Vibrational spectra and thermal decomposition of ... - ACS Publications

John J. Birtill, Paul Ridley, Stephen Liddle, Tim S. Nunney, and Rasmita Raval. Industrial & Engineering Chemistry Research 2001 40 (2), 553-557...
0 downloads 0 Views 940KB Size
J. Phys. Chem. 1992,96, 9424-9431

9424 y, &, and Av into the equation

Au=y+T&-pAv

(9)

The values are piotted against temperature in Figure 5 , where the Au values of the DOA system were calculated on the assumption that the value of pAv is negligibly small compared with those of the other two terms. It is found that the value of the water-alcohol system is very low compared with those of the waterhydrocarbon systems;I6the energy of the interfacial region is lowered by an attractive intermolecular interaction between alcohol and water such as the hydrogen bond. Further, the phase transition from the expanded film to the condensed film accompanies a large decrease in energy. In our next we will demonstrate the y vs p curves at various temperatures and prove thermodynamically the break given in Figures 1 and 2 to be due to the phase transition at the alcohol-water interface, which is analogous to the one at the oil solution of long-chain alcohol-water interfaces.13g14.25-28 Registry No. Decyl alcohol, 112-30-1; undccyl alcohol, 112-42-5; dodecyl alcohol, 112-53-8.

References and Notes (1) Hutchinaon, E. J. Colloid Sci. 1948, 3, 219. (2) Hutchinson, E.; Randall, D. J . Colloid Sei. 1952, 7 , 151. (3) Franks, F.; Ives, D. J. G . J . Chem. Soc. 1960, 741. (4) Jasper, J. J.; Van Dell, R. D. 1. Phys. Chem. 196569, 481. ( 5 ) Lutton, E. S.; Stauffer, C. E.; Martin, J. B.; Fehl, A. J. J . Colloid Interface Sei. 1969, 30, 283.

(6) Aveyard, R.; B h , B. J. J. Chem. Soc., Faraday Trans. I 1972,68, 478. (7) Sagert, N. H.; Quinn, M. K. J . Colloid Interface Sci. 1985, 105, 58. (8) Van Hunscl, J.; JW, P. Longmuir 1987, 3, 1069. (9) Villers, D.; Platten, J. K. J . Phys. Chem. 1988, 92,4023. (IO) Turkevich, L. A.; Mann, J. A. Longmuir 1990, 6, 445. (1 1) Coninck, J. D.; Villers, D.; Platten, J. J. Phys. Chem.1990,94,5057. (12) Motomura, K.; Matubayasi, N.; Aratono, M.; Matuura, R. J . Colloid Interface Sci. 1978, 64, 356. (13) Matubayasi, N.; Motomura, K.; Aratono, M.; Matuura. R. Bull. Chem. Soc. Jpn. IW8,51,2800. (14) Ikenap, T.; Matubayasi, N.; Aratono, M.; Motomura, K.; Matuura, R. Bull. Chem. Soc. Jpn. 1980,53, 653. (15) Motomura, K. J . Colloid Interface Sci. 1978,64, 348. (16) Motomura, K.; Iyota, H.; Aratono, M.; Yamanaka, M.; Matuura, R. J . Colloid Interface Sci. 1983, 69, 264. (17) Aratono, M.; Motomura, K. Bull. Chem. Soc. Jpn. 1985,58,3205. (18) Aratono, M.; Motomura, K. J. Colloid Interface Sci. 1987, I1 7, 159. (19) Matubayasi, N.; Motomura, K.; Kancahina, S.; Nakamura, M.; Matuura, R. Bull. Chem. Soc. Jpn. 1977.50, 523. (20) Motomura, K.; Aratono, M. Lungmuir 1987, 3, 304. (21) Stephenson, R.; Stuart, J. J. Chem. Eng. Dofa 1986, 31, 56. (22) Donahue, D. J.; Bartell, F. E. J . Phys. Chem. 1952, 56, 480. (23) Lin, M.; Firpo, J. L.; Mansoura, P.; Baret, J. F. J. Chem. Phys. 1979, 71, 2202. (24) Aratono, M.; Takiue, T.; Ikcda, N.; Motomura, K. To be submitted. (25) Matubayasi, N.; Dobzono, M.; Aratono, M.; Motomura, K.; Matuura, R. Bull. Chem. Soc. Jpn. 1979,52, 1597. (26) Aratono, M.; Yamanaka, M.; Motomura, K.; Matuura, R. Colloid Polym. Sei. 1982, 260, 632. (27) Motomura, K.; Iwanaga, S.; Hayami, Y.; Uryu, S.;Matuura, R. J . Colloid Interface Sei. 1981, 80, 32. (28) Aratono, M.; Uryu, S.; Hayami, Y.; Motomura, K.; Matuura, R. J . Colloid Interface Sci. 1984, 98, 33.

Vibrational Spectra and Thermal Decomposltlon of Methylamine and Ethylamhe on Ni( 111) Denis E.Cardia and Cabor A. Somorjai* Centerfor Advanced Materials, Materials Sciences Division, Lawrence Berkeley Laboratory. 1 Cyclotron Road, Berkeley, California 94720, and Department of Chemistry, University of California, Berkeley, California 94720 (Received: May 13, 1992; In Final Form: August 17, 1992)

The bonding and geometry of methylamine (CH3NH2)and ethylamine (CH3CH2NH2)on Ni(l11) have been investigated with high-resolution electron energy loss vibrational spectroscopy (HREELS). Both amines adsorb molecularly at 150 K through the nitrogen lone pair. Significant metal-hydrogen interactions in the alkyl chain were indicated by ‘softened” C-H stretching modes with frequencies shifted to 2660-2680 cm-’. Temperatureprogrammed desorption (TPD)and HREELS were used to monitor their desorption and thermal decomposition on the Ni( 111) surface. Both CH3NH2 and CH&H2NH2 are dehydrogenated in the temperature range 300-400 K. CH3NH2is dehydrogenated to HCN at about 330 K,which further d a m p a m above 360 K. CH3CH2NH2is dehydrogenated to CH3CN. initially by a-C-H bond Scission, leading to dearorption of that molecule at 350 K. On the basis of our spectra,we propose a mechanism for the dehydrogenation proc*rses of CH3NHz and CH3CH2NHzon Ni( 111).

1. Iatroduction

The adsorption of amines on metal surfaces is of considerable importance in catalysis and surface coatings chemistry. These molecules have the ability to adsorb molecularly through the nitrogen lone pair on transition-metal surfaces under ultrahighvacuum (UHV) conditions, at a temperature high enough for activating C-H, N-H, C-N, or C-C bonds. When heating the surface, the investigation of their surface reactivity by various surface science techniques is possible. The high-mlution electron energy loss vibrational specbwqy (HREELS) study of CH3NH2on Ni(100). Ni(l1 l), Cr(100), and Cr(ll1) by Baca et al.’ confirmed that methylamine adsorbs moleculiuly at 300 K like ammonia through the nitrogen lone pair. The adsorption of CH3NH2was also investigated on Ni( 111),2 Ni(100),3R(111),4R(100),Mo(100),6 5 W(100),’Rh(lll)?and R U ( ~ W ) .Methylamine ~ was found to dehydrogenate on all 0022-3654/92/2096-9424503.oo/o

surfaces. No C-N bond scission was observed on Pt( 111)5 (desorption of HCN and C2N2was detected), while some was found On Ni( 111)2and Ni( methylon Rh( 111)9and Pt( amine was totally decomposed, leaving atomic carbon and nitrogen on the surface. Ethylamine thermal decomposition was studied on W(lOO), W(lO0)-(5Xl)-C, and W(100)-(2X1)-0.’0 On W(lOO), ethylamine undergoes C-N and C-C bond scission, leading mainly to methane and ammonia desorption. In contrast, on W(100)(5Xl)-C, neither C-C nor C-N bond scission was observed after ethylamine adsorption. Acetonitrile (CH3CN) was the major product detected. An initial selective a-C-H bond activation was proposed to account for the product distribution. The W(100)-(2Xl)-O surface was inert with respect to C-H, N-H, or C-N bond scission, resulting primarily in molecular ethylamine desorption. 0 1992 American Chemical Society

CH3NH2and C2HSNH2on Ni( 111)

The Journal of Physical Chemistry, Vol. 96, No. 23, 1992 9425

Other larger amines were investigated by Inamura et al." on evaporated nickel films by X-ray photoelectron spectroscopy (XPS). It was found that all amines adsorb molecularly on the surface through nitrogen lone-pair electrons below 200 K. With the rise of temperature to 330 K,various surface complexes were observed that were similar to those obtained after adsorption of the corresponding nitriles. In the present work, the adsorption of C H J N H ~and CH3CH2NH2on Ni(ll1) at 150 K was examined by HREELS. The spectra obtained indicate that both molecules adsorbed through the nitrogen lonepair electrons. A weakming of the a-C-H bonds is indicated by a broad band at 2660-2680 an-',assigned to C-H stretching vibrations shifted to lower frequencies. The desorption and thermal decomposition were also investigated. Upon heating, the desorption of amines competes with their dehydrogenation. On Ni( 11l), both methylamine and ethylamine dehydrogenate between 300 and 400 K. Dehydrogenated methylamine completely decomposes, leaving carbon and nitrogen on the surface, while in the case of ethylamine, acetonitrile is found to desorb at 350 K. We propme that the dehydrogenation process begins with a-C-H bond scission. 2. ExperimentalSection Our experiments were conducted in an ultrahigh-vacuum (UHV) chamber equipped with a four-grid retarding field energy analyzer (RFA) for Auger electron spectroscopy ( A S ) and low-energy electron diffraction (LEED), a quadrupole mass Spectrometer for t e m p e r a t u q " m e d desorption (TPD) and residual gas analysis, and an high-resolution electron energy 108s spectrometer (HREELS) to study surface vibrations. During experiments, the residual pressure in the chamber was typically 5 X 10-lo Torr. The HREEL spectrometer12is based on other designs in w.13HREELS data were taken in the specular direction with a beam energy of 2.5-5 eV, at an overall resolution varying from 6 to 7.5 meV (50 to 60 cm-'). Most spectra were obtained with an elastic count rate of 5 X le to 5 X 10, counts/s. In most of our experiments, we observed some background CO contamination on our sample which gives a very intense peak on our spectra at about 1800 an-'.Featuns in the 1800-cm-l spectral region, such as C - N stretching modes, could be hidden by this CO contamination. During TPD experiments the crystal was resistively heated at heating rates of 8 K/s. The cooling of the crystal was achieved by circulating liquid nitrogen through the manipulator. The 99.995% pure nickel single crystal (Monocrystals) was oriented, cut, and polished, by conventional methods, to within less than 1O of the desired (1 11) face and then spot-welded to the manipulator with 1-mm-diameter nickel wires. The nickel surface was cleaned by cycles of Ar' sputtering, O2 treatments, and annealing in vacuum at 1100 K. Surface cleanliness was monitored by Auger and HREELS. Methylamine, ethylamine, and aceetonitrile were purchased from Fluka AG. Methylamine and ethylamine (anhydrous, 99% pure) were contained in 250-mL pressure bottles and were used without further purification. Acetonitrile (99%pure) was degassed and purified by a series of freeze-pump-thaw cycles. Adsorbates were admitted into the vacuum chamber through a leak valve and a 1/8-in.-o.d. stainless steel tube facing the crystal surface. Exposures of the surface were reported in langmuirs (1 langmuir = lo4 Toms), computed from the uncorrected pressures given by an ionization gauge. A maximum pressure of 5 X Torr was used.

3. ResulQdDiscPaolon vikrtkarlspechr(HREELS)ofMdEcphrM~ud EthyhmiaeAd~orbedon Ni( 111) at 150 K. Methylumine. After adsorption of 1 monolayer (3 langmuirs as shown by TPD experiments) of methylamine at 150 K on the clean Ni( 111) face, no ordered LEED superstructure is observed. The HREELS spectrum of the surface (Figure 1) is taken in the specular direction with a beam energy of 3 eV. The adsorption of methylamine on the surface decreases the elastic peak from about 5 x 104 to 8

IOPQ

C b NH;! on Ni(ll1)

I

150 K H

'

H

&r;"

z!@z

1

,

0

1000

2000 Energy loss ( cm

3000

-

4000

)

Fig" 1. HREELS spectrum and proposed geometry of methylamine adsorbed on Ni( 111) at 150 K. Exposure = 3 langmuirs (- 1 monolayer).

TABLE I: V i b r d o d "&ea (cui') of Mabyhahc mode

gas"

NHzsymstr CH3asymstr CH3 sym str NH4 def CH3 asym def CH3 5)" def CH3rock CNstr NHzwag

3360 2962 2820 1623 1474 1430 1130 1044 780

NHz asym str CH, asym str CH, asym def NHz twist CH3rock torsion metal-N str

3424 2985 1485

Ni( 111) this [ P ~ ( C H J N H Z ) Z C ~ref Z ] ~1 work

A' uI u2

u3 ~4

us V6

u7 ~8

u9

3270, 3240, 3140 3016 2928,2896 1596, 1581, 1575 1465, 1453 1429, 1417, 1405 990 1037, 1017 740

3240 2960 2820 1570 1485

3260 2920 2660 1580 1460

1195 1010 730

1200 1020 730

3305 2985

3340 3000

A" uI0

uII uI2 ~ 1 3

ul,

uII

" Reference 14.

I,

1195 265 525

Reference 15.

X lo3 counts/s. The observed vibrational frequencies are compared in Table I to those of gas-phase methylamine," to those of c ~ - [ P ~ ( C H ~ N H ~ ) and ~ C to ~ ~those ] , ' ~reported by Baca et al.' for methylamine on Ni( 111). Our spectrum differs only slightly from the one reported in ref 1. We detect a peak at 525 cm-' which we assign to the nickehitrogen stretching. We also observe a broad band at 2660 cm-'. Ethylumine. One monolayer (3-langmuir exposure) of ethylamine adsorbed on Ni(ll1) at 150 K shows a similar behavior: no ordered superstructure can be formed, and the ethylamine adsorption reduces the elastic peak intensity. However, many loa0 peaks can be obsetved on the HREELS spectrum (Figure2). In Table 11, we compare observed vibrational frequencies with the vibrational analysis of ethylamine by Hamada et aL1' In their study, by uucful comparison of the experimental data with mults of ab initio MO calculations of force constants, Hamada et al. managed to assign bonds of their infrared spectrum, in the 1002000-cm-' spectral range, to the trans and gauche rotational isomers of ethylamine. From this set of data and from those reported by Wolf'f and Ludwig"?for the 2000-3500Cm-1 spectral

Gardin and Somorjai

The Journal of Physical Chemistry, Vol. 96, No. 23, 1992

CH3 CH2 NH on Ni(ll1) 2 150 K H

H,I

H C' '

H

FSgwe 3. Trans and gauche oonformatimof ethylamine on a flat surface. The encumbranceof the CH3group and the HREEL spectrum indicate that ethylamine must be adsorbed in the trans form on Ni(ll1).

1080 L

2980

1000

0

2000

3000

4000

'

Energy loss ( cm ) Figwe 2. HREELS spectrum and proposed geometry of ethylamine adsorbed on Ni( 11 1) at 150 K. Exposure = 3 langmuirs (- 1 mono-

layer).

TABLE II: V i b r i t i d Freqoe~cies(em-') of Ethyhmk mode NH2 sym str CH, d str CHI sym str CH, sym str NH2 scission CHI d def CH3 d def CH, sym def CH2 wag CH, d rock CC str NH2 wag CN str CCN bend

gas' gauche trans 3343 2977 2885 2880 1622 1487 1465 1378 1397 1016 892 773 1086 403

Ni( 11 3200 3040 2680 2960 1540

>1440 1119 882 789 1055

>1360 1140 880 760 1080

"References 17 and 18. bThis work.

range, we are able to propose an assignment for most of our observed vibrational peaks. Similar to the case of methylamine, we also observe a broad band at 2680 cm-'. All spectra are amsistent with the interpretationof methylamine and ethylamine adsorbed molecularly on Ni( 111) at 150 K through their nitrogen lonepair electrons. The observed nickel-nitrogen stretching vibrations at 525 cm-l (methylamine) and 500 cm-' (ethylamhe) are comparable to the nickel-nitrogen stretching of NH3 on Ni( 111) at 490 cm-'.19 The interpretation of the broad bands at 2660 c m - I for methylamine and a t 2680 cm-' for ethylamine needs particular attention. Whereas a combination of the CH3 rocking and CH3 deformation modes at 1200 and 1460 cm-',respectively, could explain the presence of the band at 2660 cm-l for methylamine, no such combination is possible for ethylamine. The presence of decomposition fragments responsible for the bands at 2660 and 2680 cm-' is unlikely since very little changes in the spectra are observed up to 300 K. It is possible that these bands are due to a second layer of amines, with hydrogen bonding to the fmt layer. However, the amines' dosing was done at 200 K sometimcp, above the multilayer desorption temperature, and the bands at 2660 and 2680 c m - I were still present. We then tentatively assign the 2660- and 2680-cm-' bands to %oftened'' C-H stretching vibrations, similar to those observed with cyclohexane adsorbed on various transition-metal surfaces (ref 16 and

references therein). On Ni( 11l), for example, cyclohexane exhibits lowered frequency C-H stretching vibrations at 2720 cm-l. C-H bonds in the alkyl chain of the amine must therefore be weakened by interactionsbetween hydrogen atoms and nickel surface atoms. Raval and Chested6 have shown strong correlations between the extent of the perturbation of the C-H bonds, as indicated by the shifting in the C-H vibration frequency, and the tendency of the cyclohexane to be dehydrogenated. This suggests that the perturbed C-H bond provides a major dehydrogenationpathway on the surface, and this should be kept in mind when we study the thermal decomposition pathways of methylamine and ethylamine at higher temperatures. When methylamine is adsorbed on Ni( 11l), hydrogen atoms from the methyl group must therefore interact directly with nickel surface atoms, forming multicenter C-H-metal bonds, as suggested by the proposed geometry of methylamine adsorbed on Ni( 111) (Figure 1). With ethylamine, it is unclear which hydrogen atoms interact with the surface. Because of the size of the CH3 group, the dormation of ethylamine will most likely be the trans form, on the flat Ni( 111) surface (Figure 3). In the gas phase, the two conformers in thermal equilibrium can be distinguished by their different CH3 rocking frequencies, at 1119 cm-I for the trans and 1016 cm-'for the gauche. The higher frequency in the trans form is explained by the closer interactions of the CH3 group and the NH2 group in this conformation (Figure 3). On Ni( 11l), we note that the observed CH3 rocking frequency at 1140 cm-l agrees well with ethylamine adsorbed in the trans form. But because gas-phase CH3 rocking modes are apt to mix with C-N stretching or C-C stretching modes, as shown by MO calculations," one should be cautious about determining the ethylamine conformation only from vibrational data, especially if the C-C and C-N bonds are seriously affected by the adsorption. On the basis of steric considerations, we propose that ethylamine is adsorbed molecularly on Ni( 111) at 150 K through the nitrogen lonepair electrons with the methyl group pointing away from the nickel surface (Figure 2), while hydrogen atoms of the CHI group interact directly with nickel surface atoms. Thermal Decomposition of Methylamine and Ethylamine on Ni(ll1). (a) Methylamine Decomposition. TPD. TPD experiments were performed by adsorption of methylamine at 150 K and then resistively heating the surface to 900 K at a rate of 8 K/s. Following the exposure of the surface to 3 langmuirs of methylamine, the major products detected are methylamine (31 amu), hydrogen (2 amu), and nitrogen (28 amu) (Figure 4). A small amount of carbon monoxide (28 amu) from background adsorption is also detected. CH3NH2desorbs between -200 and -330 K, H2 desorbs with a maximum at 380 K and up to 450 K, N2 desorbs at -800 K and CO at 420 K. The evolution of the TPD spectra (31 and 2 amu) as a function of coverage is shown in Figure 5 . At low coverages (less than 1.O-langmuir exposure), we reproduce the results of Chorkendorff et al.: who found that all the adsorbed methylamine dehydrogenates on Ni( 111). leading to H2 and N2 desorption and carbon diffusion into the bulk. At higher coverage, CH3NH2desorption m u r s with a maximum shifting from 320 K for 1.0-langmuir exposure to 250 K for 3-langmuirs. Above 3 langmuirs, a lowtemperature peak at 170 K starts to appear and is attributed to the growth of a second layer of CH3NH2. The amount of desorbed H2 increases with coverage up to 1 langmuir and stays constant above 1 langmuir. H2 desorption temperature (380 K)

-

The Journal of Physical Chemistry, Vol. 96, No. 23, 1992 9421

CH3NH2and C2H~NH2 on Ni( 111)

CH3 NH2 on Ni(ll1) CH 3 NH2 1100

/

\

31 amu

1 n

100 2 0 0

00

I

2amu

3 0 0 400 500 600 700

800 900 T-270 K

>

Temperature (K)

Figure 4. TPD spectrum after adsorption of 1 monolayer (3 langmuin) of methylamine on Ni(ll1) at 150 K. Heating rate is 8 K/s.

T-150

He ( 2 amu)

n

CHsNH2 (31 amu)

0

1000

2000

3000

Energy loss ( cm

K

4000

- 1)

Figure 6. Evolution of the HREELS spectra as a function of the treatment temperature after adsorption of 1 monolayer of methylamine (CH3NH2)on Ni( 111) at 150 K.

A

2L

A

1L

-

0.5 L