3660
J . Phys. Chem. 1989, 93, 3660-3663
Exposure of the Crystal Faces of V,05 on Supported Catalysts Miki Niwa,* Yoshihito Matsuoka, and Yuichi Murakarni Department of Synthetic Chemistry, Faculty of Engineering, Nagoya University, Furo-cho, Chikusa- ku, Nagoya 464, Japan (Received: June 8, 1987; In Final Form: November 30, 1987)
The exposure of the (010) face of vanadium oxide, when supported on a variety of oxides, was measured by the combined use of benzaldehyde-ammonia titration and NO-NH3 rectangular pulse methods. For this purpose, the surface areas of the exposed support, (010) face, and other faces were individually determined. Exposure of the (010) face depended upon the kind of support, but not upon the crystal phase. Our present observations of the exposure of the (010) face on various types of support do not agree with those previously noted for the spreading or dispersion of vanadium oxide. Exposure of the (010) face was the highest on Ti02 and Zr02: it reached nearly 100% at small loadings and decreased as the loading increased. In contrast, the exposure on AI203 was negligibly small at small loadings and increased with the loading. In the case of S O 2 , the exposure was almost independent of the loading amount except for extremely small loadings where exposure was very low. On the basis of these studies, models representing the structures of these supported catalysts are proposed.
Introduction Metal oxides are often supported on suitable solids as industrial catalysts. It is therefore extremely important to determine the structure of such supported metal oxides. As for the supported vanadium oxide, a number of studies have recently been undertaken to study the structure of loaded vanadium species,Iz2 the morphology of catalyst^,^-^ and interactions between vanadium oxide and its support^.^^' Of the various types of support, titania (anatase) has been most frequently investigated because of its industrial applications. Although identification may be made by a variety of spectroscopic methods, chemical methodss are, inevitably, also used, since the catalytic functional sites are studied by using this method. Both methods should be used complementarily. Our previous papersgJOindicated that the surfaces of exposed supports and of support vanadium oxide can be distinguished by the benzaldehyde-ammonia titration (BAT) method. The support surface coverage by vanadium oxide and thickness of the metal oxide on different types of support could be measured by discrimination of the surfaces on supported catalysts. Except for small concentrations of V z 0 5 on A1203and Si02, multilayered vanadium oxide is formed, and the coverage efficiency1' and the thickness of oxide depend on the kind of support. However, this conclusion indicates only the extent of physical covering by vanadium oxide on supports and does not indicate whether or not the vanadium oxide exposed on the surface of the support is active. In this paper, the exposure of active V=O sites on supported vanadium oxides will be discussed in some detail. It is well-known that vanadium (V) oxide has a layered structure consisting of vanadium and oxygen octahedra. The vanadium (1) Kozlowski, R.; Pettifer, R. F.;Thomas, J. M. J . Phys. Chem. 1983, 83, 5172, 5176. (2) Haber, J.; Kozlowska, A,; Kozlowski, R. J . Catal. 1986, 102, 52. (3) (a) Inomata, M.; Mori, K.; Miyamoto, A.; Ui, T.; Murakami, Y . J . Phys. Chem. 1983, 87, 754. (b) Inomata, M.; Mori, K.; Miyamoto, A,; Murakami, Y. J . Phys. Chem. 1983, 87, 761. (4) (a) Bond, G. C.; Bruckman, K. Faraday Discuss. Chem. SOC.1981, 72, 235. (b) Bond, G. C.; Konig, P. J . Catal. 1982, 77, 309. (5) Bond, G. C.; Zurita, J. P.; Flamerz, S.; Gelling, P. J.; Bosch, H.; Van Ommen, J. G.; Kip, B. J. Appl. Catal. 1986, 22, 361. ( 6 ) (a) Wachs, I. E.; Saleh, R. Y.; Chan, S. S . ; Chersich, C. Appl. Catal. 1985, 15, 339. (b) Wachs, I. E.; Chan, S. S.; Saleh, Y . J . Caral. 1985, 91, 366. ( 7 ) Saleh, K.Y.; Wachs, I. E.; Chan, S. S.;Chersich, C. C. J . Cafal. 1986, 98, 102. (8) For example: Nag, K. K.; Chary, K. V. R.; Reddy, B. M.; Rao,.B. R.; Subrahmanyam, V . S . Appl. Caral. 1984, 9, 225. (9) Niwa, M.; Matsuoka, Y.; Murakami, Y . J . Phys. Chem. 1987, 91, 4519. (10) Niwa, M.; Inagaki, S.; Murakami, Y . J . Phys. Chem. 1985,89, 2550, 3869. (1 1) We used couerage efficiency to show the extent of covering of support surface by vanadium oxide. On the other hand, the thickness of oxide shows the number of layers of loaded vanadium oxide.
0022-3654/89/2093-3660$01.50/0
TABLE I: List of Supports Used in This Study support BET surface area, m2/g bulk crystal phase Ti02(a-r) 54 anatase with ca. 10% rutile
Ti02(r) Zr02 A1203(Y)
A1203(a+0)
Si02(HS) SO2(LS)
6.2 78 233 45 29 1 81
rutile baddeleyite Y
mixture of a and 0 no no
cation is surrounded by five oxygen atoms, and there is one V=O bond with a smaller bond length than the other four V-0 bonds. The V=O bond is normal to the (010) face and points either upward or downward. Faces other than (OlO), for example, (101) and (1 1 l), are also included in the s t r u c t u r e a s revealed by X-ray diffraction. Our previous study of the activity of vanadium oxide has indicated that only the V = O bonds on the (010) face become active for catalytic reactions such as NO-NH3 and hydrocarbon oxidations. Therefore, it seems that the activity of the supported vanadium oxide depends on the exposure of the (010) face. In addition to the BAT method, the authors have already proposed the NO-NH3 rectangular pulse (NARP) method as a way of measuring the number of active V=O sites on the (010) face of v205.3x'2By assuming the known surface area of one species of the V=O, the surface area of the (010) crystal face appearing on the V2Os surface can be determined. Therefore, by a combination of the BAT and NARP methods, the surface areas of the exposed support, active (010) face, and other inactive vanadium oxide surfaces can be individually determined; the exposure of the (010) face will thus be obtained. In our previous studies of supported VzOscatalysts, however, the surface area of the (010) face was measured, and by comparing it with the BET surface area, the structures of V205supported on A1203and Ti02 were ~uggested.~ It is now realized that these findings are partially invalid, because of failure to take into consideration crystal faces other than the (010) face. For example, when the ratio SOIO/SBET ( S : surface area) was small, either low spreading of V z 0 5 on the support or low exposure of the (010) face could not be concluded unambiguously. Individual determination of each of the surface areas of the exposed support, (010) face, and other faces of the VzOs is therefore necessary in order to identify the structure of the supported V2Os catalysts.
Experimental Method Four kinds of oxides were used as supports for vanadium oxide, as shown in Table I. These included different bulk crystal phases of A1203 and Ti02 and different surface areas of S O z . The (12) (a) Miyamoto, A.; Yamazaki, Y.; Inomata, M.; Murakami, Y. J . Phys. Chem. 1981.85, 2366. (b) Inomata, M.; Miyamoto, A,; Murakami, Y. J . Phys. Chem. 1981, 85, 2372.
0 1989 American Chemical Society
V205on Supported Catalysts
The Journal of Physical Chemistry, Vol. 93, No. 9, 1989 3661
--
TABLE 11: Surface Areas of Exposed Support, Vanadium Oxide, and Active (010) Face on Supported Vanadium Oxide Catalysts
s
W K
3
surface area,@m2/g exposed vanadium support oxideb (010) TiO,(a-r)
0
mol
surf. concn,
x
%
nm-2
1.o
2.9 4.7 9.1 13.1
1.6 5.4 9.8 23.0 27.9
0.1 0.5 1 .o 2.0 5.0
1.1 6.6 12.0 23.0 52.5
1 .o 2.0 5.0 10.0 25.0
0.6 1.3 3.6 8.1 26.7
75.1 64.8 43.4 20.4 11.2
1.o 5.0 10.0 25.0 50.0
0.3 1.5 3.6 12.9 49.2
AhOg(Y) 21 1 114 79.6 1.5 1.o
1 .o 5.0 10.0 25.0
0.9 8.2 16.0 42.3
AlzOp(a+B) 50.1 14.7 6.6 2.4
1 .o 5.0 10.0 25.0 50.0
0.5 2.7 5.7 14.3 38.9
10.0 25.0 50.0 75.0
12.4 31.5 77 I66
VI
e W
W
u
a LL
50
m m
I
50
V R N R D I U M SURF. CONC. L n n 1 - ~ 1 Figure 1. Relationship of the exposure of the (010) face with the surface
concentration of vanadium oxide on TiO,(a-r)(O), Ti02(r) (e),Zr02 (v),A1203(y) (A),A1203(a+0)(A),Si02(HS) ( O ) , and Si02(LS)(m). Contents of vanadium oxide in mol % are referred to in Table 11. method of preparation and the suppliers of these supports have been indicated in previous paper^.^,'^ Supported catalysts were prepared by an impregnation method. Supports were added to an oxalic acid solution, and the excess water was evaporated on a hot plate at ca. 383 K. After evaporation, the residue was dried at 393 K overnight, followed by calcination at 773 K in a stream of oxygen for 3 h. The calcination temperature was kept constant, since changes in catalyst morphology due to calcination temperature have been r e p ~ r t e d . ~ J ~ The BAT9 and NARP12 experimental methods have been described in detail previously. Throughout these measurements, vanadium oxide was fully oxidized by in situ calcination in an oxygen flow at 673 or 773 K, except in the case of the SiOzsupported catalyst. In this case, vanadium oxide was reduced at 773 K for 2 h with H2 in the BAT method, while being oxidized in the NARP method; the exposed surface area of Si02was not measured by the BAT method, but instead, that of the reduced vanadium oxide can be measured.
Results The surface areas of the exposed support and of the (010) face were measured by the BAT and NARP methods, respectively. The surface area of vanadium oxide, including not only the (010) face but also other faces, was obtained by calculating the difference between the exposed support and BET surface areas. These are shown in Table 11, and the exposure of the (010) face (in percent) is plotted as a function of the surface concentration in Figure 1. Parameters used are defined: exposure of (010) face = 100 x [(surface area of (010) face)/(surface area of V20,)] surface concentration of vanadium = (number of loaded vanadium atoms) /( BET surface area of supported catalysts) The latter parameter was used to compensate for the differences between surface areas of supports. In the case of Si02only, the surface area of vanadium oxide was measured by the BAT method, and that of the exposed support was calculated from the differences between the surface areas as obtained by BAT and BET methods. In this method, the BAT surface area of vanadium oxide, when supported on S O 2 , was obtained in the reduced condition: therefore, two kinds of correction were required, as described below. Because the surface area was usually expressed in terms of per gram of catalyst, the decrease in catalyst weight upon reduction was corrected. In this study, the averaged oxidation states of vanadium oxide were determined by measuring the amount of oxygen consumed during the temperature-programmed reoxidation experiments. Furthermore, the change in surface areas of the (13) Gasior, I.; Gasior, M.; Brzybowska, B.; Kozlowski, R.; Sloczynski, J. Bull. Acad. Pol. Sci., Ser. Sci. Chim. 1979, 27, 829.
43.7 24.7 18.6 4.9 1.3
3.4 14.4 15.4 21.8 28.9
3.8 15.1 16.1 14.5 14.3
Ti02(r) 5.89 2.70 1.45 0.77 0.37
0.57* 2.97 4.74 5.63 6.37
1.oo 3.08 4.06 4.34 3.78
Zr02 1.2* 9.1* 22.7 37.0 29.7
2.7 11.9 26.7 31.9 19.0
0 71 71 94 42
0 16.3 49.7 51.5 23.4
18.8 19.8 27.6 26.8
1.04 15.4 12.8 11.7
Si02(HS) 177 134 113 90.4 36.0
28.0 35.3 34.0 25.6 26.0
(28.0) (38.2) (38.4) (13.6) (14.3)
0.7 3.3 6.5 5.4 5.5
18.1 17.0 16.8 7.8
(15.8) (15.9) (7.1) (4.8)
4.9 5.6 2.9 3.2
Si02(LS) 47.6 34.7 14.8 9.8
DExperimentalerrors within 5% (BET) or 10% (BAT, NARP) are estimated. Because the surface area of vanadium oxide except on Si02 was calculated from the difference between BET and BAT areas, large errors are estimated as the difference was small. In cases shown by asterisks, irrational values are obtained, but, otherwise, maximum errors included in the exposure of the (010) face are ca. 20-25%. bSurfacearea of vanadium oxide on Si02was measured in the reduced condition, and the estimated value of the oxidized V 2 0 5 is described in parentheses. silica-supportedvanadium oxide after reduction had to be corrected in order to calculate the percentage exposure of the (010) face, because the surface area of the (010) face was measured in the oxidized condition, while the total area of vanadium oxide was measured in the reduced state. The surface area of vanadium oxide is usually increased by the reduction process; however, the support area is regarded as being constant, the degree of reduction being negligible. It can therefore be assumed that the difference between BET surface areas, measured in both their oxidized and reduced states, is ascribable to the increased surface area of the vanadium oxide supported. In fact, the BET surface area of the supported catalyst was increased during the reduction process: it was especially noticeable for the V205/Si02at high loading, for example, 14.3-26.0 mz/g in the VzOS/SiO2(HS) (50 mol %) and 7.1-16.8 m2/g in the V20s/Si02(LS) (50 mol %). Therefore, BET surface areas were measured in both the oxidized and reduced states, and the differences between them were used to estimate
3662 The Journal of Physical Chemistry, Vol. 93, No. 9, 1989 the surface area of the vanadium oxide in its oxidized condition. The percentage exposure of the silica-supported catalysts thus corrected is shown in Figure 1. The exposure of the (010) face on TiOz(a-r and r) and Zr02 was extremely high at the low loading, and this decreased as the loading amount increased. The exposure of more than 100% at small loadings may be caused by experimental error. In contrast, the (010) face of Al,O3(7 and a + 0) was negligibly small at the small loading; however, it increased with the increase in vanadium oxide loading and seemed to attain the maximum. In addition, it was found that the exposure of the active (010) face depended on the type of support used but remained independent of the bulk crystal phases of A1203 and TiO,. The corrected percentage exposure of active sites on the SO2-supported catalyst was almost independent of the loading except at the extremely small loading. Upon excess loading, the surface condition should become similar to that of unsupported V2O5. In fact, irrespective of the kind of support used, the (010) face exposure attained ca. 50% at high loading; this was almost in agreement with that for unsupported V20s. However, the surface concentration required for obtaining this condition was dependent on the support, and this was observed on SiO, at the smallest concentration of vanadium oxide.
Niwa et al.
Figure 2. Structure model of vanadium oxides supported on Alz03,Ti02, ZrOz, and Si02at 6 nm-2 of the V205surface concentration. Broad lines represent exposed (010) faces, and filled circles represent inactive vanadium species. Thicknesses of vanadium oxide were 2, 3, 4, and 25 on A1203,TiO,, and ZrO,, and Si02, respectively.
In this condition, the exposure decreased along with the decrease in surface concentration. Therefore, vanadium oxide on A1203 lost its active V=O sites until the multilayer was formed. Kozlowski et a1.l and Haber et al.2 have recently proposed a less active dimerized vanadium species on A1203. Furthermore, Bond et al. have also reported on the structure of an oxohydroxy vanadium species upon formation of a thin layer on Ti02.5 These studies have consistently indicated that another inactive nonlayered form of vanadium oxide is stabilized at the small loading. On the other hand, vanadium oxide on silica was regarded as having the same structure as the unsupported vanadium oxide except at extremely small loadings, since the exposure of the active (010) face was almost in agreement with that for unsupported vanadium oxide at small to high loadings. The formation of large Discussion bulk V,05 is in close agreement with the conclusion reported in a previous study by Haber et al., A tetrahedral vanadium species In our previous paper,9 it was found that the vanadium oxide supported on Si02at low loadings active in photocatalytic reactions coverage of the support surface was dependent on the type of has been reported.18 Existence of the tetrahedral vanadium species support used. The surface of the Alz03(7) was almost effectively was also confirmed by Kevan et al.I9 using electron spin-echo covered by V2O5, and the thickness obtained at 100%of coverage spectroscopy. The lower exposure of the active (010) face at small was about three layers. On the other hand, the silica surface was loadings may be ascribable to the tetrahedral structure of vananot effectively covered, and the large size of V2O5 was supported. dium oxide. Intermediate behavior between them was obtained on ZrO, and Figure 2 shows the structure model of vanadium oxides supTiO,. We will therefore discuss the exposure of the active (010) ported on A1203,ZrO,, Ti02, and SO2. This structure model face obtained in the present study as compared to the spreading accounts for the V2Os coverage of the support as well as the or dispersion obtained in the previous studies. exposure of the (010) face. X-ray diffraction and spectroscopies The highest exposure of active (010) face was observed for Ti02 such as IR, ESR, and UV-vis on these catalyst systems which and Zr02 where the coverage efficiency was smaller than for A1203 have already been published3 by us are taken into consideration. but larger than for S O 2 . Therefore, V,OS/TiO2, which is recAt less than 2 n m 2 of the surface concentration, vanadium oxide ognized as an active catalyst in various catalytic reactions (e.g., supported on A1203is inactive, forming no layer structure, while NO, removal oxidation of o-xylene into phthalic anhydride,5s6,14 on Ti02or ZrO,, the (010) face is selectively exposed. On S O 2 , from the exhaust gas15), is not characterized by the highly dissome of the vanadium oxide exist as tetrahedral species, but most persed state of vanadium oxide but by the high exposure of the parts of the remaining oxide form typical vanadium oxide. At active (010) face. Although the different property of anatase from 6 nm-,, the active (010) face also appears on A1203, where exrutile, when used as a support, was discussed from a crystalloposure is smaller than on TiO, and ZrO,. V205supported on Si02 graphic point of view by Courtine et a1.,16 these types of oxide could be regarded as being similar to unsupported oxide in more provided very similar exposure of the active (010) face, and no than this concentration. At a surface concentration of 15 nm-2, conclusion was drawn about the preferential availability of anathe support surface of alumina is completely covered with vanatase.17 dium oxide, and the surface condition is almost the same as that The interaction of vanadium oxide with titania (anatase) was for unsupported oxide. Because of the cracks and kinks on the reported by Wachs et who claimed the formation of a spisurface of V2O5/Al2O3,exposure of the active (010) face is smaller nel-like compound consisting of reduced vanadia and titania than for Ti02 and Zr02 which have relatively smooth, flat surfaces. (rutile) after calcination at temperatures above 923 K. Similar These structures could be altered to some degree, when prepared conclusions were also obtained by Gasior et al.13 However, the under different conditions, because well-dispersed monolayer formation of such a spinel compound may be disregarded in this catalysts were prepared by the i~n-exchange'~ or by grafting4*5 study, since the catalysts used were calcined at 773 K. methods, and interactions with supports appeared during calciThe exposure of active sites on V205/A1203,where vanadium nation at higher ternperat~res."',~~Therefore, the structural model oxide covered the support surface most effectively, depended upon shown in Figure 2 is proposed for the catalysts that are prepared the surface concentration of V205. As reported previously? less by the impregnation method and calcined at 773 K. In this than 2 nm-2 of the surface concentration of vanadium oxide on investigation, therefore, the differences in the structure of supalumina did not form a multilayer, but the formation of a monolayer and/or the dissolved vanadium species was e ~ t i n i a t e d . ~ ~ ported vanadium oxide can be emphasized. As discussed above, exposure of the active (010) face is not simply related to the covering of the support surface by vanadium (14) (a) Roozeboom, F.; Mittelmeijer-Hazelger, M. C.; Moulljn, J. A.; de oxide. Therefore, the surface condition of vanadium oxide should Beer, V. H. J.; Gellings, P. J. J . Phys. Chem. 1980, 84, 2783. (b) van be understood from the ability of not only spreading on supports Hengstum, A. J.; Van Ommen, J . G.; Bosch, H.; Gellings, P. J. Appl. Catal. 1983, 5 , 207 but also of exposing the active (010) face. Because of insufficient (15) Inomata, M.; Miyamoto, A,; Ui, T.; Kobayashi, K.; Murakami, Y. Ind. Eng. Chem. Prod. Res. Dev. 1982, 21, 424. (16) Vejux, A.; Courtine, P. J . Solid Srate Chem. 1978, 23, 93. ( I 7) After sending this paper, a study on ammoxidation of aromatic hydrocarbons over titania-supported vanadium oxide by Trifiro et al. was published; no substantial difference between activities of vanadium oxide supported on anatase and rutile was reported. See: Cavali, P.; Cavani, F.; Manenti, 1.; Trifiro, F. Ind. Eng. Chem. Res. 1987, 26, 639.
(18) (a) Anpo, M.; Takahashi, I.; Kubokawa, Y . J . Phys. Chem. 1980.84, 3440. (b) Yoshida, S.; Tanaka, T.; Okada, M.; Funabiki, T. J . Chem. Soc., Faraday Trans. 1 1986.81, 1513. (c) Tanaka, T.; Tsuchitani, R.; Ooe, M.; Funabiki, T.; Yoshida, S . J . Phys. Chem. 1986, 90, 4905. ( 1 9) Narayana, M.; Narasimhan, C. S.; Kevan, L. J . Chem. Soc., Faraday Trans. I 1985, 81, 137.
J. Phys. Chem. 1989, 93, 3663-3666 knowledge of the nature of supported metal oxide catalysts, it cannot yet be concluded whether or not this complex condition of the vanadium catalyst is related to the layered structure. However, we can indicate that the chemical affinity to crystal faces and/or the crystallographic fitness may influence surface condition, as mentioned below. The coverage behavior of vanadium oxide was related to the electronegativity of cations in its oxides, and some relationship with the surface acid-base properties might also be i n d i ~ a t e d . ~ Because the acidic oxide (Le., V2Os) was effectively supported on the basic support, such differences in the acid-base properties could be the cause of the high coverage efficiency of vanadium oxide. In other words, the vanadium oxide could be supported due to the different chemical property of the support. Therefore, vanadium oxide may be supported on alumina with the strongest interaction with the support. In this condition, it may not be easy to form the layer structure of vanadium oxide. On the other hand, when there is little interaction with the support as on S O 2 , va-
3663
nadium oxide behaves as if it were unsupported. After all, high exposure of the active (010) face on ZrOz and Ti02 could be understood as resulting from the medium degree of interaction with the support. The high exposure of active V=O sites on T i 0 2 (anatase) was explained on the basis of the crystallographic fitness between anatase and VlOS, as was indicated by Courtine et a1.16 This explanation is significant in the case of anatase. Such a crystallographic fitness could be observed on ZrOz and TiOz (rutile), since vanadium oxide on these supports exposed active sites similarly on the anatase. On these oxides, however, some alterations of crystal dimensions in the boundary may stimulate the fitness as was indicated for V2Os on Ti02 (rutile).3a
Acknowledgment. This work was supported by a Grant-in-Aid from the Ministry of Education, Science, and Culture, Japan (No. 59470097). Registry No. V 2 0 5 , 1314-62-1; Ti02, 13463-67-7; Zr02, 1314-23-4.
Multicomponent Cluster Ions. 2. Comparative Stabilities of Cationic and Anionic Hydrogen-Bonded Networks. Mixed Clusters of Water and Hydrogen Cyanide Michael Meot-Ner (Mautner)* and Carlos V. Spellert Chemical Kinetics Division, Center for Chemical Physics, National Bureau of Standards, Gaithersburg, Maryland 20899 (Received: December 1 1 , 1987; In Final Form: June 13, 1988) The thermochemistry of cluster ions containing nH20 and mHCN molecules, for clusters up to rank r = n + m = 5 , was obtained from equilibrium measurements. The clustering of H 2 0 about H 3 0 +and OH- and the clustering of HCN about HCNH+ show distinct shell filling effects, but the clustering of H 2 0 and HCN about CN- does not. Hydration by one water molecule eliminates the difference of 5 kcal/mol between the proton affinities of HCN and H20. Hydration also decreases the difference between the intrinsic deprotonation energies, Le., AHoacid, of H 2 0 and HCN from 37.7 to 19.8 kcal/mol in the 4-fold hydrated species. In protonated clusters of a given rank r, exchange of H 2 0 by HCN does not significantly affect the total stability of the cluster. This is in contrast to Hz0/CH30H and H20/CH3CNprotonated clusters and H20/HCN anionic clusters, where increasing nonaqueous content is stabilizing. In the latter, the most stable clusters of any size are the neat CN-SnHCN clusters.
Introduction The properties of clusters are of interest in relation to the transition from gas-phase to condensed-phase behavior, and also in relation to nucleation leading to condensation and aerosol formation in ionized planetary atmospheres. Atmospheres include a mixture of gases, and therefore the properties of multicomponent clusters are important. Water and hydrogen cyanide are common in planetary and interstellar environments and are likely to be contained in cluster ions in such environments. Therefore, we extended our studies on multicomponent clusters to the H20/HCN system. The preceding study in this series dealt with anionic and cationic clusters containing water and methanol.' In that study we compared for the first time the stabilities in anionic vs cationic clusters with increasing cluster rank r (r = n + m, where the cluster includes n H 2 0 and mHCN molecules). The observation was that the bonding energies with cluster growth showed a similar pattern both in the anionic and cationic systems. In that system, however, an infinite network of strong OH- a 0 hydrogen bonds was possible in both the anionic and cationic clusters. In contrast, in other mixed cluster systems such as Hz0/(CH,),O2 and H 2 0 / C H 3 C N ? blocking by methyl groups prohibits clusters of certain composition. The H C N molecule is intermediate in that it contains only a carbon-bonded hydrogen,
-
'Present address: Departamento de Fisica-ITA, Centro Tecnico Aerospacial 12.225, Sao Jose dos Campos, SP, Brazil.
but it is strongly acidic and therefore relatively strongly hydrogen-bonding. It will be of interest to examine therefore the development of cluster stabilities with cluster composition in this system.
Experimental Section The measurements were done by using the NBS pulsed highpressure mass spectrometer4 and procedures as described in preceding papers.'~~ For cationic clusters, the carrier gas was Ar or CH4 containing 0.1-2.4% H 2 0 and 0 . 2 4 7 % HCN. The formation of (HCN)zH+ was studied in 1.8% H C N in CH4, as well as in an H 2 0 / HCN/CH4 mixture. The formation of (HCN)3H+ was studied in pure HCN. The formation of (H20),H+ was studied in 6.4% H 2 0 in CH,. In order to avoid mass coincidence problems, D 2 0 and DCN had to be used in some experiments. Moreover, clusters involving C2H5+from methane also caused mass coincidence problems, and therefore in some experiments Ar was used as the carrier gas. For anionic clusters, CHI was the carrier gas. Trace amounts of C H 3 0 N 0 were added for electron capture, yielding CH30-, ~~~
~~~
( I ) Meot-Ner (Mautner), M. J . A m . Chem. SOC.1986, 108, 6189. (2) Hiraoka, K.; Grimsrud, E. P.; Kebarle, P. J . Am. Chem. SOC.1974, 96-,3359 - ~ -
-
(3) Deakyne, C. A.; Meot-Ner (Mautner), M.; Campbell, C. L.; Hughes, G.; Murphy, S. P. J . Chem. Phys. 1986, 84, 4958. (4) Meot-Ner (Mautner), M.; Sieck, L. W. J . Am. Chem. SOC.1983,105, 2956.
This article not subject to US. Copyright. Published 1989 by the American Chemical Society
