Diimide Formation on the Ni(100) Surface

Surface Science Letters 290 (1993) L673-L676 North-Holland surface science letters

Surface Science Letters

Diimide formation on the Ni(1OO) surface

Sean X. Huang, Tecle S. Rufael and John L. Gland The Department of Chemtstry, The Unwerstty of Mlchtgan, Ann Arbor, MI 48109, USA

Received 24 February 1992, accepted for publication 18 March 1993

Dnmlde (N2H2), an extremely reactwe species, is observed as a gas phase product from the NI(100) surface m the 200 to 450 K range dunng hydrazme thermal decomposmon and during thermal desorpUon of predlssociated ammonia These results suggest that the primary mechanism for dllm~de formation is recombination of an adsorbed NH surface Intermediate The observation that dllmlde can be formed from pre&ssocmted ammonia illustrates that a nitrogen-nitrogen bond m the precursor ts not required for dumlde formaUon Dunude formation from pre&ssoclated ammonia is enhanced by coadsorbed hydrogen, which we beheve stablhzes NI-I on the NI(100) surface In ad&tlon, the direct decomposmon of adsorbed NEH 4 contributes to the production of dnmlde at 230 K

Dnmlde, the parent of azo compounds, is of mary gas phase products are N2, NH3, and H 2 great interest to chemists because it is an ~mpor- On these surfaces (except perhaps Pt(lll)), the tant transient intermediate m both gas-phase re- extstence and tmportance of NH2(ad) and/or actions, and in the selective hydrogenation of NH(ad) intermediates is evident Thermal de- multiple bonds in organic molecules. However, It composition of hydrazlne also leads to dnmxde was not untd 1958 that dum~de was first detected production on Rh foils [2,3], and possibly on the in SltU during the discharge-induced decomposi- Pt(111) surface [9] tion of hydrazlne in solution [1], because the Ammonia adsorption and decomposition has reactwlty of dnmlde makes its isolation difficult been previously characterized on the Nt(100) [10], Only recently has gas phase dnmide formation N1(110) [11-14], and N1(111) [15] surfaces Nickel been observed from a sohd surface under UHV is a well-known ammonia decomposition catalyst con&tlons Our group has recently observed &- [16] However, several vacuum based surface lmlde formation during thermal decomposition of studies have concluded that no significant ther- ammonia and hydrazlne on polycrystalhne Rh mal decomposition can be observed on either the surfaces [2,3]. As part of a program to estabhsh N1(111) [15] or N1(110) surfaces [11,12] at temper- the generality of dumlde formation, we have un- atures below 300 K in an electron free environ- dertaken a study of dumlde formation on the ment We believe that most of the ammonia NI(100) surface using hydrazlne and predlssoo- decomposition observed in these studies ts the ated ammonia as precursors result of electron Irradiation These ammonia re- Hydrazlne adsorption and decomposition on sults may also have substantial ramifications at single crystal metal surfaces has been character- elevated pressures and temperatures, because ev- lzed previously on Ir(111) [4], Rh(111) [5], Rh(100) idence for some degree of ammonta thermal de- [6], Fe(lll) [7], N~(111) [8], and Pt(lll) [9] sur- composition clearly extsts [13,17] On the NI(100) faces Thermal decomposition studies on Rh(111), surface, molecular ammonia together with its de- Rh(100), Pt(111), and N1(111) surfaces suggest composition products (NH2, NH, N) populate the that hydrazlne decomposes completely at sub- surface at room temperature as evidenced by the monolayer coverages on these surfaces The prl- XPS spectra [10]

f ...... ] ___ I.__ The experimental apparatus utlhzed m this Products 1 ML Nell4 T,o = q5 K work has been described previously [3] It consists ot a stainless steel ultra-high vacuum (UHV) chamber In which the base pressure was 7 4 × ,,•r•... N2H4 X15] 10 ~ Torr The system ~s also eqmpped with low-energy electron diffraction (LEED) optics, Auger electron spectroscopy (AES), and a multi- >. N2~e X15 1 plexed quadrupole mass spectrometer for tem- j 6J perature-programmed reacuon spectroscopy i (TPRS) The NK100) crystal was cleaned by argon ion bombardment and subsequent flashing to 1150 g rl? K, and its cleanliness was verified by Auger elec- J tron spectroscopy The reactants (N2H 4 and NH ~) E were adsorbed at 90 K through a doser approxi- o mately 2 cm away from the NK100) surface Spe- (I) NH3 ', cial care was taken to ensure the minimum de- ~4 composition of NEH 4 by evacuating and recharg- ! ing the doser before each dose [18] TPD data L i were taken w~th a linear temperature ramp of 5 i K/s while the crystal is m line of sight with the i ~X15 oJ mass spectrometer I ! Fig 1 presents the desorpt~on spectra gener- I 1 ated following the adsorption (at 90 K) ot a II I I l I I I 0 200 400 600 800 1000 1200 submonolayer dose of hydrazme At this dosage, Temperature (K) N 2 H 4 reacts completely on the Nl(100) surface to Fig I Thermal desorptlon spectra of N2H 4 on Nl(ll0) The form NH~, N2H 2, H2, and N 2 For coverages spectra were taken m a single run of temperature-pro- above one monolayer, a molecular sublimation grammed reaction of N2H 4 (submonolayer coverage) on the peak is also identified for hydrazme at 170 K surface, with all the products momtored simultaneously Here, The sharp ammoma, mtrogen, and dnm~de peaks the major products are N2H4, N2H2, N 2, NH3, and H 2 at 230 K are believed to be the result of direct hydrazme decomposmon, as m the case of hydrazlne/Rh [3] In the temperature range 300 to 520 K, the partially dlssocmted hydrazme speoes on producmg dnmtde [18] We propose that an im- the surface go through a dehydrogenation pro- portant intermediate m th~s process is adsorbed cess, and produce gas-phase hydrogen as a result lmlde (NH) which is formed by dehydrogenation At even h~gher temperature, the only product of N2H 4 The adsorbed ~m~de continues to react observed is molecular mtrogen formed between up to 450 K, suggesting that adsorbed mude is 750 and 1100 K The broad dnmlde feature above qmte stable on the Nl(100) surface Hydrazme- the 230 K peak is quite interesting Thts peak derived NH is known to be an xmportant interme- appears m the 350-400 K range at low coverages, diate on Fe(lll) [7], and NI(lll) [8] surfaces up and broadens to lower temperature range (200- to 400 K Hydrazme-derwed NH has also been 450 K) at h~gher coverages Eventually, a clear observed on polycrystalhne Rh [2,19], AI [20], Fe peak appears at 230 K with a h~gh temperature [21], Ir [22], W [23] and Mo [24] surfaces In the shoulder extending to 450 K for coverages close 300 to 450 K temperature range, the lmade mter- to monolayer coverage of hydrazlne Based on medmte appears to dominate dmmde formauon these observauons, we believe that second-order on Nl(100) surface Above 450 K, we propose that b~molecular reacUon between adsorbed NH mter- NH dehydrogenates on the surface resulting m medmtes is the main mechamsm responsible for diminishing dlmalde yields from N2H 4 S X Huang et al / Dumtde formanon on Nz(lO0) L675

Ftg 2 shows the temperature programmed re- 20 action products from predissociated ammonia on = Ammoma Only • Ammon)a + 0 05 L Hydrogen the NI(100) surface For submonolayer coverages • Arnmoma m Hydrogen Flow of NH3, the major desorption peak is a molecular ammonia peak in the 180 to 400 K range We also observe a small amount of hydrogen which de- J~ sorbs at about 370 K Dnmide formation appears first at 350-400 K for low coverages of predisso- -o >- ciated ammonia, and broadens to a lower temper- ! ature range (180-500 K) for higher coverages of predissoclated ammonia As in the case of hydra- E zIne adsorption on the NI(100) surface, we pro- B m [] pose that the major mechanism for dnmide for- i i mation is second-order combination of adsorbed 10 20 3o lmide intermediates We beheve that adsorbed Ammoma Exposure (Arbitrary Umt) lmlde Intermediate (NH) is produced by electron Fag 3 N2H 2 yaeld from submonolayer coverage of NH 3 on Na(100), with and without coadsorbed hydrogen speoes

I I I I

Products 1 1 ML NH3 / Tad = 95 K / N)(100) bombardment of adsorbed ammonia and is stable through the 180-500 K temperature range The 8- erastence of lmlde on NI(100) was indeed verified by a XPS feature at 397 7 eV at room temperature [10] Yates et al have proved that ammonia is non-dissociative on the N1(110) surface when :~ 6- free of electron bombardment [8] The same is true on the N1(111) surface where electron-induced decomposition of ammonia may contribute N2 significantly to the formation of some intermedi- 4 ate species [6] We, too, believe there is great ,$ possibility of electron beam induced ammonia decomposition on the NI(100) surface Formation of NEH 2 is appreciable from both I=. (/) NH3 N2H 4 and predissociated NH 3 on the NI(100) 2 surface In both cases, broad dnmlde peaks ap- pear around 350-400 K at low coverages, and shift to lower temperature at higher coverages ~x3 However, ammonia is different from hydrazine in that it is a precursor which does not contain a nitrogen-nitrogen bond Dnmlde formation from predlssoclated ammonia supports our contention I I I I 0 200 400 600 800 1000 that adsorbed NH is the primary surface interme- Temperature (K) diate responsible for dumlde formation The cru- Flg 2 Thermal desorptlon spectra of NH 3 on Nl(ll0) The cial step for dnmlde formation from both NH 3 spectra were taken In a single run of temperature-pro- and N2H 4 is the formation of an adsorbed NH grammed reaction of NH 3 (submonolayer coverage) on the surface, with all the products monitored sunultaneously Here, intermediate species and the bimolecular recom- the major products are N2H z, N2, NH3, and H 2 The peak at bination of these species in the 200 to 450 K 450 K in the N 2 spectrum as hkely due to CO contammataon range Above 450 K, NH dehydrogenates further [ 67t~ ,$ A Huang et al / DnmMe ]ormatton on Nt(lO0) into Nad + Had on the surface, as indicated by the N2H 4 also contributes to the production of dJ- diminishing dnmlde yields from both N2H 4 and lmlde at 230 K Further spectroscopic characteri- predlssocmted NH~ zation of this process will allow more detailed Dnm~de from predlsSoclated ammoma ~s en- characterization of both the surface species and hanced greatly by the coadsorptlon of hydrogen surface reactions that favor the dnmlde forma- on the Nl(100) surface As shown m the dumlde tion yield curve (fig 3), the yield of dnmtde from ammonia is increased significantly m the pres- This research was partially supported by a ence of post-adsorbed hydrogen or in a 10 -7 grant from the Exxon Education Foundation Torr hydrogen flow We believe that coadsorbed hydrogen stablhzes the NH intermediate, and therefore increases the surface concentratmn of References ~mlde in the hydrogen-deficient surface enwron- ment In the 200-450 K temperature range This [1] S F Foner and R C Hudson, J Chem Phys 28 (1958) 719 result ~s in good agreement with similar results [2] J Prasad and J L Gland, J Am Chem Soc 113 (1991) from NH 3 on polycrystalhne rhodium foil [2], and 1577 serves as further support for the importance of [3] J Prasad and J L Gland, Langmmr 7 (1991) 722 lmlde as an Intermediate [4] H H Sawm and R P Merrill, J Chem Phy~ 73 (1980) The dnmlde yield from N2H 4 does not re- 996 [5] ML Wagner and LD Schmldt, Surf So 257 (1991) spond to the ad&tlon of hydrogen We propose 113 that substantial hydrogen coverages are present [6] W M Dannlel and J M White, Surf Scl 171 (1986) 289 after adsorptmn of hydrazlne because of substan- [7] M Grunze, Surf Scl 183 (1979) 603 tlal low temperature hydrazme decomposition [8] J L Gland, G B Fisher and G E Mitchell, Chem Phys Therefore, the stability and recombination of ad- Lett 119 (1985) 90 [9] D J Alberas, J Kiss, Z M Lm and J M White, prwate sorbed NH is not substantially affected by addl- communication tmnal coadsorbed hydrogen In addition, during [10] M Grunze, P A Dowben and C R Brundle, Surf Scl hydrazlne decomposition for coverages close to 128 (1983) 311 one monolayer, there is a second sharp dnm~de [11] M Grunze, M Gloze, RK Dnscoll and PA Dowben, J peak at 230 K which we believe is the result of Vac Scl Technol 18 (1981) 611 [12] C Klauber, M D Alvey and J T Yates, Jr, Surf Scl 154 dnmlde formed directly from hydrazlne This low (1985) 139 temperature dumlde is not affected by the coad- [13] I C Basslgnana, K Wagemann, J Kuppers and G Ertl, sorbed hydrogen either Surf Scl 175 (1986) 22 In summary, dnmlde (N2H 2) is observed as a [14] M Huttlnger and J Kuppers, Surf Scl 130 (1983) L277 gas phase product from the NI(100) surface in the [15] C W Seabury, T N Rhodm, R J Purtell and R P Mer- rdl, Surf So 93 (1980) 117 200 to 450 K range during hydrazlne thermal [16] A Ozakl and K I Atka, in A Treatise m Dmltrogen decomposltmn and during thermal desorptlon of FLxatlon, Eds R W F Hardy, F Bottomley and R C predlssocmted ammonia These results suggest Burns (Wdey, New York, 1979) ch 4 that the primary mechanism for dllmlde forma- [17] F P Netzer and T E Madey, Surf SCl 119 (1982) 422 tion is recomblnatmn of adsorbed NH surface [18] P A Redhead, Vacuum 12 (1962) 203 [19] A Vavere and R 8 Hansen, J Catal 69 (1981) 158 intermediates The observation that dnmlde can [20] D W Johnson and M W Roberts, J Electron Spectrosc be formed from predlssocmted ammoma illus- Relat Phenom 19 (1980) 185 trates that a mtrogen-nltrogen bond in the pre- [21] M H Matioob and M W Roberts, J Chem Res (S), cursor is not required for dilmlde formation DI- (1977) 336 lmlde formatmn from predlssoclated ammonia is [22] B J Wood and H Wise, J Catal 39 (1975) 471 [23] R C Cosser and F C Tompkms, Trans Faraday Soc 67 enhanced by the coadsorbed hydrogen which we (1971) 527 beheve stablhzes NH on the NI(100) surface In [24] R C A Contamlnard and F C Tompkans, Trans Faraday ad&tion, the direct decomposltmn of adsorbed Soc 67 (1971) 545