Adsorption and thermal stability of alkanethiol films on GaAs(110): A comparative study by TOF-DRS and TOF-SIMS

Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

Contents lists available at ScienceDirect

Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb

Adsorption and thermal stability of alkanethiol films on GaAs(110): A comparative study by TOF-DRS and TOF-SIMS Luis M. Rodríguez a, Lucila J. Cristina a, Leonardo Salazar Alarcón a, Bárbara Blum b, Roberto C. Salvarezza b, Luan Xi c, Woon Ming Lau c, Esteban A. Sánchez a, J. Esteban Gayone a, Oscar Grizzi a,⇑ a b c

Centro Atómico Bariloche, CNEA, I. Balseiro, UNC & CONICET, Bustillo 9500, R8402AGP Bariloche, Argentina INIFTA, Depto. de Química, UNLP & CONICET, CC. 16-Suc. 4, BI904DPI La Plata, Argentina Surface Science Western, University of Western Ontario, London, Ontario, Canada N6A 5B7

a r t i c l e

i n f o

Article history: Received 29 July 2010 Received in revised form 12 November 2010 Available online 17 December 2010 Keywords: Alkanethiols GaAs Ion Scattering TOF SIMS Self Assembling

a b s t r a c t We present an ion beam study of the adsorption and the thermal stability of short alkanethiol molecules adsorbed on GaAs(110). Direct recoiling spectroscopy shows that the adsorption of ethanethiol and hexanethiol proceeds directly towards a dense standing up phase without passing through a stable phase of lying down molecules as is the case for Au(111). Measurements along specific azimuths suggest that both Ga and As rows are covered by the organic molecules. Short adsorption times from the vapor phase result in films having two desorption peaks near 300 and 500 K. On the other hand, leaving the sample in a thiol atmosphere for several hours produces more stable films, similar to those produced by immersion in the corresponding thiol-ethanol solution. TOF-SIMS results confirm the C–S scission mechanism during the thermal desorption. Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction The growth of thiol based films on compound semiconductors is a growing field motivated by their applications in optoelectronics [1]. In general, the high complexity of the surfaces of thin organic films makes characterization of their chemical, electronic and thermal properties a difficult task. The high sensitivity of these films to being damaged by the irradiation of the incident beam probes complicates the field even further. In this context, high sensitive techniques based on ion probes can provide information that is complementary (and sometimes unique) to that provided by electron spectroscopies and microscopies. In this work we study the adsorption and the thermal stability of alkanethiols (Cn: HSCnH2n+1) films grown on GaAs(110) by TOF-DRS1 [2]. We compare the results to those from similar measurements performed on the better known Au(111) substrate. The thermal desorption is studied by TOF-DRS and by TOF-SIMS2 [3]. A common advantage of these techniques is the low irradiation dose required to obtain the spectra, which normally results in negligible damage to the film. This work is the continuation of a previous one carried out for hexanethiol (C6) on GaAs(110) [4]. In that work we showed that C6 molecules can chemisorb on GaAs(110) from the vapor phase, completely ⇑ Corresponding author. 1 2

E-mail address: [email protected] (O. Grizzi). Direct Recoiling Spectroscopy with Time of Flight analysis. Secondary Ion Mass Spectrometry with Time of Flight analysis.

0168-583X/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2010.12.051

covering the surface. The shape of the H and C direct recoil (DR) peaks was very similar to that corresponding to the same molecule adsorbed on Au in the dense (selfassembled) layer, suggesting that in the case of GaAs the C6 molecules go directly to a similar dense packaging without passing trough the lying down (striped) phase that is characteristic of Au [5,6]. In contrast to Au, annealing the film produced a double-peaked desorption curve, showing a first desorption process slightly above room temperature, and a second one near 500 K. Some questions that remained opened in that work are addressed here: TOF-DRS can only detect atomic elements, therefore no information about the molecular composition at the surface, and their dependence with temperature could be obtained. TOF-DRS detects the elements present at the top most layer; S, being closer to the interface could not be detected neither during the adsorption process nor during the first steps of desorption. TOF-SIMS has a higher sensitivity than TOF-DRS, a higher mass range and more molecular sub-products can be detected in the sputtering process, which allows us to discuss some of these topics in more detail. On the other hand, TOF-DRS can detect both neutral and ionic particles, making measurements independent of the local variations of the surface electronic structure. Here, we illustrate the complementary character of both techniques, and provide a comparison for films grown on Au and on GaAs, from both the vapor phase and from solution. For the case of GaAs we show that by cleaving a (100) wafer directly in the ethanol solution containing the thiol molecules a clean and passivated Cn/GaAs(110) surface can be obtained without going through complicated (in air)

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

cleaning procedures. Other topics such as the effect of surface roughness on the adsorption kinetics and the changes in the thermal stability of the film observed after leaving the GaAs surface in a thiol atmosphere for several hours are discussed.

2. Experimental details The experimental work was carried out in two laboratories: the Surface Science Laboratory of Centro Atómico Bariloche (CAB), Argentina, and the Surface Science Western (SSW) of the University of Western Ontario, Canada. The TOF-DRS measurements were performed at CAB and the TOF-SIMS at SSW. In the first case, the set up consists on a 1–100 keV ion accelerator connected to an UHV chamber with facilities for forward and backward ion scattering with TOF analysis. The ions are mass selected by a switching magnet, collimated through several apertures to 0.1 deg, and finally pulsed to 50 nsec, 30 kHz. The scattered and recoiled ions plus neutrals are detected by a channeltron electron multiplier after traversing a 109 cm collimated drift tube at a fixed scattering angle (d = 45°). Different GaAs(110) surfaces were prepared, depending on the purpose of the measurement. One of the samples consisted on a disk of 8 mm diameter, 2 mm thick, p type, cleaned by many cycles of grazing sputtering with 20 keV Ar+ ions and annealing to 720 K. During sputtering the polar incidence angle h is set to within 2–3 degree (with respect to the surface plane) and the azimuthal angle u rotated continuously within a 70 degree range (see the inset of Fig. 2b for the angle definition). This method is useful to smooth out the surface and generates a flat and well ordered GaAs(110) sample [7]. The clean GaAs(110) samples were then exposed in situ to the vapors of alkanethiols at pressures in the range of 1  109–1  104 Torr. The pure thiols (Aldrich, 98% purity) were contained in a glass tube connected to a leak valve with stainless steel (SS) tubes. A 6 mm diameter SS tube ran from the leak valve to 3 cm from the sample. Prior to exposure, the thiols were further purified by freeze–pump-thaw cycles. The thiol pressures were not corrected for the ion gauge sensitivity and do not take into account any enhancement due to the adsorption geometry, therefore, the reported dose values are not absolute values and should only be used to compare different adsorption curves since all of them have been done under the same conditions. The base pressure during the measurements remained below 1  109 Torr. To check the effect of surface roughness, adsorption measurements were also carried out on rougher samples, i.e., on samples sputtered at slightly higher polar incident angles (around 5–6 deg, with all polar angles referred to surface plane). The cleaning–smoothing method is effective and allows studying of the first adsorption stages in well reproducible samples, but requires preparation times that can extend for a week or more. For some adsorptions where full coverage was desired without going through intermediate adsorption steps we used fresh and new GaAs(110) samples prepared by cycles of low energy (600 eV), high angle (45 deg) sputtering, combined with annealing. No significant differences were detected in the thermal desorption measurements between the two preparation methods. For comparison, some TOF-DRS measurements were taken on an Au(111) single crystal obtained from MATEK. In this case the sample was a 6 mm disk cleaned by sputtering and annealing and the exposures were carried on in situ, under vacuum, as described above for the GaAs(110) single crystals. TOF-SIMS measurements were performed on both GaAs and Au samples with a reflectron-type TOF-SIMS instrument (ION-TOF TOF-SIMS IV) operating at an incident angle of 45°. In order to minimize damage the primary analysis beam was a pulsed 8 keV Ar+ [8]. The temperature dependent experiment is carried out by mounting the sample on a heating stage which is resistively heated

925

in situ at a temperature ramp of about 5.5° C/min. A time profile of the species of interest in negative mode is recorded at every step. In this case, the thiol layers on the Au samples were prepared from solution on substrates of Au evaporated on mica, purchased from the Molecular Imaging Corporation. To remove the contamination and improve the surface ordering we treated the Au surfaces by hydrogen flame annealing. After annealing, the atomic terraces of the Au (111) substrates typically evolve into a triangular shape p with straight steps. Each terrace has the well-known 3  23 herringbone structure as observed by Scanning Tunneling Microscope. The Au(111) substrates are then immersed in 5 mM C6 ethanolic solution for 3 days and rinsed with plenty of ethanol to remove weakly adsorbed molecules before introducing to the TOF-SIMS analysis chamber immediately. In the case of GaAs, we cleaved a slice of fresh GaAs(001) substrate in 5 mM C6 ethanolic solution, so that a clean (110) surface is exposed to the thiols and the formation of native oxide is avoided. The GaAs was left in the solution for 4 days and gently rinsed by ethanol and introduced to the TOF-SIMS chamber immediately to study the cross sectioned surface. Investigation of the other half of the cross sectioned surface by XPS revealed that the Fermi level position is 0.1 eV, indicating that the sample is close to having a flat band condition. This shows that the surface passivation is effective and there is a lack of band gap states induced by the thiol adsorption.

3. Results and discussion 3.1. Adsorption followed by TOF-DRS The GaAs(110) surface is an open surface presenting a strong relaxation in which the top As atoms move up by 0.2 Å and the Ga ones move 0.5 Å down from the ideal bulk termination. The surface relaxation is accompanied by rehybridization of the dangling-bonds, with a charge transfer from the top Ga atoms to the top As atoms [9]. The fact that the interatomic distances are large and that the Ga and As masses are similar and makes the GaAs(110) surface particularly interesting for TOF-DRS studies [10]. A typical spectrum for the clean surface acquired with 4.2 keV Ar+ ions along the [001] direction and at 20° polar incidence is shown in Fig. 1a. At this relatively large incident angle there are no shadowing effects for top layer atoms. The dominant peak corresponds to Ar scattering from both Ga and As surface atoms. The technique allows detection of both Ga and As recoils, which at 45° of scattering angle appear to the right side of the Ar peak. No peaks are observed due to contaminants (H, C, O) at the left side of the Ar peak. Due to the surface relaxation the Ga recoils are observed along few azimuthal directions, in most directions they are shadowed by 0 the As atoms that are lying in a plane that is slightly above (0.7 Å A) the Ga plane. The [001] direction is particularly useful because both As and Ga rows are equally exposed to the beam (inset of Fig. 1 a). When the As direct recoil (DR) intensity is recorded as a function of the azimuthal angle, keeping the polar incidence fixed at some low angle (9° in this case) there are strong changes that reflect the surface atomic structure (Fig. 1 b). Upon exposure of the clean surface to the vapors of the thiol strong changes appear in the spectra. First, two new structures due to H and C DRs are observed at the left side of the Ar peak. The latter becomes attenuated, broadened and shifted towards lower TOFs due to Ar multiple scattering on both the atoms of the adsorbed molecule and the substrate atoms (Fig. 1a). S atoms are not detected due to shadowing from the above hydrocarbon chains. The example spectra shown in Fig. 1a center and bottom correspond to two different thiols: C2 and C6, respectively. The center one is recorded before saturation while the bottom one

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

(a)

As

Ar

θ=20 φ=[001]

(b)

Ga

[110] GaAs(110)

Ga

As

C2/GaAs(110)

H

φ θ

1

[001] DR Intensity (Arb. units)

926

δ

C As

[001]

C C6/GaAs(110) 0

50

100 150 200 TOF (arb. units)

250

40

60

80

100

120

140

160

Azimuthal Angle φ (deg)

Fig. 1. (a) Typical TOF-DRS spectra taken along the [001] azimuth with 4.2 keV Ar+ ions at 20° polar incidence angle for the clean surface, the surface partially covered with C2 and at saturation with C6. The inset shows an schematics top view of the GaAs(110) surface. (b) As DR intensity for the clean surface at grazing incidence angle plotted as a function of azimuthal angle (bottom) and C DR intensity for the saturated C6 surface (top). The inset shows the definition of the angles h and u.

corresponds to saturation, i.e., further exposures (by orders of magnitude) produces almost no change in the shape of the spectra. Measurements as a function of azimuthal angle (Fig. 1b) result in a minor dependence, due either to the lack of large ordered domains for these relatively short chain alkanethiols, or to the strong vibrations of the molecules at room temperature. For larger molecules, i.e., C18, Rabalais and coworkers have been able to measure a stronger azimuthal dependence [11]. It is worth mentioning that even for C6 on Au(111), where clear STM images show well ordered and large domains, we were not able to detect a clear azimuthal dependence. This can be explained by considering that STM can sense the S heads on the Au(111) surface, even though PM-IRRAS [12] measurements of C6 SAMs on Au show that the hydrocarbon chains have significant disorder at room temperature. Integration of the DR peaks (H, C and substrate related) provides information on the adsorption kinetics. As a reference example we discuss first the case of C6 on Au(111) which has been reported before [13] (Fig. 2a). The log scale for the exposures is necessary in order to show the different scales at which the changes in the spectra take place. Initially the H and C DR intensity increases fast corresponding to adsorption at defects, then the intensity keeps increasing until a plateau is reached near 1 L, in which the striped (or parallel to the surface) phase is formed. Measurements of adsorption kinetics by other groups [5] using photoelectron spectroscopy and low energy He diffraction show similar dependence on exposure. Formation of well defined phases in this exposure range, i.e., with large domains of molecules lying near parallel to the surface has been verified by STM measurement at different laboratories [6]. At this point, since the film layer is very thin, the Ar projectiles penetrate the film and some Au recoils can still be ejected, i.e., the Au recoil intensity does not disappear completely at this film condition. Much higher exposures (not only longer times but also higher pressures) results in a new change in the shape of the spectra that can be related to formation of the denser self-assembled monolayer (SAM phase) where the molecules are oriented more perpendicular to the surface (standing-up condition) [5,6,13]. The features in TOF-DRS denoting this condition are the complete disappearance of the Au recoil peak, an increase

in the H and C DR intensity, and broadening of these peaks. A similar behavior has been observed for alkanethiols on Ag(111) and in this case confirmed by independent vacuum STM measurements performed in another chamber under similar exposure conditions. Fig. 2b shows the corresponding intensities for adsorption of C6 on GaAs(110) at room temperature. The intensities were obtained by integration of the DR peaks (after background subtraction). The spectra for this adsorption sequences were taken along the [001] direction at 20° polar incidence. First we note that the two stepped kinetics just described for Au is not observed. The intensities grow continuously till full saturation. In general, saturation takes place at lower exposures than for Au. Increased exposures by orders of magnitude do not produce a major variation in the spectra. Detailed comparison of the peak shapes for C6 on Au and on GaAs(110) led us to propose [4] that the orientation of the molecules on both substrates should be similar, or at least that the molecules are not in the parallel to the surface phase on GaAs, once saturation has been achieved. Further evidence is found here where for C6 we see complete disappearance of Ga and As recoil peaks as soon as saturation is achieved. Another interesting feature that can be observed in Fig. 2b is the fact that along the [001] azimuthal direction (and also along other directions) both the As and the Ga DR intensities decrease at the same rate, which indicates that the adsorbed molecules shadow with similar probability As and Ga target atoms since the beginning of the adsorption. This measurement alone is not enough to demonstrate that the molecules adsorb on both Ga and As atoms because the C6 molecules are long, therefore even if they are adsorbed on just one type of substrate atom they could be tilted enough from the surface normal as to shadow both As and Ga rows. To check this point we performed adsorption measurements with C2, a much shorter molecule (Fig. 3). In this case the results also show that Ga and As DR intensities decrease at the same rate. Attempts to find a difference in the evolution of the Ga and As DR intensities by changing the incident or azimuthal angles were not successful; this result would be consistent with molecules adsorbed along both Ga and As rows., i.e., with molecules interacting with the dangling-bonds of both Ga and As substrate atoms, or

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

16x10

3

(a)

C6/Au (111) 35x10

(b)

C6/GaAs (110)

Direct Recoil Intensity (arb. units)

Direct Recoil Intensity (arb. units)

Au

3

H x2 C x2

0

927

H C

As x10 Ga x10

0

10

-1

10

0

1

10

10

2

10

3

10

4

Dose (L)

0.01

0.1

1

10

Dose (L)

Fig. 2. H, C, and substrate DR intensity versus exposure for (a) C6/Au(111), (b) C6/GaAs(110).

20x10

C2/GaAs(110)

3

DR Intensity (arb. units)

As x3 Ga x3 H C

0 5 6 7 8

2

3

0.1

4

5 6 7 8

1

Dose (L) Fig. 3. H, C, Ga and As DR intensity versus dose for C2/(GaAs(110).

at least with a molecular ordering that can shadow both types of substrate atoms. For films with longer molecules, C18, prepared on GaAs(110) by immersion, a pseudohexagonal incommensurate phase was recently reported [14]. In this case, XPS measurements indicate that the molecules are adsorbed as thiolates on both As and Ga atoms, with the As–S bonding being dominating. According to this model both As and Ga top surface atoms would be shadowed by the adsorbed molecules and would explain our observation of the simultaneous decrease in the Ga and As DR intensities with exposure. However, extension of this model to short chain alkanethiols is not straightforward because the stronger Van der Waals interaction for long molecules might result in a different ordering of the molecular film. Finally, we describe briefly the effect of surface preparation in the adsorption kinetics. Fig. 4 shows four adsorption curves, two

of them taken with a surface prepared by cycles of grazing sputtering and annealing, with bombardment at 4–5° incidence (named rougher surface), and the other two, prepared with more cleaning cycles performed during 2 weeks at somewhat lower angles (between 2° and 3°, smoother surface). Measurements on the clean surface at low incident angles confirm the smoothing effect generated by the grazing bombardment (not shown), i.e., the ratio of the scattering peaks at 20° and at 5° is higher for the smoother surface. The intensities were normalized to the GaAs DR intensity of the initial (clean) surface, so they are comparable among them. The data show first the reproducibility of the measurement for a specific surface condition and second the effect of surface roughness on the adsorption kinetics. On the rougher surface the initial uptake is faster, near one order of magnitude for this specific case. Preferential C18 domain nucleation on GaAs(110) step edges has been suggested by McGuiness et al. in their structural determination [14]. Similar effects have been seen in Ag surfaces [15], where surface roughness also enhances the initial uptake, but can preclude formation of the SAM phase. Even though these measurements are not quantitative, they are useful to show the sensitivity of the adsorption curves (and film quality) to surface preparation. 3.2. Thermal stability of thiol films on GaAs(110) For most applications concerning passivation or device fabrication, it is highly desirable to have film layers that are stable at room temperatures, and preferably also well above it. Several studies of thiol desorption from GaAs have been reported in the literature by using mass spectrometers [16,17]. For CH3SH, (CH3S)2, and (CH3)2S on GaAs(110), Camillone et al. [16] proposed that CH3SH desorbs intact near 300 K (311 K for dense CH3SH layers), and that (CH3)2S desorbs from the surface at 500 K. For C1, C2, C4 and (CH3S)2 on GaAs(100), Huang et al. [17] propose that chemisorbed alkanethiolates and protons recombine and desorb as alkanethiols at around 375 K, while for the remaining alkanethiolates, S–C bond scission gives place to dialkyl sulfide desorption around 520 K and alkene desorption around 620 K, leaving S on the surface. In order to extend these results, in the following we present and describe measurements of TOF-DRS and TOF-SIMS that provide information about the film composition as a function of sample temperature.

928

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

Normalized H DR Intensity

3.5

C6/GaAs(110) rougher

smoother

0.0 10

-2

10

-1

10

0

10

1

10

2

10

3

10

4

5

10

10

6

Dose (L) Fig. 4. HDR intensity plotted versus dose for four GaAs(110) surfaces prepared by grazing sputtering (at two different incidence angles) and annealing. In the smoother surface the initial uptake is slower.

For comparison, the H DR intensity for an Au(111) substrate covered with a C6 SAM is shown in the inset. The figure shows that in contrast to the Au case, for GaAs(110) there is a faster initial decrease in the H and C DR peaks, starting already at room temperature. The derivative of the H DR intensity presents two main desorption peaks with desorption energies around 1 and 1.5 eV [4]. For Au, the coverage also decreases since the beginning of the annealing, but it is much slower, and leaving the film for hours in vacuum, near room temperature, produces no detectable change. Above 450 K the H and C DR intensities decrease faster and S starts to be detected (Fig. 6b), both as S recoils and as Ar scattering from S. The surface can be depleted from H and C by annealing, but a considerable amount of S remains at the surface and can only be removed by sputtering. Since S is not detected at the beginning of the adsorption due to shadowing from the hydrocarbon chain, it is not possible to determine accurately from these measurements the

3.2.1. TOF-DRS measurements For gas phase adsorption we have shown before [4] that C6 also has a thermal desorption peak very near to room temperature. Here we will develop this point further. Fig. 5 shows the TOFDRS spectra for C6/GaAs(110) immediately after the adsorption and after 75 min, keeping the sample at 35 °C. The initial condition corresponds to saturation of the C and H DR peaks. The decrease in the C and H DR intensities and the increase and shift in the Ar peak illustrate the desorption that takes place under this condition. The inset shows this effect versus time. Some substrate atoms start to be accessible to the beam, indicating that part of the surface became uncovered. In order to study the desorption of the self-assembled monolayer we measured TOF-DRS spectra as a function of sample temperature for C6/GaAs(110). The data is shown in Fig. 6a. The rate of desorption was set at 3 °C/min, which is slow when compared to most thermal desorption experiments reported in the literature.

3

1.5x10

T=35

o

DR Intensity (Arb. units)

C6/GaAs(110) C

H Counts

C

1.0

0.5

H 0.0 0

20

40

60

80

Time (minutes)

Initial After 75 minutes

0.0 0

100

200

300

400

500

TOF (arb. units) Fig. 5. TOF-DRS spectra immediately after adsorption and after 175 min in vacuum at 35 °C for C6/GaAs(110). (Inset) H DR intensity versus time.

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

(b)

(a) 1.0

Ga

3

Direct Recoil Intensity (arb. units)

20x10

929

S

C6/Au(111) DR Intensity H 300

350 400 450 Temperature (K)

As

500

575 K

C6/GaAs(110)

H

500 K

S x8

410 K

As+Ga x2 0 300

400

500

600

Temperature (K)

150

200

TOF (arb.units)

Fig. 6. (a) Dependence of H, C, S, Ga and As DR intensities with temperature for C6/GaAs(110). The inset shows the H DR intensity evolution for C6/Au(111) (b) Corresponding TOF-DRS spectra at some specific temperatures.

fraction of molecules that undergo S–C bond scission. The shape of the desorption curve is compatible with either two adsorption sites with different desorption energies, with two molecule adsorption configurations (thiol or thiolate, or other sub-products), or just more than one desorption route, depending on the temperature and the availability or not of H near adsorbed thiolates. Elucidation of these points will require other measurements and model calculations which are beyond the scope of this work. However, some extra information on the stability and composition of the films can be obtained from TOF-SIMS experiments, as described below. 3.2.2. TOF-SIMS measurements For illustration we describe first the results for C6 adsorbed on Au(111). In this case, the substrates were Au evaporated on mica; flash annealed, and then prepared by immersion as described in the experimental section. Fig. 7 shows the main TOF-SIMS products versus temperature together with the corresponding curve for H DR measured by TOF-DRS. The latter measured in the Bariloche set up for C6/Au(111) prepared by vapor deposition. The TOFDRS data is normalized to the initial value of the thiolate ion yield (indicated as M-1 in the figure). As we mention above, the high sensitivity of TOF-SIMS allows one to follow molecule sub-products since the beginning of the desorption. Around 450 K all TOFSIMS products, including the thiolate species show a strong change indicating that the main desorption process takes place at this temperature. Particularly, the S ion yield also presents a strong decrease near this temperature which is consistent with the near complete absence of S measured after desorption. In some adsorptions, a very small amount of S can be found at high temperatures, typically a few percent of monolayer that depends on the initial quality of the surface. TOF-DRS indicates that partial desorption already takes place below 450 K, i.e., the HDR intensity decreases continuously as a function of temperature. Due to the temperature increase, molecule chains have larger vibrations and more gauche disorder, allowing some molecules (thiolate species) to desorb and leaving even more space in the SAM for further disordering and also for changes in molecular orientation to an orientation more parallel to the surface. This is evidenced in TOF-DRS by the thinning of the DR peaks with increased temperature. As a conse-

quence of these changes the local work function may change and affect the electron transfer probabilities between the different SIMS products and the surface (neutralization of ejected products), resulting in stronger changes in the emitted ion yields with temperature. TOF-DRS measures both ions and neutrals so is less sensitive to this effect. Interestingly, the TOF-SIMS ion yield for the mass of the thiolate plus one Au atom (M Au-1) is high, even higher than that for the thiolate alone at room temperature, which is indicative of the strong interaction between S and Au. The main TOF-SIMS results for C6 on GaAs are presented in Fig. 8a. In this case a GaAs(001) was cleaved in the thiol-containing ethanol solution and introduced in the TOF-SIMS chamber immediately after rinsing. The small spot of the TOF-SIMS system allowed to perform the measurements on the cleaved face (GaAs(110)). First we note in Fig. 8a, the presence of the yield corresponding to the thiolate (M-1). This yield decreases very slowly up to 450 K and then starts to decrease more rapidly indicating desorption or molecule decomposition. The CH yield evolves with temperature in a similar way. Note that this behavior is different to that of the HDR intensity (TOF-DRS, Figs. 6 and 8b, blue curve). In the case of TOF-SIMS, the initial desorption near room temperature observed in Fig. 6 is basically absent (we will discuss this point below). Also, in contrast to the Au case, the S yield is more constant during the complete desorption process, suggesting that the C–S bond breaks during desorption, leaving the S at the surface. In agreement with this observation, the GaS and AsS ion yields are more constant, and in fact increase slightly with desorption of the hydrocarbon chains. The presence of AsS and GaS species is also in agreement with adsorption near both substrate atoms. Finally we discuss here the difference between the behavior of the M-1 ion yield in TOF-SIMS and the H DR intensity in TOFDRS. We showed above that the films grown in vacuum from vapor phase present a partial desorption near room temperature and that this could be related either to different adsorption sites or to different adsorbed species (thiol or thiolate). We prepared other films by leaving the surface in vacuum, in a thiol atmosphere (104–105 Torr) overnight (15 h). Fig. 8b shows the H DR TOF-DRS intensities for the two films together with the M-1 TOF-SIMS yield, all normalized around 370 K. Normalization at this temperature

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

Intensity

930

Temperature (K) Fig. 7. Evolution of TOF-SIMS yields with sample temperature for C6/Au. For comparison, TOF-DRS data (the H DR intensity) is shown as a continuous line.

Intensity

10

10

10

10

5

(a)

25

C6/GaAs(110)

(b)

C6/GaAs(110)

4

H DR (TOF-DRS). Vapor phase Short adsorption times After overnight exposure

Intensity (arb. units)

10

3

2

M-1(TOF-SIMS) Solution growth

CH S GaS AsS M-1

1

0 300

400

500

600

700

Temperature (K)

300

400

500

600

700

Temperature (K)

Fig. 8. (a) Evolution of TOF-SIMS yields with sample temperature for C6/GaAs(110). (b) Comparison of the H DR evolution with temperature for two films: one prepared by an exposure of the GaAs(110) surface to 160 L of C6 in vacuum and the other by leaving the same surface in vacuum overnight in an atmosphere of C6 at pressures of 104–105 Torr. The M-1 yield in the TOF-SIMS measurement for C6/GaAs(110) prepared by immersion is also shown (continuous curve).

allows comparison of the desorption curves in the high temperature region, but it does not mean that the initial coverages are different, in fact, the coverage at room temperature is very similar for all the samples, corresponding to full saturation. Comparison of the different desorption curves shows first that the thermal desorption of the film prepared overnight did not present the strong initial desorption and the H DR intensity behaved more similar to that observed in the films prepared in solution as described for the TOF-SIMS measurements. Another difference is seen in the high temperature part of the curve: the final desorption in the less stable films take place at somewhat lower temperatures than the

other ones, which is also consistent with a less dense packaging. The similarity in the overnight vacuum exposed film and the solution prepared film shows that these two methods favour the formation of films that are more stable with temperature. 4. Conclusions We present a comparative study of the growth of alkanethiol films on GaAs(110) and on the better known case of Au(111). In the study, we have used two ion based techniques that provide complementary results: TOF-DRS and TOF-SIMS. We show that

L.M. Rodríguez et al. / Nuclear Instruments and Methods in Physics Research B 269 (2011) 924–931

the coverage of GaAs(110) with C2 and C6 has a similar dependence with exposure, and that both saturate in one step, i.e., without going through the well defined lying-down phase observed in Au. The shape of the H and C DR peaks and the disappearance of the substrate signals suggest a standing-up molecule configuration. Even for the shorter C2 molecule, Ga and As recoils decrease with exposure at the same rate. The initial stages of adsorption depend on the roughness of the surface. TOF-DRS measurements versus sample temperature show the presence of S at temperatures above 450 K. The TOF-SIMS study of C6 adsorbed from solution on both Au and GaAs reveal interesting features. First, in both substrates there is an important yield of thiolate species, which suggests the presence of this species on GaAs. The S yield has a different temperature dependence in Au and in GaAs, in that it tends to remain at the surface in the latter. Short time vapor depositions result in saturation coverages with two adsorption species (or two adsorption sites), one of which desorbs near room temperature. Longer exposures (typically overnight exposures) and solution growth lead to more stable films. Acknowledgements We acknowledge funding from the CONICET-NSERC cooperation program, from SECyT: PICT06-715, PICT06-621, PAE:22711 and 22708, from UNCuyo 06/C317 and 06/C323, and from CONICET PIP 112-200801-00958. One of the authors (O.G.) acknowledges fruitful stays at the Surface Science Western Laboratory and the Chinese University of Hong Kong where part of this work has been carried out.

931

References [1] C.L. McGuiness, A. Shaporenko, C.K. Mars, S. Uppili, M. Zharnikov, D.L. Allara, J. Am. Chem. Soc. 128 (2006) 5231. [2] J.W. Rabalais, Principles and Applications of Ion Scattering Spectrometry, Surface Chemical and Structural Analysis, Wiley, Interscience, 2003. [3] A. Benninghoven, Chemical Analysis of Inorganic and Organic Surfaces and Thin Films by Static Time-of-Flight Secondary Ion Mass Spectrometry (ToFSIMS), vol. 33, Angewandte Chemie International(in English), 1994. pp. 1023. [4] L.M. Rodríguez, J.E. Gayone, E.A. Sánchez, O. Grizzi, B. Blum, R.C. Salvarezza, L. Xi, W.M. Lau, J. Am. Chem. Soc. 129 (2007) 7807. [5] F. Schreiber, A. Eberhardt, T.Y.B. Leung, P. Schwartz, S.M. Wetterer, D.J. Lavrich, L. Berman, P. Fenter, P. Eisenberger, G. Scoles, Phys. Rev. B 57 (1998) 12476. [6] H. Kondoh, C. Kodama, H. Sumida, H. Nozoye, J. Chem. Phys. 111 (1999) 1175. [7] J.E. Gayone, R.G. Pregliasco, E.A. Sánchezy, O. Grizzi, Phys. Rev. B 56 (1997) 4194. [8] N.W. Ghonaim, M. Nieradko, L. Xi, H.-Y. Nie, J.T. Francis, O. Grizzi, K.K.C. Yeung, L.W.M. Lau, Appl. Surf. Sci. 255 (2008) 1029. [9] J.L.A. Alves, J. Hebenstreit, M. Scheffler, Phys. Rev. B 44 (1991) 6188. [10] J.E. Gayone, E.A. Sánchezy, O. Grizzi, Surf. Sci. 419 (1999) 188. [11] S.S. Kim, Y. Kim, H.I. Kim, S.H. Lee, T.R. Lee, S.S. Perry, J.W. Rabalais, J. Chem. Phys. 109 (1998) 9574; L. Houssiau, M. Graupe, R. Colorado Jr., H.I. Kim, T.R. Lee, S.S. Perry, J.W. Rabalais, J. Chem. Phys. 109 (1998) 9134. [12] D. Grumelli, L.P. Mendez De Leo, C. Bonazzola, V. Zamlynny, E.J. Calvo, R.C. Salvarezza, Langmuir 26 (8226) (2010). [13] L.M. Rodríguez, J.E. Gayone, M.L. Martiarena, E.A. Sánchez, O. Grizzi, B. Blum, R.C. Salvarezza, L. Xi, W.M. Lau, Nucl. Instr. & Meth. B 258 (2007) 183. [14] C.L. McGuiness, G.A. Diehl, D. Blasini, D.-M. Smilgies, M. Zhu, N. Samarth, T. Weidner, N. Ballav, D.L. Allara, ACS Nano. 4 (2010) 3447. [15] L.M. Rodríguez, J.E. Gayone, E.A. Sánchez, H.L. Ascolani, O. Grizzi, M. Sánchez, B. Blum, Benitez, R.C. Salvarezza, Surf. Sci. 600 (2006) 2305. [16] N. Camillone III, K.A. Kahn, R.M. Osgood Jr., Surf. Sci. 453 (2000) 83. [17] P. Huang, T.H. Lin, T.F. Feng, Y.H. Lai, W.H. Hung, Surf. Sci. 603 (2009) 1244.