40 years there has been considerable interest in the chemistry of the oxides of uranium, owing mainly to their importance in the nuclear industry. One area that has been under recent investigation is concerned with the insertion compounds of these Such compounds have a general formula A,UO, where A is an electropositive species, such as hydrogen or an alkali metal, which has been inserted into the host oxide UO, with minimal structural rearrange- ment. For such reactions to occur it is necessary for the host lattice to have suitable sites for the inserted species to occupy.In addition, it must be possible for the lattice to readily accommodate electrons from the guest, which is generally present as a cation. In both respects a-U308 is well suited to this type of reaction since it possesses a rather open pillared- layer structure, and uranium has a range of available oxidation states. In a recent paper we developed an atomistic model for simulating the structural and defect properties of a-U308e7 In the present work we extend this model to study the insertion of lithium and sodium into U308. Through these calculations we are able to investigate the possible sites occupied by the alkali metal and the various mechanisms for their migration through the host matrix. Similar computer- simulation studies have been reported for analogous insertion compounds of the transition-metal oxides Fe3048 and Wo3.' 2.Experimental At ambient temperatures a-U308 adopts a pseudo-hexagonal orthorhombic structure which consists of layers of edge- sharing U05 pentagons connected by -U-0-U-0- chains, which run perpendicular to the layers." At higher temperatures this polymorph undergoes a transition to the closely related hexagonal structure a'-U3O8.'' However, the difference between the a and a' forms of U308 is minimal and both are illustrated in Fig. 1, where the hexagonal unit cell of the a' form is shown by the dashed line. The pillared- layer nature of the structure of ar-U308 suggests that the intercalation of alkali metals between the layers should be reasonably facile.Indeed, both lithium and sodium insertion compounds of this oxide have been synthesized and the materials have been characterized in terms of their thermo- dynamic, structural and transport proper tie^.^* 9 '7' The U308structure is an archetype for a number of ternary -0-U-0-chain 0oxygen Fig. 1 The structure of cl-U308(z=O plane) layer structures consisting of planes of edge-sharing UO, and MO, polyhedra connected by metal-oxygen chains. Related structures are also to be found amongst non-uranium com- pounds; LiW309F,17'18 for example, consists of layers which are packed in a similar fashion to UNb3OI0. The insertion of lithium into UV05, UTi05 and LiW309F has recently been Thus, the study of alkali-metal inser- tion into U308 will be valuable for understanding insertion processes in a wide range of materials.2.1 Lithium Insertion The insertion of lithium into a-U308 has been studied by uranium oxide phases such as UVo5,l3 UTiOS, CUU~O~~'~ Dickens et u1.293*5Electrochemical measurements indicate that and UNb3010,15~16 all of which are thought to have pillared- single-phase regions for LixU308 exist for 0.78 <x <0.88 and 416 3.6 3.2 2a B2.L 2.0 1 . 2 1 - - I I I I 1 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2-1 x in Li,U,O, Fig. 2 Voltage versus composition curve for lithium insertion into a-U308 for 1.20 <x <1.26. A typical discharge curve is shown in Fig. 2, and consists of a number of plateaux, at which two phases coexist, separated by monophasic regions, where the emf changes smoothly with x.A pure phase of composition Lio.9U308 may be prepared chemically by lithiation with anhydrous LiI. X-Ray powder diffraction measurements on this phase show that there is very little structural rearrange- ment of a-U308 on insertion. The refined lattice parameters of Lio.9U308 are given in Table 1; the increase in the c lattice parameter is consistent with the accommodation of the lithium ions in interlayer sites. However, the precise location of the lithium ions cannot be determined by X-ray diffraction since the pattern is dominated by scattering from the heavy uranium ions. Fortunately, this problem does not arise with neutron diffraction and this technique has been used in a recent study to determine the intercalation sites in Lio.9U308.12 The enthalpy of insertion of lithium into a-U308 to form Lio.9U308 has been determined by solution calorimetry to be -298k4 kJ per mole of Li.’ This value is close to the Gibbs energy of insertion as obtained from electrochemical studies.’ Thermochemical measurements have also established that Lio.9U308 is stable with respect to disproportionation into Li2U3Ol0, U308 and U02.The thermodynamic stability is also supported by Kemmler-Sack and Riidorff” who have prepared materials (Li20);UO2 *U03 in the range 0.05 <x <0.45 (x =0.33 corresponding to the composition LiU308) at a temperature of 800 “C. These phases have lattice parameters that are close to those of either a- or a’-U308, depending on the value of x.The high-temperature phase discovered by Kemmler-Sack and Riidorff is likely to be closely related to the ambient temperature phase Lio.9U308 characterized by Dickens et The introduction of lithium ~1.~9~9’ ions into the a-U308 lattice and the consequent stabilization of the lattice has also been observed by Tsvigunov and Kuznetsov.22 Crude diffusion coefficients for lithium in Lio.9U308 have been measured by current pulse method^.^ The values Table 1 Experimental cell parameters of insertion compounds of a-U308 compound alA blA 4 reference a-U,o8 6.716 1 1.960 4.147 10 Li0.88u308 6.727 11.953 4.188 5 Na1.29U308 6.875 12.630 4.267 4 J.MATER. CHEM., 1991, VOL. 1 observed indicate a rapid lithium-ion transport, which is characteristic of the insertion compounds of layered materials. 2.2 Sodium Insertion The sodium insertion compound Na1.3U308 has been pre- pared4 and the X-ray diffraction pattern showed, as might be expected, that the increase in lattice parameters from the parent oxide is larger than that observed for lithium insertion (see Table 1). No thermochemical measurements have so far been carried out on Na,U308. 3. Insertion Calculations In the previous section it was seen that lithium and sodium can be inserted into a-U308 under ambient temperature conditions. Theoretical calculations can provide a valuable insight into these processes by examining the sites occupied by the guest species and the mechanisms for their migration through the host matrix.3.1 Theoretical Methodology The lattice simulation techniques employed in this work were described in detail in a previous paper.’ The defect calculations are based on a generalized Mott-Littleton procedure which divides the lattice into two regions: an inner region I, which incorporates the defect, and an outer region 11. In region I, the interactions between ions, as defined through interatomic pair potentials, are calculated explicitly, whereas in region I1 forces are calculated using a continuum model. Ions in region I are relaxed to zero force using a Newton-Raphson minimiz- ation procedure. An essential part of the procedure is the specification of the interatomic potentials.In this work short- range pairwise interactions are combined with long-range Coulombic interactions. The short-range repulsive interac- tions were calculated as a function of the interionic distance, r, by means of an electron-gas method.23 Each set of short- range interactions were then fitted to a Born-Mayer potential form: V”.‘*(r)=A exp (-r/p) to obtain the parameters A and p. The many-electron disper- sion interactions not accounted for in the electron-gas method were estimated using the Slater-Kirkwood formula24 and incorporated into the potential function as a -C/r6 term. The calculated interatomic potential parameters used in this study are given in Table 2. Alkali-metal-oxygen interatomic potentials have also been calculated by a similar method by Le~is,~’and Freeman and Catlow26 who obtained parameters which are close to the ones used in this work.Any slight differences between the two sets of potentials are attributable to the different oxide-ion densities used in the two studies. Ion polarization effects are taken into account by means of the shell This treats an ion as a massless shell of charge Y coupled to the core by means of a harmonic spring of force constant k. Shell-model parameters for the alkali metals were estimated from the free-ion polarizabilities, a, by assuming a shell charge of 1.0 atomic units and using the relationship: to obtain the force constants. The shell-model parameters used in this work are given in Table 3.Shell-model parameters for alkali-metal cations in the alkali-metal halides have also .~~been derived by Catlow et ~2 by empirical fitting techniques. J. MATER. CHEM., 1991, VOL. 1 Table 2 Potential parameters used in this study u5.33+ -ox 3023.3 1 0.33889 48.2 1 u5.33+ -0P 2963.89 0.34 165 48.2 1 u5+-ox 2866.86 0.34280 50.56 U~+-OP 2899.06 0.34184 50.56 ox-ox 154.30 0.43456 35.30 ox-OP 149.47 0.43842 35.30 OP-OP 144.78 0.44141 35.30 u5.33+ -u5.33+ 3 1002.05 0.24421 67.97 US+ -u5+ 27406.37 0.25 128 74.59 u5.33+-Li+ 3427.57 0.2 160 1 0.97 US+-Li+ 3282.06 0.21993 1.01 Li+ -OX 308.07 0.34619 0.63 Li+ -0P 322.80 0.34286 0.63U5.33+-Naf 11412.60 0.22198 3.86 U5+-Na+ 10452.24 0.22675 4.04 Na+-OX 876.08 0.33574 2.66 Na+ -0P 894.13 0.3 3465 2.66 NB OX denotes chain oxygens [0(1) and 0(2)] and OP denotes oxygens in the planes [0(3), O(4) and 0(5)].Table 3 Shell model and polarizability parameters u5.33+ 6.68 21 1.41 10.67 3.04 us+ 6.35 181.40 11.0 3.20 O2- -3.00 49.49 4.5 2.62 Li + I .oo 506.1 1 1.o 0.03" Na+ 1.oo 96.98 1.o 0.15" " Ref. 35. These alternative parameters were not used in the present work since, for lithium and sodium, they somewhat over- estimate the ionic polarizability which will be manifested in the C6van der Waals term of the interatomic potential. It should be noted, however, that a few test calculations were performed with the shell parameters of Catlow et al.but the results differed little from those presented in this paper. The model for U308 employs an average uranium-ion oxidation state of 53 to reproduce the nearly identical environ- ments observed around each uranium-ion site. The interac- tions calculated between isolated intercalated ions and the lattice therefore represent an average over the uranium sublat- tice. This is a valid approximation so long as the residence time of the guest ion in any one position is greater than the frequency of the charge transfer which is presumed to occur between the uranium ions. This is likely to be true of the guest ions when they are within their equilibrium sites but may not be true when they are migrating between sites. We therefore expect the calculation of structural properties to be more reliable than the calculation of the activation energies for migration.3.2 Intercalation Sites In the U308 structure we can identify two categories of site that may be occupied by the guest ions. With reference to Fig. 1, if the alkali metal sits directly between two oxygen ions in adjacent layers then it will occupy a distorted five-co- ordinate trigonal bipyramidal site as in Fig. 3. There is one such site above each of the three crystallographically distinct oxygen ions in the planes. The alternative site is the nine-co- ordinate position illustrated in Fig. 4. In this site the interca- lated ion sits centrally above the triangle of oxygen ions defined by O(4)-O(4)-O(3). Using the Madelung potentials calculated at the various intercalation sites and reported in 0 Po @ interstitial alkali metal Fig.3 Trigonal bipyramidal interstitial site uranium int erst it ial alkali metal Fig. 4 Nine-co-ordinate interstitial site Table4, it is clear that the Madelung energy favours the occupation of the trigonal bipyramidal sites by the intercalated cations. However, the Madelung energy does not provide a complete description of the processes involved in incorporat- ing the interstitial ion; we must also consider the effect of relaxation of the lattice around the guest ion. In Table 5 we present the results of calculations of the energy to bring an alkali-metal ion from infinity and place it at an interstitial site in the crystal.These results show that the smaller lithium ions will preferentially occupy the trigonal bipyramidal sites (which have a more favourable Madelung potential) rather than the nine-co-ordinate position. However, since the ener- gies for a lithium ion at each of the trigonal bipyramidal sites are rather similar, we predict that there will be no strong preference for the occupation of any particular trigonal bipyr- amidal site. This suggestion is supported by a recent neutron Table 4 Madelung potentials at the various intercalation sites site V,leV five-co-ordinate site apical oxygen O(3) -5.48 O(4) -5.45 O(5) -7.18 nine-co-ordinate site -2.86 Table 5 Formation energies of isolated alkali-metal ions" intercalation site calculated energies/eV lithium insertion five-co-ordinate site apical oxygen O(3) (44) O(5)nine-co-ordinate site -7.61 -7.63 -7.63 -7.36 sodium insertion five-co-ordinate site apical oxygen O(3) O(4)00)nine-co-ordinate site -3.81 -3.83 -3.21 -5.26 " Calculations performed with a region I size of ca.425 ions. diffraction study of Lio.gU308,12 which shows that if the insertion is carried out at room temperature, the lithium ions are randomly distributed over the trigonal bipyramidal sites. At higher temperatures, where there are fewer kinetic con- straints, the lithium ions were observed to occupy only the trigonal bipyramidal sites above the O(5)oxygen ions. This observation is also consistent with the calculated relative preference amongst the sites (see Table 5). In contrast to the results for lithium, the results in Table 5 indicate that isolated sodium ions would favour occupation of the nine-co-ordinate sites. This site is more able to accommo- date the larger sodium ion than the trigonal bipyramidal sites.We must be aware that calculations based on isolated intercalated ions may not be comparable with experimental measurements performed on phases containing a substantial concentration of guest ions. The preference for one ordering pattern over another, for example, could lead to different sites being occupied from those predicted above. It is therefore prudent to carry out additional calculations on the possible ordered phases of MU308 and to compare these with the available experimental observations. 3.3 Modelling of MU308 In MU308 we assume that all of the uranium ions have an equal oxidation state of I.: The structures of these phases can then be modelled by taking an initial structure, consisting of an unperturbed U308 framework with the alkali metals in the appropriate sites and with all uranium ions having a charge of 5+, and relaxing the configuration to a minimum lattice energy.This minimization occurs at constant pressure, allowing the alkali-metal insertion to cause a change in both the U308 cell volume and the lattice vectors. It should be noted that this procedure will not sample structures that are beyond an activation energy barrier from the initial structure.However, for our purposes this should not be a problem since we are mainly interested in topotactic processes, which involve relatively small perturbations to the U308 structure. For both LiU308 and NaU308, individual calculations were performed with the alkali metal ordered on each of the four sites. For LiU308 the lattice energies corresponding to the three structures in which the lithium ions order on the three sets of trigonal bipyramidal sites were very similar and significantly more negative than the structure in which the lithium ions occupy the nine-co-ordinate sites. This is in agreement with the conclusions of the previous section. The cell parameters and lattice energy for LiU308 with the lithium ions on the trigonal bipyramidal sites above the O(5)oxygen ions are given in Tables 6 and 7, respectively. Comparing the J.MATER. CHEM., 1991, VOL. 1 Table 6 Calculated cell parameters of insertion compounds of a-U308 a-U,O, 6.96 12.06 4.14 LiU,O, 7.10 12.19 4.25 NaU,O, 7.05 12.21 4.27 Table 7 Calculated lattice energies of insertion compounds of a-U30, compound HJeV a-U,O, -507.97' LiU,O, -465.74 NaU,O, -463.67 " Ref. 7. calculated cell parameters for LiU308 with the experimental cell parameters for Lio.88U308 given in Table 1, we predict a similar expansion for both the a and the c parameters. For NaU308 it was found that the lowest-energy configur- ations involved significant deviations from the U308 structure. This is contrary to experimental observations and the topotac- tic nature of the insertion reaction.For the cases in which the U308 structure is essentially preserved, the minimum lattice energy occurs for the structure having the sodium ions in the nine-co-ordinate sites. Furthermore, for the calculation with the sodium ions initially on the trigonal bipyramidal sites above the O(4) oxygen ions, all the sodium ions relaxed into the nine-co-ordiante sites. These results suggest that the sodium insertion compounds of U308 with compositions around NaU308 (which, at ambient temperatures, have been observed to occur with only a slight modification of the U308 structure) are metastable. However, in these metastable com- pounds we predict that the sodium ions will occupy the nine- co-ordinate sites and the calculated lattice parameters and lattice energy for this compound are given in Tables 6 and 7, respectively.The relaxed coordinates for the MU308 phases are reported in Table 8, along with those for a-U308 itself. It is clear that the internal structure of the U308 unit cell is essentially preserved during the insertion reaction. In most inorganic compounds the sum of the bond strengths around any particu- lar atom is expected to be within ca. 5% of the formal oxidation state of that atom.29 The bond strength can be calculated according to:30 Sij=(;) -N where sii is the bond strength between atoms i and j which are separated by rip N and ro are empirical parameters which are derived by fitting to a large range of structural data.The oxidation state, 6,for a particular atom is therefore estimated by summing the bond strengths over the nearest neighbours to the atom: This method can be used to check the structure of a particular compo~nd~~.~~or to assign oxidation states to the ions.33 It has been shown that the parameters ro and N vary little with oxidation and since few Uv-containing oxide struc- tures have been investigated, the values used here are those determined for the Uv' ion, rather than for the Uv ion. The oxidation states for the metal ions in U308, LiU308 and NaU308 calculated by this method are given in Table 9. For U308 itself the oxidation states are ca. 5% lower than the ones estimated by the same method but using the experimental J.MATER. CHEM., 1991, VOL. 1 Table 8 Comparison of calculated structures of a-U308 and MU308 (M =Li, Na) (space group C2rnm) a-U308 LiU308 NaU308 A, b= 12.21 A, ~=4.27A)(a=6.96 A, b= 12.06 A, ~=4.14A) (~=7.10A, b= 12.19A, ~=4.25A) (~7.05 atom symmetry xla Ylb zlc xla Ylb zlc xla Ylb ZIC 2a 0.961 0.000 0.000 0.961 0.000 0.000 0.96 1 0.000 0.000 4d -0.017 0.326 0.000 -.0.053 0.327 0.000 0.0 12 0.3 16 0.000 2b 0.934 0.000 0.500 0.925 0.000 0.500 0.97I 0.000 0.500 4e 0.996 0.3 13 0.500 0.992 0.316 0.500 1.007 0.321 0.500 2a 0.569 0.000 0.000 0.558 0.000 0.0o0 0.584 0.000 0.000 4d 0.179 0.130 0.000 0.173 0.131 O.OO0 0.200 0.128 0.000 4d 0.309 0.333 0.000 0.282 0.338 0.000 0.328 0.333 0.000 4e" ---0.274 0.326 0.500 2b ------0.328 0.000 0.500 " Half occupied.Table 9 Bond-strength calculations for U308 and MU308" U308 (calculated) around U( 1) around U(2) 2~ U(1)-0(1) 2.079 1.919 2 x U(2)-0(2) 2.078 1.923 2 x U(1)-0(4) 2.182 1.558 1 x U(2)-0(3) 2.182 0.779 2 x U(l)-0(5) 2.272 1.310 1x U(2)-0(4) 2.181 0.781 1 x U(1)-0(3) 2.728 0.298 1 x U(2)-0(4) 2.729 0.298 1 x U(2b--0(5) 2.271 0.656 ISij=5.09 LiU308 (calculated) around U( 1) around U(2) around Li 2 x U(1)-0(1) 2.139 1.698 2 x U(2)-0(2) 2.152 1.654 1 x Li-O(1) 2.380 0.108 2 x U(1)-0(4) 2.199 1.507 1 x U(2)-0(3) 2.254 0.678 2 x Li-0(2) 2.012 0.429 2 x U(1)-O(5) 2.352 1.129 1x U(2)-0(4) 2.883 0.235 2 x Li-O(5) 2.129 0.341 1 x U(1j-0(3) 2.870 0.240 1x U(2)-0(4) 2.018 1.090 1 x U(2)-O(5) 2.390 0.527 csij=4.57 1 x U(2)-O(5) 2.331 0.587 csij=4.77 ISij=0.88 NaUJ08 (calculated) around U( 1) around U(2) around Na 2 x U(1)-O(1) 2.138 1.701 2 x U(2)-0(2) 2.138 1.701 1 x Na-O(1) 2.516 0.152 2 x U( 1)-0(4) 2.298 1.247 1 x U(2)-0(3) 2.303 0.618 2 x Na-0(2) 2.523 0.301 2 x U( 1)-O(5) 2.241 1.389 1 x U(2)-0(4) 2.650 0.338 2 x Na-0(3) 2.797 0.193 1 x U( 1)-0(3) 2.657 0.334 1 x U(2)-0(4) 2.303 0.618 4 x Na-0(4) 2.797 0.386 1 x U(2)-O(5) 2.237 0.700 ISij== 1 x U(2)-O(5) 2.237 0.700 csij=4.68 CSij =1.03 Values of ro: U, 2.059 A; Li, 1.378 A; Na, 1.622 A.Values of N: U, 4.300; Li, 4.065; Na, 4.290. (See ref. 30) coordinate^.^ This reduction of the mean uranium oxidation and can be split up into several contributions: state reflects the expansion of the unit-cell volume, and therefore bond lengths, in going from the experimental to the calculated U308 structure.An expansion of this order is frequently observed when using electron gas potentials. Bear- ing this in mind, the changes in bond lengths, and therefore where HL(U3o8) and HL(MU3o8) are the lattice energies of oxidation states, on forming the MU308 phases are in good U308 and MU308 respectively, E,(M) is the enthalpy of agreement with the assumption of v and I oxidation states sublimation of M, Z,(M) is the first ionization energy of M for uranium and the alkali metal, respectively. The calculated and Ee,(U"+) is the gas-phase electron affinity of the U308structural changes are therefore in accordance with empirically uranium ions.However, calculating AinsH in this way isderived trends observed over a wide range of inorganic unreliable since we are taking the difference between twocompounds. large quantities [HL(U308)and HL(MU308)]. In our earlier Having calculated lattice energies for the phases MU308, paper' we noted that the experimental and calculated lattice we should now be in a position to calculate enthalpies of energies for U308 agreed to within 4%. This is a goodinsertion (AinsH).This enthalpy change corresponds to the agreement but, because of the large numbers involved, the process numerical difference amounts to some 20eV. Such problems M(s)+u308(s)--' MU308(S) are much less acute if we calculate instead the enthalpy change for the process Li+(g)+NaU,O,(s)-,Na+(g) +LiU,O,(s) In this case, there is some cancellation in the errors associated with the individual lattice-energy terms.From Table 7, the enthalpy change for this process is found to be -2.1 eV. This value is consistent with the value of -1.1 eV4 for the analogous process Li+(g)+NaUO,(s)-,Na+(g) +LiUO,(s) and reflects the greater difficulty in inserting the larger sodium ions into the host matrix. 3.4 Migration Mechanisms Having established the possible intercalation sites we now turn to the problem of alkali-metal migration through the host lattice. The insertion reaction depends very much on the availability of favourable diffusional routes for the guest ions in the host matrix. Since we predict that the lithium and sodium ions will occupy different sites in U308, their migration behaviour might also be quite different.For an alkali-metal ion occupying a trigonal bipyramidal site, an obvious mechanism for its migration is a direct jump to a neighbouring trigonal bipyramidal site. The saddle point for such a move will be approximately midway between the two sites, as shown in Fig. 5. From Fig. 1 we can see that there are four distinct jumps possible between neighbouring trigonal bipyramidal sites. If we label the sites by their apical oxygen ions then the four jumps are O(4) +0(4), 0(4)-+0(5), 0(5)+0(3) and 0(3)-,0(4). The calculated activation energy barriers for each of these jumps can be illustrated by plotting the energy profile for alkali-metal migration along the path 0(4)+0(4)+0(5)+0(3)-+0(4) as shown in Fig.6. The calcu- lated energy profiles for lithium and sodium migration are quite different. For lithium migration, the plot is simplified by the fact that the energies of the lithium ion at each site are very similar. The activation energies for lithium-ion migration are greatest for the 0(4)+0(4) and 0(3)+0(4) steps, reflecting the exposure to the uranium cation charge field that occurs along these direct routes. The energy barriers Q Q QI Q-1 J. MATER. CHEM., 1991, VOL. 1 -6 6(4) O(4) O(5) O(3) O(4) position of intercalated ion Fig. 6 Energy profile for alkali-metal migration between trigonal bipyramidal sites. (a)Sodium, (b)lithium for the 0(4)+0(4) and 0(3)+0(4) steps can, however, be reduced if the lithium ion migrates via the nine-co-ordinate position, an option that is not available for the 0(4)+0(5) and 0(5)+0(3) jumps.The energy profile for this route is shown in Fig. 7. It therefore appears from the calculations that the rate-determining activation energy for lithium migration will be ca. 0.7 eV. This activation energy is consist- ent with the high lithium mobility observed experimentally in U308e3 For sodium, the activation energies for direct migration between two trigonal bipyramidal sites (see Fig. 6) are lower than those for lithium owing to the release of steric forces when the sodium ion moves out of the trigonal bipyramidal position. The activation energies for these steps lie between 0.43 and 1.05 eV.The greatest barrier to sodium diffusion comes, however, from the activation energy of 1.52 eV for the -3.5 -,(iXTX\ 0.12-j ----x-\ X -4.0'-\ T-7/ 4.5 -1.52 "\ i -5.0 -% \ -6 -5.5 a a --6.0 @ migrating alkali metal I-8.0 IX trigonal bipyramidal O(3) 9-coordination site O(4) site position of intercalated ion Fig. 5 Saddle point for migration between two trigonal bipyramidal Fig. 7 Energy profile for alkali-metal migration between trigonal sites bipyramidal and nine-co-ordinate sites. (a)Sodium, (b)lithium J. MATER. CHEM., 1991,VOL. 1 421 migration of the ion out of the nine-co-ordiante site into a neighbouring trigonal bipyramidal site. Since the nine-co- ordinate sites are far apart, this step is always required and is therefore rate-determining.Consequently, the migration of isolated sodium ions in U308 is predicted to be much slower than that for lithium ions. 7 8 9 10 11 12 R. G. J. Ball and P. G. Dickens, J. Muter. Chem., 1991,1, 105. M. S.Islam and C. R.A. Catlow, J. Solid State Chem., 1988, 77, 180. J. C. Newton-Howes and A. N. Cormack, J. Solid State Chem., 1989, 79, 12. B. 0.Loopstra, Acta Crystallogr., 1964, 16, 651. B.0.Loopstra, J. Appl. Crystallogr., 1970,3, 94. P.G. Dickens and A.V. Powell, J. Solid State Chem., 1990,in 4. Conclusions 13 the press. R. G. J. Ball, P. G. Dickens, S. Patat and S. Hull, 1991,to be submitted. In this paper we have presented the results of a theoretical study of the insertion of alkali metals into ct-U308.Calcu- lations on the energetics of isolated alkali-metal ions suggest that lithium will preferentially occupy five-co-ordinate trigonal bipyramidal sites, whereas sodium will occupy nine-co-ordi- nate positions. In the lithium intercalate there does not appear 14 15 16 17 H. R. Hoekstra and R. H. Marshall, J. Znorg. Nucl. Chem., 1965, 27, 1947. R. Chevalier and M. Gasperin, C.R. Acad. Sci., 1968, C267,481. A. Deschanvres, L. Leparmentier and B. Raveau, Bull. SOC.Chim. Fr., 1971,3460. J. M. Moutou, M. Vlasse, M. Cervera-Marzal, J. P. Chaminade and M. Pouchard, J. Solid State Chem., 1984, 51, 190. to be a strong preference for one trigonal bipyramidal site over another. Simulation of the stoichiometric insertion com- pounds MU308 reinforces these predictions.The insertion process is demonstrated to occur with the preservation of the U30s structure but with a slight expansion in the lattice parameters. The sodium insertion compound prepared under ambient temperatures, however, does appear to be metastable. Calculations of the migration of isolated alkali-metal ions within the U308lattice predict that lithium will have a much higher mobility than sodium. 18 19 20 21 22 23 J.P. Chaminade, J.M. Moutou, G. Villeneuve, M.Couzi, M. Pouchard and P.Hagenmuller, J. Solid State Chem., 1986, 65, 27. P. G.Dickens, A. V. Powell and G. P. Stuttard, Proc. Znt. Con$ Electron. Ceram. Muter., Wyoming, USA, 1990,in the press. S. H. Chang, C. Delmas, J. P. Chaminade and P. Hagenmuller, Solid State Zonics, 1990, 39, 305.S. Kemmler-Sack and W. Rudorff, Z. Anorg. Allg. Chem., 1967, 354,255. A. N. Tsvigunov and L. M. Kuznetsov, Radiokhimiya, 1974, 16, 882. J. H.Harding and A. H. Harker, U.K.A.E.A. Harwell Report, 1982,AERE R-10425. We thank Professor C. R. A. Catlow, Dr. R. W. Grimes and Dr. A. H. Harker for useful discussions throughout the course of this work. All defect calculations were performed on the Cray-XMP computer at the University of London Computer Centre. Some of this work was carried out within the Corpor- ate Research Programme of AEA Technology. 24 25 26 27 28 29 J. C. Slater and J. G. Kirkwood, Phys. Rev., 1931, 37, 682. G. V. Lewis, Ph.D. Thesis, University of London, 1984. C. M. Freeman and C. R.A. Catlow, J. Solid State Chem., 1990, 85, 65. B.G.Dick and A. W. Overhauser, Phys. Rev., 1958, 112,90. C. R.A. Catlow, K. M. Diller and M. J. Norgett, J. Phys. C, 1977,10, 1395. 1.D. Brown and R. D. Shannon, Acta Crystallogr. Sect. A, 1973, 29, 266. References 30 I. D. Brown and K. K. Wu, Acta Crystallogr. Sect. B, 1976, 32, 1957. P. G. Dickens, S. V. Hawke and M. T. Weller, Muter. Res. Bull., 1984, 19, 543. P. G. Dickens, S. D. Lawrence and M.T. Weller, Muter. Res. Bull., 1985, 20, 635. P. G. Dickens, D. J. Penny and M. T. Weller, Solid State Zonics, P. G. Dickens, A. V. Powell and A. M. Chippindale, Solid State Zonics, 1988,28-30, 1123. P. G. Dickens, S. D. Lawrence, D. J. Penny and A. V. Powell, Solid State Zonics, 1989, 32-33, 77. 1986, 18-19, 778. 31 32 33 34 35 D. Altermatt and I. D. Brown, Acta Crystallogr. Sect. B, 1985, 41,240. A. M. Chippindale, P. G. Dickens and W. T. A. Harrison, J. Solid State Chem., 1989, 78, 256. I. D. Brown, J. Solid State Chem., 1989, 82, 122. E. H. P. Cordfunke and W. Ouweltjes, J. Chem. Thermodyn., 1981, 13, 187. P. W. Fowler and N. C. Pyper, Proc. R. SOC. London, A, 1985, 398,377. R. G. J. Ball, U.K.A.E.A. Harwell Report, 1989,AERE R-13396. Paper 0/05584K;Received 12th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100415

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 423-427 Progress towards Preparation of High-surface-area Rare-earth Oxides Linda A. Bruce, Simon Hardin, Manh Hoang and Terence W. Turney CSIRO, Division of Materials Science and Technology, Locked Bag 33, Clayton, Victoria 3168, A ustra lia A series of high-surface-area rare-earth oxides has been prepared by precipitation from aqueous solution as either the hydroxide or the carbonate, followed by controlled dehydration, acetone washing, vacuum drying and calcination at ca. 600 "C. The materials have been characterised by X-ray diffraction (XRD), transmission electron microscopy (TEM), infrared spectroscopy (IR) and surface analysis. Precipitation by hydroxide produces a higher surface area for the lighter rare earths, whereas precipitation by carbonate is superior for the heavier members.By judicious choice of method, pure oxides (containing only trace amounts of surface carbonate) with surface areas >50 m2 g-' can be obtained for all rare earths studied. No microporosity was observed in any of the oxides. Keywords: Rare-earth oxide; B.E.T. surface area; Catalyst synthesis Rare-earth oxides are of considerable current interest, both as oxidation catalysts for reactions such as methane coup- ling'.2 and alcohol dehydr~genation,~ and as supports for transition metals in, for example, the hydrogenation of carbon However, commercially available rare-earth oxides possess very low surface areas, generally <10 m2 g-', which restricts their usefulness as catalysts or as catalyst supports.Published efforts to increase the surface area of rare-earth oxides fall into two categories, one commencing from a precursor rare-earth salt and the other from a pre- formed oxide. Each has generally led to only modest improve- ment in surface areas. Thus: (i) Work by several groups has shown that calcination of the voluminous hydrogel of rare- earth hydroxide obtained by reaction of rare-earth ions with hydroxide ions usually leads to an oxide surface area of <40 m2 g-1.8-'1 An exception is Yb203, which has been reported with B.E.T. surface areas of up to 54 m2 g-', after calcination at 600 0C.9 Very slow, isobaric thermal dehy- dration of cerium(1v) hydroxide has yielded CeO, with areas of 58 m2 g-' and 23 m2 g-' after calcination at 500 and 650 "C, respectively.'2 However, the preparation time of 24 days makes this method rather impractical.Other preparative routes to rare-earth oxides have been tried, including decomposition of salts, such as the trichloroacetate, carbonate, oxalate or nitrate, but the surface area is generally under 10 m2 g- '.I3-" (ii) The most commonly employed technique to increase the area of preformed rare-earth oxides has been repetitive hydrolysis and recalcination of commercially avail- able oxide sample^.^.'^*'^ Thus, increases in area from <1 to 28.5 m2 g-' have been reported for La203, and from 2.8 to 13 m2 g-' for Sm203. This paper reports two studies on the preparation of high-surface-area oxides in the series Y and La through Lu.One technique, hereafter termed the 'hydrox- ide method', involves dehydration of rare-earth hydroxide gels by washing with a water-miscible solvent, followed by calci- nation to the oxide. The other, termed the 'carbonate method', consists of a similar dehydration of rare-earth 'carbonate' gels, followed by decomposition to the oxide. In addition, some preliminary work is included on the effect of the temperature of precipitation and the order of mixing of the reactants. Controlled dehydration of hydrogels by employing water- miscible, non-aqueous solvents is a technique which has been reported to produce high-area alumina,,' silica2' and zir- conia-titaniat2 from their respective hydrogels. The method has also been used as a step in the production of high-density (i.e.low-area) rare-earth oxide ceramics from Gd, Er, Y, Eu, Hf, and Zr23 at temperatures in excess of 1800 "C. However, the nature of the products in the temperature region generally of interest for catalytic materials (<go0 "C) has not been studied before. Experimental Materials The rare-earth source materials were stated to be of at least 99.9% purity by the suppliers as follows: Sm203, Ho203, Yb203, y2°3, Nd203, Dy203, pr6011, Gd(N03)3*5H20 (Aldrich); Er2O3, Tb407 (Cerac); La203 (Strem); La(N03)3* 6H20 (Baker); Ce(N03)3 6H20, (NH4),[Ce(N03),] (99%), 'ammonium carbonate' (ca. 5 :1 mixture of ammonium bicarbonate and ammonium carba- mate) (Ajax). Hydrated nitrate salts of the rare earths were prepared in quantitative yields from the commercial oxides following the procedure of Marsh,24 by dissolution in excess nitric acid, evaporation, recrystallization from water and drying over phosphorus pentoxide in vacuo.Methods Hydroxide hydrogels were prepared by slow addition of ammonia (1 mol dm-3) via a dropping funnel (ca. 1 dm3 h-') to a stirred solution of metal nitrate (0.3-0.4 or 0.025-0.05 mol dmP3), under a nitrogen atmosphere. Gels based on carbonate-type precipitation were prepared by similar addition of 'ammonium carbonate' solution (0.3-1 mol dm- 3, to a stirred metal nitrate solution (0.025-0.05 mol dm-3) through which carbon dioxide was bubbled. In each case the pH was monitored during the precipitation.Preparative pro- cedures are typified by the following examples. Preparation of Sm203by the Hydroxide Method Sm(N03)3-6H20 (1 1.11 g), dissolved in distilled water (1 dm3), was titrated with ammonia (500 cm3, 1 mol dm-3) with vigor- ous stirring under a nitrogen atmosphere at room temperature, to a final pH of 10.5. The product gel was separated from the supernatant liquid by centrifugation, redispersed in water (2 dm3) and centrifuged five times to a final conductivity of <2.5 mS m- 'and then centrifuged with acetone (2 dm3) three times. Preliminary drying in a stream of dry nitrogen was followed by drying in vacuo at room temperature for 1 day to a residual pressure of <1 Pa, and then with a stepwise increase in temperature, as the residual vacuum improved, to a final temperature of 400 "C and a pressure of <0.5 Pa.Final calcination to 620 "C was in air in a muffle furnace for 4 h: yield 3.8 g (86%). Preparation of Sm203by the Carbonate Method Sm(N03)3*6H,0 (13.8 g) dissolved in water (3 dm3) was added dropwise over 30 min to a solution consisting of commercial 'ammonium carbonate' (24 g) in water (1 dm3). During the addition, carbon dioxide was bubbled vigorously through the mixture, which was also stirred with a paddle stirrer. Precipitation commenced at pH 4.3; at the end of the addition the pH was 7.3. After bubbling COz for a further 2 h, the pH was 6.7. The hydrogel was separated by centrifug- ation and redispersed and recentrifuged four times with water (1 dm3), and then three times with acetone (0.5 dm3).Prelimi- nary drying was in uucuo at room temperature, followed by heating in air at 120 "C overnight. The sample was then calcined at 400 "C for 2 h followed by 600 "C for 2 h: yield 4.9 g (90%). A number of the precipitations were made by adding the nitrate solution to the stirred base (referred to as 'reverse order') and, in others, the precipitation was carried out at 60 "C instead of at the ambient temperature (ca. 21 "C). Characterisation Adsorption-desorption isotherms of nitrogen at -196 "C were obtained using a Carlo Erba 1800 instrument. XRD data were obtained using a Siemens D-500 diffractometer (filtered Cu-Ka radiation, scan steps 0.04").Infrared spectra were obtained as KBr discs (ca.2% m/m oxide concentration) using a Perkin-Elmer 577 infrared spectrometer; KBr (Ajax AR) was recrystallised from water with addition of conc. HBr and dried at 300 "C before use. Selected area electron diffrac- tion (SAD) and TEM studies were made using a JEOL- 1OOCX electron microscope. Samples were dispersed ultrasonically in hexane, and examined on holey carbon films. X-Ray photo- electron spectra (XPS) were recorded on a Vacuum Gener- ators' ESCALAB with an aluminium anode at 150 W (pass energy 30 eV, 4 mm slits). Nitrate analysis was made using an ion-selective electrode from EDT Research. Results and Discussion Hydroxide Gel Synthesis The initial pH of rare-earth nitrate solutions was usually low (PH =3-4) owing to traces of free nitric acid from the prepara- tive procedure and hydrolysis of the hydrated rare-earth cations. On addition of 5-10cm3 of aqueous ammonia, the pH rose sharply to ca. 8 for La to Sm, or to ca.7 for Gd to Yb. At the same time, the onset of precipitation was evidenced by some opalescence of the solution. The bulk of the precipi- tation occurred within a narrow pH range. At [OH-]/ [RE3'] x 3, there was a rapid further rise in pH. On com- pletion of ammonia addition, the suspension gave no indi- cation of substantial settling, and was intractable to filtration. Centrifugation (2500 rpm for 30 min) gave a voluminous hydrogel with a slight opalescence of the supernatant liquid in some cases, indicating product loss.In subsequent water washes to remove soluble ions, it was sometimes necessary to add up to 10 vol% of conc. ammonia solution to facilitate isolation of the gel. After acetone washes, the gel settled more J. MATER. CHEM., 1991, VOL. 1 readily. Yields by this method were variable, and ranged from <50% for LazO3 to 90% for Smzo3. In preparing the rare-earth oxides, acetone was chosen as the solvent for controlled dehydration (uide infra). As removal of acetone by direct calcination frequently led to partial coking of the product, the gel was first dried in a stream of nitrogen; further acetone and less volatile condensation prod- ucts of acetone were removed in uacuo, first at ambient temperature and then at successively higher temperatures to 400 "C, prior to calcination.Water and nitrogen oxides were also condensed in the cold trap owing to the partial dehy- dration of hydroxide to oxide and decomposition of occluded nitrates, respectively. The evolution of nitrogen oxides in the course of vacuum decomposition of acetone-treated gels was not surprising. The occlusion of substantial nitrate-ion concentrations (0.06-0.25 mol residual nitrate per mole of oxide) in the rare-earth hydroxide gel has been noted in earlier st~dies.'~*'~ The conditions required to free the product from residual nitrate were determined. On dissolution of samples in 0.1 mol dm-3 sulphuric acid, analysis (using an ion-selective electrode) showed that occluded nitrate was present (ca. 3% m/m) after calcination at 420 "C, but was below detection limit (<0.06% m/m) after calcination at 620 "C.Surface nitrate was not detectable by ESCA analysis after calcination at 620 "C in air for 4 h. After heating to 620 "C, only Ce02 as prepared from the hydroxide was still amorphous to X-rays. However, SAD showed the material to be microcrystalline and no amorphous material was evident by TEM. All sesquioxides were obtained as the low-temperature C phase, except Laz03 and Nd203, which were A phase; ceria was f.c.c. These structures are in agreement with literature expectation^.'^ Terbium and praseo- dymium are known to form a range of non-stoichiometric oxides; those obtained here were Tb407 (JCPDS File #32- 1286), for which a defect fluorite structure has been sug- gested,'* and Pr6011 (JCPDS File #6-0329) with an f.c.c.structure. Infrared spectra of freshly prepared samples (as KBr discs) all displayed broad, weak absorption bands in the 3450 and 1630 cm-' regions, attributable to v(H-0-H) and 6(0-H) vibrations of adsorbed water molecules. Dehydration to the oxide was complete under the heating regime used; none of the preparations displayed the strong, sharp band at 3600 cm- which is characteristic of lattice hydr~xide.'~ How-ever, aged samples of oxides, especially Laz03 and Ndz03, displayed this band as a result of the well known hydroxyl- ation reaction of rare-earth oxides with atmospheric water vapo~r.~O The hydroxide preparations were performed in an inert atmosphere to minimise adsorption of COz by the gel.Even with this precaution, all of the samples displayed a weak band at 1060 cm-I and, with the exception of Ce02 and Tb407, weak bands also near 1500, 1460 and 860 cm-'.The bands at 1060, 1500 and 146Ocm-' arise from C-0 stretching modes of unidentate carbonate and that at 860 cm-' from the out-of-plane bending mode of the bidentate carbonate group.31 The interaction of atmospheric carbon dioxide with Tb407 and CeOz to form carbonate is evidently considerably weaker than that for the other rare earths studied. Samaria and La203 qualitatively gave the strongest bands. The interac- tion of rare-earth oxides with carbon dioxide has been exten- sively studied.30 Quantitative temperature-programmed adsorption-desorption measurements, examining the forma- tion and decomposition of the carbonate phases formed from the current oxide preparations, have been reported separ- at el^.'^ Calcination near 800 "C is necessary to ensure com- plete removal of surface carbonate from La203.However, J. MATER. CHEM., 1991, VOL. 1 425 Table 1 TG of carbonate gels weight loss (YO) observed calculated to M203 rare earth T< 600 "C 600 cT/"C <800 M,(CO3)3 MOHC03 M2°(C03)2 M202C03 La 18.6 8.6 28.8 24.5 21.3 11.9 Sm 21.4 2.4 27.5 23.3 20.1 11.2 Tb 20.7 2.9" 26.5 22.4 19.4 10.7 Er 16.2 2.5 25.5 21.6 18.7 10.2 Ce 19.0b 0 - 26.5' - 20.4' ~~ ~ ~ T< 700 "C; Tc450 "C; 'calculated weight losses for 'Ce(OH),CO3' and 'CeOCO,' converting to CeO,.this step results in a considerable lowering of surface area (to 27 m2 g-', vide infra). Carbonate Gel Synthesis As in the hydroxide method, the addition of the first 5-10 cm3 of base often led to an initial sharp rise in pH due to neutralisation of residual free nitric acid. Precipitation occurred at pH 4.5-5.5, and a further sharp rise in pH was observed at [ammonium carbonate]/[M3+] ratio of ca. 3. In contrast to the hydroxide gel, carbonate gels settle spon- taneously to a significant extent, such that ca. 80% of the supernatant liquid may be removed by decantation, and a short centrifugation suffices to give good separation of the gel. No addition of ammonia is necessary in subsequent washing steps.In contrast to the hydroxides, the carbonate precipitates were readily dried in uucuo at room temperature only. This difference may partly be due to formation of more of the less volatile acetone condensation products in the ide route. The carbonate route did not produce La203 and Nd203 as pure products; carbonate phases were observed as substantial components. Infrared spectra confirmed that both La and Nd gels underwent only partial transformation to oxide, even at 700 "C, with strong carbonate bands persisting. All other samples appeared to convert almost completely to oxide, leaving only weak bands in similar positions to those observed from the hydroxide route. Retention of carbonate in the lighter rare earths is in accordance with their higher ba~icity.~.'~Hence, lighter rare-earth oxides are better pre- pared by the hydroxide method.Surface Areas Adsorption-desorption isotherms at -196 "C were measured for rare-earth oxides obtained by each route under a variety of preparative and calcination conditions. They were found to be either type I1 or type IV (both of which are amenable to B.E.T. analysis), depending on the particular method and hydroxide method, by a base-catalysed aldol c~ndensation.~~ member of the series. B.E.T. parameters were determined by Final oxide yields from the carbonate method were typically >90%. After the oxide samples had been dried to 120 "C, XRD traces were featureless, with the exception of that from the Pr preparation which corresponded to * 8H20 (JCPDS File #3 1-1143).Infrared spectra from all specimens showed broad intense carbonate bands. Possible compositions of the solid at this stage might be M,(CO3)3, MOHC03, M20(C03),, M202C03, or a mixed hydroxycarbonate such as M2(OH)6 -2x(co3),.34 Thermogravimetric analysis (TG), using N2 as flow gas, was carried out on samples predried at 120 "C, for Sm, Er, Tb, CeIV and La. Table 1 shows the experimental weight losses observed to 600 "C, and any additional weight loss to 800 "C. Also shown in Table 1 are calculated weight losses from the various carbonate species, assuming conversion to the appropriate oxide. If the original precipitate were M2(C03)3, the observed weight loss would be substantially larger than that found.The figures for Sm and Tb are very close to those for hydroxycarbonate. How- ever, the analysis for Tb is complicated by the variable oxidation states of the precipitate and decomposition product. For Er, the losses are too low for hydroxycarbonate, yet too high for oxycarbonate. Thus, either a mixed hydroxy-oxycarbonate formulation or one corresponding to M2(OH)6-2x(C03)xappears likely.34 For Ce, on the other hand, oxycarbonate losses come closest to the observed fig- ures. The shape of the TG curve for lanthanum shows a two- stage decomposition. The overall weight loss is close to that of La2(C03)3, decomposing via the oxycarbonate to the oxide. After they had been heated to 600 "C, the oxides of Ce, Tb, Er, Ho and Yb prepared by the carbonate route were X-ray amorphous, but SAD showed that they were microcrystalline; again, no amorphous material was detected by TEM.The pure-oxide phases obtained were the same as from the hydrox- least-squares analysis of plots in the range 0.05 <p/po<0.35; 'c' values were typically between 100 and 150, justifying application of the B.E.T. method. Confirmation that the preparations were not microporous was obtained from both a, and t plot^,^',^^ which invariably gave zero intercept on the adsorption axis. A separate study of the relationship of particle size and morphology to surface area and mesoporosity is published elsewhere.37 Tables 2 and 3 list the highest areas obtained by acetone dehydration and calcination of the hydrogels precipitated at Table2 B.E.T.surface area of rare-earth oxides prepared by the hydroxide route oxide y203 Ce02 Pr,O 11 Nd203 Sm203 Gd203 Tb40, Dy2°3 Ho203 Er*O3 Yb203 surface area/m2 g- [MI" B/Mb B.E.T.' ref? 0.44 56.4 63 0.385 13.0 69 0.050 10.0 55 0.025 20.0 86 0.30 13.5 60 0.025 20.0 53 0.025 20.0 60 0.025 20.0 34 0.025 20.0 50 0.025 20.0 50 0.025 20.0 25 0.025 13.9 26 " Initial molar concentration of rare-earth ion, M3+. B/M = ammonia/metal molar ratio; ammonia added dropwise as 1 mol dm-' solution to stirred solution of M3+ at ambient temperature. 'After calcination to 620 "C for 4 h, following acetone dehydration pro- cedure.d 600 "C<calcination temp. <650 "C; highest area found in literature. From ref. 11. From ref. 9. From ref. 10. From ref. 6. Table 3 B.E.T. surface area of rare-earth oxides prepared by carbonate route oxide [MI" [Bib B/M' B.E.T. area/m2 g-' y2°3CeO, 0.010 0.016 1.oo 0.303 16.7 12.Od 98 180 Pr,O 11 Sm203 0.020 0.010 1.oo 0.303 6.3 10.0 <10 52 0.010 0.303 10.0 53 Hoz03 0.0 16 0.303 6.3 56 Er203e 0.010 0.303 10.0 58 Yb203e 0.010 0.303 12.0 59 L'203 0.010 0.303 13.1 59 "Initial molar concentration of rare earth ion, M3+ or 'Molar concentration in titrant of commercial (NH,),[Ce(NO,),]. 'ammonium carbonate', B. Final molar ratio of 'ammonium carbon- ate': metal; carbonate solution added dropwise to stirred solution of metal salt at ambient temperature.Reverse order of addition, with Ce" solution added dropwise to stirred solution of carbonate. 'Precipitation at 60 "C. ambient temperature through the series, by the hydroxide and carbonate routes, respectively. It should be emphasised that not all of the preparative variables have been optimised, and so further improvement in areas can confidently be expected. These tables also give for comparison the previous best reported value under similar final calcination temperature. As with lanthanum, the low-area (22 m2g-') product from neo- dymium by the carbonate method was shown to be impure by TEM and IR studies, and these are therefore omitted from Table 3.The B.E.T. surface area of samples of La203 prepared by the hydroxide method is given in Table4. The results show that the area obtained was increased from 37 to 69 m2 g-' by the introduction of acetone dehydration to otherwise identical drying treatments. The value of 37 m2 g-' alone is comparable to the highest literature value we have found (Table 2). Substitution of methanol for acetone was found to give a product of lower surface area, possibly due to polar interactions with the gel in a similar manner to water. Washing with higher alcohols, such as propan-2-01, gave results similar to acetone, as shown in Table 4, but greater solvent retention made drying in uucuo more difficult.The use of chemical dehydrating agents, such as 2,2-dimethoxypropane (Table 4) or trimethyl orthoformate, offered no advantages over acetone. It appears that removal of water at low temperatures with organic solvents discourages agglomeration of the primary rare-earth hydroxide or carbonate particles, ultimately resulting in high-surface-area material. The effectiveness of gradual removal of water at low temperatures has been shown Table 4 Effect of dehydration medium on surface area of La203" dehydration medium [La]' B/Lac B.E.T. area/m2 g-water 0.385 13.0 36 acetone 0.385 13.0 69 acetoned 0.025 20.0 65 acetone 0.020 5.0' <10 propan-2-01 0.040 12.5 67 2,2-dimethoxypropan& 0.385 7.6 63 " After calcination to 620 "C for 4 h, following dehydration procedure. Initial molar concentration of La3+.'B/La =ammonia/La3+ molar ratio; ammonia added dropwise as 1 mol dm-' solution to stirred solution of La3+ at ambient temperature. Reverse order of addition, with La3+ added dropwise to the stirred solution of I mol dm-j ammonia at ambient temperat~re.~~ Using 'ammonium carbonate' instead of ammonia. Refluxing with 2,2-dimethoxypropane before calcination. J. MATER. CHEM., 1991, VOL. 1 previously for Ce02 by isobaric dehydration. l2 However, the procedure is time consuming. Non-aqueous solvents presum- ably decrease Ostwald ripening by lowering hydroxide solu- bility as well as lowering surface tension, leading to reduced agglomeration of primary crystallites during drying.Similar solvent effects have been observed in the formation of con- trolled-porosity aluminas.20 The addition of La3+ to the ammonia solution instead of vice versa caused an insignificant change in area (65 m2 g- '). On the other hand, attempted preparation of La203 by the carbonate method, rather than the hydroxide method, gave a product (after heating to 600 "C) of <10 m2 g-'; TG, SAD and IR studies showed this material to be still heavily contami- nated with carbonate. Thus, preparation of high-area La203 by the carbonate method is not considered feasible. In Table 5, a comparison is given of the areas obtained for Yb203 and Er203 by the hydroxide and carbonate routes. For these later rare earths, precipitation via the carbonate route gave larger areas than via the hydroxide. Examination of the table also shows that further gains were made by carrying out the precipitation at 60 "C instead of ambient, from 49 to 59 m2 g-' in the case of Yb2O3 and from 36 to 58 m2 g- ' in the case of Er2O3. The results also suggest that a slightly higher-area product was obtained if a lower B/M ratio was used.Optimal reagent concentrations and tempera- tures have yet to be established. Cerium differs from the other rare earths in giving a dioxide on heating. It is found that, with carbonate precipitations, the final surface area of the oxide obtained is strongly dependent on the initial concentration of ammonium ceric nitrate.38 Conclusion Rare-earth oxide powders with relatively high surface area have been prepared by two methods.The two routes are complementary in that the hydroxide method is more effective for the preparation of the lighter (earlier) rare-earth oxides and the carbonate method for the heavier (later) members of the series. Further improvement in the areas obtained so far should be possible for at least some of the series. Catalytic applications of these materials are currently under study. In particular, they have proven to be efficient catalyst supports in carbon monoxide hydrogenation, and active catalysts in methane coupling. Table 5 Effect of preparative conditions on surface area" of rare-earth oxides B.E.T. oxide methodb [MIc initial [Bid B/Me area/m2 g-' 0.010 1.o 16.7 39 0.016 0.303 6.7 48 0.010 0.303 12.0 49 0.010 0.303 12.0 59 0.025 0.016 1.o 0.303 13.9 6.3 26g36 0.010 0.303 10.0 58 0.025 1.o 20.0 2Y " After calcination at 600 "C for 2 h, following acetone dehydration.'C =carbonate method; H =hydroxide method. 'Initial molar con- centration of rare-earth ion, M3+. Molar Concentration of commer- cial 'ammonium carbonate', B. Molar ratio of 'ammonium carbonate'/metal; carbonate solution added dropwise to stirred solu- tion of metal salt at ambient temperature. f Precipitation at 60 "C. * Calcination at 620 "C. J. MATER. CHEM., 1991, VOL. 1 427 References 1 2 K. Otsuka, K. Jinno and A. Morikawa, J. Catal., 1986, 100, 353. A. Ekstrom, R. A. Regtop and S.K. Bhargava, Appl. Catal., 1990, 62, 253. 21 22 23 24 R. K. Iler, The Chemistry of Silica, Wiley, New York, 1979. R. A. Dombro and W. Kirch, Eur. Pat. Appl. 84 110 078.7, 1984. S. J. Dole, R. W. Sheidecker, L. E. Shiers, M. F. Berard and 0. Hunter, Muter. Sci. Eng., 1978, 32, 277. J. K. Marsh, J. Chem. SOC.,1941, 561. 3 4 5 6 7 8 9 10 I1 12 M. P. Rosynek, Catal. Rev. Sci. Eng., 1977, 16, 111. R. F. Hicks and A. T. Bell, J. Catal., 1985, 91, 104. R. Pierantozzi, US.Pat. 4508846, 1985; Chem. Ind., 1985, 22, 115. M. D. Mitchell and M. A. Vannice, Ind. Eng. Chem. Fund., 1984, 23, 88. L. A. Bruce, S. Hardin, M. Hoang and T. W. Turney, in Studies in Surface Science and Catalysis, ed. D. Bibby, C. D. Chang, R. F. Howe and S. Yurchak, Elsevier, Amsterdam, 1988, vol.36, p. 529. J. L. G. Fierro and A. M. Olivan, J. Colloid Interface Sci., 1984, 100, 303. R. Alvero, J. A. Odriozola, J. M. Trillo and S. Bernal, J. Chem. SOC., Dalton Trans., 1984, 87. J. L. G. Fierro and A. M. Olivan, J. Less-Common Metals, 1985, 107, 331. Y. Imizu, K. Sat0 and H. Hattori, J. Catal., 1982, 76, 65. J. L. G. Fierro, S. Mendioroz and A. M. Olivan, J. Colloid Interface Sci., 1985, 107, 60. 25 26 27 28 29 30 31 32 33 34 C. J. Hardy, S. R. Buxton and M. H. Lloyd, ORNL-4000, 1967. J. P. McBride, ORNL-TM-1980, 1967. L. R. Eyring, in Handbook on the Physics and Chemistry of Rare Earths, ed. K. A. Gschneidner and L. R. Eyring, North-Holland, Amsterdam, 1979, vol. 3, p. 337. P. E. Caro, J. Less-Common Metals, 1968, 16, 367. Ya. I. Ryskin, in The Infrared Spectra of Minerals, ed.V. C. Farmer, Mineralogical Society, London, 1974, p. 138. S. Bernal, F. J. Botana, R. Garcia and J. M. Rodriguez-Izquierdo, J. Muter. Sci., 1988, 23, 1474. L. H. Little, Infrared Spectra of Adsorbed Species, Academic Press, London, 1966, p. 88. K. Foger, M. Hoang and T. W. Turney, J. Mater. Sci., in the press. J. March, Advanced Organic Chemistry: Reactions, Mechanisms and Structure, McGraw-Hill Kogakusha, Tokyo, 2nd edn., 1977, p. 849. S. Bernal, J. A. Diaz, R. Garcia and J. M. Rodriguez-Izquierdo, J. Muter. Sci., 1985, 20, 537. 13 14 15 16 17 18 19 20 R. G. Charles, J. Inorg. Nucl. Chem., 1963, 27, 1489. S. T. King and E. J. Strojny, J. Catal., 1982, 76, 274. R. Alvero, A. Bernal, I. Carrizosa, A. Justo, J. A. Odriozola and J. M. Trillo, J. Less-Common Metals, 1985, 112, 347. M. Akinc and D. Sordelet, Adv. Cerarn. Muter., 1987, 2, 232. I. Akalay, M-F. Guilleux, J-F. Tempere and D. Delafosse, J. Chem. SOC., Faraday Trans. 1, 1987,83, 1137. M. P. Rosynek and D. T. Magnuson, J. Catal., 1977, 46,402. C. Sudhakar and M. A. Vannice, Appl. Catal., 1985, 14, 47. A. White, A. Walpole, Y. Huang and D. L. Trimm, Appl. Catal., 1989, 56, 187. 35 36 37 38 B. C. Lippens, B. G. Linsen and J. H. de Boer, J. Catal., 1964, 3, 32. S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Porosity, Academic Press, London, 2nd edn., 1982, p. 93. L. A. Bruce, S. Hardin, M. Hoang and T. W. Turney, Aust. J. Chem., 1991, in the press. L. A. Bruce, M. Hoang and T. W. Turney, J. Muter. Chem., submitted. Paper 01056254 Received 14th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100423

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991,1(3), 429-435 Reorientation of Mo-co-ordinated Water Molecules in High- hydrogen-content oxide bronzes H,,M003 and H,,Mo03 A Neutron Scattering Studyt Robert C. T. Slade,* Paul R. Hirst and Helen A. Pressman Department of Chemistry, University of Exeter, Exeter EX4 4QD, UK Variable-temperature incoherent quasielastic neutron scattering (QENS) measurements have been used to investigate motions of hydrogenic species in the oxide bronzes H1.7M003 and H,,,Mo03. In both phases there are distinct (slowly exchanging) populations of H undergoing reorientational (those present in Mo-co-ordinated H,O) and self-diffusive (those present in Mo-co-ordinated hydroxyl) motions. In H,,,MoO, the H,O molecules undergo four-fold reorientation [€,= 26+3 kJ mol-', log (z:&) = -14.4k0.71about the Mo-0 bond.In H,.,MoO,, reorientation of H,O is restricted to a 180" flip about the bisector of the H-0-H bond angle. Keywords: Quasielastic neutron scattering spectroscopy; Hydrogen molybdenum bronze; Reorientation; Self- diffusion Hydrogen molybdenum bronzes H,Mo03 are a series of single phases derived from parent layered Moo3 by topotactic insertion of hydrogen.' Interest in these materials stems from possible applications e.g. in catalysis and in electrochemical devices. Nuclear magnetic resonance ~tudies~-~ revealed high H-atom mobilities in the highest hydrogen-content phases [x = 1.7 (red monoclinic) and x =2.0 (green monoclinic)]. Incoherent inelastic neutron scattering (IINS) vibrational spectroscopy identified the presence of both co-ordinated water molecules and hydroxyl groups in the red monoclinic phase (x =1 .7).5 A neutron diffraction study determined the H-atom positions to be associated with terminal oxygens [0(3), Fig.l(a)] at the top and bottom of the Moo3 layers.6 Mo-co-ordination of water molecules is similar in the related molybdic acids Moo3 nH,O (n= 1, 2).798 In the bronze, projection of the interlayer region onto the interlayer plane (x/a =0.5) results in an approximately square grid [Fig. I@)], the O(3)atoms lying at the intersections and the H-atom sites lying on the lines. A simplistic square-grid model had been proposed previously in interpreting 2H NMR ~pectra.~ IINS spectroscopy also identifies the presence of Mo-co- ordinated water molecules in the green monoclinic phase (x = 2.0)' No diffraction study of this phase has been published, but bond strength and structural considerations" indicate that the water molecule H atoms are located in the interlayer region as in the red monoclinic phase.We have previously reported use of incoherent quasielastic neutron scattering (QENS) techniques in characterisation of two (slowly exchanging) dynamic hydrogen populations in H1.7M003.1 One population (hydrogen in hydroxyl groups) undergoes rapid self-diffusion [Dt (295 K)=4 x cm2 s-'1. The second (hydrogen in water molecules) has a higher activation barrier to self-diffusion and undergoes reorien- tational motion. We now report detailed QENS investigations of reorientation of water molecules in H1.7M003 and H2,,Mo03 and that the mechanisms for this differ in these closely related phases.Materials All sample handling was carried out under nitrogen, the materials being susceptible to oxidation by air. t Neutron scattering experiments carried out at the Institut Laue- Langevin, Grenoble, France. 0 xx a xx 0 xx X X X X "t X a xx X o XI X a xx I o X X X X o xx xx o xx a X X X X X X X "1 b Fig. 1 (a)The layer structure of molybdenum trioxide represented in terms of linked MOO, octahedra. In high-hydrogen-content hydrogen molybdenum bronzes H,MoO,, inserted H atoms are located in the interlayer region, bonded to terminal O(3) atoms.(b)The interlayer region in high-hydrogen-content hydrogen molybdenum bronzes: 0,terminal O(3) (upper layer); 0, terminal O(3) (lower layer); x, H-atom sites (no more than one occupied per 0(3)...0(3) line, no more than two occupied per 0(3), average occupancy 0.425 in H1.,Mo03 and 0.5 in H2.0M~03). The projection onto a plane midway between the layers approximates to a square grid, with H-atom sites on the lines and O(3) atoms at intersections Green monoclinic H2 .oMo03 was prepared by Zn/HCl(aq) reduction of Moo3 under flowing nitrogen.I2 Red monoclinic H1.7M003 was prepared from H2.0MoO3 by heating at 120 "C under dynamic vacuum.' X-Ray powder diffractometry (Philips diffractometer, CU-Ka radiation) confirmed both materials to be single monoclinic phases, with diffraction patterns and unit cells in good agreement with literature data.' The formulae for the materials were determined by chemical reducing-power analysis.Samples were dissolved in an excess of 0.05 mol dm-3 cerium(1v) ammonium sulphate (in 2 mol dm-3 aqueous sulphuric acid). Excess cerium(1v) ion was determined by potentiometric titration against standardised 0.05 mol dm-3 iron@) ammonium sulphate (in 2 mol dm-3 aqueous sulphuric acid). The amount of cerium(rv) ion con- sumed is related to the number of reducing electrons in HxMo03 by MoVI -X + xCerV+xCe"' + MoV' (1) Analysis of the materials studied gave x = 1.97f 0.03 (green monoclinic) and x = 1.68f0.02 (red monoclinic). Samples for neutron-scattering studies were sealed (indium gaskets) in slab-shaped rectangular cross-section aluminium cans (window thickness 0.1 mm).Chemical and X-ray analyses were reconfirmed after neutron investigations. Experimental Quasielastic neutron scattering spectra were recorded at the Institut Laue-Langevin (ILL). Measurements on the back- scattering spectrometer IN13 were made with an elastic energy resolution AEo = 8 peV (FWHM) and over an elastic scat- tering-vector magnitude range 1.19 < Qel/A-' < 5.02. Sample cans were inclined at 135" to the incident beam. Sample temperatures were controlled to fl K using standard ILL cryostats (thermostatted centre stick). Temperatures at which quasielastic scattering data were to be collected were deter- mined using constant-energy window experiments.For H~.,MoO~, measurements of scattering spectra were made at T=60, 304 and 317 K. Prior studies of samples heated in neutron scattering cans revealed partial conversion to H1.7M003 at higher temperatures. For H1.7M003, measurements of scattering spectra were made at T= 90, 255, 306,328 and 349 K. In studies of each sample the instrumental resolution function was determined using a similarly mounted vanadium sheet and empty-can scattering was also recorded. After subtraction of the empty-can scattering and the background, and correction for absorption and slab-sample geometry, sample spectra were normalised by comparison with vanadium spectra and converted to the symmetrised scattering law S(Q, o)form (each step using standard local ILL procedures).Results Self-diffusing Populations We have previously described the use of observed neutron- scattering intensities (after background subtraction) in deter- mination of the relative magnitudes of two H-atom populations separately undergoing self-diffusion and reorien- tation." At low temperature, all the H present (P'") contrib-utes only to elastic scattering [observed intensity Z(Q, Tow)]. At higher temperature, 17: quasielastic broadening occurs and, if we denote the H population observed in the instrumental energy window as PObs, =In C~bs/p'O'lln [Z(Q, Wl(Q,Tow)] + SQZi (1) where 6 is the difference in the mean-square H-atom displace- ments at the two temperatures.Plots of In [Z(Q, T)/Z(Q, Tow)] uersus QZl are shown in Fig. 2 for H1.7MO03 (ToW=9O K) and H2.,Mo03 ( qOw=60 K). Non-zero intercepts on the vertical axis arise from a quasielastic component with neutron energy transfers at high Qel far outside the instrumental energy- transfer window (broadening to become effectively additional background at high Qel).That component is the Qcl-dependent quasielastic component of the self-diffusing population." The percentages of H present undergoing self-diffusion can be evaluated from the y-axis intercept of the high-& linear dependences and are given in Table 1. The value cited for H1,7M0O3 at 255 K is the percentage present as hydroxyl at low temperature as evaluated from broad-line NMR.2 At that J.MATER. CHEM., 1991, VOL. 1 -1 4 04-*.. 306 K P -1 4 04 Q:,/A-2 3 -0.51 -0.54 Fig. 2 Intensity ratios as a function of Qcrfor measured (IN13) QENS spectra (less apparent backgrounds) for (a) H1.7Mo03 at high tem- peratures and 90K and (b) H2.,Mo03 at high temperatures and 60K. Straight lines shown are those giving the percentages of the total 'Hpresent that are undergoing self-diffusion (see Table 1) Table 1 Self-diffusing H atom population as a percentage of total H atoms present in high-hydrogen-content hydrogen molybdenum bronzes 1.7 H2.0M003 T/K population (YO) TIK population (%) 255 10" 304 4f2 306 22f4 317 7f2 328 33f3 349 45f5 ~ " Estimated from broad-line 'H NMR spectra at low temperature' (see text).temperature diffusion has slowed to become a purely elastic component at all Qcl values." H2.oMoO3 is very nearly stoichiometric [with two H atoms per terminal O(3) in that limit] with a consequential very low proportion of H present in hydroxyl. For H1.7M003 the J. MATER. CHEM., 1991, VOL. 1 temperature dependence of the relative population diffusing (present in hydroxyl) reveals an energy preference for H-atom sites corresponding to co-ordinated water molecules, with the H-atom distribution in the interlayer region [the grid of Fig. l(b)] becoming more random at higher temperature. Scattering Spectra Spectra at the lowest temperatures (Tow)were indistinguish-able from the instrumental resolution function.Spectra at higher temperatures all showed quasielastic broadenings for both materials. These spectra were initially fitted individually to a simple analytical form, consisting of a simple scattering law s(Q, o)=Bo(QV(o)+F(Q, 0) (2) convoluted with the instrumental resolution function. The quasielastic component F(Q, o)was taken to be adequately represented by a single Lorentzian (15).The empirical elastic incoherent structure factor [EISF(Q)] is the ratio of the elastic to the total (elastic+quasielastic) intensity in the incoherent scattering spectrum EISF(Q)=Bo(Q)/CBO(Q)+ 7 F(Q, 4dol --co =Bo(Q) for normalised S(Q, o) (3) and is a measure of the time-averaged spatial distribution of the proton (incoherent scattering being dominated by the 'H present), while the time-dependent proton position is in the quasielastic term F(Q, o).EISF values at Qe, contaminated by coherent scattering (known from X-ray diffraction and confirmed using the ILL neutron diffractometer DlB) were corrected following Richardson et all3.For H1.7M003 at T>300 K a decreasing EISF at Qel<2.0 A-was a consequence of the diffusive motion, two Lorentzian quasielastic components (corresponding to diffusion and reorientation) being appropriate when broaden-ing due to diffusion is not background. EISFs Appropriate to Reorientational Motion When separate self-diffusing and reorientating populations are being observed, the observed scattered intensity is the sum of three components Iobs =12'+ +Id (4) where Z2' and ]:'are the elastic and quasielastic contributions from the rotational scattering law and Id is the (quasielastic) contribution from the scattering law for translational diffusion.The EISF appropriate to the reorientating population alone (EISF'"') is EISF'"' =Zzt/(ZS'+Z;')= zZ'/(I,,bs -Id) (5) Fig. 3 shows the reorientational EISF'"' appropriate to H2,0Mo03(317 K) and H1,7M003(328 K) and compares them to the predictions of various reorientational models for H20 molecules (see below). In the case of H2.0M~03the very small diffusing population renders Idnegligible (i.e.no correc-tions to the empirical EISF were necessary). In the case of H1.7M003,Id=O at Q2>6A-2 (see the discussion in the preceding section) and Id values at lower Qel values were evaluated from Fig.2 using deviations from the back extrapo-lated dependence at high Qel. At the lowest Qel values, corrections applied (for Bragg scattering and diffusion) are not completely adequate to remove the effects of self-diffusion. 43 1 0.51 A\\ 0 u. 0.5-0-0 2 4 6 Q,, /A -' Fig. 3 Comparison of empirical (IN13) Q,,-dependent elastic incoher-ent structure factors for reorientation (EISF"') with the predictions of various models for reorientation of terminal water molecules (see text and Fig. 4):(a)H,.,Mo03 at 328 K; (b) H,.,Mo03 at 317 K Data Analysis Reorientational Models In considering possible rotations it is pertinent to consider rotational scattering laws arising from motion of H atoms between N equivalent sites on a circle (following Barnes14)or on the surface of a sphere (isotropic rotational diffusion, following Sears'').For a population reorientating about a single axis (Barnes model) the scattering law Srot(Q,o)is then in the form of eqn.(2) (convoluted with the instrumental resolution function) with where n B,,(Qa)=N-' j0[2Qasin(np/N)] cos(2nnplN) (7) p= 1 and jo(x)=(sin x)/x for a powder sample, a is the radius (of gyration) of the circle, N is the number of sites and zn= z1 sin2(n/N)/sin2(nn/N).z1 is the half width at half maximum (HWHM) in angular frequency for the first Lorentzian and is related to the mean residence time on a site z,,, by z,,, =z1[ 1-cos (2n/N)] (8) For reorientations in which only one of the water H atoms changes position, the theoretical EISF'"' is simply related to that for both moving (EISFboth)on the same circle in the same way by EISF'"' =0.5 (1.O +EISFhth) (9) Various models for the reorientation of terminal water molecules in the interlayer region are conceptually possible.432 These can be idealised as rotations on a circle and 'theoretical' values of EISF'O' calculated. Taking the 0-H bond length as 1.0 A and the bond angle as 11 5.7" (values derived from the structure of HI MOO^^), the following models (illustrated in Fig. 4) involving O(3) atoms in adjacent layers were con- sidered: Model A, interchange of H atoms by two-fold reorien- tation about the axis bisecting the bond angle H-O(3)-H (N=2, all H moving), a 180" flip of H20; Model B, a two- fold reorientation of the water molecule about the 0(3)-H...0(3) axis, one H atom not being moved (N=2, one H atom static); Model C, four-fold reorientation about the Mo-0(3) axis (N=4, all H moving); Model D, uniaxial rotational diffusion about the Mo-0(3) axis (model C with N increased to infinity); Model E, isotropic rotational diffusion on a sphere centred on 0(3), random reorientation of H20.Models involving rotational diffusion have been included for completeness but are unlikely because (i) there is directional H-bonding (inhibiting D and E) and (ii) the impossibility of random reorientation in the presence of co-ordination to Mo (eliminating E).The variation of EISF'"' with Qelis only slight for small variations in bond length or bond angle. Examination of Fig.3 shows that, for each phase, the empirical Q,,-dependence of EISF'"' adequately matches (par- ticularly in view of the number of corrections made) the predictions of a reorientational model, but that the appropri- ate model differs for the two phases. In the case of the higher H-content phase, H2.-,Mo03, the appropriate model is a two- fold reorientation (model A) corresponding to a 180" flip of co-ordinated H20. In the case of H1.7M003, however, a A Mo 0 V m I C MO Fig. 4 Three possible reorientational motions of terminal water mol- ecules (co-ordinated to Mo) in high-hydrogen-content hydrogen mol- ybdenum bronzes.These and further models are discussed in the text. Neighbouring 0 atoms are on an adjacent layer (see Fig. 1) J. MATER. CHEM., 1991,VOL. 1 switch of observed reorientational mechanism to four-fold reorientation (model C) is evident. Modelling the Scattering Law for H,,Mo03 The empirical S(Q,o)data were fitted to a convolution of the instrumental resolution function with a scattering law of the form S(Q,a)=U~(O)+VSrot(Q,a)+WPans(Q,o) (10) as previously." U corresponds to an elastic contribution from coherent (Bragg) scattering (at known Qel values). The rotational scattering law was in the form of eqn. (2) and (6) as appropriate to model C (four-fold reorientation). The percentage of the total population undergoing self-diffusion at a given temperature was given in Table 1.For a model of self-diffusion by discrete jumps Strans(Q,o)is a Lorentzian with Q,,-dependent halfwidth r(Q).I6 As Qel-+O, Strans(Qw) becomes identical to the scattering law for continuous diffusion and where D is the H self-diffusion coefficient. In spectra at T> 300 K, r(Q)increases rapidly with Qel so that 9""(Q,o)becomes effectively an additional background in spectra at high Qel. Spectra were initially fitted individually to the form of eqn. (7) as follows. At Qe,>2.5 A-': The diffusive contribution was taken to be background (W=0). Reorientational parameters evaluated for this high Qe, range are given in Table 2. Poor counting statistics in spectra at 349 K are reflected in the large relative uncertainty in derived values.At Qel< 2.5 A-': Ps(Q,o)was taken to be a single Lorentzian with r(Q)varying according to eqn. (1 1) and reorientational parameters fixed to the mean values in Table 2. Derived self-diffusion coefficients were 4.8 & 0.9 and 6.8k1.3 cm2 s-' at 306 and 328 K, in reasonable agreement with previous reported values."*'7 The use of eqn. (1 1) to describe the variation of r(Q)in this study will be less accurate than in previous work on IN5" at lower Qel values (Qel<1.1 A-I). A more accurate form for variation of r(Q)at the higher Qel values in this work would require a detailed model (not known) for the diffusion pathwayI6 and is not warranted by the data. In spectra at 255 K, the predicted r(Q)values are very much less than the instrumental resolution (D being very much lower) and, when convoluted with the instrumental resolution function, S(Q,a)is then indistinguishable from S(Q,W)=X~(O)+ VSrot(Q,O) (12) Spectra were initially fitted individually to this form.The evaluated reorientational parameters are given in Table 2. Fig. 5 presents final fits to S(Q, o)as a function of Tand Table 2 Reorientational parameters for water-molecule reorientation in high-hydrogen-content hydrogen molybdenum bronzes TJK HWH M/peVa 7res/lo-11s H,.,MoO,: four-fold reorientation 349 26f5 2.5 f0.6 328 15+l 4.4 f0.3 306 7.5 &0.5 8.8 k0.6 255 0.99f0.25 66k2 317 H2.,Mo0,: two-fold reorientation 8.3 k2.0 1.6k0.6 304 3.2 &1.O 4.1 & 1.4 a For the first Lorentzian in the appropriate Barnes model14 (see text).J. MATER. CHEM., 1991, VOL. 1 0.17 0 ElmeV 0.17 n i, 0.21 i i, L 0.21 I 0 ElmeV 0.21 0 ElmeV 0.21 0.26 0.26 0.26 I 10 ElrneV 0.26 0 ElmeV 0.26 Fig. 5 Fits to the empirical (IN13) scattering laws S(Q, o)obtained for H1.,Mo03as a function of Tand Qel, Top solid lines are fits to the experimental data. Fitting methods are discussed fully in the text. Other lines have the following meanings. Spectra at 255 K (a):the upper dashed line separates the elastic and quasielastic (broadened) contributions; the lower dashed line shows the background; spectra shown correspond to (left to right) eel=1.19, 1.59, 2.37, 3.44, and 4.82 A-’.Low Qel spectra at higher temperatures [(b)306 K, (d) 328 K]: the upper dashed line separates the elastic and quasielastic (broadened) contributions; the central solid line shows the contribution arising from self- diffusion; the lowest dashed line shows the background; spectra shown correspond to (left to right) Qel= 1.59, 1.99 and 2.37 kl.High Qel spectra at higher temperatures [(c) 306 K, (e) 328 K]: the upper dashed line separates the elastic and quasielastic (broadened) contributions.the lowest line shows the ‘background’ (the true background plus a contribution from self-diffusion that is ‘flat’ in the instrumental energy transfer window); spectra shown correspond to (left to right) Qe1=2.37, 3.44 and 4.71 A-’ Qel(HWHM and D fixed to the mean values of each tempera- Modelling the Scattering Law for H2@MoO3 ture), which appear satisfactory at all temperatures and Qe, values.Assumption of Arrhenius temperature dependence Attempts to fit the empirical S(Q,o)data following the for reorientational zreS gives E, =26 & 3 kJ mol-and approach used above for HI.,Moo3 were unsuccessful owing log (zrOes/s)= -14.4& 0.7. to the small size of the self-diffusing population and the relatively poor counting statistics. S(Q,o) was taken as a convolution of the instrumental resolution function with a scattering law in the form of eqn. (10) with Sro'(Q,o)appropri-ate to model A (180" flip of co-ordinated H20). D values evaluated at low Qel varied considerably at different Qel.Fixing D at a given temperature to a value appropriate to H1.7M003 did not give discernible improvements in the fits at low Qel. The empirical S(Q,o)data were fitted to a convolution of the instrumental resolution function with a scattering law of the form S(Q, 0)= U6(o)+ VSr0'(Q,o) (13) (ie.a nil self-diffusing population). Spectra at high Qel were initially fitted individually and derived reorientational param- eters are given in Table 2. Fig. 6 presents final fits to S(Q, w) as a function of Tand Qel (HWHM fixed to the mean value . I 10 EIpeV 70 (b) J. MATER. CHEM., 1991, VOL. 1 at each temperature), which are satisfactory at al) but the lowest Qel values (where a discernible, as opposed to back- ground, self-diffusive contribution would be anticipated).A wider temperature range would be required for accurate evaluation of a reorientational activation energy and prefactor for this phase. Change in Mechanism The change in mechanism from four-fold reorientation [about the Mo-0(3) axis] in H1.7M003 to two-fold reorientation [about the line bisecting the H-O(3)-H angle] in H2.0Mo03 merits discussion. Two-fold reorientation (1 80" flips) of Mo-co-ordinated water in molybdic acids Moo3 nH20 (n= 1,2) has been characterised by variable- temperature 'H NMR relaxation-time mesurements'* and is slower than the two-fold reorientation in H2.0M003. An 1 70 170 Fig. 6 Fits to the empirical (IN13) scattering laws S(Q, w) obtained for H,.oMoO, as a function of Tand Qel: (a) 304 K; (b) 317 K.Top solid lines are fits to the experimental data. The fitting method is discussed fully in the text. Other lines have the following meanings; the upper dashed line separates the elastic and quasielastic (broadened) contributions; the dotted line shows the background; spectra shown correspond to (left to right) Qel= 1.29, 2.09, 3.99 and 4.98 A-' J. MATER. CHEM., 1991, VOL. 1 butions from self-diffusing H (present in hydroxyl groups). Fitting of empirical scattering laws indicates different reorien- fHg" 6" PHH H tational motions to occur in the two phases. For H1.7M003 the spectra were satisfactorily modelled assuming a four-fold reorientation about the Mo-0 axis [an approximate four-fold symmetry is evident in models of the interlayer region, see Fig.l(b)] with E,=26+ 3 kJ mol-I andrHiAHP" HrHH H log (zrOes/s)= -14.4& 0.7. For the higher-H-content H2.0M003, H/!IH 0-H / TH HTHH H rH Fig. 7 An ordered arrangement in the interlayer region of H,.,MoO,. Filled and open oxygens are co-ordinated to Mo atoms in different layers (see Fig. 1). H20molecules can reorient only via rotation about the 'molecular C, axis' (model A). Reorientation about the Mo-0 bond (four-fold reorientation, model C) is prevented by the restriction that no O..-O line may be occupied by more than one H approximate four-fold symmetry is, however, apparent in the interlayer region in the bronzes studied in this work [see Fig.l(b)]. The interlayer separation is smaller in H2,0M~03 (in which it is slightly smaller than in MOO, itself),' which may restrict higher-order reorientation in the higher H-con- tent phase. The near-complete co-ordination of each O(3) atom by two H atoms in H2.,Mo03 may lead to local ordering of the interlayer grid as in Fig. 7, with consequent inhibition of higher-order reorientation [no grid line may be associated with two H atoms, see Fig. I@)] and with non- stoichiometry (co-ordinated hydroxyl groups in place of HzO) arising at a boundary between differently ordered regions. Conclusions The high-hydrogen-content hydrogen molybdenum bronzes H1.7M003 and H2.0M~03 both contain Mo-co-ordinated H20 molecules. The 0 atoms of H20 are at the top and bottom of Moo3-like layers, with the H in the interlayer region.Incoherent QENS spectra (instrument IN1 3 at ILL) detect reorientation of HzO in both phases, and additional contri- HrHd-reorientation is restricted to 180" flip of H20 about the line bisecting the bond angle H-0-H. We thank the Institut Laue-Langevin for access to spec-HrHtrometer IN13. We thank the SERC for grants in support of the Exeter neutron scattering programme and for studentships for P.R.H. and H.A.P. We thank Drs. A. Magerl, C. Ritter and R. C. Ward for practical assistance and helpful discussions. References 1 J. J. Birtill and P. G. Dickens, Muter. Res. Bull. 1978, 13, 311. 2 R. C. T. Slade, T. K. Halstead and P. G. Dickens, J.Solid State Chem., 1980, 34, 183. 3 A. C. Cirillo, L. Ryan, B. C. Gerstein and J. J. Fripiat, J. Chem. Phys., 1980, 73, 3060. 4 C. Ritter, W. Miiller-Warmuth and R. Schollhorn, J.Chem. Phys., 1985,83, 6730. 5 P. G. Dickens, J. J. Birtill and C. J. Wright, J. Solid State Chem., 1979, 28, 185. 6 P. G. Dickens, A. T. Short and S. Crouch-Baker, Solid State lonics, 1988, 28-30, 1294. 7 B. Krebs, Acta Crystallogr., Sect. B, 1972, 28, 2222. 8 J. R. Gunter, J. Solid State Chem., 1972, 5, 354. 9 R. C. T. Slade, T. K. Halstead, P. G. Dickens and R. H. Jarman, Solid State Commun., 1983, 45, 459. 10 P. G. Dickens, R. H. Jarman, R. C. T. Slade and C. J. Wright, J. Chem. Phys., 1980,77, 5500. 11 R. C. T. Slade, P. R. Hirst, B. C. West, R. C. Ward and A. Magerl, Chem. Phys. Lett., 1989, 155, 305. 12 0. Glemser and G. Lutz, 2.Anorg. Allg. Chem., 1957, 264, 17. 13 R. M. Richardson, A. J. Leadbetter, D. H. Bonsor and G. J. Kruger, Mol. Phys., 1980,40, 747 14 J. 0. Barnes, J. Chem. Phys., 1973, 58, 5193. 15 V. F. Sears, Can. J. Phys., 1973, 58, 5193. 16 M. BCe, Quasielastic Neutron Scattering. Principles and Appli- cations in Solid State Chemistry, Biology and Materials Science, Adam Hilger, Bristol, 1988, ch. 5. 17 R. E. Taylor, M. M. Silva-Crawford and B. C. Gerstein, J. Catal., 1980, 62, 401. 18 R. H. Jarman, P. G. Dickens and R. C. T. Slade, J. Solid State Chem., 1981, 39, 387. Paper 0/05628F; Received 14th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100429

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 437-440 Luminescence Properties of A,ReCI, Crystals Marco Bettinelli", Colin D. Flint*band Gianluigi Ingletto" a Dipartimento di Chimica lnorganica, Metallorganica e Analitica, Universita di Padova, Via Loredan 4, 35131 Padova, Italy Laser Laboratory, Department of Chemistrx Birkbeck College, Gordon House, 29 Gordon Square, London WClH OPP, UK lnstituto di Biochimica, Facolta di Medicina Veterinaria Universita di Parma,' Via del Taglio 43100 Parma, Italy At temperatures below 60 K A,ReCI, (A= K, Rb, Cs) crystals emit intense, well resolved luminescence in the 13 900-12 800 cm-' region under blue-green excitation. The emission is strongly temperature dependent and also depends somewhat on the excitation wavelength and crystal quality.It is here assigned as the vibronic structure associated with the r7(,T2J -+Ts(4A,g) transition of centrosymmetric and essentially octahedral ReCIg- ions at up to nine distinct defect sites. The compounds (NR,),ReCI, (R=CH,, C2H5) also luminesce in the same spectral region but more weakly than the alkali-metal salts. [N(n-C,H,),], ReCI, does not luminesce under these conditions. The nature of the defect sites and their excitation processes are discussed. Keywords: Luminescent material; Alkali-metal salt; Defect site The luminescence behaviour of the ReC1;- ion diluted into transparent hosts and K2PtC16 has been studied in some detail.'*2 Emission occurs from the r7(2T2g) and r8(2T1,) states to the ground state T8(,A2,) and any intermediate states.Although numerous vibrational and electronic spectral st~dies,~,~as well as magnetic,'^^ and calori- metric" measurements have been reported for the undiluted K2ReCl, crystals and this compound has been used as a model for a number of related theoretical studies, no lumi- nescence from this class of compounds has been reported. During careful measurements of the absorption spectra of K2ReC1, and Cs2ReC16'l at liquid-helium temperatures, we detected intense luminescence to low energy of the r8(4A2g)+r7(2T2,)electronic origin. In this paper we report a detailed analysis of the vibrational structure and temperature dependence of the emission from these crystals and also from the analogous Rb2ReC1, compound. Interionic coupling in these compounds is rather strong, as evidenced by the intensity of the co-operative absorption bands and the observation of magnon sidebands in absorption below the Nee1 temperature of K2ReC16." Energy migration through the lattice is efficient until the excitation becomes trapped at a defect site.In contrast no co-operative absorptions, magnon sidebands or trap emission is observed for K2ReF612.'3 and the vibrational structure in the absorption spectrum demonstrates a remark- ably complex vibronic behaviour. We observe rather weaker emission from (NR,)2ReC1, (R=CH3, C2H5); in this case the emission is from non-centrosymmetric ReCl; -traps, the weakness of the emission no doubt reflecting the contribution of the C-H vibrational motion to the non-radiative decay rate.[N(n-C4H9),],ReC1, does not emit significantly from the r7(2T2g) state. Structural, Magnetic, Electronic and Vibrational Data Cs2ReC16 and K2ReC16 are cubic at room temperature (space group Fm3m 0h5).9,14The Re4+ ions occupy sites of perfect Oh symmetry. K2ReCl, undergoes several structural phase transitions at low temperatures and becomes antiferromag- netic at 12 K.6 No structural or magnetic phase transitions have been reported for Cs2ReC1, in the 300-4K range but there is some evidence for crystal strain at low temperatures." No details about the crystal structure of Rb2ReCI6 are known, but it is reasonable to assume that it is isostructural with K2ReC16 and Cs2ReC16 at room temperature.[N( CH,),] ReC1, and [N( c2H5)4] ReC1, are cubic (Frn3rn from X-ray powder diffraction data, isostructural with K2ReC16)15 and monoclinic (C2/c from single-crystal data, isostructural with [N(C,H,),], SnCl,),', respectively. The ReC1;-complex ions show only very small deviations from a perfect Oh symmetry in the latter crystal. The reported Weiss constants derived from bulk magnetic susceptibility measurements are higher for K2ReC16 (88 K) than for Cs2ReC1, (45 K).5 It is possible to predict, on the basis of a crude molecular-field model, that the Nkel temperature for Cs2ReC1, will be significantly lower than 12 K. Tetraalkylam- monium hexachlororhenates have been reported to be mag- netically di1~te.I~ The electronic structure of the ReCl2- ion is well under- stood. The t;, strong-field configuration gives rise to the six states Tg(4A2,), T8(2T1,)(7643 cm- l); T8(2E2,) (8906 cm-'); r6(2T1,) (9344cm-'); T7(2T2g)(13 841 cm-') and (1 5 299 cm- ').The energies are given for the excited states for ReC1;- in K2PtC1, at 4 K.2 The corresponding absorption energies in undiluted A2ReC16 are closely similar." Experimental Good-quality crystals of K2ReC16 and Cs2ReC1, were grown as previously described. Rb2ReC16 crystals, also of high optical quality, were grown by the same method as was used for Cs2ReC1,. Large crystals of (NR,),ReC16 (R =CH3, C2H5, n-C,H9) could not be grown owing to their low solubility but these compounds were obtained as microcrystalline powders by the addition of excess of an aqueous solution containing (NR,)' to a solution of K2ReC1, in 2moldrnb3 aqueous HCl.Luminescence spectra were recorded using excitation from an argon laser and a 1 m monochromator as previously described.l8 For most measurements the laser power incident on the sample was <25 mW and since the absorbance of the crystals is low in the 454-514 nm region, no sample heating was detected. For the alkali-metal salts the intensity of the emission under these conditions was such that the mono- chromator slits were closed to their mechanical stop at less than 10 pm. The measured spectral band width of the mono- chromator under these conditions is less than 1 cm-'. Absorp- tion spectra, lifetimes and excitation spectra were measured as previously described." Results The luminescence spectrum of a crystal of Rb,ReC16 at 4 K using 457.9 nm excitation is shown in Fig.1. More than 20 features with half-height widths of ca. 10cm-' are readily resolved together with some broader features. The intensity of these features decreases rapidly as the temperature is raised, with the higher-energy bands showing the strongest tempera- ture dependence. The relative intensity of the lines also depends on the wavelength of line used for excitation. Careful examination of a large number of spectra measured under different conditions enables many of the more prominent lines to be divided into nine pairs where the relative intensity of the two lines is invariant with temperature and excitation wavelength.The separations between the members of these pairs are between 22 and 36 cm-' and the higher-energy member is, generally, 1.5-2 times more intense than the lower member. We propose that these pairs correspond to transitions occurring at nine distinct but closely related sites in the crystal; many of these pairs are identified in the figure. The features in the absorption spectrum of the A2ReC16 crystals in the region of the Ts(4A2g)+r7(2T2g) transition are similar to those in the absorption spectrum of the ReC1;- ion diluted into cubic crystal^,^." but less well resolved. They have a close mirror symmetry with the luminescence from the r7(2T2g)state of the diluted ReCl;-. From the absorption spectrum we estimate, using the ground-state values of v6 and v4 (131 and 165 cm-' respectively),' that the intrinsic elec- tronic origin of the Ts(4Az,)-+r7(2T2,) transition of Rb2ReC16 (unobserved) is close to 13 875 cm-'; this estimate is likely to be correct within a few wavenumbers since the vibrational wavenumbers of the internal motions are expected to be XI J.MATER. CHEM., 1991, VOL. I closely similar in the ground and excited electronic states as they are both derived from the t;, configuration. The luminescence of Rb2ReC16 does not closely resemble that of the ReC1;- ion, diluted into cubic crystals. The highest- energy pair is weak and observed only with 514.5 nm exci- tation at temperatures near 4 K. However the wavenumbers of this pair, 13 759 and 13 724 cm- ',are within a few wave- numbers of the expected position of the v6 and v4 vibronic origins of the intrinsic emission from the r7(,Tzo) state.A difficulty with assigning these bands to the intrinsic emission is that the band that would be assigned as vg is more intense than that due to v4, whereas the reverse is the case for the intrinsic absorption. This intensity distribution matches that of the defect emission to low energy (see below). We cannot rule out the possibility that the intrinsic emission makes at least some contribution to the intensity of the bands, but it seems more likely that the emission is due to a very shallow trap. The prominent features at 13 731 and 13 702 cm-' are assigned as v6 and v4 vibronic origins on an unobserved electronic origin, which is calculated to occur near 13 862 cm- using the ground-state vibrational frequencies. The difference between the wavenumber of this origin and that of the intrinsic origin is too large to be attributed to the sum of experimental error plus possible errors in the excited- state wavenumbers of the vibrational modes.It is likely, therefore, that this prominent emission arises from a shallow trap rather than the intrinsic level. As before, support for this view again comes from the relative vibronic intensities of v6, v4 and v3. Because the r7(2T,,)+r,(4A2g) transition is intra- configurational t;,-+t$, vibronic intensity distributions for the same species are expected to be similar. The v3 vibronic origin is expected to be near 13 555 cm- 'but is not unambigu- ously observed owing to the presence of other weak emissions near this wavenumber, but the intensity of this mode is clearly much lower than that of the other two vibronic origins.In the absorption spectrum v4 is somewhat more intense than v3 whereas v6 is weaker than either v4 or v3. This is similar to the intensity distribution for the r7(2Tz,)-r,(4A2,) emis-7 I I 1 II I I I I Fig. 1 Luminescence spectra of Rb,ReCl, at 4.0 K excited at (a)514.5, (b)457.9 nm. Note the changes in the abscissa scale. The v6, v4 vibronic origins (and vj where observable) on most of the traps are indicated by the bars J. MATER. CHEM., 1991, VOL. 1 sion of the dilute crystals containing ReCl;-.For the emission based on the 13 863 cm-' origin and indeed for all other emissions in the pure crystals, the intensity distribution is v6>v4 whereas v3 is much weaker and difficult to resolve from the other emission in the 13 550 cm-' region. The analysis of the remainder of the spectrum follows similarly; the seven further pairs of features identified to low energy are each readily assigned as v6, v4 vibronic origins of the r7+r8 luminescence transition derived from different traps, and study of the weaker features enables v3 vibronic origins to be identified for most of these. In three cases very weak progressions in v1 can be located based on the v6 and v4 vibronic origins. The resultant analysis (Table 1) accounts for the position and relative intensity of all observed spectral features. For most of the nine traps the relative intensity distribution (v6 >v4 > >v3) is similar, but different from that of the same transition in both the intrinsic absorption spec- trum and the emission from ReClg- in dilute crystals.Where bands are well separated v6 is significantly sharper than v4. All the vibrational frequencies are similar on the eight sites but the small differences appear to be outside experimental error. In no case is the pure electronic origin observed, nor are any splittings of the vibronic origins detected. Thermal cycling of the crystal to 80 K followed by cooling to 4 K resulted in a marked broadening of all spectral features so that the weaker traps could not be resolved and on examination at room temperature the crystal was found to be shattered.Presumably there is a structural phase transition in Rb2ReCl6 in the 4-80 K region analogous to those observed in K2ReC16 at 110 and 113 K.4 A second crystal gave only the broadened spectrum suggesting that it underwent the phase transition during the initial cooling. The luminescence behaviour of Cs2 ReCI, is closely similar to that of the Rb salt, although somewhat less well resolved and with corresponding features at 50-75 cm- ' lower energy. The v6, v4, v3 intensity distributions of the trap emission are comparable to those of the Rb salt although v6 is even more prominent. For K2ReCl6 all features are very much broader so that only the most prominent trap gave a clearly defined maximum but the luminescence behaviour was consistent with that of the Rb salt.Somewhat unexpectedly, both (NR4)2ReC16 (R =CH3, C2H5) luminesce at 4K, although more weakly than the alkali-metal salts. The spectra are different in that the pure electronic origins associated with the various traps are the most intense features, although relatively weak v6 and v4 vibronic origins could be detected. No emission from [N(n- C4H9)4] 2 ReC1, was observed. Attempts were made to elucidate the excitation process by measuring the excitation spectrum and decay curves of the emission. No sharp features were observed in the excitation spectrum between 520 and 470 nm (19 230 and 21 276 cm-') Table 1 Vibrational analysis for the r,(2T2&-*r,(4AzB) trap emission for Rb,ReCl, at 4 K (457.0 and 514.5 nm excitation) no.trap energy/cm-'" '6 v4 v3 "1 13 875 126' 15Ib 13 862 131 160 315, 326 360 13 812 134 157 322 -13 721 135 160 322 -13 640 131 160 ---13 494 133 159 319 13 480 129 161 3 14 343 13 428 129 165 315 352 13 410 133 155 "Derived from the vibronic structure, in no case is the pure electronic origin observed; bobserved only in the 514.5 nm excited spectrum. for any of the emission bands. Each emission had an exponential decay curve with a strongly temperature-depen- dent lifetime in the region of 100 ps at 4 K decreasing to a few ps at 30 K. Discussion The lowest-energy charge-transfer state2' of the ReCl2- ion lies at ca.29 700 cm-'. The 4T2, crystal-field state probably lies in the same spectral region but has not been observed. The highest ti, single-ion crystal-field state is T8(2T1,) at ca. 15400cm-'. Between these regions we have identified a number of co-operative absorptions in K2ReCl6 and Cs2ReC16 involving simultaneous excitation of two adjacent ions to combinations of the five ti, excited state in Oh*symmetry." Similar co-operative absorptions here have also been observed for Rb2ReC16. The resultant 15 pair states are likely to contribute weak absorption bands in the 16 000-31 000 cm-' region. There are even weaker features due to v1 progressions based on these pair states. The highest-energy feature observed in our single-crystal absorption spectra is a very weak feature at 18 602 cm-'.Subsequently, two further pair transitions at 21 570 and 21 920cm-' have been resolved21 [I'8(2T1,)+ I'7(2T2B) plus the v1 progression on this state]. A band due to the 2r8(4A2g)+2r6(2T1g) transition expected at 18 900 cm-' was not observed. The absorbance of the crystals at 19 436 cm-' (the 514 nm argon laser line) is extremely low; at higher energy there is a weak rising background but the absorbance of this does not become comparable to the co- operative absorbance until 21 300 cm- '. Nevertheless, both the 514 and the 457 nm (21 839 cm-') laser lines excite intense and generally similar luminescence (but with some variation in the intensity due to individual traps).We assume that the excitation process in our experiments is into weak, broad, defect-induced absorption bands. These defect levels relax rapidly to the r7(2T2g) band and thence to the observed traps. We emphasize the remarkable intensity of the emission when the excitation process is into such weak absorption bands. Nine different traps up to 365cm-' below the T7(?Tig) band have been identified in Rb2ReC16, and it is likely that a similar range of traps occur in both Cs2ReCl6 and K2ReC16. In each case, trap emission is characterized by a vanishingly weak electronic origin, a vibronic origin intensity distribution of the type v6>v4> >v3, no observable splittings of the vibronic origins, a v6 vibronic origin which is significantly sharper than v4 and v3, and very weak v1 progressions. These features show that the defects are essentially centrosymmetric ReC1;-ions.The trap depth is an approximate measure of the decrease in Re4+ interelectron repulsion at the trap site relative to the intrinsic site. The traps are therefore likely to be due to ReC1;-ions at compressed lattice sites near interstitial species or dislocations. These dislocations may also be the source of the absorption in the 19436cm-' region. As the temperature is raised, the intensity from the shallower traps decreases more rapidly than that from the deeper traps, as expected. The vibronic intensity distribution of the trap emission is of particular interest. It is quite different from that of isolated ReClg- ions.This raises a number of intriguing questions concerning the source of the vibronic intensity.22 We thank G. Sperka for preparing the crystals of Rb2ReC16. We also thank NATO and CNR for the award of fellowships to M.B. References 1 A. M. Black and C. D. Flint, J. Chem. SOC., Faraday Trans. 2, 1977, 73, 877. 440 2 C. D. Flint and A. G. Paulusz, Mol. Phys., 1981,43, 321. 3 P. B. Dorain, in Transition Metal Chemistry, ed. R. L. Cariin, Dekker, New York, 1968,vol. 4,p. 1. 4 G.P. O’Leary and R. G. Wheeler, Phys. Rev. B, 1970, 1,4409. 5 B. N. Figgis, J. Lewis and F. E. Mabbs, J. Chem. SOC., 1961, 3138. 6 R. H.Busey and E. Sonder, J. Chem. Phys., 1962,36, 93. 7 H. G. Smith and G. E. Bacon, J. Appl.Phys., 1966,37, 979. 8 V. J. Minkiewicz, G. Shirane, B. C. Frazer, R. G. Wheeler and P. B. Dorain, J. Phys. Chem. Solids, 1968,29, 881. 9 H. D. Grundy and I. D. Brown, Can. J. Chem., 1970,48,1151. 10 R. H. Busey, H. H. Dearman and R. B. Bevan Jr., J. Phys. Chew., 1962,66, 82. 11 M. Bettinelli and C. D. Flint, J. Phys. C, 1988,21, 5499. 12 M. Bettinelli, L. Di Sipio, G. Ingletto, A. Montenero and C. D. Flint, Mol. Phys., 1985,56, 1033. 13 M. Bettinelli, L. Di Sipio, G. Ingletto and C. D. Flint, Chem. Phys. Lett., 1970,138, 361. J. MATER. CHEM., 1991,VOL. 1 14 G. Sperka and F. A. Mautner, Cryst. Res. Technol., 1988, 23K, 109. 15 K. W. Bagnall, D. Brown and R. Colon, J. Chem. SOC., 1964, 30 17. 16 M. Bettinelli, L. Di Sipio, G. Valle, C. Aschieri and G. Ingletto, 2. Kristallogr., 1989,188, 155. 17 V. Spitzyn, A. 1. Zhirov, M. Yu Subbotin and P. E. Kazin, Russ. J. Inorg. Chem. (Engl. Transl.), 1980,25, 556. 18 G. Sperka, H.P. Fritzer, M. Bettinelli and C. D. Flint, Spectro-chim. Acta, Part A, 1988,44,1377. 19 M. Bettinelli and C. D. Flint, Chem. Phys. Lett., 1990, 167, 45. 20 J. C. Collingwood, S. B. Piepho, R. W. Schwartz, P. A. Dobosh, J. R. Dickinson and P. N. Schatz, Mol. Phys., 1975,29, 793. 21 K.Gatterer, personal communication, 1990. 22 R. Acevedo and C. D. Flint, Theoret. Chim. Acta, in the press. Paper 0/05647B;Received 17th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100437

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 441-445 Generation of Charge Carriers and an H/D Isotope Effect in Proton-conducting Doped Barium Cerate Ceramics Robert C. T. Slade* and Narendra Singh Department of Chemistrx University of Exeter, Stocker Road, Exeter EX4 4QD, UK Ionic conduction by H+ and D+ in doped perovskites BaCe,-,M,O,-, (M=Y, Gd; x=0.05, 0.10) at high temperatures in H,O-and D,O-moistened atmospheres is investigated. Conduction by hydrogen ions was confirmed by emf measurements on various gas cells using specimen ceramics as the solid electrolyte membrane separating moist and 'dry' nitrogen atmospheres. Ionic charge carrier densities in these materials are very low, as demonstrated by thermogravimetric and infrared measurements. Electrical conductivity (a) measurements were carried out in the temperature range T= 600-900 "C using complex impedance techniques.Plots of log aT versus 1/T showed Arrhenius behaviour in all cases. Attempt frequencies for charge-carrier migration correspond to deformation of the Ce/M-0-H bond angle, with H+ migrating between sites corresponding to attachment to neighbouring oxide ions. An isotope effect is seen in the activation barriers, €, (ca.33 kJ mol-' for H+ and ca. 43 kJ mol-' for D+). This results from tunnelling of charge carriers through a time-dependent barrier arising from coupling to lattice phonons. Keywords: Perovskite; Ionic conduction; Thermogra vimetric analysis; Infrared spectroscopy Doped barium cerate ceramics BaCe, -,M,03 (M =Y, Yb, Nd, La, Gd; x=O.O5, 0.10) exhibit protonic conduction (i.e.migration of H+ ions) in moist atmospheres at high tempera- tures (600-1000 OC).' These materials have a perovskite-type structure with Ce/M being octahedrally co-ordinated by oxygens. CI includes oxygen-ion vacancies. Other perovskite oxides have also been demonstrated to conduct ionically (when doped) via proton formation/incorporation, e.g. SrCe03,2-5 BaCe03,6 KTa03.'v8 The formation of con-ducting protons has also been studied in several other types of doped oxide in H2-containing atmospheres (e.g. Zr02,9.10 A1203," Y203,12 Th0213); these studies do not all involve a moist atmosphere. Protonic conductivities in other ceramic systems appear considerably lower than those associated with doped A"B"03 perovskites.l4 Protonic conduction in doped perovskites results from the oxygen vacancies which are introduced on substitution of lower valency cations M"' on some Ce" sites.The protons are said to arise from high-temperature equilibria between the condensed phase and a moist (H20-containing) atmos- phere:" viz. Vo" +to2400+2h' (1) H20+2he+2H.+302 (2) H2O+V0'*42H'+Oo (3) V;; denotes an oxygen vacancy, Oo an oxide ion at a normal lattice site, H' a proton (H') and h' a hole (electronic). Reaction (2) is the reaction of water with holes (localised on Ce", reaction giving Ce"') to generate H', which will co- ordinate to oxides to give hydroxyl ions (co-ordinated to octahedral metals, uiz. Ce/M). If this were a significant route for generation of H+ charge carriers, undoped BaCe03 would have a protonic conductivity comparable to the doped mater- ials; this is not the case.The generation of charge carriers in these systems can be described simply as water molecules reacting to fill vacant oxygen sites in the doped perovskites, thereby generating further hydroxyl ions as follows 00+Vo'. +H20 =2[OH]o (4) The protons then migrate/conduct by motion between sites, corresponding to attachment to adjacent oxygens. Studies of protonic conduction in solids have excited con- siderable interest internationally owing to the potential appli- cations of proton-conducting solid-electrolyte membranes in many novel electrochemical devices, e.g. advanced fuel cells, H2 sensors, steam electrolysers for H2 production and elec- trochromic optical displays.Proton-conducting materials at near ambient temperatures are commonly hydrated, e.g. solid acids and acid salts (acid phosphates, acid sulphates, hetero- polyacids) or layered double hydroxides. The hydrating water is easily lost when the temperature is raised, with a subsequent drop in the conductivity of the electrolyte by several orders of magnitude (the hydrogen-bonded conduction pathway, and subsequently the entire structure, is destroyed). Oxide ceramics in which H+ conduction can be induced, on the other hand, offer structural and chemical stability at high temperatures. In the present study, the electrical properties of the doped ceramics BaCeo~g5Yo~050, -,have-,and BaCeo,90Gdo~lo03 been studied at high temperatures in atmospheres moistened with H20 and D20.Thermogravimetric analyses have detected changes in mass on reaction of powders in atmos- pheres and at temperatures characteristic of sintered device components, and the concentrations of charge carriers have been determined. An H/D isotope effect arising from the protonic nature of the conduction is reported. Experimental Doped perovskites BaCeo.95Yo.0503-, and -aBaCeo~goGdo~lo03 were prepared as described previously.' Y203 or Gd2O3, Ce02 (Aldrich, 99.9%) and BaC03 (BDH) were mixed in the stoichiometric ratios accord- ing to XBaCO +(1-x)Ce02+-M203 = 2 BaCe1-,M,03-a+C02 (M=Y, Gd) The powdered dry raw materials were mixed, ground and calcined in air in alumina crucibles at 1250 "C for 15 h.The products were characterised using X-ray powder diffraction (Philips diffractometer, Ni-filtered Cu-Ka radiation) to estab- lish the formation of single perovskite phases. Values of the cubic unit-cell constant a were 4.414(3) (BaCeo.95Yo.0503 -a) and 4.422(2) A (BaCeo~90Gdo~lo03-J. Thermogravimetric measurements on perovskite powders employed a Stanton Redcroft TG-750 thermobalance with flowing N2(g). Infrared spectra (KBr discs) were recorded using a Perkin- Elmer 88 1 instrument for (i) the dry 'raw' perovskite powders, (ii) powders treated in flowing moist N2(g) (passed through an H20 or D20 bubbler at ambient temperature) at 900 "C for 10 h.To form ceramic discs for the electrochemical studies the calcined powders were reground and pressed hydrostatically (10 tonnes) into pellets (16 mm diameter, ca. 1 mm thickness), with 1% by mass of camphor as volatile binder. Pellets were sintered in air at 1500 "C for 10 h. Porous platinum paste (Engelhard A4338) was applied (as a paint) to pellet faces as electrodes. Electrical conductivity measurements employed a.c. impedance techniques using a Hewlett-Packard HP 4192A LF impedance analyser (operating in the range 5 Hz-100 kHz) programmed via an IBM-compatible computer for data collec- tion and analysis (employing the EQUIVCRT modelling software of Boukamp). The purpose-built electrochemical cell assembly is shown in Fig.1. The whole assembly was mounted horizontally in an electric-tube furnace programmable down to 1 "C min-'. Two cell compartments were separated by the ceramic disc and individual controlled gas flows [either moist or 'dry' N2(g)] fed to each compartment. 'Moist' gas was produced by passage through a water (H20 or D20)bubbler at ambient temperature, whereas 'dry' gas was produced by passage through an activated molecular-sieve column (resulting in a very much lower partial pressure of water). A gold wire ring gasket provided the seal between the compart- ments, requiring the electrode/electrolyte assembly to be heated to 950 "C for ca. 1 h to obtain a gas-tight seal prior to electrochemical studies. Electrochemical potentials were measured with a Kiethley 6 14 electrometer.Electroly Disc Pt wire FURNACE ~ a Tube Au Ring Gasket Fig. 1 Schematic diagram of the electrochemical cell used in emf and conductivity measurements in the temperature range 600-900 "C(see text). Top, the gas flow paths; bottom, the electrode assembly J. MATER. CHEM., 1991, VOL. 1 Thermogravimetry Perovskite powders were heated (5 "C min-') to 900 "C and then soaked at that temperature for 10 h, all in moist flowing N,(g) (passed through an H20 bubbler at ambient tem-perature). The mass of each material initially decreased rapidly on heating to 900 "C. In an earlier study of doped strontium cerate, O2evolution on heating was detected by gas chroma- tography and was assigned to the reversal of reaction (1) on heating in the absence of 02(g), thereby introducing more oxide vacancies and reducing the mean oxidation state of Ce. This explanation is unsatisfactory in that it implies a higher mobility for 02-ions than for the H+ involved in the subsequent equilibration (see below) which generates charge- carriers.An alternative explanation is that some oxygen is incorporated in dioxygen anions such as 0;-(barium per- oxide is a known compound) according to a reaction such as VO" +00 +302 =[02]20" (5) Self-diffusion of excess 0 atoms could then occur via a rapid bond-exchange (Grotthus-like) mechanism, which would not be a conduction mechanism (0;-+02-=02-+0;--transfer of neutral atoms). Generation of peroxide ions would not be accompanied by reduction of the cerium oxidation state.The amounts of oxygen evolved during heat- ing corresponded to a loss of 4.0 & 0.1 mol% of incorporated 0 atoms for BaCeo.90Gdo~lo02.95 and 3.0k0.1 mol% for BaCeo~95Yo~0502.975.No electronic-conductivity contribution was detected in the impedance studies described below. This is an important observation supporting the postulate of the incorporation of excess oxygen without reduction in the mean Ce oxidation state; the presence of an appreciable concen- tration of Ce"' would be expected to lead to significant electronic conductivity (by electron hopping) analogous to that in the defect fluorites Ce02-x. During soaking at 900 "C, the powders slowly increased in mass over a few hours to a stable final value, this arising from reaction (4) (reaction with H20).The observed mass changes corresponded to H contents for the final products at 900 "C of (i) 0.90f0.09 mol% for perovskite -aBaCeo~90Gdo~,o03 and (ii) 0.64 f0.09 mol% for BaCeo.95Yo.0503-a(the relatively large errors are due to the very small mass changes involved in this second stage, which introduces the charge carriers). The ionic charge-carrier con- centrations in these materials are thus very low. Analogous uptake by doped strontium cerate has been demonstrated by subsequent heating and gas chromatographic detection of evolved H20.I6 The thermogravimetric (TG) studies above were then repeated with replacement of H20 (in the bubbler) by D20.Initial mass losses (due to O2 loss) were very similar to the values above. Observed mass increases on high-temperature soaking were higher, for each material, for uptake of D20 than for uptake of H20. For instance, the observed mass change for perovskite BaCe0.,,Gdo. -a corresponded to a D content for the final product at 900°C of 1.43f0.09 mol%. The charge-carrier concentrations are thus higher in deuterated materials than in hydrogenated ones, indicating a shift to the right in equilibrium (4) on moving from H20 to D20 moistening of the flow gas (all other conditions being the same). Infrared Spectra Infrared spectra for the dry as-synthesised powders were very similar to those of the materials soaked at high temperature in moist flowing N2(g).Fig. 2 shows the spectra for as- J. MATER. CHEM., 1991, VOL. I I....!.,....I 4000 3200 2000 1600 1000 600 wavenumber/cm -' Fig. 2 Infrared spectra of as-synthesised BaCe,~,,Gd,~,,03 -a (top) and of the same material soaked for 10h in H,O-moistened flowing N,(g) and then for a further 5 h in D,O-moistened flowing N,(g) (bottom), both at 900 "C synthesised BaCeO~90GdO~1003 -aand that of the same material soaked for 10 h in H,O-moistened flow gas and then for a further 5 h in D,O-moistened flow gas, both at 900°C. Electrical Conductivity Conductivity measurements for ceramic discs BaCeo~90Gdo~lo03 -a were made in -a and BaCeo.95Yo.0503 flowing N2(g) in the temperature range 600-900 "C after pretreatment at T>600 "C in the same N2(g) (moistened with H20 or D20, as appropriate for studies of Hf or D+ conductivity) for 10 h.The values of the bulk electrolyte resistances were evaluated from complex impedance plots. Some typical impedance spectra are shown in Fig. 3 (for BaCeO~,,GdO~,,O,-a with D+ conduction). Spectra in the case of BaCeo~g5Yo~os0,-a were very similar in form. Such spectra are typical of ionic motion in solids with reversible electrodes. At T<700 "C a well defined depressed semicircle at higher frequencies is seen, with a linear impedance variation at lower frequencies. The impedance variations at low fre- quencies indicate a diffusion-related phenomenon, assigned in 15 10 . 5 ,/" Contrary to the assertion of Shin et a1." in their studies of 01'1.Y20,-doped SrZrO, and SrCe03, no information results as 30 40 soto the location of H in the conducting ceramic.This is not surprising in view of the very low H contents evident in the TG studies above. EMF Measurements Emfs of the cells using sintered discs of BaCeo~90Gdo.,o0,-a and BaCeo.95Yo.os03 -a were measured using the electro- chemical cell assembly (Fig. 1) at T2600 "C. Stable emfs were ru Iobserved on passage of moist (H,O) and 'dry' nitrogen gas through different compartments of the cell for a few hours. The negative electrode was that in the 'wet' compartment and the polarity of the cells reversed on interchanging the moist and 'dry' gas flows. This behaviour of the cells is explicable only in terms of the ceramic diaphragm acting as a protonic 16 20 24 28conductor.The emf arises from the different partial pressures of H20 in the two cell compartments and the following electrode reaction: H20(g)=2H+ (electrolyte) +302(g)+2e-(6) 4 At high temperature the equilibrium H20(g) =H2(g)+302(g) 3might be established and 02(g) could be available for gener- ation of possible 02-charge carriers by the reaction 2302(g)+2e-=O2-(electrolyte) (7) 1If this were the dominant electrode reaction and conduction was by 02-migration, the compartment with the higher partial pressure of water would have the higher partial pressure of oxygen and contain the positive electrode (contrary to experimental observation). Observed emfs decreased as the temperature increased, showing an electronic contribution to the conductivity to increase at higher temperatures. The observed emfs were for BaCeo~90Gdo~ro03-a: 65, 40, 21 mV at 600, 700,800 "C, and for BaCeo.95Yo.os03-a: 45,38, 18 mV at 600,700, 800 "C.I .0 1 I * 10 12 14 16 Z'/R Fig. 3 Typical complex-plane impedance plots as a function of tem- perature for BaCeo~,,Gd,-,,,03 -a in flowing N,(g) moistened with D,O (by passage through bubblers at ambient temperature). (a)600, (b) 700, (c) 800 "C 444 this case to diffusive processes in the electrode (Pt) or at the electrode/electrolyte interface. The depressed semicircular arc at higher frequencies arises from electrical charge transfer at the electrode/electrolyte interface, this giving rise to a charge- transfer resistance in parallel to the electrode double-layer capacitance at the interface.As anticipated, the value of the charge-transfer resistance gradually decreases with increasing temperature and finally vanishes at higher temperatures (T> 700 "C). No additional arcs corresponding to grain-boundary conductivity were observed. The bulk electrolyte resistance at a given temperature differed when measurements corresponded to H+ and D+ migration, i.e. in H20-and D20-moistened N2(g) flows. From the value of the bulk electrolyte resistance, R, the electrical conductivity, Q is evalu- ated using the expression IS=t/RA (8) where t is the thickness of the ceramic disc and A is the electrode cross-sectional area.Temperature dependences of the conductivities are shown in Fig. 4 for discs of BaCeo~90Gdo~1003-aand BaCeo~,,Yo~o,O,-a and for H+ and D+ conduction. Measure- ments refer to conductivities of the same discs in H20- and D20-moistened flowing gas and were fully reproducible on (i) temperature cycling and (ii) multiple cycles between H20 and D20 moistening of flow gas. Furthermore, very similar 1.2I I ' I0.2 0.8 0.9 1.0 1.1 1.2 103 KIT 1.0 0.8 0.6 0.4 0.2 I.0.0' * ' . ' 0.8 0.9 1.0 1.1 1.2 1.3 103 KIT Fig. 4 Temperature dependences of conductivities arising from H + and Df migration in ceramics BaCe,-,M,O,-, in moist (H20 or D20) flowing N2(g) (see text). Top, BaCeo~goGdo,,o03~Q; bottom, BaCeo,g5Yo~0503~~,.Solid lines correspond to the Arrhenius param- eters E, and A given in Table 1 H+ in the same host material (see Table 1).H+/D+ transfer J. MATER. CHEM., 1991, VOL. I data were obtained for different discs of the same materials. In all cases, the conductivity exhibits an Arrhenius-type T dependence explicable in terms of a thermally activated pro- tonic charge migration with aT= A exp (-E,/R'I) (9) The evaluated activation energies, E,, and pre-exponential factors, A, are given in Table 1. It is clear that an isotope effect is present in the ionic conduction occurring in both ceramic membranes investigated, with E, depending strongly on the migrating species (H+ or D'). Origins of Proton Mobility The H+ conductivity is related to the H+ self-diffusion coefficient D by the Nernst-Einstein equation Q =(Nq2/kBT)D (10) where N is the number density of charge carriers, q is the charge per carrier and kB is the Boltzmann constant. For a three-dimensional random walk, D is related to the jump frequency v by the equation D =A2v/6 (1 1) where 2 is the distance between neighbouring sites (taken as the O...Oseparation in this case).For an activated jump v =v, exp (-E,/R7) (12) Combining eqn. (8)-(lo), QT=(Nq2A2V,/6k~)exp (-E,/R'I) (13) Comparison of eqn. (9) and (13) gives an expression for the Arrhenius prefactor A, linking it to structure, migration pathway and jump-attempt frequency (also given in ref. 18) A =Nq2A2v,/6kB (14) The attempt frequency v, is thus given by V, =6AkB/Nq2i2 (15) For BaCeO~,,GdO~,,O, the experimental data give v, = 1.3 x lOI3and 1.8 x lOI3 Hz for H+ and D+ migration, respect- ively.This frequency is comparable to that for vibrations of metal-co-ordinated hydroxyl groups, and in particular to that for deformation of the Ce/M-0-H bond angle ('wagging', hydroxides being co-ordinated to the octahedrally co-ordi- nated metals in the structure). One might therefore anticipate that v, would be slightly lower (by a factor of ca. J2) for D than for H, a consequence of the higher reduced mass in the D case. However, both A and N are subject to experimental errors (A,in particular, being deduced by extrapolation) and the agreement between the deduced values is thus remarkable.It is evident in Table 1 that A is higher for D+ conduction than for H+; this peculiarity is a simple consequence of the higher carrier density, N, in the case of D', the attempt frequencies v, being similar. Activation energies for D+ migration are higher than for Table 1Arrhenius prefactors A and activation energies E, for H+/Df conduction in doped barium cerates BaCe, -xM,O, -a proton conduction deuteron conduction ceramic E,/kJ mol-' log (A/K S cm-') E,/kJ mol- log (A/K S cm-') ~ ~ BaCe0.90Gd0.1003 -01 35.6 0.9 2.63 k0.05 43.9 k0.7 2.96 f0.04 BaCe0.95Y0.0503 -a 30.6 k0.6 2.24 k0.04 43.2 1.7 2.78 f0.09 J. MATER. CHEM., 1991, VOL. 1 is believed to occur by tunnelling between sites corresponding to attachment to different oxygens.The question of the origin of E, therefore necessitates consideration of activated phenom- ena enabling tunnelling. Co-operative motions of the structure (phonons) will lead to modulation of O...O separations, with consequential modulation of the barrier to migration and hence of the frequency of tunnelling between sites (i.e.through the barrier). Therefore, we have combined variations of both the O...O separation (by phonons) and the Ce/M-0-H bond angle (the 'wagging'), the nature and potential associated with the latter differing in H and D cases. Phonon spectra are temperature dependent, as required. It is well known that the proton tunnels more readily than the deuteron (see for example ref.19), the tunnelling probability decreasing rapidly with mass. Combination of the above factors leads to a stronger temperature dependence for D+ of the frequency of achieving a critical O...O separation (to enable D+ to tunnel with a given probability), and hence to the observed relative activation energies. The discussion of the preceding two paragraphs omits consideration of the effects of charge-carrier introduction uia an equilibrium [reaction (4)]. A small variation in charge carrier (H+) concentration with temperature has been noted in studies of analogous doped strontium cerates.16 It follows that E, values are likely to contain a small contribution from the equilibrium. Further consideration (e.g. by computational chemistry) of the origins and variations of E, are beyond the scope of this study.Shin et all7 studied H+ and D+ migration in Yz03-doped SrZrO, and SrCeO,. From conductivity ratios at given tem- peratures, they postulated an isotope effect, but dismissed the possibility of quantum effects (tunnelling). Their discussion and study did not, however, include consideration of vari- ations of activation energies E, with isotope exchange, nor the values of Arrhenius prefactors A. In particular, single- temperature ratios of conductivity have little meaning in the presence of E, variations. Scherban and Nowick" working on Yb,O,-doped SrCeO, at T1400"C (lower than in this study aimed at likely operating temperatures for devices) detected higher activation energies for D+ than for Hf migration, closely parallelling the results of this study.The observed isotope effects are, in themselves, primary evidence that the migrating species is H+, rather than OH-or even 0'-. It should be noted, however, that these systems are closely related to other defect perovskites which have been characterised as O2-ion conductors." Under appropriate conditions and at higher temperatures, the materials in this study would be expected to also conduct 02-ions (with E, >75 kJ mol- for that process2o), as postulated by Bonanos et a1." It has been the purpose of this study to work under conditions stimulating conduction by H only.+ Conclusion H conduction in M20,-doped barium cerates arises from + incorporation of water into the perovskite structure, H20 reacting with oxygen vacancies and an oxide ion to form two hydroxyl groups.Whereas others have discussed the low electronic conductivities of these materials in terms of low concentrations of Ce"' and small polaron motion,18 it is likely that excess oxygen is incorporated in these materials in dioxygen anions (without reduction of the mean Ce oxidation state). H+ migration is by hopping between sites correspond- ing to attachment (in hydroxyl groups) to neighbouring oxy- gens. The attempt frequencies measured correspond to deformation of the Ce/M-0-H angle, necessary for migration of H+ over the oxide network. An isotope effect is seen, activation energies E, being higher for D+ migration than for H +.This arises from interaction with lattice phonons, the jump involving tunnelling (with H tunnelling more + readily than D').A conductivity-limiting feature of these materials is the low charge carrier concentration. H -conductivities would be expected to rise at high steam pressures (e.g.in water electroly- sers) by driving reaction (4) further to the right, thereby increasing the charge-carrier density. Studies are in hand to investigate optimisation of the charge carrier density. We thank SERC and British Gas plc for co-funding this project and for permission to announce these results. N.S. thanks Gaya College (Magadh University, India) for study leave. References 1 R. C. T. Slade and N. Singh, Solid State Ionics (Proceedings of the 5th European Workshop on Solid State Protonic Conductors- Assisi, September 1990), in the press.2 H. Iwahara, T. Esaka, H. Uchida and N. Maeda, Solid State Zonics, 1981, 3/4, 359. 3 H. Uchida, N. Maeda and H. Iwahara, J. Appl. Electrochem., 1982, 12, 645. 4 H. Uchida, H. Yoshikawa and H. Iwahara, Solid State Ionics, 1989, 35, 229. 5 H. Uchida, H. Yoshikawa, T. Esaka, S. Ohtsu and H. Iwahara, Solid State Ionics, 1989, 36, 89. 6 H. Iwahara, H. Uchida, K. Ono and K. Ogaki, J. Electrochem. Soc., 1988, 135, 529. 7 W. Lee, A. S. Nowick and L. A. Boatner, Solid State Ionics, 1986, 18/19, 989. 8 S. Q. Fu, W-K. Lee, A. S. Nowick, L. A. Boatner and M. M. Abraham, J. Solid State Chem., 1989, 83, 221. 9 S. Strotz and C. Wagner, Ber. Bunsenges. Phys. Chem., 1966,70, 781. 10 C. Wagner, Ber Bunsenges. Phys. Chem., 1968, 72, 778. 11 M. M. El-Aiat and F. A. Kruger, J. Appl. Phys., 1982, 53, 3658. 12 T. Norby and P. Kofstad, Solid State Zonics, 1986, 20, 169. 13 D. A. Shores and R. A. Rapp, J. Electrochem Soc., 1972, 119, 300. 14 T. Norby, Solid State Zonics, 1990, 4/41, 857. 15 H. Uchida, H. Yoshikawa and H. Iwahara, Solid State Ionics, 1989,34, 103. 16 H. Uchida, H. Yoshikawa, T. Esaka, S. Ohtsu and H. Iwahara, Solid State Zonics, 1989, 36, 89. 17 S. Shin, H. H. Huang, M. Ishigame and H. Iwahara, Solid State Zonics, 1990, 4/41, 910. 18 T. Scherban and A. S. Nowick, Solid State Ionics, 1989, 35, 189. 19 P. W. Atkins, Physical Chemistry, Oxford University Press, Oxford, 3rd edn., 1986, p. 320. 20 X. Turillas, A. P. Sellars and B. C. H. Steele, Solid State Zonics, 1988,28-30,465. 21 N. Bonanos, B. Ellis, K. S. Knight and M. N. Mahmood, Solid State Zonics, 1989, 35, 179. Paper 0/05806H; Received 27th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100441

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 447-450 Matrix Surface Modification by Plasma Polymerization for Enzyme Im mo bi Iization Mehmet Mutlu,a*b Selma Mutlu,"Bb Mark F. RosenberglaPc John Kane,*' Malcolm N. Jonesc and Pankaj Vadgama" " University of Manchester, Medical School, Dept. of Clinical Biochemistry, Hope Hospital, Manchester M6 8HD, UK Hacettepe University, Chemical Engineering Dept., Beytepe Campus, 06532 Ankara, Turkey University of Manchester, Medical School, Dept. of Biochemistry and Molecular Biology, Manchester M13 9PT, UK Polycarbonate (PC) membranes have been treated with dimethylamine (DMA), mpentylamine (PA) or mheptylamine (HA) in a glow-discharge apparatus. Amino group concentrations on plasma-polymerized PC membranes were then assayed by binding radiolabelled [l-'4C]acetic anhydride to the membrane, followed by scintillation counting.For the different plasma-polymerized membranes different degrees of radiolabelled (1251) glucose oxidase or rennet binding were observed and this was found to be directly related to the surface amino-group concentrations for the appropriate membrane. Keywords: Surface modification; Plasma polymerization; Enzyme immobilization; Radiolabelled enzyme; [ 7-'4C]acetic anhydride Enzymes are biocatalytically active entities upon which the metabolism of all living organisms is based. With the technical methods in common use it is generally impossible to separate efficiently the dissolved or finally suspended native biocata- lysts from the products of bioconversion. Therefore, the loss of these valuable materials results in high product cost.The problem can be resolved by the use of immobilized biocatalysts, which, in addition, allow the process to be carried out continuously. A biocatalyst is termed 'immobilized' if its mobility has been restricted by chemical or physical means. This artificial limitation of mobility may be achieved by widely differing methods, such as binding the biocatalysts to one another or to carrier substances, by entrapment in a network of a polymer matrix or by membrane confinement. A variety of techniques have been developed for the immobi- lization of enzymes. 1-4 These methods find practical appli- cation in the fields of medicine, food, agriculture, chemistry and other disciplines.Covalent binding is one of the most popular immobilization techniques, as it is expected to pro- duce a close association between one biocatalyst and another, or between a biocatalyst and a carrier. Covalent binding is usually employed for coupling enzymes but not whole cells,5 the latter being too sensitive to the harsh reagent conditions demanded. A range of functional groups can be used for immobilization including, amino groups, as well as carboxy, mercapto, hydroxy, imidazolyl and phenolic groups. Some of these groups, e.g. SH and amino groups can react directly with appropriate groups of the carrier. Others, e.g. OH groups, have to be activated before they can react with a carrier group. The active site of an enzyme is often located deep within the molecule, and a solid phase prepared by coupling small ligands directly to a matrix can exhibit a low capacity due to steric hindrance between the matrix and the substance binding to the ligand.In these circumstances a 'spacer arm' can be interposed between matrix and ligand to facilitate effective binding. The aim of this study was to modify the matrix surfaces of PC membranes by exposure to plasma glow discharge of amino-group-containing entities (DMA, PA, HA) and to immobilize enzymes (glucose oxidase and rennet) using glutar- aldehyde as a coupling agent. Experimental Surface Modification of PC Membranes by Plasma Polymerization High-density PC membranes (HDPCM) with nominal pore sizes ranging from 0.01 to 0.05 pm (rated by manufacturer, 8 x 10' pores cm-2) were supplied by Poretics (Livermore, CA).DMA, PA and HA were obtained from Sigma (Poole, Dorset). The plasma polymerization was carried out in the system shown schematically in Fig. 1. The apparatus consisted of a glass-tube reactor with two copper electrodes mounted on the outside and a radiofrequency generator (RFG).The reactor was evacuated to 10-3-10-4 mbar. A radiofrequency of 13.6 MHz was applied, its power ranging from 0 to 100 W. Power losses were kept to a minimum by a matching network. The membranes were exposed to plasma media for 5 min at 10 W with DMA or PA or HA monomer fed at a flow rate of 80 cm3 min-'. Preparation of Radiolabelled Enzymes Glucose oxidase (E.C.1.1.3.4. from Aspergillus niger: 292 IU mg-' protein) and rennet [extracted from calf stomach, containing 80% NaCl] were obtained from Sigma (Poole, Dorset). The enzymes were labelled by 1251to a specific activity of 7 mCi mg-' by the Chloramine-T method described by Hunter and Greenwood.6 The radiolabelled needle mopnrerelectrode valve /--S I meter plasma reactor \ -=-vacuum electrode pump Fig. 1 Schematic representation of plasma-polymerization apparatus enzyme was purified by Sephadex-G25 column chromotogra- phy. A series of different concentrations for both enzymes (0.002, 0.02, 0.2, 0.6 and 2.0 mg cm-3) was prepared and 12'1- glucose oxidase and 1251-rennet were added to these solutions as a tracer just before commencement of the experiments.Determination of the Surface Amino-group Concentrations For each amine-modified PC membrane and a control mem- brane of unmodified PC, three sections of equal weight and surface area were placed in glass containers (7 cm3) taking care to ensure the membranes were as dry as possible. An equal volume of dry dimethyl sulphoxide (3 cm3) was added to each membrane. For each membrane type a range of [l-14C]acetic anhydride (Radiochemical Centre, Amersham, Bucks., UK; specific activity 30 mCi mmol -I) activities were added: 2.5, 5, 10 and 15 pCi corresponding to 0.008, 0.016, 0.032 and 0.048 mg of acetic anhydride, respectively. The membranes were agitated continuously for 3 h and washed exhaustively by rinsing with water overnight to remove any non-specifically adsorbed acetic anhydride.The extent of acetylation on each membrane, which should in turn reflect the availability of amino groups, was assessed by counting the 14C attached to the washed membrane. This was under- taken by adding 4 cm3 of scintillation fluid to each membrane followed by counting on a scintillation counter. To allow for [1-14C]acetic anhydride non-specifically adsorbed to the membranes, the counts bound to the control membrane were subtracted from those bound to the activated membranes prior to the calculation of the amino-group concentrations. The calculation involved relating the number of disinte-grations per minute (DPM) to the specific activity of the [1-'4C]acetic anhydride from which the number of moles of bound acetic anhydride and therefore the number of moles of amino group could be established.The calculation took into account that only one side of the membrane was coated with the amine and that only one of the terminal carbons of the [1-'4C]acetic anhydride was radiolabelled. Immobilization of Enzymes The mechanism of the immobilization of the enzymes on the plasma-treated surfaces is shown schematically in Fig. 2. The plasma-treated PC membrane surfaces (16 cm2 surface area) were activated by incubation with 10cm3 of 12.5% (v/v) glutaraldehyde solution for 24 h at 37 "C. In order to prevent stagnant film formation on the membrane surfaces, activation tubes were continuously mixed (Luckham, Rotatest Shaker, Model R 100, UK).The surface-activated membranes were then washed with distilled water to remove any unbound glutaraldehyde present on the surfaces. A series of enzyme solutions were prepared with different enzyme contents (0.02, 0.2, 2, 6 and 20 mg) in 10cm3 of buffer solution, and after the addition of the radiolabelled tracers ('251-glucose oxidase or 1251-rennet) the membranes were placed into the tubes and left to interact for 24 h at 37 "C. Again, in order to prevent stagnant film formation on the membrane surfaces, activation tubes were continuously mixed during the coupling process. At the end of the incubation period the membranes were removed and washed with a sequence of buffers, 0.1 mol dm-3 acetate buffer (pH 4), 0.1 mol dmP3 borate buffer (pH 8), 0.1 mol dm-3 acetate buffer (pH 4) and 0.5 mol dm-3 NaCl, to remove non-specifically bound protein.Quantification of Immobilized Enzyme on the Surface The assessment of the immobilized enzyme on the plasma- treated surfaces was carried out by counting the radioactivity J. MATER. CHEM., 1991, VOL. 1 DMA or -[-PAplain membrane or HA plasma polymerization a membrane glutaralde hyde surface activation nu membrane enzyme enzyme attachment nU membrane enzyme immobilized enzyme Fig. 2 Mechanism of enzyme immobilization on plasma-treated surfaces of each membrane on a y-scintillation counter (LKB, Wallac, Finland). In addition the activity of each enzyme solution was counted and the amount of the enzyme bound on eachcm2 of the membrane was thereby calculable.Results Determination of Surface Amino-group Concentration for Plasma-polymerization-treatedPC Membranes The binding of [1-14C]acetic anhydride of increasing concen- trations to various plasma-polymerization(P1zP)-treatedmem-branes is shown in Fig. 3. It should be stressed that at saturation the surface density of amino groups in PlzP(HA) and PlzP(AA) surfaces are ca. 3 and 4 times higher than the background levels, respectively. The results also indicate that 10 pCi of [l-14C]acetic anhy- dride was sufficient to saturate the reactive sites and therefore this activity was used to evaluate the amino-group concen- trations for subsequent membranes.The number of moles of acetic anhydride bound to the membrane was calculated and thus the number of moles of amino group present on each membrane was established after correction for non-specific binding to plain PC membranes. The results are shown in Table 1. It is clearly shown in the table that, the amino-group density is highest for the PlzP(DMA)-treated surface and lowest for the PlzP(HA) surfaces. 449J. MATER. CHEM., 1991, VOL. 1 I . . . I...I..._ 4 8 12 16 activity of [1-"'CJacetic anhydride/pCi Fig. 3 The effect of initial activity of [l-14C]acetic anhydride on disintegration rate for plasma-polymerization-treatedsurfaces. Type of surface: 0,control; 0,PlzP(DMA); 0,PlzP(PA); A, PlzP(HA) Table 1 The concentration of amine groups bound to various plasma- polymerized PC membranes plasma-polymerized PC membrane NHJmmol cm-2 HA 1.6 x lo-' PA 3.0 x lo-' DMA 16.0 x lo-* Immobilization of Enzymes The amount of two different enzymes (glucose oxidase and rennet) immobilized on the plasma-polymerization-treated surfaces are shown in Fig.4(a) and (b), respectively. For all types of surface, it can clearly be seen that the amount of bound enzyme increases as the initial enzyme content is increased without evidence of saturation. Fig. 4(a) shows that the plasma-polymerized DMA, PlzP(DMA), treated surface has the largest capacity to bind glucose oxidase. This capacity is ca. 18 times greater than for the plain mem- brane at the highest level of initial enzyme content (20 000 pg).N I E 0)a \8 E *-OI1.50 Q) R0 3-0 W 000L 3 250 initial glucose oxidase contentlpg 5.0 N I 4.5 ;(b) E 4.0 .-B , , , , . , , , ,a::-, .I o.oo 5000 10000 15000 20000 25000 However, this capacity is only ca. 3.2 times greater for rennet 7000, 1 [Fig. 4(b)]. In addition, rennet has a higher tendency to bind non-specifically to the plain PC surface than has glucose oxidase [Fig. 4(a)and (b)]. The correlation between the surface amine concentration of the various membranes at saturation with the amount of bound glucose oxidase at the surface (20 000 pg initial enzyme content) is shown in Fig. 5. In addition, Fig. 5 shows the percentage of available amino groups occupied by glucose oxidase.It can be seen that the amount of glucose oxidase bound increases with amino concentration. However, the glucose oxidase bound to the membrane, expressed as a percentage of the total available amino-group binding sites, decreases with increasing amine concentration. Discussion Plasma-plymerization-treated PC Membranes PC membranes are commonly used in biosensor technology as selective barriers to contaminating species. For this reason, these membranes were selected for surface modification by plasma polymerization and subsequent enzyme immobil-ization. In order to investigate the effect of the chain length, three different types of amino-group-containing monomers, DMA, PA and HA, were selected. In the coupling processes of enzymes or different types of protein, the length of the spacer arm is critical.If it is too short, the arm may be ineffective and the ligand may fail to bind substances in the sample. If it is too long, non-specific effects become pronounced and reduce the ~electivity.~ Plasma polymerization was selected as the method of surface modification for enzyme immobilization. The advan- tage of plasma polymerization is that a simple derivation step with high surface enzyme loading can be achieved without the use of complicated chemistry or the use of destructive reagents, which might compromise the structural integrity of the membrane. A satisfactory application of plasma polymeriz- ation for surface modification has been elaborated by one of the authors'.' and These studies conducted with blood-plasma proteins showed that plasma polymerization can be used to modify the surface of polyurethane bio- materials.Determination of Surface Amino-group Concentration of Plasma-plymerization-treatedPC Membranes Many organic species can form surface polymers under 'glow discharge' conditions, so-called 'plasma polymerization'. The application of this technique for coating metal surfaces was N-'E lo-' (400 0D surface amino-group density/mmol cm-* initial rennet content/pg Fig. 5 The relationship between surface amino-group density and (-) the amount of bound glucose oxidase per unit area of mem- Fig. 4 The effect of the type of animated surfaces by plasma polymer- brane and (---) percentage occupancy of amino groups by glucose ization on (a) glucose oxidase immobilization, (b) rennet immobiliz- oxidase (GOD) (at 20 000 pg for 24 h reaction).0,HA; .,AA; A, ation. Symbols as for Fig. 3 DMA initiated about two decades ag0.’~9l~ However, little is known about the mechanism of plasma polymerization, and this is still under intense investigation. Commonly accepted mechan- isms involve the formation of radicals which initiate polymer- ization and subsequent attachment to the surface and/or attachement of the radicals directly to the surfaces and formation of highly cross-linked polymer^.'^ The studies shown here (Fig. 3 and Table I), involving analysis of free amino groups on the plasma-treated surfaces suggest that plasma polymerization may involve C-H bonds, rather than N-H bonds, in the formation of radicals as suggested by the presence of free amino groups on the membrane surfaces.The higher level of amine immobilization found for DMA compared to PA and HA suggests that the secondary amine reacts more readily with the PC surface, possibly because of increased charge-transfer effects which would be favoured by the inductive effect of the methyl groups. It is also possible that the primary amines might react more via the N-H bonds and so render the residues unreactive to acetic anhydride. The concentration of primary amino groups on an appro- priate plasma-polymerized membrane was calculated to assess whether there was any relationship between this value and the degree of glucose oxidase binding.Initially the reagent 2,4,6-trinitrobenzenesulphonicacid (TNBS) was employed, since this compound has been found to react specifically and under mild conditions with amino groups to give trinitro- phenyl derivatives.” Unfortunately, however, these derivatives are formed on the membrane surface making assessment of the amino-group concentrations difficult by absorbance measurements. For this reason an alternative technique was investigated involving acetylation of the amino groups on the membrane with radiolabelled [l-14C]acetic anhydride fol- lowed by heterogeneous scintillation counting of the respective membranes. Adams et ~1.’~have acetylated lysozyme to enable the behaviour of this protein at an air/water interface to be evaluated.A similar procedure was employed for the acety- lation of the membranes, although in this case separation of the non-bound from the bound acetic anhydride was made easier since the unbound material could simply be washed Off. Immobilization of Enzymes The choice of the enzyme proteins used was dictated by availability and low cost. To study the efficacy of the washing step in removing non-specifically bound enzyme, plain mem- branes were counted before and after the washing cycle. For glucose oxidase 85% of the bound radioactivity (i.e. enzyme) could be removed. However, for rennet this was only 20%. Since, only 0.24 and 0.19% of the glucose oxidase enzyme was attached to the PlzP(DMA)-treated PC membrane at an initial enzyme content of 6000 and 20 000 pg, respectively, [Fig.4(a)] and the surface did not saturate, it follows that the reaction of glucose oxidase with the membrane is very slow and possibly does not reach equilibrium during the 24 h reaction time. The amount of enzyme binding increases with, but is not proportional to, the surface amino-group density (Fig. 5). This may be due to steric hindrance inhibiting maximal binding of glucose oxidase to higher membrane densities of amino J. MATER. CHEM., 1991, VOL. 1 groups. However, the percentage of available amino groups occupied by glucose oxidase is greater for the surfaces treated with larger spacer arms.This may be due to the fact that these amino groups are separated from the surface by the spacer arm and therefore less susceptible to steric hindrance. Conclusion Fig. 4(a) and (b)indicate kinetically slow bulk-concentration- driven reactions. The surface reaction is not zero order with respect to the bulk reactants, despite the extended incubation periods employed. However, binding of the proteins does reflect the concentrations of the original surface amino groups and confirms that increased amounts of enzyme can be bound to membranes by increasing the surface amino-group density. The reduced binding efficiency associated with the decreased length of the spacer arm (Fig. 5) could have been due to steric limitation of glucose oxidase molecules on the membrane surface.It may also be possible that surfaces that contain longer spacer arms, such as PlzP(HA)-treated surfaces, have a proportionally higher tendency to bind very large enzyme molecules because of the increased availability of amino groups for coupling. Rennet attachment, both specific and non-specific, is greater than that of glucose oxidase. Both effects may have been due to the greater availability of binding sites for the smaller, and therefore sterically less hindered rennet molecule. It is unlikely that the membrane pores had a significant quantitative effect on this differential binding. M. M. gratefully acknowledges financial support from the International Atomic Energy Agency (IAEA) (Grant No: C6/TUR/88 13).References 1 W. Hartmeier, Immobilized Biocatalysts, Springer-Verlag, Berlin, Heidelberg, 1988. 2 D. R. Zaborsky, Immobilized Enzymes, CRC Press, Cleveland, 1973. 3 T. M. S. Chang, Biomedical Applications of Immobilized Enzymes and Proteins, Plenum Press, New York, 1976, vol. I and 11. 4 B. Mattison, Immobilized Cells and Organelles, CRC Press, Florida, 1983. 5 D. Thomas and J. P. Kervenez, Analysis of Control oflmmobilized Enzyme Systems, North-Holland, Amsterdam, 1976. 6 W. M. Hunter and F. C. Greenwood, Nature (London), 1962, 194, 495. 7 P. O’Carra, S. Barry and T. Griffin, Biochem. SOC. Trans., 1973, 1, 289. 8 M. Mutlu, M. T. Ercan and E. Piskin, Clin. Muter., 1989, 4, 61. 9 M. Mutlu and E. Piskin, Med. Biol. Eng. Comp., 1990, 28, 232. 10 H. Yasuda, J. Polym. Sci. Macromol. Rev., 1981, 16, 199. I1 A. S. Chawla, Polymeric Biomaterials, ed. E. Piskin and A. S. Hoffman, NATO AS1 Series, E 106, Martinus Nijhoff, Dordrecht, 1986, pp. 231-243. 12 A. Bradley and J. P. Hommes, J. Electrochem. Soc., 1963, 110, 15. 13 J. Goodman, J. Polym. Sci., 1960, 44,551. 14 H. Yasuda, J. Macromol. Sci. Chem. A, 1976, 10, 383. 15 K. Satake, T. Okuyama, M. Ohashi and T. Shinoda, J. Biochem., 1960, 47, 5. 16 D. J. Adams, M. T. A. Evans, J. R. Mitchell, M. C. Phillips and P. M. Rees, J. Polym. Sci., Polym. Symp., 1971, 34, 167. Paper 0/05810F; Received 28th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100447

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 451-455 451 Infrared Reflectance Spectra of Sb-doped SnO, Ceramics C. Stephan Rastomjee," P. Anthony COX,"Russell G. Egdell,*" Jeremy P. Kemp" and Wendy R. Flavellb a Inorganic Chemistry Laboratory, South Parks Road, Oxford OX1 3QR, UK Department of Chemistry, UMIST, P.O. Box 88, Manchester M60 1QQ UK Infrared specular reflectance spectra of Sn,-$b,O, ceramics have been measured over a range of bulk doping levels extending up to x=O.O3. At low doping levels conduction electrons screen out coupling of the infrared radiation to bulk phonon modes in a way expected from a model in which reflectivity is calculated from the bulk dielectric function. However, at higher doping levels a reflectivity minimum develops at ca. 800 cm-'. This is interpreted in terms of a model where a surface layer with low carrier concentration sits on top of an undepleted bulk.Keywords: Antimony-doped tin oxide ceramic; Infrared reflection spectroscopy; Depletion layer 1. Introduction Stoichiometric SnO, is a semiconductor with a bandgap of 3.89 eV, adopting the tetragonal rutile structure. The oxide may be doped n-type by oxygen deficiency or substitutional replacement of Sn with Sb or of 0 with F.' At doping levels above 5 x lo'* cm-3 the material becomes a metallic conduc- tor and carriers occupy an essentially free-electron-like con- duction band of dominant Sn 5s atomic character. The plasma energy of Sb-doped SnO, increases with doping level, reaching ca. 0.55 eV (4400cm-') in 3% Sb-doped Sn02.2-4 Owing to the low plasma frequency, doped SnO, retains good trans- parency in the visible regi~n.~ This property, in combination with high reflectivity at lower frequencies, leads to applications in infrared reflecting windows for both domestic and special- ised applications.Technological interest in tin oxide also includes its application as a sensor material for reducing gases such as CO and CH4.6 The generally accepted model for sensor activity is as follows. Chemisorbed oxygen on the SnO, surface under ambient conditions depletes carriers in the near- surface region. Surface oxygen is removed by catalytic oxi- dation of reducing gases, thus releasing carriers back into the surface and producing a rise in conductivity.In previous publications we have reported the use of techniques including ultraviolet photoemission (UPS) and high-resolution electron energy loss spectroscopy (HREELS) to measure the concentration of n-type carriers at doped SnO, surfaces under ultra-high vacuum In the present study we extend the scope of the earlier work by applying infrared reflectance spectroscopy to examine phonon structure in Sb-doped SnO, ceramics as a function of bulk doping level. Screening of phonons by the carriers provides a probe of carrier distribution close to the surface under ambient conditions. Evidence emerges from our work that for highly doped Sn0, there is significant carrier depletion in the near-surface region for samples held in air. 2. Experimental Samples were prepared by a coprecipitation method.Weighed quantities of Sn shot (Aldrich Gold Label, 99.999%) and Sb powder (Koch Light, 99.999%) were dissolved separately in aqua regia made from BDH AnalaR grade acids. The solutions were then mixed to give the required doping level and made alkaline with excess AnalaR grade ammonia solution. After the solutions had been boiled for at least 1 h the white gelatinous precipitate was allowed to settle overnight before collection and washing in a Buchner funnel. The precipitate was dried overnight at 100°C and then slowly heated to 1000 "C in recrystallised alumina crucibles. After sintering for 14 days this temperature with occasional regrinding, samples were finely ground in an agate mortar and pressed into 13 mm diameter pellets between optically smooth tungsten carbide dies. They were then sintered in air for at least 24 h to yield mechanically robust ceramic discs.X-Ray powder diffractometer traces from samples prepared in this way contained only peaks attributable to a well crystallised rutile phase with lattice parameters not signifi- cantly different from those of tin dioxide itself. Analytical electron microscopy and atomic absorption spectroscopy con- firmed that the Sb doping level was close to the nominal value for the more highly doped samples. Some of the samples were studied by a range of surface analytical techniques including X-ray photoelectron spectroscopy (XPS), UPS and HREELS, as detailed Specular infrared reflectance spectra were measured using unpolarised radiation incident at 20" to the surface normal in a Mattson Alpha Centauri FTIR spectrometer, incorporat- ing a Nichrome wire radiation source, KBr beamsplitter and DTGS detector.The experimental resolution was set at 16 cm- I. Background interferograms were recorded from an aluminium mirror, which was assumed to be 100% reflecting over the region of interest. 3. Results The experimental infrared spectra are shown in Fig. 1A. Undoped SnOz displays a spectrum with two reflectivity maxima above 400 cm-', peaking at 620 and 480 cm- '. The higher-energy peak is the stronger and has a pronounced shoulder on its high-energy side. The reflectivity declines to a minimum at 800 cm-I and remains low at higher frequency. The drop in reflectivity below 800 cm-' corresponds to the longitudinal optical (LO) phonon frequencies of the highest- energy E, and the A," modes, which are at 757 and 704 cm- ', re~pectively.~The initial effect of doping is to produce an increasing reflectivity across the infrared region, especially at low frequency.Maxima in reflectivity corresponding to those of undoped SnO, persist in the experimental spectra, although the maxima are less pronounced as the doping level increases. At higher doping levels a shallow trough in the spectra develops at CQ. 750 cm-l. In making a comparison between experimental reflectance data and simulations of the spectra (section 4, below) it should be remembered that the absolute 452 A B 20 40 10 20 20 30Y860 o b 20401 80Ol--+-+- 10 20 iJ!? 0 n 1300 800 400 1200 800 400 waven urn ber/cm -Fig.1 A, Experimental infrared specular reflection spectra of Sb-doped SnO, ceramics as a function of bulk doping level; B, simulated spectra, using Dresselhaus averaging procedure and flat-band carrier distribution (see text). Doping level, Sb (YO):(a) 3.0; (b)0.6; (c) 0.1; (40 value of the reflectivity is lower than expected when dealing with ceramic pellets owing to optical imperfection of the sample surface. 4. Discussion 4.1 Modelling of Data with Reflectivity from Bulk Dielectric Function Tin(1v) oxide adopts the tetragonal rutile structure with space group Dit.There are two formula units per cell and the 15 optical modes span irreducible representations r as follow^:^^* r=Ai,+ A2,+ A2, +Big+ B2,+2Bi,+ E,+ 3E, (1) The dipole-moment operator spans representations E, and A2, and hence only the four optical modes of these symmetries involve the dipole-moment change necessary for infrared activity. Moreover, inspection of the basis vectors shows that the E, modes involve atomic displacements perpendicular to the c axis and the A2, mode displacements parallel to the c axis. The infrared reflectance R(o)at normal incidence is given by R(4= IC1-&4I/C1 ++M2 ={[n(o)-1l2 +k(o)2)/{[n(o)+1]2+k(o)2) (2) where &(a) =[n(o)+ik(o)12 (3)=~(0)~ and the complex dielectric function &(a)is given by n =&(a)~(o) + 1 R:/(w? -o2-ioy,) (4)i= 1 J.MATER. CHEM., 1991, VOL 1 The summation in eqn.(4) extends over the dipole active modes and Ri, mi and yi are respectively the dipole strength, transverse resonance frequency and damping constant for the ith oscillator. For single-crystal SnO, with the electric vector polarised parallel to the c axis, only the one term associated with the Azu mode contributes to the summation (4), giving a dielectric function E,,(o). However, with the electric vector polarised perpendicular to the c axis the summation extends over the three E, modes, giving the function E,,(w). Addition-ally, it is necessary to select the value of E(m) appropriate to the relevant polarisation.The ‘parallel’ and ‘perpendicular’ reflectivities of single-crystal Sn02 calculated from eqn. (2) are shown in Fig. 2, the relevant parameters in eqn. (4)being obtained from the experiments of S~mmitt.~For both polaris-ations a single broad Reststrahlen band appears above 400 cm-’, two of the E, bands in the parallel polarisation being below the wavenumber range available on our spec-trometer. Evidently, the experimental spectrum for our polycrystalline ceramic sample is completely unlike the single-crystal spectra and cannot be obtained by a suitably weighted summation of the reflectances for the two polarisations. We have explored two approaches for calculation of spectra for ceramic samples. The first uses an ‘average’ dielectric function given by &a,(@) =$zz(a) +&xx(a) (5) cav(o)is then inserted into eqn.(2) to calculate the reflectivity. This approach gives two reflectivity maxima above 400 cm-’, although the two peaks are of comparable intensity and the trough between them is much more pronounced than found experimentally [Fig. 3(b)]. An alternative to this ad hoc approach, taking proper account of the random orientation of crystallites, has recently been suggested by Dresselhaus and co-worker~.~This assumes that the grain size is larger than the wavelength of the infrared radiation and involves an integration over the angle 8 between the c axis of a crystallite and the Poynting vector of the 80 60 40 20 hZo 0 80 -c’60 1200 1000 800 600 400 wavenurnber/crn -Fig.2 Simulated infrared reflection spectra for single-crystal SnOJ 1lo), (a)with electric vector polarised parallel to c axis, (b) with electric vector polarised perpendicular to c axis J. MATER. CHEM., 1991, VOL 1 80 60 40 20 50 a, c 80 a, c '60 40 20 0 1200 1000 800 600 400 wavenum ber/cm -' Fig. 3 Simulated infrared reflection spectra of polycrystalline SnO, (a)using the Dresselhaus method for averaging reflectivity over crystal orientation (b)using ad hoc effective medium method for averaging the dielectric function before calculating the reflectivity. See text for details of these procedures where K~(o) =[E,,(o) COS~e +E,,(o) sin2 (7) "&4=C&XX(4l 1'2 (8) This approach gives two Reststrahlen peaks of roughly the correct relative intensity [Fig.3(a)] and the overall spectral profiles are in reasonable agreement with experiment, except that the shoulder on the higher-frequency peak in the model spectrum is less pronounced than that found experimentally, and the peak itself is less sharp than in the experimental spectra. The reasons for the discrepancies are unclear to us at present, although one possibility is that there is significant texturing of the ceramic close to the pellet surface so that the averaging procedure used in eqn. (6)is not quite correct. To model the effects of doping on experimental spectra it is necessary to introduce a term into eqn. (4)to represent the dielectric response of the conduction electrons introduced by doping.In place of eqn. (4)we write n =~(m)&(a) + 1 QT/(o:-o2-ioy,) -sZi/(02-ior,) (9)i= 1 where rpis the plasmon damping parameter and the unscre- ened plasma frequency, R,, is given by R, =(ne2/&om*)1/2 (10) Here, e is the charge on the electron, E~ is the permittivity of free space, n is the carrier concentration and m* is the electron effective mass. The value of m* was fixed by reference to our own previous HREELS data2-4 for 3% Sb doped SnO,. Here a surface plasmon loss is observed at 0.55eV. The surface loss condition Re E(o)= -1 (11) leads to the following expression for the surface plasmon energy ESP: ESP=hR,/[&(m) +1]'12 =(ne2/com*[&(.o)+1]}1/2 (12) Hence we find m*/me=0.75, where me is the electron rest mass.We assume an average value for &(a)in eqn. (12); the possibility of anisotropy in m* does not enter into our analysis. However, some anisotropy in the plasma response is incorpor- ated into our calculations of the reflectivity by the use of different values of &(a)in E,,(o) and E,,(o). Calculated reflectivities for the different doping levels are shown in Fig. 1B. The plasmon energy for 0.1% doped SnO, is only ca. 1000 cm-', so that at this doping level the reflectivity drops markedly across the infrared region extending up to 1200cm-'. The drop becomes less pro- nounced as the plasmon energy increases with increasing doping level. Phonon structure is superimposed on the plas- mon background and becomes progressively weaker with increasing doping level. At higher doping levels, the phonon structure takes the form of two fairly sharp maxima, the highest energy peak occurring at the E, transverse optical (TO) phonon frequency.No well defined structure is found around the LO frequency, corresponding to the minimum in reflectivity for undoped Sn02. Qualitatively, the disappear- ance of structure at the LO frequency can be understood in terms of screening of the long-range coupling that leads to the LO-TO splitting. In view of the model calculations, the experimental data for the highly doped samples, showing a pronounced trough in the reflectivity around the A,,/E, LO phonon frequency, must be regarded as extremely puzzling. 4.2 Modelling of Data using Two-layer Model The discrepancy between experimental data and results from model calculations using the bulk dielectric function leads us to consider analysis of the data using a model where a layer with a carrier concentration lower than that of the bulk sits on top of the undepleted bulk.Details of this model have been given in a recent publication." The rationale for intro- ducing this approach is as follows. It is well known from experiments on thin layers of dielectrics on metals that, at non-normal incidence, reflectivity minima corresponding to LO frequencies of the dielectric overlayer can be observed in experimental spectra." In the present context we suppose that a carrier-free layer can arise from band bending at the surface, mimicking the dielectric layer of the classical experi- ments.The undepleted bulk then functions as the highly reflecting metal underlayer. In experiments on Sb-doped Sn02 using electron spectroscopic techniques under ultrahigh vac- uum (UHV) conditions we find no evidence of surface carrier depletion so that the origin of depletion layers seen in the reflectance experiments carried out in air must lie in depletive adsorption of oxygen on the ceramic surface. The model for the reflectivity is based on a general theory of surface response to electromagnetic fields advanced by Flores and Garcia-Moliner.12 They introduce a surface impedance tensor 2 which can be defined through the equa- tion (SI) EllS=Z*(nx Hs) (13) where c is the speed of light, n the unit outer normal vector from the surface, HSthe magnetic field at the surface and EllS the electric field parallel to the surface, evaluated at the surface.Matching conditions applied to electric and magnetic 454 fields for incident, transmitted and reflected waves at the air/ depletion layer, and depletion layer/bulk interfaces leads to a reflectivity R for p-polarised radiation (H parallel to surface): R= ICzeff-(1/c&o) cosei)/CZeff +(I/CEO) cos @ill2 (14) where Bi is the angle of incidence of the radiation relative to the surface normal and Zeff is the effective impedance of the two-layer system. This is given in terms of impedances of surface and bulk layers 2, and zb, respectively: Zeff=Z,(1 + D2A)/(1 -D2A) (15) where A = (zb -zs)/(zb + zs) (16) and D is the damping factor, D =exp [-dlIm E~(O)O~/C' -q211/2] (17) where E, is the dielectric function for the surface layer of thickness d, and q is the x component of the wavevector of the incident radiation.Note the expected limiting behaviour d+O, Zeff+Zb and d+m, Zeff+Zs. Similar expressions can be derived for s-mode radiation. Results of a calculation based on this model assuming unpolarised radiation (i.e. mixed s and p polarisation) are shown in Fig. 4 for a 200 A depletion layer on 3% Sb-doped SnO,. This thickness is based on the following considerations. I I I 76 t I I I 1 I 1200 1000 800 600 400 wavenumber/cm -' Fig.4 Simulated infrared reflection spectra of 3% Sb-doped SnO, with surface depletion layer 200 8, thick. (a) Carrier free depletion layer, (b)depletion layer supporting plasmon mode of wavenumber 484 cm-' (60 meV), (c) depletion layer supporting plasmon mode of wavenumber 726 cm-I (90 meV). In (b)and (c) it is assumed that the plasmon damping parameter r in the surface layer is given by hT= 50 meV J. MATER. CHEM., 1991, VOL 1 At the (001) surface of SnO, there is one cation per cell. Suppose that oxygen chemisorption on these surface cations depletes four electrons per adsorbed molecule (corresponding to reduction to 02-).Given that there are two cations per cell in the bulk and that the lattice parameter c= 3.185 A, at 3% doping the carrier depletion will extend over a length range d=(4 x 3.185 x 100/3x 2) 8,~200A.Of course, we are not implying that the polycrystalline surface is strongly (001) textured in adopting this approach to estimating the depletion layer thickness. For simplicity, the dielectric function for bulk and surface layers is averaged using the ad hoc procedure of eqn. (5). As expected, a dip in the experimental spectra develops at the LO frequency at the highest E, mode just below 800cm-', but this is much sharper and weaker than in the experimental data. One possibility is that the relaxation time for LO phonons in the surface layer is much shorter than for bulk LO phonons, thus broadening the absorption dip. However, the intensity of the dip still remains too low.This leads us in turn to consider an alternative model where the surface layer supports a reduced (but non-zero) carrier concentration as compared with the bulk. If the plasmon frequency in the surface layer is similar to that of the LO phonons, strong coupling occurs to give modes of mixed plasmon-phonon character. In the reflection experiment, absorption then occurs at the longitudinal frequency of the coupled modes within the surface layer. For plasmon frequen- cies ca. 560-800 cm- (70-100 meV) the simulated spectra bear a striking qualitative resemblance to those measured experimentally. We thus identify the unexpected broad dip in reflection spectra of highly doped SnO, as being associated with a surface layer of reduced carrier concentration.In speculating as to the mechanism of carrier depletion, we are guided by the following considerations. First, as noted above, the depletion layer is not found in UHV experiments and so appears to be related to oxygen adsorption on the ceramic surface. Secondly, major qualitative discrepancies between experimental spectra and simulations based on the bulk dielectric function only become apparent at high doping levels. Previous work has shown that at the higher doping level surface cation sites are increasingly occupied by segre- gated Sb ions, the fractional Sb surface coverage being close to unity at 3% Sb bulk doping It thus appears that segregated Sb cations act as sites for the depletive adsorption of oxygen.This in turn can be rationalised in terms of competition between atmospheric 0, and H20 for surface cation sites. Segregated Sb ions are essentially Sb"' species3 carrying a directional sp hybrid lone pair of electrons. Such ions are likely to be weak Lewis-acid centres, with a low heat of adsorption for HzO. Thus 0, is able to adsorb on the surface Sb sites, producing the depletion layer. 5. Concluding Remarks The present study clearly shows that infrared reflectance spectra of tin dioxide are very sensitive to the bulk carrier concentration, the major effect of doping being to screen out coupling between lattice phonon modes and external electro- magnetic fields. However, the existence of surface depletion layers allows appearance of new reflectance features associated with longitudinal optical modes in the surface layer.Ani- sotropy of the dielectric function introduces further complex- ities into analysis of the reflectance data. C.S.R. is grateful to British Gas for the Award of a Research Scholarship. J. MATER. CHEM., 1991, VOL 1 455 References 7 R. Summitt, J. Appl. Phys., 1968, 39, 3762. Z. M. Jarzebski and J. P. Marton, J. Electrochem SOC., 1976, 123, 8 P. A. Cox, R. G. Egdell, W. R. Flavell and R. Helbig, Vacuum, 1983, 33, 835. 229C. 9 G. L. Doll, J. Steinbeck, G. Dresselhaus, M. S. Dresselhaus, A. J. P. A. Cox, R. G. Egdell, A. F. Orchard, C. Harding, W. R. Pat-terson and P. W. Tavener, SoIid State Commun., 1982,44, 837. P. A. Cox, R. G. Egdell, C. Harding, W. R. Patterson and P. W. 10 Strauss and H. J. Zeiger, Phys. Rev. B, 1987,36, 8884. J. P. Kemp, P.A. Cox, R.G. Egdell and K. Kang, J. Phys. Condens. Matter, 1989, 1, SB123. Tavener, Surf: Sci., 1982, 123, 179. R. G. Egdell, W. R. Flavell and P. W. Tavener, J. Solid State 11 C. Kittell, Introduction to Solid State Physics, John Wiley, New York, 5th edn., 1976. Chem., 1984,51, 345. Z. M. Jarzebski and J. P. Marton, J. Electrochem. SOC., 1976, 123, 333C. 12 F. Flores and F. Garcia-Moliner, Introduction to the Theory of Solid Surfaces, Cambridge University Press, Cambridge, 1979. e.g. J. F. McAleer, P. T. Moseley, J. 0.W. Norris and D. E. Williams, J. Chem. SOC.,Faraday Trans. I, 1987, 83, 1323. Paper 1/00073J;Received 7th January, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100451

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 457-460 Formation of Ordered Multilayers by Stepwise Oligomerisation Joseph Y. Jin and Robert A. W. Johnstone* Department of Chemistry, The University of Liverpool, P.O. Box 147, Liverpool L69 3BX, UK By stepwise oligomerisation, multilayer structures have been laid down on molecularly ordered substrates. The stepwise process utilises the reaction of alkenyl isocyanates with surface hydroxy groups. Keywords: Oligomerisation; Multilayer; Langmuir-Blodgett film; Alkenyl isocyanate Highly ordered monolayers of organic materials can be assembled through use of the Langmuir-Blodgett technique.' Subsequently, multilayers can be built up by repetitious deposition of such monolayers.' Although this is an excellent method for preparing molecularly highly ordered films of material, the substances used are constrained to be amphi- philic, and this requirement implies limitations on the types of film that can be assembled.Usually, the amphiphile contains a long hydrocarbon chain which has insulating properties and tends to isolate the individual layers from each other. A further problem with such multilayer structures lies in their relatively fragile nature due to the lack of chemical bonding between the substrate and the first layer, between any two layers and between molecules within the layers themselves. More recently, methods of laying down mono- and multi- layers of 'self-ordering' substances have been described. Alkenyltrichlorosilanes have been used to react with surface hydroxy groups on aluminium,2 glass2 and silicon3 to form a chemically bonded monomeric film showing considerable molecular ordering.The so-formed new upper surface contains vinyl groups which can be readily transformed into terminal hydroxys through treatment with B2H6 and basic H202.4 Also, hydroxy groups have been introduced through reduction of surface methyl ester functions with LiAlH4.3 Subsequently, self-ordered monolayers have been laid down on gold surfaces through reaction with alkanethiols,' although it is not clear whether there is any formal chemical bonding between the gold and the thiol groups. In developments of the work with trichlorosilanes, attempts have been made to assemble multilayers through stepwise iterative reactions.In particular, conversion of vinyl groups into hydroxy groups on the upper surface of a monolayer, was followed by further reaction with alkenyltrichlorosilane to lay down a second layer, etc6 It appears that the second and subsequent layers are not so highly ordered as the first layer, probably because of the need for each silicon atom to be tetrahedrally bonded to three oxygens, residing either on the surface or between silicon atoms. This requirement appears to lead to the formation of defects in the second and sub- sequent layers. Characterisation of the mono- and multi-layers has been in terms of the contact angle with infrared spec- tros~opy,~*~-'~ and X-ray diffraction.8 ellips~metry,~?~ In the present work, new methods for laying down layers on top of an ordered monolayer are being investigated.In many ways, the difficulty of bonding subsequent layers to a monolayer in a stepwise oligomerisation can be compared with the difficulties experienced in solid-phase peptide syn- thesis, where each reaction needs to proceed in as close to a 100% yield as is possible, and byproducts of reaction must be easily removable. '' However, unlike solid-phase peptide synthesis in which the growing peptide chains are kept as far apart as possible, stepwise reactions on an ordered solid surface need to proceed under conditions in which the reacting molecules become packed close together. This last requirement means that the kinetics of film formation will be controlled, especially in the terminal stages, by the degree of difficulty in access of the reactant molecules to the surface that is being covered and the need to remove byproducts from the reacting centres.Indeed, the distribution and separation of reactive sites across the upper layer may control the degree of orderli- ness achieved by oligomerisation. Too close a packing density of sites could make it extremely difficult to obtain complete coverage. Therefore, it would be useful to alter the density of reactive sites on the upper surface in a controlled manner and examine the degrees of order in subsequently formed layers. Work on this aspect of the oligomerisation method is underway. Another area which could affect order in the film lies in the choice of solvent, either through its entry into the layers or through its tendency to promote conformations that may not be consistent with a high degree of order.Thus, stepwise oligomerisation on a highly ordered surface preferably utilises fast, high-yielding reactions having either no or easily removable byproducts. One such reaction reported here involves the use of isocyanates, which react with alcohols, thiols and amines to give urethanes, thioure- thanes and ureas without byproducts RN=C=O+ R'XH-RNHCO-XR' (1) An added advantage of this reaction lies in the formation of a secondary amide bond, similar to that which exists in proteins, nylons and polyurethanes, and which is known to form strong intermolecular hydrogen bonds.I2 Where many such bonds occur, as in protein structures, considerable physi- cal strength is imparted to the structure.For multilayers formed through reaction of an isocyanate with a reactive substance [R'XH, eqn. (l)], this hydrogen bonding should help both to stabilise the structure and to encourage self- ordering within the layers. In the present work, such amide bonds have been formed by iterative (stepwise) reactions of alkenyl isocyanates with surface hydroxy groups. These step- wise reactions (oligomerisation) have given thin multilayers. Experimental 'H NMR spectra were obtained in CDCl,, standard TMS, on a Bruker AC 200 spectrometer, and IR spectra of liquids or solids (as Nujol mulls) were measured on a Perkin-Elmer 1720-X FTIR instrument.Water/Surface Contact-angle Measurements The method for measurement of static advancing contact angle has been well de~cribed.'~ Contact angles were deter- mined on sessile drops of doubly distilled water at room temperature (20-25 "C). The sessile drops were applied by first forming a 10-3cm3 drop at the tip of a syringe needle and allowing it to contact the surface, after which the drop size was increased to ca. 3 x cm3 and the needle with- drawn. Contact angles, measured within 1 min of applying the drop to the surface, were determined by measuring the tangent to the drop at its intersection with the surface, using a telescopic sight on a graduated (degrees) circular scale. All reported values are averages of at least three measurements.Although drop sizes were mostly ca. 3 x cm3, it was determined that contact angles were independent of drop size, within the range (1-20) x cm3. Ellipsometer Measurements An Applied Materials Laser Ellipsometer I1 was used, with an He-Ne laser (A=632.8 nm) as the light source. The incident angle was 70" and the quarter-wave compensator was set to -45". Each set of polariser and analyser angles was the result of at least three measurements taken at different locations on the sample. Film thicknesses were determined through calcu- lations based on previously derived equation^.'^ The refractive index of the base silicon oxide surface was calculated to be 3.902-0.152i. For calculation of film thickness of the organic layers, the refractive index was assumed to be 1.45.15 Deposition of Silane Monolayer This was carried out according to a reported procedure.2 A clean silicon wafer (2cm x 3 cm), for which the surface was believed16 to hold ca.5 x 1014 %-OH groups per cm2, was immersed in a solution of n-undec-10-enyltrichlorosilane (0.04 g, 0.21 mmol) in hexadecane (15 cm3) at 20-25 "C for 15 min. The wafer was removed from the solution, rinsed first in hexadecane at 20°C and then in light petroleum (b.p. 60-80 "C) several times before being dried off in a stream of nitrogen at room temperature. This coated water was used in the next process. Layer thickness was obtained by ellipso- metry, and water contact angle was measured. Conversion of the Upper Surface Vinyl Groups into Terminal Hydroxy Groups This type of reaction has been well doc~mented.'~The following method is closely similar except that the reaction was monitored with time to ensure complete conversion of vinyl to hydroxy.The silicon wafer, coated with an undecenylsiloxane monolayer was immersed in a solution of diborane in tetra- hydrofuran (1 mol dm -3, for 2 min under a dry argon atmos- phere at room temperature to allow tetrahydroborate addition to the terminal vinyl. The wafer 'was then dipped into an alkaline solution of H202 [l :l(v/v) of 30% H202 (m/v) and 0.1 mol dm-3 aq. NaOH] for 2 rnin at room temperature to convert the tetrahydroborate groups into hydroxy groups. The wafer was rinsed several times in ethanol and distilled water and then dried in a stream of nitrogen, ready for the next coupling stage.Layer thickness and water contact angle were measured at different times to determine optimum immersion time in the reaction medium. Oligomerisa tion Steps In the first stage, the silicon wafer, supporting its silanised organic monolayer in a terminal hydroxy form, was immersed in a solution of n-dec-9-enyl isocyanate (0.04 g, 0.22 mmol) in hexadecane (15 cm3) for 40 rnin at 75 "C. After this, the wafer was washed in hexadecane at room temperature for 1 rnin and then rinsed several times in light petroleum (b.p. 60-80 "C) J. MATER. CHEM., 1991, VOL. 1 before being dried finally in a stream of nitrogen. Layer thickness and contact angle were measured.In the second stage, the new terminal vinyl groups were converted into terminal hydroxy groups by the same pro- cedure as described above for the first silanoxy layer, using diborane and alkaline hydrogen peroxide. After the final drying stage with nitrogen, layer thickness and contact angle were measured. The above two steps were repeated to add two further layers to give a total of four organic layers on top of the original silicon oxide surface layer of the silicon wafer. Preparation of n-Dec-9-enyl isocyanate (a) n-Undec-10-enoic acid (1 5.0 g, 0.08 mol) in dry ether (50 cm3) was refluxed with thionyl chloride (1 1.0 g, 0.092 mol) for 2 h. The solvent and excess of thionyl chloride were evaporated to give undec- 10-enoyl chloride which was used without further purification in the next stage. (b) The crude undec-10-enoyl chloride from stage (a) was added dropwise to concentrated aqueous NH, (d =0.88, 60 cm3) with stirring over 20 min. A solid separated immedi- ately and was filtered off to give undec-10-enamide (8.4 g, 56%), m.p.85 "C (from light petroleum, b.p. 60-80 "C; lit.'' m.p. 82 "C); 6, 1.3 (10 H, br s), 1.5 (2 H, m, H2C=CHCH,CH,-), 2.09 (2 H, 4, H2C=CHCHz--), 2.24, (2H, t, H2NCOCH2-), 4.95, 5.03 (2H, m, H2C=CH-), 5.55 (2 H, br d, NH2), 5.82 (1 H, m, H2C-CH-). (c) n-Undec-10-enamide (2.6 g) in dichloroethane (75 cm3) was reacted with oxalyl chloride (2.6 cm3) in dichloroethane (15 cm3) with stirring at 4 "C (i~e-water).'~The cooling bath was removed and stirring was continued for 1 h, after which the mixture was refluxed for 5 h.The solvent was removed to give a liquid, n-dec-9-enyl isocyanate (1.5 g, 58%), b.p. 68 "C at 0.1 mmHg.? Because of its reactivity to water, this isocyan- ate was characterised through its large IR band at 2243 cm-', which disappeared in moist air, being replaced by new bands at 3356 and 3192cm-' (-NH2); its conversion into the corresponding methyl carbamate was effected by reaction with methanol. Thus, the isocyanate (0.5 g) was dissolved in methanol (5 cm3) at room temperature with stirring for 10 min and the methanol was evaporated to give N-n-dec-9-enyl methyl carbamate, m.p. 101 "C (from MeOH); aH 1.30 (10 H, br s), 1.66 (2 H, m, H2C=CHCH2CH2-), 2.05 (9, H2C=CHCH2-), 2.75 (2 H, t, --CH,NH-), 3.79 (3 H, S, CH30-), 4.95, 5.05 (2 H, m, H2C=CH-), 5.80 (1 H, m, H2C=CH-), 7.85 (1 H, br s,-NH-).(Found: C, 67.3; H, 9.9; N, 5.7%. C13H23N02 requires: C, 67.6; H, 10.8; N, 6.6%). Preparation of n-Undec-lO-enyltrichlorosilane (a) Undec-10-enoic acid (58 g, 0.32 mol) was refluxed with LiA1H4 (12 g, 0.32 mol) in diethyl ether (400 cm3) for 18 h, after which excess of ethyl acetate was added cautiously to decompose any remaining hydride reagent. The mixture was washed with dil. H2S04 and water and dried (MgS04). Evaporation of the solvent and distillation of the residual oil gave n-undec-10-enol as a colourless liquid (40.2 g, 92%), b.p. 9 1-93 "C at 0.5 mmHg (lit.20 132- 133 "C at 15 mmHg).(b) To a solution of n-undec-10-enol (8.6g) in CC14 (100 cm3) was added dropwise tri-n-butylphosphine (30 cm3) over a period of 20min at room temperature. The solution became warm and was left to stir for a further 2 h. Evaporation of the solvent gave a residual oil which was distilled to give n-undec-10-enyl chloride as a colourless oil (58% yield), b.p. 71.5 "C at 0.3 mmHg (lit.2' 105-107 "C at 6 mmHg). t 1 mmHgx133.322 Pa. J. MATER. CHEM., 1991, VOL. 1 (c) A solution of n-undec-10-enyl chloride (3.14 g) in diethyl ether (20 cm3) was added to magnesium turnings (0.95 g) in diethyl ether (1 5 cm3) to which had been added a small crystal of 12.The mixture was stirred and heated under reflux for 18 h to give a solution of n-undec- 1 0-enylmagnesium chloride which was transferred to a dropping funnel, under N2; this Grignard reagent was added dropwise to a solution of SiC1, (6 g) in diethyl ether (30 cm3) with stirring and then the whole was heated under reflux for 18 h. The resulting mixture was filtered and the solvent was evaporated to yield an oil which was distilled to give n-undec- 10-enyltrichlorosilane as a colourless oil (1.25 g, 26%), b.p.33 "C at 0.05 mmHg (lit.4 110 "C at 20mmHg); 8H 1.2 (14H, br), 1.48 (2 H, m, H2C=CHCH,CH,--), 1.95 (2 H, 9, H2C=CH-CH2), 4.84, 5.94 (2 H, m, H2C=CH-), 5.71 (H, m, H2C=CH-). Results and Discussion The experiments were carried out on pure silicon wafers, which had been left in contact with air for a sufficient period that an oxidised surface layer was produced, containing multiple hydroxy groups.I6 A monolayer was deposited on this silicon 'hydroxide' surface by immersing the silicon wafer in a solution of undecenyltrichlorosilane in hexadecane.6 Measurements of water contact angle and ellipsometry con- firmed that a monolayer having a hydrophobic surface had been deposited.By immersion of such silicon wafers in unde- cenyltrichlorosilane solution for periods of 2-15 min, it was determined that contact angle and film thickness reached limiting values of ca. 105' and 15 81, respectively, after 10 min (Table 1). The surface vinyl groups were converted into ter- minal hydroxy groups using B2H6 and alkaline H202. During this hydroxylation, measurements over a period of 5 min indicated that the water contact angle reached a limiting value of ca.46" after <2 min. This value is in accord with those found by other workers* and indicates the presence of hydroxy groups on the upper surface of the monolayer. The monolayer film on its silicon substrate was then immersed in a solution of n-dec-9-enyl isocyanate in hexa- decane at 75 "C for 40 min. Again, contact angle measurements indicated that a minimum time of 30 min was needed to reach a limiting value of ca. 82 "C, indicating a change in surface character from hydrophilic to hydrophobic as the hydroxy groups reacted with the isocyanate. Ellipsometry showed a new total film thickness of ca. 35 81, which indicates an added thickness of 1981 on top of the original monolayer.From molecular models, an extended (uncoiled) n-dec-9-enyl carba- mate chain would be expected to have a length of ca. 1681. Thus, allowing for errors in measurement, the added film thickness corresponds to a layer consisting of chains of n-dec- 9-enyl carbamate groups extending vertically from the original monolayer surface. These results are consistent with the formation of a close-packed second layer having terminal vinyl groups on its upper surface. As described above for the first layer, the vinyl groups of this second layer were converted into terminal hydroxy groups using B2H6 and alkaline H202. Table 1 Limiting values of water contact angle with time for reaction of n-dec-9-enyl isocyanate with surface hydroxy groups reaction time"/min contact angleb/" film thickness'/A 5 77 20.2 15 77 23.1 30 21.3 "Time of immersion of film in a solution of the isocyanate in hexadecane at 75 "C.bMean static advancing water contact angles. 'Estimated error 12 A. Table 2 Water contact angles and film thickness of multilayers ellipsometry measurements layernumber" contact angleb/" P/" Adlo total film thickness'/A 0 c25 42 10.7 - 1 105 39.8 10.8 15.7 1' 46 39.7 10.7 16.1 2 82 37.1 11.0 35.5 2' 53 37.0 11.2 37.0 3 78 35.2 11.1 49.6 3' 53 34.8 11.2 52.9 4 77 31.2 11.3 79.4 "The numerals indicate the number of each layer counted from the silicon oxide base with 0 =silicon dioxide surface alone.Thus, 1=first monolayer on the surface having vinyl groups and I' is the same layer except that the surface vinyl groups have been converted into terminal hydroxys. Layers 2,2' and so on (n, n') correspond to success- ive reactions with isocyanate, followed by conversion of vinyl to hydroxy, as described in the discussion section. Mean static advanc- ing contact angle, measured as described in the experimental section. 'P =polariser azimuth angle. A =analyser azimuth angle. See exper- imental section. Error in measurement of film thickness is estimated to be ca. +2 A, for each layer. For the purposes of this table, the silicon oxide surface is given a zero thickness, although it was actually 20A thick. Refractive index for the organic layers was taken to be 1.45.14 The water contact angle now changed to ca.52', consistent with the presence of hydroxy groups on the surface. These surface hydroxy groups were reacted with n-dec-9-enyl isocy- anate by repetition of the procedure described above. The water contact angle changed to 78" and the film thickness increased again by ca. 13 81. These stepwise reactions with the alkenyl isocyanate gave a thin film of four layers, the first being bonded to silicon through Si-0-Si bonds and the remainder by interlayer urethane (-NHC02-) linkages. Table 2 shows that the first layer gives a water contact angle (105") close to those reported for a highly ordered film having an upper hydrocarbon-like surface.2,6 As each of the layers, 2, 3,4, is prepared, the upper surface, which should also be hydrocarbon-like (terminal vinyl groups), gave water contact angles of 78-80', considerably less than that of the first layer.This result could be interpreted as being due to a film containing many defects, as demon- strated in earlier work with mixed thiols on a gold ~ubstrate.~.~ However, there is a consistency about the contact angles for layers 2, 3 and 4, which indicates that, if the result is due to defects, then the same degree of defect structure must be present in each layer except the first. Although this cannot be ruled out, it is also possible that the new urethane linkages, having strong dipoles and capable of very strong hydrogen bonding, may be able to make their presence felt at the surface.It has been demonstrated previously7 that any dipolar or other effect of ether linkages buried below a hydrocarbon surface become negligible beyond a depth of 5-10 81. The much stronger dipole of the urethane bond (amides have dipoles of ca. 3.7-4.1 D compared with ethers at ca. 1.1 D) may increase this depth effect markedly, and their hydrogen- bonding nature may lead to some water being entrained within the film. Further examination of this discrepancy, compared with earlier published data, is being carried out. Examination of the hysteresis of the contact angle should enable differentiation between a dipole effect and possible defects in the film.? An attempt was made to attach the alkenyl isocyanate ~~ ~ ~ t Constant hysteresis between the first and following layers could rule out an interpretation in terms of defects.460 directly to the silicon oxide surface, instead of onto a pre- viously deposited organic monolayer. Despite prolonged immersion of a silicon wafer (prepared as for deposition of alkenyltrichlorosilane, i.e. with surface hydroxy groups) in a solution of n-dec-9-enyl isocyanate, no organic film could be detected on the silicon by contact angle or ellipsometry after washing it with solvent and drying. It seems that, if the isocyanate had reacted with surface hydroxy groups, as seems likely, then the resulting carbamic acid derivatives probably decarboxylated readily to return the surface to a hydroxide state which desorbed the resulting amine; decarboxylation of free carbamic acids is well known.22 0 I I.I1-Si-OH + O=C=NR -+ -Si-O-CNHR I I I I -Si-OCONHR + H20 -+ -Si-OH + C02+ H2NR (2)I I Conclusion Thin multilayers can be assembled on molecularly ordered organic monolayers deposited on a silicon oxide surface by stepwise reactions (oligomerisation) using isocyanates to form urethane linkages. An advantage of this reaction is the absence of any byproducts, its high yield and ease of operation. Additionally, there is strong hydrogen bonding between adjacent urethane linkages, giving intralayer stabilisation. Differences between water contact angles for hydrocarbon surfaces and those reproted here may be due to this hydrogen- bonding character and/or the strong dipolar effects of urethane bonds.The authors thank SERC for a grant (Y.J.) and Dr. R. Greef for his kind help with calculating film thickness from our ellipsometer readings. J. MATER. CHEM., 1991, VOL. 1 References 1 See Memorial Edition, Thin Solid Films, 1980, 68, 1-133. 2 R. Maoz and J. Sagiv, J. Colloid Interface Sci., 1984, 100, 465; J. Gun, R. Iscovici and J. Sagiv, J. Colloid Interface Sci., 1984, 101, 201; J. Gun and J. Sagiv, J. Colloid Interface Sci., 1986, 112, 457. 3 S. R. Wasserman, Y. Tao and G. M. Whitesides, Langmuir, 1989, 5, 1074; N. Tillman, A. Ulman and T. L. Penner, Langmuir, 1989, 5, 101. 4 L. Netzer, R. Iscovici and J. Sagiv, Thin Solid Films, 1983, 99, 235.5 C. D. Bain, J. Eva11 and G. M. Whitesides, J. Am. Chem. SOC., 1989, 111, 7155. 6 L. Netzer, R. Iscovici and J. Sagiv, Thin Solid Films, 1983, 100, 67. 7 G. M. Whitesides and H. Biebuyck, in Molecular Recognition. Chemical and Biochemical Problems, ed. S. M. Roberts, Royal Society of Chemistry, 1989, pp. 270-285. 8 M. Pomerantz, A. Segmuller, L. Netzer and J. Sagiv, Thin Solid Films, 1985, 132, 153. 9 J. Israelachvili and M. L. Gee, Langmuir, 1989, 5, 288. 10 G. H. Davies and J. Yarwood, Langrnuir, 1989, 5, 229. 11 R. C. de L. Milton, S. C. F. Milton and P. A. Adams, J. Am. Chem. SOC., 1990, 112, 6039. 12 R. W. Seymour, S. M. Estes and S. L. Cooper, Macromolecules, 1970, 3, 579. 13 S. R. Holmes-Farley, R. H. Reamey, T. J. McCarthy, J. Deutch and G. M. Whitesides, Langmuir, 1985, 1, 725. 14 F. L. McCrackin, E. Passaglia, R. R. Stromberg and H. L. Stein-berg, J. Res. Nut. Bur. Stand. Sect. A, 1963, 67A, 363. 15 Handbook of Chemistry and Physics, ed. R. C. Weart, CRC Press, Cleveland, 1976-77, 56th edn., E221. 16 A. H. Carim, M. M. Dovek, C. F. Quate, R. Sinclair and C. Vorst, Science, 1987, 237, 630; T. Zhuravlev, Langmuir, 1987, 3, 316. 17 H. C. Brown and B. C. Subba Rao, J. Am. Chem. SOC.,1959,81, 6428. 18 F. Becke and J. Gnad, Annalen, 1968, 713, 212. 19 For a comparable method see, Organic Syntheses, ed. E. C. Horning, Wiley, New York, 1955, vol. 3, p. 846. 20 R. Toubiana and J. Asselineau, Ann. Chem., 1962, 7, 593. 21 W. J. Gender and E. M. Behrmann and G. R. Thomas, J. Am. Chem. SOC., 1951,73, 1071. Paper 1/00237F; Received 17th January, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100457

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(3), 461-468 46 1 Thermal Decomposition of Cobalt(i1) Acetate Tetrahydrate studied with Time-resolved Neutron Diffraction and Thermogravimetric Analysis Robin W. Grimes*a and Andrew N. Fitchb a Davy Faraday Research Laboratory, The Royal Institution of Great Britain, 21 Albemarle Street, London WlX 4BS, UK Department of Chemistry, University of Keele, Keele, Staffordshire ST5 5BG, UK The thermal decomposition of cobalt acetate tetrahydrate has been studied using time-resolved powder neutron diffraction. By using selectively deuterated samples, the loss of water or the breakdown of the acetate group can be identified by following the decrease in the incoherent background of the diffraction pattern as the hydrogen atoms are lost. The results suggest that by 150 "C dehydration is complete and a glass-like phase is formed.Crystallization of this anhydrous acetate occurs at 200 "C. Further heating initiates a two-stage decomposition of the anhydrous acetate terminated by the formation between 275-310 "C of a tetrahedrally co- ordinated cubic zinc blende form of COO. This transforms at 310°C to a rock-salt structure. The neutron diffraction data have been complemented by thermogravimetric and chemical analyses from which we have been able to propose some possible intermediate decomposition products and suggest an explanation for the formation of the unusual zinc blende form of COO. Keywords: Cobalt oxide; Cobalt acetate; Thermogravimetric analysis; Polymorph; Neutron diffraction 1.Introduction The thermal decomposition of cobalt acetate is of particular interest as it can lead to the formation of three polymorphs of COO. The most common form exhibits the rock-salt (sodium chloride) crystal structure (space group Fm3m). The two additional polymorphs have the cubic zinc blende (Fa34 and hexagonal wurtzite (P6,mc) structures.' With respect to co- ordination, the difference between the polymorphs is that in the zinc blende and wurtzite forms both oxygen and cobalt ions are tetrahedrally co-ordinated, whereas in the sodium chloride structure they are octahedrally co-ordinated. The zinc blende and wurtzite forms of COO can be synthe- sized via the decomposition of cobalt(I1) acetate, cobalt(I1) butanoate or cobalt@) ~ctanoate.'-~ In all cases, the decomposition proceeds either in ~)acuo~-~or under an inert atmosphere such as nitrogen'.* or argon.3 Particularly careful temperature control is necessary since the zinc blende and wurtzite structures are only formed in the range 290-310 "C, whereas the rock-salt material forms above 320 "C.After they are quenched to room temperature, the tetrahedrally co-ordinated oxides are completely stable in air. In the first study of COOpolymorphs, Redman and Stew- ard,' using X-ray diffraction, were able to identify zinc blende and wurtzite forms of COO as the decomposition products of cobalt(I1) acetate. In more recent work,, transmission electron microscopy (TEM) techniques were used to establish the morphology of the small zinc blende and wurtzite crystallites formed as the end products of acetate decomposition.The morphology of the polymorphs could be distinguished by TEM and the low-energy planes were identified in each case. For the cubic zinc blende material, which forms at the lowest temperature, the particles take on a characteristic 'tetrahedral' shape. The hexagonal wurtzite polytype, formed at slightly higher temperatures, exhibits 'lath'-shaped crystallites, one end of which is flat, the other pointed. The 'lath'-shaped crystallites were always associated with small clusters of the zinc blende polymorph. It was suggested that growth faults between the zinc blende crystallites would provide low-energy nucleation sites for wurtzite crystallites.The structural study was complemented by lattice energy calculation^,^ which were used to predict cohesive energies and relative permittivities for different cobalt oxide structures. It was found that the total lattice energy favoured the rock- salt structure over both the wurtzite and the zinc blende polymorphs. These theoretical and experimental observations suggest that details of the thermal decomposition process particular to carboxylates are responsible for the formation of a less stable tetrahedral co-ordinated form of COO. The thermal decomposition of anhydrous cobalt@) acetate has been investigated using thermogravimetric analysis.' Edwards and Hayward reported that the acetate decomposed to COO at ca. 350 "C.In addition to the thermal decompo- sition, these workers measured the diffuse visible spectrum of the anhydrous acetate and noted that it was typical of a material with an octahedral co-ordination of metal ions. This hypothesis was supported by measurements of the magnetic moment of the anhydrous material, which were characteristic of an octahedrally co-ordinated high-spin cobalt(I1) ion. Since on average there are only two acetate ligands per cobalt ion, to achieve octahedral co-ordination, the acetate bonding must be bidentate and is probably polymeric. The co-ordination of cobalt ions in cobalt(I1) acetate tetra- hydrate is also octahedral.6 However, in this material the cations are surrounded by four water molecules and by two oxygen ions from unidentate acetate groups. Hence, the way in which the acetate groups co-ordinate around cobalt ions in the acetate tetrahydrate must be different from in the anhydrous material. Cobalt(I1) acetate tetrahydrate has been studied using elec- tron ionization spectroscopy.' In addition, the mass spectrum of the basic form of cobalt(I1) acetate, tetracobalt(I1) hexa- acetate oxide Co,O(Ac),, has also been investigated.* In both cases, the majority gas-phase species were clearly +[Co,O(Ac),]+, [Co,O(Ac),] and the molecular ion [Co,O(Ac),] +.This was interpreted as indicating that the hexaacetate oxide is a stable gas-phase Although the structure of tetracobalt(I1) hexaacetate oxide has not been determined, the structure of the related material Zn,O(Ac), is well known.g It has an analogous structure to J.MATER. CHEM., 1991, VOL. 1 form of COO, a second CO(CH&O~)~.~H~Osample was heated in the neutron beam until just after the single peak was observed. From this temperature, the sample was quenched rapidly, thereby providing a method of synthesising the material, monitored in situ in the neutron beam. This material was subsequently analysed by powder X-ray diffrac- tion and the resulting pattern is shown in Fig. 2. Although the diffraction peaks are broad, presumably a consequence of the small particle size,4 the pattern was easily indexed and the calculated lattice parameter of 4.544(3)A is in good agreement with previous value^.'^^ Above 3 10 "C the diffrac- tion profile is consistent with a rock-salt COO structure.' Thus, the results presented in Fig.1 suggest that by 295 "C the starting material has decomposed to an oxide and that at 310 "C a phase transition occurs between the zinc blende and rock-salt oxide structures. 10000 -111 9000 -8000 -7000 -6000 v)c.5 5000 0 4000 3000 2000 1000 0 Fig. 3 shows a neutron diffraction profile from the sample in which the acetate ligands were deuterated but the water molecules were not. Since deuterium does not create the large incoherent scattering that hydrogen does, the diffracted inten- sity is now discernible above the background. Nevertheless, the background caused by the hydrogen of the water is still considerable and these results show quite clearly that between 100 and 155 "Cthere is a complete loss of water.Also, within this temperature range, as the water is lost, new diffraction peaks are observed in addition to those from the starting material. This suggests that an alternative hydrated acetate structure can be formed which presumably has less than four water molecules. We have noted that there is a considerable difference between the incoherent scattering profiles reported in Fig. 1 and those in Fig. 3; there is also a change in the coherently Fig. 2 X-Ray powder diffraction pattern of the zinc blende form of COO synthesised in the neutron beam and taken on a Philips PW 1050 diffractometer with Cu-Kcc radiation. The calculated lattice parameter is 4.544(3)A.The small peak at 44" is aluminium (200) from the sample plate h I I I I 1 I I 1 6oo 10 20 30 40 50 60 70 80 281" Fig.3 Neutron diffraction profiles of CO(CD,CO~)~*4H20decomposition as a function of heating diffracted profiles. This is due to the different scattering lengths of hydrogen and deuterium, which results in different relative intensities of the diffraction peaks but does not affect the angular distribution. If we now consider in detail the diffraction profiles between 190 and 295 "C it is apparent that a number of different anhydrous forms of cobalt acetate are formed before complete decomposition to the oxide. This is an important observation to which we shall refer in our discussion of subsequent data.We also note that the formation of the zinc blende form of COO can clearly be seen to begin at 270 "C although decompo- sition of the acetate is not complete until 295 "C. In addition, the increase in the diffraction intensity from the zinc blende phase occurs in parallel with a decrease in the diffraction intensity from the last and most well defined of the anhydrous acetate phases (we shall refer to this highest-temperature anhydrous phase as the HT anhydrous phase). As in Fig. 1, the zinc blende to rock-salt transition is observed to occur at 310 "C. In Fig.4 by considering the reduction in the background for the diffraction profiles of decomposing CO(CH,CO~)~ 4D20, we are able to investigate the temperature dependence of acetate loss.The results show that acetate loss begins at 230 "C and is complete by 295 "C as it was in Fig. 1. These temperatures also correspond to the range in stability of the HT anhydrous phase. More detailed analysis of Fig. 4 reveals that there is an initial slow loss of acetate which is accompanied by an increase in the diffraction intensity from the HT anhydrous phase. More rapid acetate loss occurs as the diffraction intensity of the HT anhydrous phase decreases, with zero intensity occurring at the same temperature as the minimum in the incoherent background. Thus, both the formation and the decomposition of the HT anhydrous phase seem to be associated with acetate loss. This observation would be consistent with the HT anhydrous phase being a product of acetate decomposition intermediate between CO(CH~CO~)~and COO.We shall test this hypothesis in sections 3 and 4. The last feature in Fig. 4 to which we wish to draw attention is the apparent lack of any sharp Bragg peaks at a temperature of ca. 170 "C. A similar situation occurs in Fig. 1 and 3 but it is obscured by adjacent data at higher and lower tempera- tures. The feature is more visible in Fig. 4 because of a sudden small drop in the incoherent background which occurs at 170 "C. This single isolated reduction in the background is a typical consequence of sample movement in the neutron beam J. MATER. CHEM., 1991, VOL. 1 and is not due to any physical or chemical change in the sample.The last sample (Fig. 5) has deuterated acetate and water ligands. Consequently, there is very little incoherent scattering and hence a very low background. This results in clear diffraction profiles. Therefore, having determined from the data presented in Fig. 1, 3 and 4 the temperatures at which the acetate and water ligands are lost, the detailed effect that such chemical changes have on the diffraction profiles can be realised in Fig. 5. To capitalise on this further, the data in Fig. 5 have been reformulated in such a way that the diffrac- tion profiles are viewed in the reverse direction to the previous figures, that is in order of descending temperature (ie. the highest temperature profile is now at the back of the plot).The data have also been partitioned into three overlapping temperature ranges: (a) 252-25 "C;(b) 295-167 "C; (c) 376-273 "C. Together these constitute Fig. 6. It is important to be aware that the diffraction intensities of each set of profiles in Fig. 6 have been normalised to the largest diffraction peak of each set. As such, the three sets of profiles (a),(b) and (c) no longer connect in an obvious way. However, this transform- ation has the advantage that more of the detailed diffraction data is discernible, particularly in the high-temperature pro- files which have lower absolute intensities. In describing the data of Fig. 6, we can also summarise the results of the neutron diffraction experiments as a whole. As heating proceeds, there is little change in the diffraction profile until 155 "C when the intensity drops almost to zero.The relatively small modulations in diffraction-peak heights that precede 155 "C might justifiably be attributed to movements of the material in the sample holder. The extra diffraction peaks that were noted during the decomposition of CO(CD~CO~)~*4H20 were absent with this last fully deuter- ated sample. Fig. 6(b)begins just after the water ligands have been lost and during the period in which there are no well defined diffraction peaks. The broad undulations of the diffraction intensity that are seen between 160 and 200 "C are typical of a material in a glassy phase. Beyond 200 "C, the broad profile transforms into a well defined crystalline pattern.This anhy- drous phase rapidly reduces in intensity as the HT anhydrous phase begins to appear at 250 "C and as acetate ligands start to devolve. The maximum intensity of the HT anhydrous phase occurs at 270 "C. In Fig. 6(c), at 273 "C, the diffraction intensity of the HT anhydrous phase is falling and the zinc blende form of COO I I I I I I I 1 €jo0 10 20 30 40 50 60 70 80 2810 Fig. 4 Neutron diffraction profiles of Co(CH,COJ2 *4Dz0decomposition as a function of heating J. MATER. CHEM., 1991, VOL. 1 Fig. 5 Neutron diffraction profiles of CO(CD,CO~)~*4D20decomposition as a function of heating is starting to appear. The maximum for the zinc blende structure occurs at 310 "C after which COO quickly takes on the rock-salt form. This result is supported by previous work1l4 and is consistent with a solid-state phase transition.We note that during the decomposition of all four differently deuterated samples, no neutron diffraction evidence was found for the formation of the wurtzite form of This is in fact not a particularly surprising result as the wurtzite form of COO is not always observed during decomposition. It has been suggested that wurtzite formation requires overlapping zinc blende structured crystallites4 and these may well not be present in sufficient density to permit the formation of wurtzite in the relatively short time before the rock-salt form of COO becomes overwhelmingly favoured. 3. Thermogravimetric Analysis Thermogravimetric analysis (TG) was carried out under flow- ing argon using a Stanton balance.While the sample was heated at 1 "C min-' up to 500 "C the weight was monitored continuously. As the final product of heating was COO, the total weight lost during the experiment implied that the starting material had a molecular weight of 279.3. This corresponds to a material with excess surface water or acetic acid equivalent to 1.35 water molecules per formula unit. The thermogravimetric data were also used to calculate the weight loss over 5 min intervals. This resulted in the difference thermogravimetric profile for cobalt@) acetate tetrahydrate decomposition shown in Fig. 7. From the diagram it is immediately apparent that the decomposition proceeds in two stages.The first weight loss (54% of the total) occurs between 20 and 150 "C. In the neutron diffraction study, the equivalent temperature range corresponds to the loss of the hydrogen associated with water ligands. The second decomposition stage occurred between 230 and 315 "C (46% of the total weight loss). By comparison with the neutron diffraction results this decrease in sample weight can be related to the loss of the acetate hydrogen. The important point that we can now make is that the weight changes determined from TG are consistent with the hypothesis that the loss of hydro- gen associated with either the water or the acetate implies the loss of the whole ligand. In addition the results suggest that if there is excess acetic acid left over from the sample prep- aration, this is lost with the water.Lastly, we note that decomposition temperatures determined from the thermograv- imetric and diffraction experiments disagree by k10 "C. This is expected since somewhat different temperature gradients will be in evidence across the samples in the different experi- mental set-ups. Between 150 and 230 "C the thermogravimetric data (see Fig. 7) show that the sample is stable to decomposition. However, at 200 "C the diffraction data suggest that the sample changes from being glass-like to crystalline. Therefore, we may conclude that the phase change at 200 "C is physical and not driven by a decomposition process. In Fig. 7, it can be seen that the decomposition of the anhydrous acetate (between 230 and 315 "C) occurs in two steps.The first stage between 230 and 270°C accounts for 28% of the acetate weight loss. The second process starts at 250 "C,overlapping slightly with the first, and finishes at 315 "C with the formation of COO. From a comparison with diffraction data, it would seem logical to suggest that the first stage of acetate loss results in the formation of the HT anhydrous phase (discussed in section 2.2 while commenting in Fig. 3-5). On the basis of previous data concerning the decomposition of cobalt(I1) a~etate,~we suggest that the formula of the intermediate decomposition HT anhydrous phase is CO~(CH&O~)~O Ctetracobalt(I1) hexaacetate oxide]. As such, we would expect its formation to occur with the loss of 25% of the total acetate weight.Although this is slightly lower than the experimental value for first-stage anhydrous acetate decomposition (28%) the discrepancy could easily be accounted for if just 4% of the hexaacetate oxide had decomposed to COO (see Table 1). This possibility is not inconsistent with diffraction results since in Fig. 6(c) we noted that the initial formation of COO occurred very early in the stability range of the HT anhydrous phase. Nevertheless, on the basis of these data alone we cannot rule out alternative explanations for the discrepancy. For example, the incorporation into the HT anhydrous phase of basic cobalt acetate fragments such as those observed in mass-spectroscopic studies.'** However, it is not clear if these fragments would be stable in the solid state.The thermogravimetric data in the temperature range 20- 150°C suggest that during the loss of water and/or excess acetate intermediate products are also formed. In section 2.2 we commented that extra peaks observed during the decompo- sition of CO(CD&O~)~ *4H20 might be associated with the formation of a secondary hydrous phase. If such a material did form, the thermogravimetric results would be consistent with the formation of CO(CH~CO~)~*2H20 at 100 "C. We note that an analogous phase is known for zinc.'0b However, on the basis of our results this suggestion must regarded as simply speculative. 4. Carbon/Hydrogen Analysis Additional support for our hexaacetate oxide model of anhy- drous acetate decomposition has been obtained through CHN analysis.The results presented in Table 1 are for a partially decomposed anhydrous acetate sample which was prepared in a tube furnace, under flowing argon, from cobalt(I1) acetate 295 2 16?\ 1 I 20 J. MATER. CHEM., 1991, VOL. 1 tetrahydrate heated at 1 "C min-' to 260 "C. By weighing the sample before and after CHN analysis a molecular weight of 143.3 was determined for the partially decomposed acetate. This value implies that the decomposition of the acetate has proceeded to 33%, that is slightly beyond the first stage of acetate loss. Therefore, from the diffraction results we can assume that some CO~(CH~CO~)~O has decomposed to COO.By fitting a combination of these two compounds to yield an average molecular weight of 143.3 the hypothetical carbon/ hydrogen content of the mixture can be calculated. The results of this process, presented in Table 1, show an excellent agree- I I I 1 I I 40 60 80 281" J. MATER. CHEM., 1991, VOL. 1 (c1 /_nT I I I I 80 I 20 I II 40 60 Fig. 6 Neutron diffraction profiles of Co(CD,C0,),*4D20 decomposition as a function of inverse heating. (a) 252-25 "C;(b)295-167 "Cwhere G-L is the glass-like phase and HT is the HT anhydrous phase; (c) 376-273 "C where ZB is the zinc blende COO phase and RS is the rock- salt COOphase Tl°C 20 50 100 150 200 250 300 350 I I I I I I I I 80 7 60 .-c E \F rn' E 40 CJ).-P 20 I\.-._._.-.:-.-.-.-..-.--: I I '.\. I I 1 I I 1 \I0 0 30 60 90 120 150 180 210 240 270 300 330 Fig. 7 Difference thermogravimetric profile determined for the decomposition of cobalt(1x) acetate tetrahydrate under flowing argon at a heating rate of 1 "Cmin-'. (a) Loss of water; (b)loss of acetate ment with the experimental CHN analysis. It is gratifying intermediate decomposition products for cobalt@) acetate that such a simple model is consistent with diffraction, thermo- tetrahydrate. Initially, water and/or excess acetate is lost gravimetric and CHN results. resulting in a glassy-phase anhydrous acetate. A glassy phase is able to form since, as explained in the Introduction, there 5.Summary is a complete change in the co-ordination of the cobalt ion as a consequence of dehydration. When it is heated further, From the complementary use of diffraction and thermogravi- the glassy phase first crystallises and then begins to decom- metric techniques we have been able to identify a number of pose. We suggest that the anhydrous acetate decomposition J. MATER. CHEM., 1991, VOL. 1 Table 1 Analysis of hexaacetate oxides as partial decomposition products of anhydrous cobalt acetate C/H analysis Co(Ac),Co,(Ac),Omodel fitted to a weight loss of 28% 0.96 Co,(Ac),O+0.16 COO experimental values for a partially decomposed acetate sample model fitted to a molecular weight of 143.4 0.89 Co4(Ac),0+0.42 COO occurs via CO~(CH~CO~)~Oand that this is a stable, crystal- line, intermediate.This result is consistent with diffraction, thermogravimetric and chemical analyses. Further heating to 290 "C results in the total decomposition of the oxyacetate to form the zinc blende form of COO.This unusual form of COO transforms at 310 "C to the rock-salt polymorph. Previous theoretical calculations4 suggest that the rock-salt form of COO is more stable than the zinc blende. We wish, on the basis of our present observations, to offer one possible explanation for the formation of the zinc blende polymorph. By inference from experimental structures of similar hexaacet- ate oxides, it would seem that our basic acetate intermediate decomposition product is composed of structural units of CO~(CH~CO~)~Oin which the four cobalt ions tetrahedrally co-ordinate a central oxygen ion and are themselves tetra- hedrally co-ordinated by oxygen.During the final decompo- sition, this structure might provide a stable nucleus which would lead to a tetrahedrally co-ordinated zinc blende form of COO rather than necessitate a complete change in co- ordination in forming the octahedrally co-ordinated rock-salt structure. We are currently attempting to determine the structure of the intermediate decomposition product by using powder X-ray diffraction techniques. molecular wt. per Co ion C (%) 177.0 27.1 151.5 23.8 148.4 23.3 143.3& 0.2 22.2& 0.3 143.4 22.4 H (YO) weight loss (YO) 3.4 0 3.0 25 2.9 28 2.9& 0.3 33 2.8 33 We thank the Institut Laue-Langevin for the provision of the neutron beam time on DlB, the Department of Chemistry, University of Keele for the CHN analysis and Richard Catlow for useful discussions. References 1 M.J. Redman and E. G. Steward, Nature, (London) 1962, 193, 867. 2 M. J. Redman, Ph.D. Thesis, Royal Society of Chemistry, 1961 (available at City University Library, London, UK). 3 A. N. Fitch, R. W. Grimes and C. R. A. Catlow, I.LL Report 88CA03G, 1987, p. 149. 4 R. W. Grimes and K. P. D. Lagerlof, J. Am. Ceram. SOC., 1991, 74, 270. 5 D. A, Edwards and R. N. Hayward, Can. J. Chem., 1968, 46, 3443. 6 J. N. van Niekerk and F. R. L. Schoening, Acta Crystallogr., 1953, 6, 609. 7 G. C. DiDonato and K. L. Busch, Inorg. Chem., 1986, 25, 1551. 8 J. Charalambous, R. G. Copperthwaite, S. W. Jeffs and D. E. Shaw, Inorg. Chem. Acta, 1975, 14, 53. 9 H. Koyama and Y. Saito, Bull. Chem. SOC.Jpn., 1954,27, 112. 10 R. W. G. Wykoff, Crystal Structures, John Wiley, New York, 2nd edn., 1968, vol. 5, (a)p. 336, (b)p. 329. Paper 1/00661 D; Received 12th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100461

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991,1(3), 469-472 Surface Structure of a Glassy Carbon Scanning Tunnelling Microscopy Study Norman M. D. Brown* and Hong Xing You Surface Science Laborators Department of Applied Physical Sciences, University of Ulster, Coleraine, Co. Londonderry BT52 ISA, UK Scanning tunnelling microscopy (STM) has been used to study the topographical details of the surface structure of a glassy carbon at ultra-high resolution. In addition to small, localised graphite-like domains <50 A x 50 B( in area, the STM images of the glassy carbon studied, taken in the constant-tunnel-current mode in air, are in general characterized by two topographical features: (i) a granular structure with grain sizes typically in the range 80.0-250.0A in some surface locations and (ii) straight or curved fibrillar structures lying side by side, in other surface locations.The latter are considered to reflect the residual structure of the polymer chains initially present during processing of the material. Keywords: Scanning tunnelling microscopy; Glassy carbon; Surface structure; topography 1. Introduction Glassy carbons, which exhibit uniqueness and diversity in their physical, chemical and mechanical properties, such as high mechanical strength, thermal stability, inertness to strong oxidising agents, high biocompatibility, good electrical con- ductivity etc., have started to find wide application in industry, engineering and scientific The physical, chemical and mechanical properties of interest are heavily dependent on the microstructure of the glassy carbon.So, for a better understanding of the relationship between these properties and structure, and to broaden the areas of application, glassy carbons have been studied widely by X-ray diffra~tion,'.~ small-angle X-ray scattering,' vs low-energy electron diffrac- tion: X-ray photoelectron ~pectroscopy,~ electron microscopy [SEM (scanning) and TEM (transmi~sion)]'~~~~ and Raman ~pectroscopy.~The focus of such structural studies is on establishing whether glassy carbons have a random disordered structure or are more graphite-like. However, the fundamental understanding of the structures of glassy carbons has been hampered by the lack of techniques for revealing local struc- tural details at ultra-high resolution, namely on the iO-9-iO-'0 m scale.With the invention of the scanning tunnelling microscope,1o it became possible to study the structure of such surfaces directly in real space, even down to atomic resolution. Even so, little is known about the surface structure of glassy carbons at scales below a nanometre. So far, the sensitivity of the STM tip to loose particular carbon on the surface appears to have prevented topographical details from being obtained at ultra-high resolution, especially when operating in air."*" The first report of the use of STM techniques to a commercial pitch-based glassy carbon has been published recently." The STM images obtained with the STM tip and the sample under paraffin oil showed ordered regions extending over no more than ca.20A. These regions consist of atomic rows spaced at either cu. 1.5 A or ca. 2.2 A; no explanation was given for the occurrence of the two different periodicities on this particular glassy carbon surface. Atomic resolution STM image, i.e. periodic rows, has also been observed from a further commercial glassy carbon surface, again in an oil environment, but in this case no structural details were presented.'* Furthermore, even though small graphitic domains in a pyrolysed polyimide matrix have been observed by STM, no other detailed surface structures were revealed in that case.13 In this paper, we present topographical details observed by STM in air of the surface of a commercial glassy carbon.First, it is shown that small, graphite-like domains are present on the surface of the glassy carbon sampled. Secondly, evi- dence for the existence of a residual polymeric structure in the form of straight or curved fibrils in some surface locations is presented. Finally, a more granular structure, frequently observed at other surface locations, is also shown. 2. Experimental The glassy carbon samples used (Cardio Carbon, Swansea) were prepared by heating a newly developed phenolic resin in shaped pieces to temperatures in excess of 1000 "C in an inert atm0~phere.l~ These were then ground and polished using a 4 pm diamond particle-impregnated lead pad to give a uniform thickness and surface finish, as required. As- received plates of this glassy carbon were cut into 10 mm x 8 mm x 1.0 mm pieces prior to mounting on the sample holder of the scanning tunnelling microscope (W.A. Technology, Cambridge). This consists of a scanning head suitable for operation in either air or vacuum, an electronic control unit, a Tandon 386 microcomputer with VGA colour monitor and frame store facility, and a monochrome monitor for image display. The STM tips were prepared by the electrochemical etching of a 0.5 mm diameter tungsten wire in an aqueous 1 mol dm-3 KOH ~olution.'~These were then rinsed in distilled water and finally in redistilled acetone before use. The STM images, composed of 256x256 pixels, were obtained in air in the constant-tunnel-current mode with tip transit speeds between 1 and 100 A s-'. Typical tunnelling conditions were 1 nA for the tunnel current and 800-1500 mV for the sample bias (with the sample positive).3. Results In the first instance, much effort was devoted to the search for graphite-like structures on the surface of the glassy carbon examined. Thorough and careful imaging in various surface locations showed that local graphite-like topography could be detected reproducibly by the STM tip. Fig. 1 illustrates the two kinds of topographical features repeatedly observed on the surface of the glassy carbon sampled. One of these is characterized by a mixture of incomplete hexagonal arrange- ments and periodic rows; the other is characterized by atomic Fig. 1 Typical STM images of the glassy carbon taken in the constant- current mode (1.0 nA) showing (a)a mixture of incom lete hexagonal arrangements and periodic rows (image dimension 20 1x 20 A x 5 A; sample bias, 1006 mV) and (b) the zigzag atomic rows (image dimen- sion 21 x 21 A x 7 A;sample bias, 1006mV) rows running in a zigzag manner.Typically, both graphite- like topographies appear, as shown in Fig. 1, on small scan areas, i.e. 20 A x 20 A and 21 A x 21 A. The spacings of the atomic rows in Fig. l(b) are measured to be 4.2f0.1 A, the same as the distance between two adjacent atomic corru- gations of the incomplete hexagonal arrangements in Fig. l(a).’ Fig. 2 shows further graphite-like surface details over a larger scan area, i.e. 39 A x 39 A.Upon close inspection of Fig. 2, it can be seen that there exists a small-scale hexagonal ordering in the right front part of the image with the distance between two adjacent atomic corrugations of 4.2 & 0.1 A. The atomic rows running in a zigzag manner characterise the rest of the image, with the same row spacing (4.2 f0.1 A). The second STM topographical feature frequently observed J. MATER. CHEM., 1991, VOL. 1 on the surface of the glassy carbon studied here, when scanning over areas of some hundreds of square Angstroms, is that of a straight or curved fibrillar structure. Fig. 3 and 4 show images typical of those observed. In Fig. 3(a), there are several aligned fibrils in an uneven form, with a typical fibrillar separation of ca.44.0 A.In Fig. 3(b), the aligned fibrils are rather smoother in outline with a different fibrillar separation of ca.34.0 A. In addition, somewhat curved fibrils, sitting side by side, could occasionally be observed on the surface of the glassy carbon, as shown in Fig. 4. Fig. 4(a) and 4(b),obtained from different positions of the sample surface, show two kinds of curved fibril. In Fig. 4(a), they have an uneven form and different corrugation heights, whereas in Fig. 4(b) they are markedly smooth with nearly the same corrugation height. The fibrillar separations are measured to be ca.32.0 A for Fig. 4(a) and ca. 46.0A for Fig.4(b). By changing the tip scanning directions, the alignment of the fibrillar structure is not changed in any one location.Correspondingly, the fibrillar structure was found to have different alignments on moving the tip to another location on the sample surface. Thus, the possibility of the fibrillar structure arising from the tip-induced artefacts can be discounted. A third STM topographical feature frequently observed on the surface of the glassy carbon examined is the granular structure found over rather larger scan areas. As shown in Fig. 5, the granules with different sizes, typically ca. 168.0 A and ca.230.0 8, [Fig. 5(a)], and with nearly uniform size, typically ca. 165.0 A or 180.0 A [Fig. 5(b)],are distributed on either side of a boundary. Similarly, Fig. 5(c) shows an island on the surface of the glassy carbon composed of granules of various shapes and sizes, in the range 80.0-170.0 A.The STM images showing the above topographical features Fig. 3 Typical STM images of the glassy carbon taken in the constant- current mode showing the straight fibrillar structure (a) with an uneven form (image dimension, 380 A x 380 A.x 49 A;sample bias, Fig. 2 An STM image of the glassy carbon with a somewhat larger 1050 mV; tunnel current, 0.8 nA) and (b) with a more smooth form graphite-like structure (dimension, 39 8,x 39 A x 8 A;sample bias, (image dimension, 219 A x 219 A x 78 A; sample bias, lo00 mV; tunnel 1000 mV; tunnel current, 1.0 nA) current, 1.0 nA) J. MATER. CHEM., 1991, VOL. I Fig. 4 Typical STM images of the glassy carbon taken in the constant- tunnel-current mode showing the curved fibrillar structure (a) with an uneven form (image dimension, 500 8,x 500 8, x 19 8,;sample bias, 1045 mV; tunnel current, 1.0 nA) and (b)with a more smooth form (image dimension, 282 8, x 3 14 8, x 35 8,;sample bias, 1034 mV; tunnel current, 0.8 nA) were routinely checked for reproducibility by scanning in either the x or the y direction with different sample biases, in the range 800-1500mV in both bias directions, and with different tunnel currents.Thus, all the topographies presented here were found to be reproducible. In addition, while it was found that the fibrillar structure referred to was usually not tangled in appearance, some STM images did show regions probably with tangled fibrils, existing where individual fibrils could not be distinguished clearly.Furthermore, the individual straight fibrils identified were found to extend over several hundreds of Angstroms without a change in their alignment. The curved fibrils were often observed in more isolated local areas. Occasionally, some curved fibrils were found with one end running into a straighter fibrillar region. Finally, it should be noted that the fibrillar structure and the granular structure identified were seldom found side by side on the sample surface. In the above survey of the samples, some of the STM images obtained did not show any resolvable topographical features. These might arise from the contaminated locations on the sample surface, or from surface regions of poor conductivity, or more likely from high local disorder.4. Discussion Now the topography of graphite has been studied extensively and in depth by the scanning tunnelling microscope in under water16 and in UUCUO.~~The occurrence of so-called giant corrugations within the unit cell, often close to or exceeding the lattice constant in the graphite (0001) plane (2.46 A), particularly when operating in air and water, is generally attributed to elastic deformations induced by the interatomic forces between tip and ~urface.'~ In the work 47 1 Fig. 5 Typical STM images of the glassy carbon taken in the constant- current mode (1.0 nA) showing the granular structure. (a)The granules with two different sizes sitting on either side of a boundary (image dimension, 2000 8, x 2000 8,x 200 8,;sample bias, 1000 mV); (b) the granules with nearly the same size on either side of a boundary (image dimension, 2000 8, x 2000 8, x 196 8,; sample bias, 1200 mV) and (c) an island containing the granules described with various sizes (image dimension, 1000 8,x 1000 A x 104 8,;sample bias, 1 100 mV) described here, the giant corrugations of 2.0-5.0 8, are in the same range as those of an earlier graphite study.I6 It is found that the STM tip sometimes cannot resolve the individual atoms of the graphite unit cell, particularly in the case of more amorphous carbon materials.11*12*20 Here, the zigzag rows of poorly resolved atoms appear in the locally ordered regions imaged, as shown in Fig.l(b) and Fig.2. The periodic phenomena shown in the various figures might be considered to arise from multiple tunnelling as a consequence of the condition of the tip used, e.g. the tip may have extraneous molecules adsorbed on it or have closely adjacent high spots. On the other hand, the periodicities of 4.2 0.1 8, found are larger than the lattice constant of the normal graphite unit cell (2.468,), but they are very close to J3 x2.46 A (4.26A), the lattice constant of the unit cell of the J3 x J3 graphite structure. Since multiple-tip effects, which sometimes occur in STM graphite studie~,'~,'~ seem unlikely to match these dimensions, they are discounted here, particularly since such image details were reproducible with different tips. In addition to the above, high-resolution electron micro~copy'*~-~has revealed that glassy carbons consist of a random arrangement of labyrinth-like carbon chains or rib- bons of near polymeric chain dimensions, i.e.a few tens of square Angstroms in cross-section with several tens or hun- dreds of Angstroms between cross-links. Here, the straight and curved fibrillar structures (shown in Fig. 3 and 4)lying side by side on the surface of the glassy carbon are assumed to be derived from a single polymer chain or possibly from small bundles of a few polymer chains.21i22 The uneven or smooth forms of the fibrils might be associated with pro- cessing-dependent changes of the local surface of the glassy carbon studied and/or be related to the structure of the polymer precursor used. From an earlier electrical study of a glassy ~arbon,~ the mean free path of conducting electrons has been estimated to be ca.15 A, again on the scale of the width of a polymer chain. Multiples of this width are matched reasonably well by the fibrillar separations or widths shown in Fig. 3 and 4. Hence, the significance of the mean free path as a regular fraction of the fibrillar features is stressed. Depending on the processing temperatures and the polymer resins used, the yields of glassy carbons from polymer pyrolysis are usually 40-60%.',23 Accordingly, some of the polymer resins do not given an amorphous glassy carbon but retain their own general structure, as suggested by Fig. 3 and 4. Other starting materials generally transform to a disordered non-graphitized carbon with glass-like structural properties.The crystallite sizes of such glassy carbons measured by X-ray diffracti~nl-~~~are generally no more than 100 A. Therefore, the granular sizes in the range 80.0-250.08, found in this work (Fig. 5) are somewhat larger than those indicated by X-ray diffraction. In this regard, it should be noted that the scanning tunnelling microscope can measure the distribution of granular sizes concerned, whereas X-ray diffraction can only give a mean value, because the X-ray beam samples a significantly larger sample cross-section. 5. Conclusion The STM images of glassy carbon show the following topo- graphical features: (1) the surface of this glassy carbon is generally characterised by a non-graphitic str~cture.~*~*' How-ever, in very local positions, small graphite-like domains of <50 A x 50 A are observed.The 4.2k0.1 A periodicity of these domains is commensurate with the lattice constant of the 43xJ3 graphite structure unit cell (4.26 8,). (2) The aligned fibrillar structures with a straight or curved shape are observed in some locations of this glassy carbon surface. This fibrillar structure is believed to reflect the structure of the J. MATER. CHEM., 1991, VOL. 1 precursor polymer chains. (3) The granular structure found at other surface locations suggests a crystallite size generally larger than the mean derived by X-ray diffraction for such materials.This work provides impetus for further STM studies of the relationship between the surface structures of glassy carbons, the precursor materials and their processing. Likewise, STM offers much in the investigation of. the surface-modification work on such systems. H-X. Y. thanks the University of Ulster for the postgraduate research studentship. Thanks are also due to the IRCSS at the University of Liverpool and to the International Fund for Ireland for support. The author thanks Dr. W. D. Unsworth of Cardio Carbon Company Ltd. for the supply of samples. References 1 G. M. Jenkins and K. Kawamura, in Polymeric Carbon-Carbon Fibre, Glass and Char, Cambridge University Press, Cambridge, 1976. 2 J. Rautavuori and P. Tormala, J. Muter.Sci., 1979, 14, 2020. 3 P. Delhaes and F. Carmona, in Chemistry and Physics of Carbon, ed. P. L. Walker and P. A. Thower, Marcel Dekker, New York, 1981, vol. 17. 4 D. F. Baker and R. H. Bragg, Phys. Rev. B, 1983, 28, 2219. 5 R. Perret and W. Ruland, J. Appl. Crystallogr., 1972, 5, 116. 6 S. R. Kelemen, H. Freund and C. A. Mims, J. Vac. Sci. Technol. A, 1984, 2,987. 7 A. A. Galuska, Appl. Sure Sci., 1989, 40, 19. 8 R. H. Bragg, Synth. Met., 1983, 7, 95. 9 A. V. Baranov, Ya S.Bobovich and V. I. Petrov, Opt. Spectrosc., 1987, 63, 606. 10 G. Binnig, H. Rohrer, Ch. Gerber and E. Weibel, Phys. Rev. Lett., 1982, 49, 57. 11 V. Elings and F. Wudl, J. Vac. Sci. Technol. A, 1988, 6, 412. 12 T. Endo, H. Yamada, T. Sumomogi, K. Kuwahara, T. Fujita and S. Morita, J. Vac. Sci. Technol. A, 1990, 8, 468. 13 C. Z. Hu, L. Feng and J. D. Andrade, Carbon, 1988, 26, 543. 14 W. D. Unsworth, Technical note (CC067) on TURBOFORM carbon, Cardio Carbon Company Ltd. July, 1990, 15 N. M. D. Brown and H-X.You, Sure Sci., 1990,233, 317. 16 J. Schneir, R. Sonnenfeld and P. K. Hansma, Phys. Rev. B, 1986, 34,4979. 17 R. J. Colton, S. M. Baker, R. J. Driscoll, M. G. Youngquist, J. D. Baldeschwieler and W. J. Kaiser, J. Vac. Sci. Technol. A, 1988, 6, 349. 18 L. L. Soethout, J. W. Gerritsen, P. P. M. C. Groeneveld, B. J. Nelissen and H. Van Kempen, J. Microsc., 1988, 152, 251. 19 J. M. Soler, A. M. Baru and N. Garcia, Phys. Rev. Lett., 1986, 57,444. 20 B. Marchon and M. Salmerson, Phys. Rev. B, 1989,39, 12907. 21 N. M. D. Brown, H-X.You, R. J. Froster and J. G. Vos, J. Muter. Chem., in the press. 22 T. R. Albrecht, M. M. Dovek, C. A. Lang, P. Grutter, C. F. Quate, S.W. J. Kuan, C. W. Frank and R. F. W. Pease, J. Appl. Phys., 1988, 64, 167. 23 D. W. McKee, Annu. Rev. Muter. Sci., 1973, 3, 195. Paper 0/05458E; Received 4th December, 1990

ISSN:0959-9428    

DOI:10.1039/JM9910100469

出版商:RSC    

年代:1991

数据来源: RSC