8968
J . Phys. Chem. 1991, 95, 8968-8972
from the difference between the calculated and experimental potential energy [-8.74 f 0.12 (Table I) and -9.9 kcal/mol,28 respectively], this difference arising primarily from use of a cutoff and from the omission of many-body interactions. However, the calculated value (-4.46 f 0.12 kcal/mol) agrees with the results of other computational methods with the same (CI) water model (-4.3 1 f 0.07 kcal/mol in the a b initio perturbation method29 and -4.30 f 0.06 kcal/mol in the recursion method22). From Table V, the avera e molar excess free energies of the solvent water molecules (AH2o) e# are higher than that in pure liquid water (-4.46 f 0.1 2 kcal/mol) for all of the aqueous solutions considered here; i.e., the solvent water molecules become more or less unstable compared with the situation in pure liquid water. However, in considering the total exfree energies per one water niolecule in these systems (A:t,/N), the value for the CH4 solution is higher than that in pure liquid water, whereas it is lower for the other solutions with polar solute molecules; Le., the CH4 solution is less table than pure liquid water, whereas the other solutions are more stable.
Note Added in Proof. Recently, F. Huisken and M. Stemmler (Chem. Phys. Lett. 1991,180,332) used molecular beam depletion spectroscopy to study the methanol-water complex, with experimental results in agreement with the a b initio calculations of Kim et al.12that were used in the present paper for the methanol-water system. Acknowledgment. This work was supported by the US.-Korea Cooperative Science Program between the National Science Foundation and the Korea Science and Engineering Foundation (NSF Grant INT-87-05307), by the Korea Research Center for Theoretical Physics and Chemistry, by the National Science Foundation (NSF Grant DMB84-0181 l ) , and by the National Institute of General Medical Sciences, National Institutes of Health (NIH Grant GM-14312). Registry NO. HZO, 7732-18-5; CH,, 74-82-8; CH3OH, 67-56-1; CH,NH,,74-89-5; CH3COOH. 64-1 8-6; CHINH3+, 17000-00-9; CH3COO-, 7 1-50- I .
Complications in the Measurement of the Magnetic Field Penetration Depth in Y Ba,CuSO, and Bi,Sr,CaCu,O, Superconductors by Electron Spin Resonance Line Broadening of Surface Paramagnetic Probes Jerzy T. Masiakowski, Micky Pun, and Larry Kevan* Department of Chemistry and The Texas Center for Superconductivity, University of Houston, Houston, Texas 77204-5641 (Received: February 1, 1991)
The electron spin resonance line broadening of a paramagnetic probe on an oxide superconductor below the superconducting critical transition temperature is shown to depend on the parallel or perpendicular orientation of the probe to the applied magnetic field. This effect must be considered when using the line broadening associated with the magnetic flux lattice to measure the magnetic field penetration depth. The appropriate values at 0 K of the magnetic field penetration depth in YBa2Cu30xand Bi&CaCu20x sintered superconductors are 1400 and 2850 A, respectively. A strong effect of silver doping in YBaCuO superconductors on the magnetic field penetration depth is also shown. It is shown that the temperature below the superconducting critical transition temperature at which significant line broadening from the flux lattice occurs is better identified as the magnetic flux lattice melting temperature. This is particularly clear in the BiSrCaCuO superconductor in which the flux lattice melting temperature lies significantly below the superconducting critical transition temperature. This also supports that line broadening associated with the magnetic flux lattice can be isolated from other sources of broadening.
Introduction Recently, Rakvin et al.' have used electron spin resonance (ESR) line broadening of a paramagnetic probe to measure the magnetic field penetration depth X by adsorption of diphenylpicrylhydrazyl (DPPH) on the surface of a superconducting sample. The method is based on the original nuclear magnetic resonance methodology proposed by Pincus et aL2 for conventional type 11 superconductors. For type I1 superconductors, below the superconducting critical transition temperature T, and between the lower critical magnetic field (Ifcl)and the upper critical magnetic field ( I f c 2 ) ,both superconducting and normal states coexist in a mixed state forming a regular periodic structure of the magnetic field distribution which is called a vortex lattice or a magnetic flux lattice. Below T, the ESR line width of a paramagnetic probe is broadened by the inhomogeneous magnetic field associated with the emergence of a magnetic flux lattice in the superconductor. ( I ) Rakvin, B.; Pozek, M.; Dulcic, A. Solid Srare Cmmun. 1989, 72, 199. (b) Rakvin, B.;Mahl, T. A.; Bhalla, A. S.;Sheng, Z. Z.; Dalal, N . S. fhys. Rev. B 1990, 41, 769. (2) Pincus, P.; Gossard, A. C.; Jaccarino, V.; Wernick, J. H. f h y s . Leu. 1964, 13, 21.
0022-3654/91/2095-8968$02.50/0
For high-T, oxide superconductors the ESR resonant field of a paramagnetic probe (around 3000 G for X-band ESR) lies between H,, and Hc2in the mixed-state region. The variance of the magnetic field distribution caused by the vortex lattice is field independent between HCIand Hc2and is given by eq 1 where A ( A f l ) = A@02X4 (1) is a constant and O0 is the magnetic flux quantum. Thus, ESR line broadening can be used to monitor the inhomogeneity of the magnetic field a t the surface of a superconducting sample ifthe line broadening associated with the flux lattice can be isolated. From an appropriate temperature dependence of X in eq 1, one can determine the magnetic field penetration depth a t 0 K, &, by measuring the temperature dependence of the ESR line width of a probe on the surface of a superconducting sample. This method has the important advantage of easy sample preparation. The spin probe in a solution can be easily deposited on the surface of a superconducting sample by evaporation of the solution. The probe, however, should cover only a small fraction of the entire surface of the sample. Rakvin and co-workers' prepared their samples by immersing pieces of oxide superconductors in a spin probe solution and hence coated the entire superconducting surface with the spin probe. For such a prepa0 1991 American Chemical Society
Y Ba2Cu30, and Bi2Sr2CaCu20xSuperconductors
The Journal of Physical Chemistry, Vol. 95, NO. 22, 1991 8969
r
I
I
40
I
I
I
60
80
100
T, K
Figure 1. A spherical superconducting sample showing a DPPH probe in H, and H,orientations. A schematic representation of the vortex latticc is shown as dashed lines parallel to H applied. The inhomogeneity of the magnetic field, AH, due to the vortex lattice broadens the DPPH ESR line only when the probe is in the H, orientation. The periodicity of the vortex lattice is d .
Figure 3. Shifts of the ESR resonant field as a function of temperature for DPPH adsorbed on the surface of a YBaCuO sphere. The negative shift is observed for the parallel orientation of the DPPH probe with respect to the external magnetic field H. The positive shift is observed for the perpendicular orientation of DPPH relative to H.
be reliably estimated. We also show that the flux lattice melting temperature can be detected if it is significantly different from T,. This suggests that we can isolate the broadening due to the magnetic flux lattice.
Experimental Section The experiments were carried out on a Bruker ESP 300 ESR spectrometer (X band) with an Oxford Instruments Model 900 variable-temperature liquid helium flow system. To fit the experimental data to the theoretical curve from eq (l), the constant A is 3.71 X for a triangular flux lattice6 and the temperature dependence of X is assumed to be described by the two-fluid model for type I1 superconductors.7 Then eq 1 becomes eq 2. Figure 2.
ESR spectrum at 43 K of DPPH coating an entire sphere of
Bi$r2CaCu20x. ration there is another cause of line broadening, namely, the partial expulsion of the magnetic flux out of the superconducting sample (Meissner effect).).‘ To explain this, consider two sites for the spin probe, one perpendicular to the external magnetic field and the other parallel to the field, as shown in Figure 1. The probe site perpendicular to the field experiences, due to the Meissner effect, a magnetic field lower than the applied value which implies that the resonance shifts toward higher magnetic field. For the parallel site of the probe the Meissner effect field expulsion increases the magnetic flux line density so that the probe experiences a magnetic field higher than the applied field. Thus, the resonance for the parallel position shifts toward lower magnetic field. The size of the positive and negative shifts of the resonant fields depends on the effective susceptibility of the material as well as on the demagnetization factor of a particular sample shape.’.4 When the spin probe coats the entire sample surface, there is a continuous distribution of resonant fields, from parallel to perpendicular sites, with more of the probe sites occupying the parallel locations. This results in a broad asymmetric ESR line (Figure 2). This effect can introduce a large error if the assumption is made that the broadening is only due to the inhomogeneity of the vortex lattice. In the most recent paper by Rakvin et al., they recognize this site broadening effect on a disk sample of the YBa2Cu30xsuperconductor.s In this work we demonstrate the site broadening effect discussed above for spherical samples and show how the broadening due to the magnetic flux lattice may be isolated. From this the magnetic field penetration depth in ceramic superconductors may (3) Frait, 2.;Fraitova, D.; h s t , L. J . fhys.. Co11oq. 1988, 8. 2235. (4).Farach, H.A.; Quagliata, E.;Mzoughi, T.; Mesa, M. A.; Poole, C. P.; Cmwick. R. Phys. Reo. B 1990.41, 2046. (5) Rakvin, B.; Pozek, M.; Dulcic, A. Physica C 1990, 170, 166.
(A@) = 3.71
X lO-’@O2[l
- (T/Tc)4]1/2&,4
(2)
Two methods of paramagnetic probe introduction were used. In one method a drop of 10 mM acetone solution of DPPH was used to deposit the DPPH paramagnetic probe as a small spot on the surface of the superconducting samples. A microsyringe was used to put a drop of DPPH solution on the surface. The drop then dried in air. In the other method a speck of powdered DPPH was pressed on the surface of the superconductor with a spatula. For some experiments a spherical superconductor sample was chosen. The spheres were prepared by polishing on a rotating base covered with fine sandpaper. The initial cubical sample was placed in a tube perpendicular to the rotating base mounted just above the surface of the sandpaper. The inner diameter of the tube was chosen to allow free, random tumbling of the sample. The sphere radii varied from 1.3 to 1.7 mm, the sphericity was fairly good (is%maximum variation in the radius), and the surface was fairly smooth. The details of sample preparation for the YBa2Cu30,,, Ag,Y Ba2Cu30, and Bigr2CaCu20, superconducting compounds and their characterization are given
Results and Discussion Spherical Samples. In order to distinguish the broadening due to the vortex lattice which affects only the ESR line of the spin probe in the H, orientation (see Figure 1) from the broadening due to differences in flux density on the surface caused by flux (6) Brandt, E. H. Phys. Reo. B 1990,41, 2046. (7) Tinkham, M. Introduction to Suprconductioity; McGraw-Hill: New York. 1915. (8) Cuvier, S.; Puri, M.; Bear, J.; Kevan, L. J . Phys. Chem. 1990,943864. (9) Puri, M.; Masiakowski, J. T.; Marrelli, S.;Bear, J.; Kevan, L.J . Phys. Chem. 1990, 94,6094. (10) Puri, M.; Masiakowski, J. T.; MaTlli, S.;Bear, J.; Kevan, L. J. Phys. Chem. 1991, 95, 1152.
8970
The Journal of Physical Chemistry, Vol. 95, No. 22, 1991
Masiakowski et al. g = 2.0036
71K
87 K
,--F T>Tc
t
g = 2.0036
Figure 4. ESR spectra of a solid DPPH probe in a parallel orientation
(the most intense low-field line) and of finc polycrystallineDPPH spread at other orientations on the surface of a YBaCuO sphere as a function of temperature. The dashed line shows the maximum line broadening due to magnetic flux explusion for a paramagnetic probe dispersed over the perpendicular orientation region on the surface of the sample (see text). expulsion by the superconductor, experiments with spheres of superconducting materials were performed. A small spot of DPPH was located on the surface of the sphere, and the line shift and the line broadening of the ESR resonance were measured for perpendicular and parallel orientations with respect to the applied magnetic field as shown in Figure 1. A spherical shape was chosen to avoid any orientational differences in flux line distribution on the surface after rotation by 90°, since only for a sphere is the demagnetization factor isotropic and equal to one-third. Figure 3 shows the resonant field shifts for both the perpendicular and parallel positions of the paramagnetic probe on the surface of a YBa2Cu30, sphere. The sample was field cooled in 3350 G. The sample must be cooled below T, in a magnetic field. Zero-field cooling and then applying a magnetic field causes nonreproducibility of the results. The reason is that for zerofield-cooled samples there are vortices pinned in grain boundary regions which never pass through grains. The vortex lattice is then irregular. In contrast, field cooling results in flux line pinning by pinning centers within the grains to provide a regular vortex lattice. I I In Figure 3 the data plotted for the parallel site are taken from an experiment with solid DPPH pressed on the surface of the sample. When DPPH is added via a drop of solution, it penetrates into the surface pores and is actually located within a certain depth of the sphere's surface. In this case for a parallel orientation with respect to the external field the probe does not sample the expelled flux, and experimentally no shift of the resonance field is observed. However, when the temperature is lowered, the line width slightly broadens. This broadening is probably due to the vortex lattice since the anisotropy of the magnetic field penetration depth X implies that the vortex lines inside the crystallites of a ceramic sample will not necessarily be parallel to the applied field.'2-15 ( I I ) Ji, L.; Rzchowski, M. S.; Tinkham, M.Phys. Rev B 1990,42,4838. ( 1 2 ) Celio, M.;Riscman, T. M.;K i d . R. F.;Brewer, J. H.; Kossler, W. J. Physica C 19811, 153-155, 753. ( I 3) Barford, W.; Gunn, J. M. F. Physica C 1988, 153-155, 691. ( I 4) Plonus, M. A. Applied Electromagnetics; McGraw-Hill: New York, 1978; p 491. ( 1 5 ) Gammel, P. L.; Scheenmeyer, L. F.; Waszczak, J. V.;Bishop, D. J. Phys. Rev. Lett. 1988, 61, 1666.
g = 2.0036 Figure 5. Temperature dependence of the ESR spectrum of DPPH in
the H, orientation deposited on the surface of a YBa,Cu30, sphere.
For a perpendicular orientation of the probe the solution preparation seems preferable because then the probe is more homogeneous and localized and does not extend away from the surface. Figure 4 shows ESR spectra of a solid DPPH probe in a parallel orientation (the most intense low-field line) and of fine polycrystalline DPPH spread at other orientations on the surface of the YBaCuO sphere. The spectra are shown as a function of temperature. The outer low-field line corresponds to the parallel site while the outer high-field line is assumed to correspond to the perpendicular site. Around 80 K the latter line disappears due to broadening by the vortex lattice. The line width of this can be used to determine X,from eq 2. However, for a continuous distribution of DPPH particles over the spherical surface, even below 80 K one expects a broad line due to a superposition of resonances at nonperpendicular sites. The separation between the outer low-field and the high-field lines increases as the temperature is further lowered. This effect we call broadening due to flux expulsion or "site-splitting" broadening. We believe this is what Rakvin and co-workers observed in their earlier papers.l Figure 5 shows the broadening of the ESR spectrum of DPPH in a perpendicular orientation on the surface of a YBaCuO sphere as a function of temperature. The broadening of the ESR line seems to be uniform, which indicates that the inhomogeneous magnetic field causing the broadening penetrates throughout the DPPH layer. A well-defined surface requires that the thickness of the DPPH layer is much smaller than the vortex lattice constant d (Figure 1). The magnetic field inhomogeneity extending out from the surface caused by the vortex lattice decreases as e-2rkr, where z is a distance and k = 2 r / d is the wavevector of the vortex lattice. For X-band ESR with the magnetic field around 3300 G the vortex lattice constant is of the order of 0.1 pm.I6 Thus, the DPPH apparently does not extend out from the surface beyond this distance. This is supported by scanning electron microscopy. The ceramic samples are porous, and the DPPH solution penetrates somewhat into the surface. This is seen by scanning electron microscopy which indicates that the DPPH is deposited within a depth of 4-10 pm. The thickness of the DPPH layer on the surface is difficult to determine but certainly is less than 1 pm. Hence, the fraction on or outside the surface is negligible compared with the amount of DPPH filling the pores within 4-10 Mm of the surface. This is why all the DPPH probe senses approximately a uniform distribution of magnetic field inhomogeneity due to the vortex lattice (Figure 5 ) . By modifying the equation for effective magnetic susceptibility4 for a spherical YBa,Cu,O, sample, we determine x = -3/2[(HR(I, ~~
~
(16) Malozemoff, A. P. In Physical Properties oJ High Temperature Superconductors I; Ginsberg, D. M., Ed.; World Scientific: Singapore, 1989; p 120.
Y Ba2Cu30, and Bi2Sr2CaCu20, Superconductors
I
The Journal of Physical Chemistry, Vol. 95, No. 22, 1991 8971
\
\
20
0
40 Ag, %
Figure 7. Magnetic field penetration depth, &, as a function of silver weight percent in Ag-doped YBaCuO. The dashed line at 6400 A is the microwave skin depth for pure silver at 9 GHz. T,K
Figure 6. Temperature dependence of the line broadening, AH,:, of a DPPH probe in a H , orientation on the surface of a YBaCuO sample. The best fit of the theoretical curve (eq 2) was obtained for X, = 1400 & 100 A. The line width extrapolated to 35 K is about 62 G. The inset curve shows the superconducting critical transition temperature T,.
- HR(l[)/HR,)]for a sample a t 35 K to be equal to -0.01 where and HRo,) denote the ESR resonant field of DPPH at perpendicular and parallel sites, respectively, as taken from Figure 2. The 'site splitting" at 35 K for a YBa2Cu30ysample equals 24 G (Figure 2). When this value is substituted into eq 2, it results in Xo 2 2600 A, which is artifactually high. The extrapolated broadening from the vortex lattice at 35 K is actually 62 G, which is given in Figure 6, which shows the best fit of the experimental points to the theoretical curve. This corresponds to a correct value of A,, = 1400 A 100 A. Similar experiments were performed for Bi2Sr2CaCu20, samples. The effective magnetic susceptibility at 35 K was x = -0.002, and Xo was 2850 f 200 A from vortex lattice broadening. The higher value of A,, for BiSrCaCuO compared to that for YBaCuO causes the line due to vortex lattice broadening to not disappear, and it can be followed down to 35 K. For BiSrCaCuO if the "site-splitting" broadening of -5 G is not separated from the vortex lattice broadening of -8 G, one finds an artifactually lower value of X,GZ 2550 A. Silver-DopedSamples. An important feature of the spin-probe technique for measuring X,is that it determines the magnetic field penetration throughout the sample. This is particularly important for porous, doped, or impure materials where the inhomogeneity of the magnetic field is significantly weakened by flux lines penetrating the sample through nonsuperconducting regions. A series of YBaCuO samples doped with different silver contents were investigated to determine the magnetic field penetration depths. A reduction in the normal state resistance is observed with an increase of silver concentration in such silver-doped samples. This is considered to be due to silver forming improved metallic conductance by the silver filling the intergrain boundaries.Io The samples had slab shapes with dimensions approximately 2 X 2 X 1 mm. The DPPH probe was adsorbed from an acetone solution in the center of the 2 X 2 mm face and placed perpendicular to the external magnetic field. For the samples with lower silver content a significant shift of the ESR line toward higher field was observed, which supports a correct perpendicular orientation of the probe with respect to the external magnetic field. All samples were field cooled in 3350 G. Figure 7 shows the penetration depths as a function of silver content. The penetration depth increases smoothly from 1400 A for pure YBaCuO (0% Ag) to about 6500 A for YBaCuO with 35 wt % silver. The error increases with higher silver content which probably reflects the inhomogeneous silver distribution within the pellet from which the slabs were cut. For a sample containing 40 wt% silver no low-field microwave absorption characteristic of superconducting material was detected below the superconducting critical transition temperature. It is interesting that the penetration depth of the Ag-doped samples at high silver concentrations approximates the
I 0 0
I
\\I
50
100
TK Figure 8. Temperature dependence of the FSR line broadening of DPPH in a H , orientation on the surface of BiSrCaCuO. The inset curve shows the vortex lattice melting temperature TM and the superconducting critical transition temperature T,. The best fit to the curve was obtained for X, = 2850 f 200 A. (TMwas used instead of Tc in eq 2).
skin depth of the microwave magnetic field at 9 GHz in pure silver
of about 6400 A.14 Flux-Lattice Melting Temperature. Melting of the flux lattice in single crystals of YBaCuO and BiSrCaCuO was studied by Gammel and co-workers." They found experimentally that in a finite magnetic field the flux lattice in YBaCuO melts a t T, for H perpendicular to the c axis and about 3 K below T, for H parallel to the c axis. The situation is different in BiSrCaCuO in which the flux lattice melts much below T, near 30 K for both parallel and perpendicular orientations of the c axis to H. Figure 8 shows the DPPH line broadening due to the vortex lattice in BiSrCaCuO. The curve shows two characteristic temperatures. The ESR line starts to broaden below the critical transition temperature T, = 81 K, but the broadening is small because the inhomogeneity of the magnetic field is small due to vortex mobility so that no discrete vortex lattice exists in this temperature range. Below -40-35 K the line broadens greatly. It is noted that no such broadening in the 40-35 K range is seen for DPPH adsorbed on a nonsuperconductor. This broadening is interpreted as the formation of a stable flux lattice. Since the experimental frequency of 9.5 GHz is comparable to the characteristic attempt frequency of 10 GHz for high-T, superconductors,16 the vortex dynamics is best considered in terms of flux melting rather than as temperature-activated flux flow. Thus, the flux lattice can be said to melt above -40 K. In eq 2 the flux lattice melting temperature TMshould actually be used instead of T, if TMis significantly less than T,. The best fit of the experimental points to eq 2 on this basis gives a flux lattice melting temperature of T M = 44 K. This is in moderate agreement with the 30 K value of Gammel et aLi5 This also means that when T M is significantly lower than T,, T M can be determined by ESR line broadening of paramag-
J . Phys. Chem. 1991, 95, 8972-8975
8972
netic probes on high-T, superconductors.
Conclusions The results show that the ESR line broadening of an adsorbed spin probe on the surface of an oxide superconductor due to vortex lattice emergence below T,can be separated from Meissner effect flux exclusion and hence can lead to a valid determination of the magnetic field penetration depth. The spin probe on the surface of a superconductor, however, has to be deposited in a proper way and oriented perpendicularly to the external field. The detection of vortex lattice melting in BiSrCaCuO is strong evidence that the ESR line broadening is due to the flux lattice that causes inhomogeneity of the magnetic field. Hence, the spin probe method can be used to determine the magnetic field pen-
etration depth and the flux lattice melting temperature if it is significantly lower than T,. The results obtained for Ag-doped YBaCuO confirm the assumption that the magnetic field inhomogeneity caused by the vortex lattice is significantly weakened by the magnetic field penetrating through nonsuperconducting regions, in this case Ag. The values of ho consequently increase with Ag content.
Acknowledgment. This work was supported by the Texas Center for Superconductivity at the University of Houston under Grant MDA972-88-6-0002 from the Defense Advanced Research Projects Agency and by the State of Texas. Registry No. YBa2Cu30,, 107539-20-8; B,,Sr2CaCu20,, 11490161 -0;Ag, 7440-22-4.
Mass Spectrometrlc Searches for Gaseous Sodium Carbonates D. L. Hildenbrand* and K. H. Lau SRI International, Menlo Park, California 94025 (Received: January 7, 1991)
In connection with modeling of the chemistry of meteor-deposited sodium in the upper atmosphere, we have examined the mass spectra of several pertinent sodium-containing effusion sources for evidence of gaseous NaHC03, Na2C03,or any other Na-C-O-H species. Although clear and unambiguous evidence for the presence of Na2C03(g) was obtained with various sources involving solid and liquid Na2C03and NaOH in the presence of gaseous C02, H20, and H2,all searches for NaHC03(g) were negative. An upper limit of about 29 f 5 kcal mol-' was estimated for the bond strength D(NaOH-C02), based on the detection limit for NaHC03 and measured abundances of NaOH and C02. Preliminary thermodynamic quantities were derived for Na2C03(g). The results are discussed in terms of atmospheric chemistry models in the literature.
Introduction Gaseous NaHC03 has been proposed' as a possible terminal reaction product for meteor-deposited sodium in the upper atmosphere and also as a potential source of 'sudden sodium layers" in the E region of the ionosphere by means of the dissociative attachment process NaHC03
+e
-
Na
+ HC03-
(1)
which is expected to be efficient.z Beams of HC03- have been generated by the interaction of 0-and OH- with CO2-H20 mixture^,^.^ but the electron affinity and related thermochemical properties have not been established experimentally. To our knowledge nothing is known about the gaseous species NaHCO,. The rate constants for the gaseous reaction NaOH C 0 2 M, in which NaHCO, is the expected gaseous product, were reported,s but no evidence regarding product identity was presented. And for HC03-, the geometrical structure parameters and binding energy of the gaseous ion were calculated from molecular theory,6 but there has been no experimental confirmation. We are not aware of any other observations on gaseous bicarbonates. To gain more information about these species, we have examined the mass spectra of vapors emitted from an effusion cell source containing Na2C03or NaOH in the presence of COz, H20, and Hz.Pressures of the added gases were increased to maximum permissible levels to optimize conditions for NaHC03 formation.
+
+
( I ) Murad, E.; Swider, W. Geophys. Res. Le??.1979, 6, 929. (2) yon Zahn, U.; Murad, E. Geophys. Res. Leu. 1990. 17, 147. (3) Keesee. R. G.;Lee. N.; Castleman, A. W.. Jr. J. Am. Chem. soc.1979, 101, 2599. (4) Hierl, P. M.; Paulson, J. F. J . Chem. Phys. 1984, 80, 4890. ( 5 ) Ager, J. W., 111; Howard, C. J. J . Geophys. Res. 1987, 92, 6675. (6) Jonsson, E. G.:Karlstrom, G.; Wennerstrom, H. J . Am. Chem. SOC. 1978, 100, 1658.
0022-3654/91/2095-8972%02.50/0
The results of these experiments are described below.
Experimental Section All parts of the tubular effusion cell source were fabricated from platinum. The main part of the cell was a 0.99-cm4.d. thin-walled tube with an internal 0.32-cm diameter tube for gas addition; the tube cap contained a 0.1 I-cm diameter effusion orifice. A schematic of the cell and heater arrangement is shown in Figure 1. The Na2C03samples were contained in a Pt cup in the hot zone of the cell, as in a normal effusion experiment with gas addition. With NaOH, which has a higher vapor pressure, the sample was supported on a Pt screen below the heater zone. A ball of Pt wire placed in the sample region was effective at the higher temperatures in confining the molten samples by surface tension. Another ball of Pt wire placed near the exit orifice served as a baffle to increase gas-surface collisions and to promote equilibration. The reactive gases could be added singly or in premixed combinations. Mass spectra were obtained with the magnetic-sector instrument and experimental technique described in previous publications.'~* As always, observed ion signals were checked for response to displacement of the neutral beam-defining slit to ascertain their effusion cell origin. The cell configuration is expected to yield an essentially equilibrium distribution of species. A pressure calibration based on the established vapor pressure of gold was combined with estimated ionization cross-section ratios to evaluate sDecies Dartial Dressures from measured ion intensities. The Na2C03*powder and NaOH pellet samples were reagent grade materials used without further purification. C 0 2 and H2 reagent gases were obtained from the Matheson Co., while lab(7) Hildenbrand, D. L. J . Chem. Phys. 1968,48, 3657; 1970,52, 5751. (8) Kleinschmidt, P. D.; Lau, K. H.; Hildenbrand, D. L. J. Chem. Phys. 1981, 74, 653.
0 1991 American Chemical Society
