Chemical reaction of transition metal clusters (Fe, Co, Ni...

27
Chemical reaction of transition metal clusters (Fe, Co, Ni) with ethanol by using FT-ICR mass spectrometer Shuhei INOUE a) and Shigeo MARUYAMA b) a) Department of Mechanical System Engineering, Hiroshima University 1-4-1 Kagamiyama, Higashi Hiroshima-shi, 739-8527, Japan b) Department of Mechanical Engineering, The University of Tokyo 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan Chemical reactions of transition metal cluster ions (Fe, Co, Ni) with ethanol were investigated by using an FT-ICR mass spectrometer. Metal clusters of about 10-20 atoms were generated by a pulsed laser-vaporization and supersonic-expanded cluster beam was directly introduced into FT-ICR mass spectrometer. Observed reactions are simple chemisorptions of ethanol and dehydrogenated chemisorptions strongly depend on metal species and their cluster sizes. In case of Co cluster, detailed dehydrogenation reaction steps were elucidated through several isotope experiments using C2H5OD, CD3CH2OH, C2D5OD in addition to C2H5OH. A qualitative tendency that the reaction pattern changed along the periodic table (Fe → Co → Ni) was observed. The dependency of chemisorption rate on metal cluster size were similar for 3 metals, however, the cluster sizes at the maximum reaction rate are in the order of Fe < Co < Ni, which also obeys the order in the periodic table.

Transcript of Chemical reaction of transition metal clusters (Fe, Co, Ni...

  • Chemical reaction of transition metal clusters (Fe, Co, Ni) with ethanol by

    using FT-ICR mass spectrometer

    Shuhei INOUE a) and Shigeo MARUYAMA b)

    a) Department of Mechanical System Engineering, Hiroshima University

    1-4-1 Kagamiyama, Higashi Hiroshima-shi, 739-8527, Japan

    b) Department of Mechanical Engineering, The University of Tokyo

    7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan

    Chemical reactions of transition metal cluster ions (Fe, Co, Ni) with ethanol

    were investigated by using an FT-ICR mass spectrometer. Metal clusters of about 10-20

    atoms were generated by a pulsed laser-vaporization and supersonic-expanded cluster

    beam was directly introduced into FT-ICR mass spectrometer. Observed reactions are

    simple chemisorptions of ethanol and dehydrogenated chemisorptions strongly depend

    on metal species and their cluster sizes. In case of Co cluster, detailed dehydrogenation

    reaction steps were elucidated through several isotope experiments using C2H5OD,

    CD3CH2OH, C2D5OD in addition to C2H5OH. A qualitative tendency that the reaction

    pattern changed along the periodic table (Fe → Co → Ni) was observed. The

    dependency of chemisorption rate on metal cluster size were similar for 3 metals,

    however, the cluster sizes at the maximum reaction rate are in the order of Fe < Co < Ni,

    which also obeys the order in the periodic table.

  • 1. INTRODUCTION

    The carbon nanotube (CNT) 1 discovered in 1991 has the network structure

    with hexagonal rings and pentagonal rings of carbon atoms. The carbon nanotube is

    classified into two types; one is single-walled carbon nanotube (SWNT) 2 and the other

    multi-walled carbon nanotube (MWNT). 1 CNTs are expected as new materials from

    various physical and chemical characters based on the geometric unique structure. So

    far, a number of researches attempting to apply SWNTs to e.g. electrodes, 3, 4 AFM tips,

    5, 6 nano-wires, 7 and field emitters 8, 9 have been performed actively.

    At present, mainly due to its scalable nature, CVD 10-15 method is regarded as

    the most suitable technique to produce SWNTs industrially. Among them, the HiPco

    (High-Pressure CO) method 16 has been recognized as a most commercially feasible

    approach to produce large amount of relatively high quality SWNTs continuously.

    Recently, Maruyama et al. 17 proposed a new CVD method where alcohols (e.g. ethanol

    and methanol) were first employed as a carbon source for the synthesis of SWNTs. The

    method, named as alcohol CCVD (ACCVD) method, is superior to other CVD methods

    since they can produce high-quality SWNTs at relatively low temperature. However, its

    synthetic mechanism has not made clear yet, so fundamental investigations for growth

    mechanism is essential for the further improvement of ACCVD method.

    In CVD methods, transition metals such as Fe, Ni, Co are used as a catalyst,

    and Co are also employed in other fields, and many researches 18-24 are proceeded on a

    cluster level. However, most of existing researches are restricted to small size clusters

    (less than 10 atoms) mainly because generating large-sized cluster is rather difficult. In

    this report, we have explored a basic reaction mechanism of larger size clusters (Fe, Co

    and Ni) with ethanol, in which we simulated actual catalyst size used in the CCVD

  • processes for SWNT production.

    2. EXMERIMENTAL

    The experimental apparatus are based on the same design concept as the

    apparatus in Smalley's group 19 and the detailed characteristics are described in

    elsewhere. 26-29 Figures. 1 and 2 show the cluster beam source and direct injection

    FT-ICR apparatus, respectively. The metal cluster ion beam was generated outside of

    magnetic field by the laser-vaporization cluster beam source shown in Fig. 1. A pulsed

    gas valve, the sample motion mechanism and a skimmer were installed in a 6-inch

    6-way UHV cross. A solid sample disk was vaporized by the focused beam of Nd: YAG

    laser (2nd harmonics) while timed pulsed gas was injected to the nozzle. In the

    atmosphere of helium gas, vaporized atoms condensed to clusters, and then, were

    carried and cooled by the supersonic expansion of helium gas. The cluster beam was

    directly injected to the magnetic field through a skimmer with the opening diameter of 2

    mm and a deceleration tube.

    The FT-ICR is the unique mass spectroscopy based on the ion-cyclotron motion of

    clusters in a strong magnetic field. The ion cyclotron frequency f is inversely

    proportional to the ion mass M as follows.

    M

    Bqf

    π2=

    Extremely high mass-resolution at high mass-range such as resolution of 1 amu at

    10,000 amu range can be obtained. Furthermore, since the ions can be trapped in the

    vacuum for a few minutes, it is possible to perform the chemical reaction experiments.

    The ICR cell, 42 mm I.D. 150 mm long cylinder was placed in a stainless-steel tube

  • (SUS316) of 84 mm I.D. which penetrated the homogeneous 5.826 Tesla

    super-conducting magnet commercially available for NMR. Two turbo-pumps

    (300 s/l ) fore-pumped by a smaller turbo-pump of 50 s/l were placed at the floor in

    order to avoid the effect of strong magnetic field. The typical background pressure was

    3×10-10 Torr.

    For the chemical reaction experiments, ethanol gas was supplied to the cell by a

    pulsed valve for a fixed period. The pulse value was adjusted so that the pressure at the

    ICR cell chamber became at 1-2×10-8 Torr. After pumping out, cluster ions were excited

    to detect the mass distribution.

    3. RESULTS AND DISCUSSION

    3.1. COBALT CLUSTERS

    Figure 3(a) shows a typical spectrum of cobalt clusters as injected from the cluster

    beam source. In this figure, a horizontal axis denotes cluster size (upper measure is a

    number of cobalt atom and lower measure is atomic mass unit), and a vertical axis

    expresses signal intensity. There are small peaks between pure cobalt clusters, which are

    water molecule (H2O, 18 amu) chemisorbed on the clusters. Fig. 3(b-d) shows results of

    reaction with ethanol (at R.T., 1×10-8 Torr ) for 0.2, 0.5 and 1.0 second each. After 0.5 s

    reaction products begin to be observed, and there are many reaction products seen in Fig.

    3(d). It is very interesting that there are two reaction types. In case of Con (12

  • (C2H5OH) is 46 amu, so this reaction must be simple chemisorptions, however it is not

    clear from this data whether an ethanol molecule is dissociated or not on the cluster

    surface. We guess the ethanol atom is chemisorbed dissociatively on the cluster surface,

    as the reason discussed later. Fig. 4 shows the expansion view of fig. 3(d). Almost the

    same reaction patterns are seen in each graph. In Type-A reaction main products are

    almost four species where each products are shifted from pure cobalt clusters by 9, 11,

    42, 46 amu. All of other signals show intermediate products. Type-A can be divided into

    two kinds, one is one molecule reaction (Type-A1) and the other is two-molecules

    reaction (Type-A2). Signals at 42 and 46 amu are classified into Type-A1, where 46 amu

    means simple chemisorptions and 42 amu means dehydrogenation. In Type-A2 reaction,

    on the other hand, two ethanol molecules are reacted on cobalt clusters and H2O or H2

    molecules are dissociated, as expressed below:

    Type-A1: Con+ + C2H5OH → Con(C2H6O)

    + or Con(C2H2O)+ + 2H2

    Type-A2: Con+ + 2C2H5OH → Con(C4H6O)

    + + H2O + 3H2

    or (Eqs. 1)

    Con+ + 2C2H5OH → Con(C4H4O)

    + + H2O + 4H2

    Type-B: Con+ + C2H5OH → Con(C2H6O)

    +

    It is should be important to know which hydrogen atoms are dissociated. Figure 5 shows

    the isotope experiment of cobalt clusters reacted with ethanol. The ethanol isotopes in

    which one or more hydrogen atoms in the ethanol molecule are replaced by deuterium

    atoms (a: normal ethanol, CH3CH2OH, b: ethanol-d, CH3CH2OD, c: ethanol-d3,

    CD3CH2OH, d: ethanol-d6, CD3CD2OD) are used in the experiment. In Fig.5 spectrum

    are expanded from Co14 to Co15, upper horizontal axis means cluster number and

  • numbers in smaller font size (18 and 42) indicate the distance from Co14 in atomic mass

    unit. The “18” usually means water chemisorption. The triangle-marked peaks

    correspond to the chemisorption of an ethanol molecule followed by dehydrogenation,

    and square-marked peaks correspond to a simple chemisorption of an ethanol. Captions

    in this figure such as "4amu", "5amu" and so on are distance between these triangle- and

    square-marked peaks. Figure 5(a) shows the reference reaction spectrum that is the

    expansion view of Fig. 4(d). In Fig. 5(a) the mass difference between simple

    chemisorption and dehydrogenate chemisorption is 4 amu, indicating that there must be

    4 hydrogen atoms dissociated. In Fig.5 (b), since ethanol-d has one deuterium atom, the

    mass change depends on weather D atom is dissociated or not. In this reaction when it

    presumes from the dominant spectrum(←?), 3 H atoms and 1 D atom are dissociated.

    The small peaks appeared at 42 amu and 46 amu positions from Co14 are the products

    of H/D exchange (shown in Eqs.2).

    ( )( ) ( )

    +→+

    ++→+++

    ++

    HHDOCCoDOHCCo

    HDHOHCCoODHCCo

    2n22n

    222n52n Dehydrogenated and H/D exchange

    (Eqs.2)

    ( )( ) ( )

    +→+

    →+++

    ++

    DOHHCCoHODHCCo

    ODHCCoODHCCo

    52n52n

    52n52n Simple and H/D exchange

    In case of fig.5(c), following to the same theory, dissociated atoms are 2D and

    2H atoms. The signals appeared at 42 amu and 50 amu positions are also the products of

    H/D exchange. Comparing these dehydrogenated chemisorption and H/D exchange

    results the reaction mechanism can be clear. At first the position of dissociated atoms

    are known, two of them are from the first carbon atom (methyl), one is from second

  • carbon atom, and the last is from hydroxyl. The H/D exchange often seems to occur,

    however exchanged atoms are different between dehydrogenation chemisorption and

    simple chemisorption.

    In case of simple chemisorption reaction, the mass of H/D exchanged signal is

    smaller than original by 1 amu (shown fig.5(b)) that is why D atom is dissociated and H

    atom is chemisorbed, on the other hand as shown in fig.5(c) the mass of H/D exchanged

    signal is larger than original by 1 amu so only hydroxyl bond (shown in fig.6(a)) can be

    exchangeable. Though there are five C-H (or C-D) bonds that are CH3 and CH2, CH3

    bonds cannot be exchangeable. If this process takes place at CH3 (or CD3) site, final

    products must be shown at larger position in Fig.5 (b) and smaller position in Fig.5 (c).

    If there are some exchanges happen as a result signals are shown as Fig. 5, intermediate

    products should be shown. That is why this exchange can happen at only hydroxyl bond

    site.

    In case of dehydrogenation chemisorption, original H or D atoms are shown Fig.

    6(b), and H/D exchangeable site might be methyl bond, because H/D exchange signals

    are shown at larger position in Fig. 5(b), and shown at smaller position in Fig. 5(c).

    These are the results of being replaced by D from H in Fig. 5(b) and of being replaced

    by H from D in Fig. 5(c).

    Figure5 (d) shows the result of chemical reaction with C2D5OD. In this figure

    dehydrogenation signal must be shown at 8 amu from simple chemisorbed signal,

    however the signal peak is located at 9 amu from simple chemisorbed signal. It is

    possible to say that all D atoms have been replaced by H atoms since there are many

    hydrogen atoms as this reason. To see lower mass there are water chemisorbed signals,

    usually 19amu signal (* shown in fig.5(d)) should not appear based on its natural

  • isotope existence. However in this reaction D atoms are so rich that all H atoms can be

    replaced as shown in Eq. 3.

    OHDCHODODDCOH 52522 +→+ (3)

    Though D atoms are rich enough, there is no signal that is replaced water

    molecule by Two D atoms. If water molecules keep molecular state, both H atoms are

    still the same condition, so it is implausible that only one atom can be replaced by D

    atom. That is the reason why water molecules must not remain in the molecular state on

    the cobalt cluster, but they are chemisorbed dissociatively.

    Figure7 is the chemical reaction with ethanol-d3. In this figure intermediate stage

    signal is weak but obviously observed. From this result the first step of dehydrogenation

    is a dissociation of hydrogen (H-D), because the mass of them are 3 amu. Since the

    hydrogen atom of a hydroxyl bond is easy to be replaced, it seems that hydrogen atoms

    marked by *1 shown in Fig. 8 are first removed by dehydrogenation. From these

    experiments, as shown in Fig. 8, at first hydrogen atoms (shown as *1) are dissociated

    as hydrogen atom, and O and C atoms (shown as C2) may be adsorbed on metal. Then

    H atoms (shown as *2) are dissociated as a molecule, and 2 C atoms would form double

    bond with metal.

    Generally reaction mechanism is different between cation and anion clusters, so

    that it is important to know both reaction mechanisms. Figure 9 shows cobalt anion

    clusters reacted with ethanol (a: normal, b: -d, c: -d3), obviously different reaction

    mechanism is seen. It turns out that the hydrogen atom of a hydroxyl has been

    dissociated judging from a reaction result. There are no simple chemisorptions and

    dehydrogenation chemisorptions, but only either one of H or D atom is dissociated and

    H/D exchange does not happen. Because anion clusters might be generated under the

  • low energy condition, ethanol molecule is not bonded dissociatively but associatively

    with cluster side. The ethanol molecule may retain its structure and bonded with cluster

    side through the oxygen atom. If the ethanol molecule chemisorbed dissociatively with

    cobalt clusters, H/D exchange may occur like the reaction of cation clusters.

    3.2. NICKEL CLUSTERS

    Figure10(a) shows the typical spectrum of as-injected nickel clusters. Since

    there are five isotopes in case of nickel, the spectra become very complex. However,

    FT-ICR mass-resolution is high enough to compare with their mass distribution based

    on natural existence. Figures 10(B, C) show the result of reaction, where all products

    are seen just before the next nickel cluster signal as shown in gray. It is difficult to

    identify its reaction mechanism because of its variety of isotope. However it can be

    surmised that it is an dehydrogenation chemisorption reaction as shown in Fig. 10(B, C)

    judging from the distribution of an isotope. Fig. 10(B, C) are expansion view of Fig.

    10(A-b, A-c) each that are surrounded with square. It can be clear that four hydrogen

    atoms are dissociated from a cluster side like the case of cobalt shown in Fig. 10(B).

    These dissociated four atoms may be the same as the cobalt case because 2H and 2D are

    dissociated, which is the same as cobalt case. In case of cobalt H/D exchange has

    occurred, and this also occurred in the case of nickel. Comparing with large peaks to

    small peaks that consist isotope distribution of nickel clusters, the height of large peaks

    must be about five times higher than that of small peaks in subjected to natural

    existence shown in Fig. 11(a). If about 20 % of deuterium atoms are replaced by

    hydrogen atoms, spectrum shapes are in good agreement with experimental data shown

    in Fig.11(b, c). This assumption will lead the same reaction mechanism that

  • exchangeable atom is deuterium.

    3.3. IRON CLUSTERS

    Figure12(a) shows the typical spectrum of as-injected iron clusters. Since iron

    also has isotopes, the signal is complicated too. As shown in Fig.12(b) , the reaction

    products have appeared just beneath the strong iron signal. In order to investigate the

    atomic weight of reaction products, Fig. 13(a) shows an expanded view of the portion

    surrounded by square in Fig.12(b). Spectra of iron clusters may show unique isotope

    distribution, but these spectra are in good agreement with ideal distribution shown in

    Fig. 13(b). All of representative signal peaks marked with ▲, ■, ● and ○ show

    perfect consistencies. In this reaction ethanol-d3 molecule shows simple chemisorption,

    because the mass difference between the representative signals is 49 amu shown in Fig.

    13(a). As to the H/D exchange process it is difficult to identify, but the peaks observed

    between large peaks are so weak that H/D exchange might not occur.

    3.4. COMPARISON OF THREE METALLIC KINDS

    As we have discussed through this paper, Fe (4s2, 3d6), Co (4s2, 3d7) and Ni

    (4s2, 3d8) are considered to have almost the same properties. Here, there are two

    interesting phenomena we have noticed. One is about the reaction mechanism and the

    other is about the reaction rate. In this experiment, the reaction mechanism of transition

    metal clusters (Fe, Co, Ni) also change in order of atomic number. In the case of iron,

    the reaction with the ethanol molecule was nothing but a simple chemisorption, and in

    case of nickel, four hydrogen atoms were dissociated from the clusters. However in case

    of cobalt, which is situated between iron and nickel on the periodic table, two kinds of

  • reaction pattern were shown. In this result, indicating the reaction mechanism also

    changes with atomic number. Interestingly, the reaction mechanism changes in

    accordance with the order in the periodic table. Figure.14 exhibits relative rate constant

    of each transition metal cluster with ethanol. In case of alkali metal clusters, they show

    magic number behavior based on super shell theory, however there is no report of magic

    number for transition metal clusters. Figure 14 presents the relationship between

    chemical reactivity and the cluster size, for the cases of Fe, Co, and Ni. Although their

    curves shows qualitatively same tendencies, this comparison clearly reveals that the

    most reactive cluster size shift larger as the atomic number is increased even by one or

    two. Authors speculates that this could be a partial reason of why the adequate catalysis

    different along atmospheric temperature and reaction species.

    ACKNOWLEDMENT

    The authors greatly thank Mr. Y. Murakami (The University of Tokyo) for

    revising English. This work was supported by Grant-in-Aid for Scientific Research

    (#12450082) from MEXT, Japan. S. Inoue was financially supported by Grant-in-Aid

    for JSPS Fellow from MEXT, Japan.

    REFERENCES

    1 S. Iijima, Nature, 354, 56 (1991).

    2 S. Iijima and T. Ichihashi, Nature, 363, 603 (1993).

    3 S. J. Tans, M. H. Devoret, H. J. Dai, A. Thess, R. E. Smalley, L. J. Geerligs and C.

    Dekker, Nature, 386, 474 (1997).

    4 M. Bockrath, D. H. Cobden, P. L. McEuen, N. G. Chopra, A. Zettl , A. Thess , R. E.

  • Smalley , Science, 275, 1922 (1997).

    5 H. J. Dai, J. H.Hafner, A. G. Rinzler, D. T. Colbert and R. E. Smalley, Nature, 384,

    147 (1996).

    6 H. Nishijima, S. Kamo, S. Akita, Y. Nakayama, K. I. Hohmura, S. H. Yoshimura and

    K. Takeyasu, Appl. Phys. Lett., 74, 4061 (1999).

    7 B. I. Yakobson, M. P. Campbell, C. J. Brabec, J. Bernholc, Comp. Mat. Sci., 8 ,

    341(1997).

    8 W. A. de Heer, A. Chatelain, and D. Ugarte, Science, 270, 1179 (1995).

    9 Y. Saito, and U. Mori, Jpn. J. Appl. Phys., 37, 346 (1998).

    10 H. J. Dai, A. Rinzler, P. Nikolaev, A. Thess, D. T. Colbert and R. E. Smalley, Chem.

    Phys. Lett., 260, 471 (1996).

    11 H. M. Cheng,F. Li, G. Su, H.-Y. Pan, L. -L. He, X. Sun and M. S. Dresselhaus, Appl,

    Phys. Lett., 72, 3282 (1998).

    12 K. Mukhopadhyay, A. Koshio, T. Sugai, N. Tanaka, H. Shinohara, Z.Konya and J. B.

    Nagy, Chem. Phys. Lett., 303, 117 (1999).

    13 A. M. Rao, E. Richter, S. Bandow, B. Chase, P. C. Eklund, K. A. Williams, S. Fang,

    K. R. Subbaswamy, M. Menon, A. Thess, R. E. Smalley, G. Dresselhaus, M. S.

    Dresselhaus, Science, 275, 187 (1997).

    14 M. Yudasaka, R. Yamada, N. Sensui, T. Wilkins, T. Ichihashi and S. Iijima, J. Phys.

    Chem. B, 103, 6224 (1999).

    15 H. Kataura, Y. Kumazawa, Y. Maniwa, Y. Ohtsuka, R. Sen, S. Suzuki and Y.

    Achiba,Carbon, 38, 1691 (2000).

    16 P. Nikolaev, M. J. Bronikowski, R. K. Bradley, F. Rohmund, D. T. Colbert, K. A.

    Smith and R. E. Smalley, Chem. Phys. Lett., 313, 91 (1999).

  • 17 S. Maruyama, R. Kojima, Y. Miyauchi, S. Chiashi and M. Kohno, Chem. Phys. Lett.,

    360, 229 (2002).

    18 Åse Marit Leere Øiestad, Einar Uggerud, Chem. Phys., 262, 169 (2000).

    19 J. Conceicao, R. T. Laaksonen, L.-S. Wang, T. Guo, P. Nordlander and R. E. Smalley,

    Phys. Rev. B, 51, 4668 (1995).

    20 R. Georgiadis, E. R. Fisher, and P. B. Armentrout, J. Am. Chem. Soc., 111, 4251

    (1989).

    21 M. P. Irion and A. Selinger ,Ber. Bunsenges. Phys. Chem. 93, 1408 (1989).

    22 W. D. Vann , R. C. Bell and A. W. Castleman Jr., J. Phys. Chem. A, 103, 10846

    (1999).

    23 M. Ichihashi , T. Hanmura , R. T. Yadav and T. Kondow ,J. Phys. Chem. A., 104,

    11885 (2000).

    24 T. Hanmura , M. Ichihashi and T. Kondow , J. Phys. Chem. A,106, 4525 (2002).

    25 S. Maruyama, L. R. Anderson and R. E. Smalley, Rev. Sci. Instrum., 61, 3686

    (1990).

    26 M. Kohno, S. Inoue, A. Yabe and S. Maruyama, Micro. Thermophys. Eng., 7, 33

    (2003).

    27 S. Maruyama, M. Kohno and S. Inoue, Therm. Sci. Eng., 7, 69 (1999).

    28 S. Maruyama, Y. Yamaguchi, M. Kohno and T. Yoshida, Fullerene Science and

    Technology, 7, 621 (1999).

    29 A. G. Marshall and F. R. Verdun, "Fourier Transforms in NMR, Optical, and Mass

    Spectrometry" Amsterdam, Elsevier, (1990).

  • Window

    To ICR Cell

    Fast Pulsed Valve

    Expansion

    Cone

    “Waiting” Room

    Target Disc

    Gears

    Gears

    Window

    Feedthrough

    for Up-down

    Feedthrough

    for Rotation

    Vaporization

    Laser

    FIG. 1. Laser-vaporization cluster beam source.

  • Turbopump

    Gate Valve

    Cluster Source

    6 Tesla Superconducting Magnet

    100 cm

    DecelerationTube

    Front Door

    Screen Door

    Rear Door

    Excite & DetectionCylinder

    Electrical Feedthrough

    Gas Addition

    Ionization Laser

    Probe Laser

    FIG. 2. FT-ICR apparatus with direct injection cluster beam source.

  • 800 1200

    12 16 20 24

    Mass (amu)

    Intensity (arbitrary)

    (a)as injected

    (b)0.2s

    (c)0.5s

    (d)1.0s

    Number of Cobalt Atoms

    C2H5OH (46amu)

    C2H2O (42amu)

    FIG. 3. Chemical reaction of cobalt clusters with ethanol.

  • Mass (amu)

    (a)Co13–14

    (b)Co14–15

    (c)Co15–16

    Co13+

    Co14+

    Co14+

    Co15+

    Co15+ Co16+

    1842

    46

    119 25 27

    FIG.4. Expansion view of FIG.3(d).

  • 820 840 860 880

    14 15

    Mass (amu)

    Intensity (arbitrary)

    (a)C2H5OH

    (b)C2H5OD

    (c)CD3CH2OH

    (d)C2D5OD

    18

    Number of Cobalt Atoms

    42

    4amu

    5amu

    6amu

    9amu

    *

    FIG.5 Chemical reaction of Co14+ with ethanol (normal, -d, -d3, -d6).

  • H - C - C -O -H*

    H H

    H H

    H - C - C -O -H*

    H H

    H H

    (a) H* is exchangeable site. (Simple chemisorption)

    - C - C -O -

    H* H

    - C - C -O -

    H* H

    (b) H* is exchangeable site. (Dehydrogenation chemisorption)

    FIG.6 H/D exchangeable site.

  • 940 960 980 1000

    Mass (amu)

    Intensity (arbitrary)

    46amu

    Co16+

    intermediate

    FIG.7 First step reaction of Co17+ with ethanol-d3.

  • C1 C2*2

    *2

    *1

    *1

    Hydrogen

    Carbon

    Oxygen

    C1 C2*2

    *2

    *1

    *1

    Hydrogen

    Carbon

    Oxygen

    FIG.8 Dehydrogenation Reaction.

  • 1180 1200 1220 1240

    20 21

    Mass (amu)

    Intensity (arbitrary)

    (a) CH3CH2OH

    (b) CH3CH2OD

    (c) CD3CH2OH

    45amu

    45amu

    48amu

    Number of cobalt atoms

    FIG.9 Chemical reaction of Co20- with ethanol (normal, -d, -d3).

  • 500 600 700 800

    9 10 11 12 13 14

    Mass (amu)

    Intensity (arbitrary)

    (a)as injected

    (b)with CH 3CH 2OH

    (c)with CD 3CH 2OH

    Number of Nickel Atoms

    (A) Reaction with ethanol and ethanol-d3.

    520 540 560 580

    Mass (amu)

    Intensity (arbitrary) Ni9

    +

    42amu

    700 720 740

    Mass (amu)

    Intensity (arbitrary)

    43amu

    Ni12

    (B) Expansion view of A-b. (C) Expansion view of A-c.

    FIG.10 Chemical reaction of nickel clusters with ethanol.

  • (a) calc. Ni12

    (b) 20%H/D exchange

    (c) experimental data

    FIG.11 The ratio of H/D exchange for Ni12+.

    (a) This is ideal distribution calculated from isotope ratio.

    (b) Assuming 20% of D atoms are replaced by H atoms.

    (c) This is the reaction product of Ni12+ + ethanol-d3.

  • 400 600 800 1000

    8 10 12 14 16 18

    Mass (amu)

    Intensity (arbitrary)

    (a)as injected

    (b)with CD 3CH 2OH

    Number of Iron Atoms

    FIG.12 Chemical reaction of iron clusters with ethanol-d3.

  • 620 640 660

    Mass (amu)

    Intensity (arbitrary)

    (a) exp.

    (b) calc.

    49amu

    Fe11+

    FIG.13 Expansion view of Fe11+.

  • 5 10 15 20Number of Atoms

    Relative Reaction Rate (arb. unit)

    CobaltNickel

    Iron

    FIG.14 Relative rate constant of Fe, Co and Ni clusters with ethanol.