Bekir
Salgın†
*a,
Diego
Pontoni
b,
Dirk
Vogel
a,
Heiko
Schröder‡
c,
Patrick
Keil§
a,
Martin
Stratmann
a,
Harald
Reichert
bc and
Michael
Rohwerder
a
aMax-Planck-Institut für Eisenforschung GmbH, Max-Planck-Strasse 1, 40237 Duesseldorf, Germany. E-mail: salgin@mpie.de
bEuropean Synchrotron Radiation Facility, 6, rue Jules Horowitz, 38043 Grenoble, France
cMax-Planck-Institut für Metallforschung, Heisenbergstrasse 3, 70569 Stuttgart, Germany
First published on 28th August 2014
Materials science in general, and surface/interface science in particular, have greatly benefited from the development of high energy synchrotron radiation facilities. Irradiation with intense ionizing beams can however influence relevant sample properties. Permanent radiation damage and irradiation-induced sample modifications have been investigated in detail during the last decades. Conversely, reversible sample alterations taking place only during irradiation are still lacking comprehensive in situ characterization. Irradiation-induced surface charging phenomena are particularly relevant for a wide range of interface science investigations, in particular those involving surfaces of solid substrates in contact with gaseous or liquid phases. Here, we demonstrate partially reversible radiation-induced surface charging phenomena, which extend far beyond the spatial dimensions of the X-ray beam mainly as a consequence of the interaction between the surface and ionized ambient molecules. The charging magnitude and sign are found to be surface chemistry specific and dependent on the substrates' bulk conductivity and grounding conditions. These results are obtained by combining a scanning Kelvin probe with a synchrotron surface diffractometer to allow simultaneous in situ work function measurements during precisely controlled hard X-ray micro-beam irradiation.
It is well known that a sample surface charges when it is directly exposed to an X-ray beam.8,16 In the hard X-ray regime, surface charging is the result of a complex interplay between photoelectric absorption, Auger emission and Compton scattering processes. The electric fields building up as a consequence of surface charging can directly affect the orientation and packing of adsorbed molecules and molecular aggregates.17–19 For instance, the structure of double layers in electrochemical systems is governed by beam induced interface charging, as e.g. observed for ionic liquids on sapphire.20 X-ray-induced surface charging can be completely or partially reversible, therefore ex situ surface analysis after synchrotron experiments is of limited interest and potentially misleading. We therefore implemented a scanning Kelvin probe (SKP) specifically designed for the in situ monitoring of surface and interface electrostatic properties during synchrotron investigations.21
The Kelvin probe is a non-destructive, non-contact technique22 for measuring the work function. Considering the sensitivity of the method for even the slightest net charging,23 it is a suitable tool for detecting work function variations induced by surface charging phenomena. Although, by definition, charging should not affect the work function of a sample, for insulating or semiconducting samples it causes the misalignment of the Fermi level, which manifests itself in the Kelvin probe measurement as an apparent decrease/increase of the work function for positive/negative net charging of the surfaces24 (see discussion below). The technique was already used for sensing surface photovoltage variations at the X-ray beam illumination footprint on GaAs25–28 and GaP(110)/Ag films.29 In a recent ex situ study, significant surface potential changes, after vacuum ultraviolet (VUV) exposure, were reported for various dielectric materials.30,31 However, detailed in situ studies featuring sufficient lateral resolution for resolving the zone affected by the X-ray beam are still missing. In addition, the role of different surface chemistries on beam induced charging effects was not to date investigated quantitatively and systematically.
In this study, we demonstrate that X-ray induced surface charging extends spatially well beyond the beam footprint region mainly as a consequence of the interaction of ionized air molecules with the surface. Variation of the surface chemistry, substrate bulk conductivity, and grounding methods affect the charging phenomena both qualitatively and quantitatively. Our approach enables detailed studies of all factors involved in irradiation induced surface charging phenomena, thus providing the means to control and monitor in situ their magnitude and spatial extent.
Just before the synchrotron experiments, silicon and sapphire samples were subjected to solvent cleaning followed by treatment in acidic or alkaline piranha solutions. Treating amphoteric metal oxides in solutions having pH values different from their isoelectric point (IP) induces fixed charged surface functional groups. For silica (IP at pH ∼ 2) acidic piranha creates positive fixed surface charges, while alkaline piranha induces negative surface charges. This change of the surface chemistry is clearly demonstrated by X-Ray Photoelectron Spectroscopy (XPS) analysis of the surface oxygen presented in Fig. 2. The alkaline treated sample features a distinct low energy shoulder related to negatively charged functional groups (SiO−).34,35 The intensity of this feature increases at lower take-off angles, therefore confirming that it originates from the surface. On the other hand, the low energy XPS component is absent in acidically treated silicon, because in this case protonated groups (SiOH2+) carry the fixed positive surface charges.34,35 The experiments were conducted at ambient condition (air conditioned, i.e. RT and low relative humidity), which enables the compensation of the charging resulting from these treatment-induced functional groups by counter charges adsorbed from the surrounding atmosphere.36 Therefore in the initial stage, (i.e., before the beam exposure) both surfaces are net uncharged, although they exhibit slightly different work functions as revealed by ex situ Kelvin probe measurements (−130 ± 50 mVSHE and +60 ± 20 mVSHE, for the acidically and alkaline treated surfaces respectively). These differences are small compared to the variations caused by exposure to the X-ray beam. More detailed sample preparation procedures are reported in the Materials and methods section.
The phosphorus-doped n-type Si wafer had a resistivity of 7–13 Ω cm and (111) surface orientation. It underwent the same acidic piranha pretreatment described previously.
The polished NiAl single crystal alloy had (110) surface orientation. The sole pretreatment consisted in dipping consecutively for 10 minutes in isopropanol, acetone and chloroform. The sample was subsequently rinsed with ultrapure water and dried with a stream of ultrapure argon.
The sapphire single crystal was first dipped consecutively in isopropanol, acetone, chloroform for 15 min each. It was then pretreated for 3 min with acidic piranha, as described previously. Finally it was rinsed for 5 min with ultrapure water and dried with a stream of ultrapure argon.
All chemicals were supplied by Sigma-Aldrich Chemie GmbH (Munich, Germany), if not otherwise stated.
The above observations can be rationalized by considering the main types of charges involved in the experimental system: (a) the fixed, charged surface functional groups created by the chemical pre-treatment, (b) the abundant additional ionized molecules created in the surrounding atmosphere once the X-ray beam is switched on, (c) the highly mobile majority charge carriers present in the bulk semiconductor – positive holes in p-type silicon – which act as mirror charges with respect to the surface-adsorbed ionized air molecules, and (d) the emitted photoelectrons inducing positive surface charging once the beam reaches the surface (Fig. 3).
When the X-ray beam is first switched on, ionized air molecules bearing a charge opposite to that of the fixed functional groups rapidly adsorb on the surface. This explains the different sign of the initial work function variations for acidic versus alkaline pre-treatments (blue and orange arrows in Fig. 1). A similar effect is observed in charged polymer systems, where charge carriers of sign opposite to that of the polymer exhibit high mobility,24,37 while charge carriers with the same sign experience the so-called Donnan exclusion.38 As already mentioned above, although, by definition, charging should not affect the work function of a sample, for insulating or semiconducting samples it causes the misalignment of the Fermi level, which manifests itself in the Kelvin probe measurement as an apparent decrease/increase of the work function for positive/negative net charging of the surfaces,24 as observed here. The mirror charges present in the bulk silicon beneath the native silicon oxide layer compensate the surface charges absorbed from air (Fig. 3A) by delocalizing the net charge to the bottom of the sample (insulating oxide layer in contact with the base plate).
The abrupt work function decrease observed in both samples when d = 0 μm indicates positive surface charging that originates from the emission of photoelectrons (Fig. 3B). Although the work function decrease is slightly more pronounced at the beam location, it actually extends over the entire probed area (1 mm), well beyond the beam footprint lateral width of ∼20 μm, as observed also in silicon-supported lipid multilayers.39 This demonstrates that local photoemission alone is not sufficient to explain the observed charging phenomena. In addition, Compton scattering in the bulk substrate can enlarge the lateral width of the X-ray dose profile delivered by the direct micro-beam, but it cannot be invoked as the main origin for the macroscopically extended (∼1 mm) surface charging detected by the Kelvin probe. Finally, after beam switching off, the work function at the beam position is always slightly lower than anywhere else. This indicates permanent residual positive charging of the surface area that is directly irradiated by the X-ray beam.
While it is not surprising that intense high energy X-ray irradiation can cause permanent positive charging at the beam footprint, it is not immediately apparent why surface charging induced by ionized air molecules is almost reversible/irreversible for acidic/alkaline pre-treatments. We attribute this phenomenon to differences in surface charge mobility. As known from the ion transport mechanism at insulator/metal interfaces,24 mobile surface charges move together with mirror charges present in the bulk semiconductor beneath the native oxide layer. While the majority charge carriers of p-type silicon (holes) facilitate the discharging of the negative charges absorbed from air on the acidically treated surface, the dissipation of positive mobile ionized air charges absorbed on the alkaline treated surface requires negative mirror charges, which are lacking in p-type Si. In order to confirm this hypothesis, the same X-ray beam exposure protocol was applied to an acidically treated n-type Si wafer, where the majority charge carriers (electrons) cannot contribute to the dissipation of the mobile negative surface charges absorbed from air. Indeed Fig. 4 demonstrates that while the charging behavior is similar, the discharging of n-type Si after beam switching off (Fig. 4B, t ∼ 3200 s) is much slower than the instantaneous discharging exhibited by p-type Si (Fig. 4A, t ∼ 3400 s).
Our investigations of X-ray-induced surface charging phenomena were completed by applying the same experimental protocol to other samples belonging to three different groups in terms of conductivity: a conductor (NiAl alloy), a semiconductor (p-type silicon) with improved grounding ensured by an evaporated gold layer at its bottom, and an insulator (sapphire). The corresponding work function traces at beam (red) and off-beam (black) positions for different beam depths d are presented in Fig. 5. For the NiAl alloy (Fig. 5A), no significant response is detected as charges quickly dissipate or are compensated by corresponding counter-charges. Small effects are however still discernible, thus confirming the high sensitivity of our approach. The gold-coated p-type silicon as well exhibits only minor variations (Fig. 5B), indicating the role of the improved grounding for enhanced charge dissipation. However, after beam switching-off, the photoemission-induced permanent positive charging at the beam footprint is more evident than in the un-grounded case (Fig. 1C), due to the improved dissipation (by grounding) of the mobile negative charges absorbed from air. In the insulator case represented by sapphire (Fig. 5C), the acidic treatment (pH ∼ 8 for sapphire IP) induces a positive work function shift at beam switching-on as observed for Si (Fig. 1B and 4). However, two differences with respect to Si samples are noticeable when the beam reaches the surface of sapphire: the work function decrease is stronger (∼700 meV) at the beam footprint (x = 500 μm) and weaker (∼200 meV) at the off-beam position (x = 950 μm). Fig. 6B reports the resulting deeper and more localized work function trenches measured for sapphire during direct surface irradiation, as compared with the corresponding data for p-type Si (Fig. 6A and C). The broader Si profiles are due to the higher mobility of the negative ionized air charges, attracted to the surface by photoemission-induced positive surface charging, and mirrored by mobile holes in the bulk Si substrates. Contrary to the silicon case, at beam switching-off (t ∼ 3200 s in Fig. 5C) no changes are observed in the sapphire work function: the negative charging of the surface is completely retained since there are no mobile positive charge carriers in the bulk sapphire insulator.
Footnotes |
† Present address: Stryker Trauma GmbH, Prof.-Küntscher-Strasse 1-5, 24232 Schönkirchen, Germany. |
‡ Present address: Tesat-Spacecom GmbH & Co. KG, Gerberstrasse 49, 71522 Backnang, Germany. |
§ Present address: BASF Coatings GmbH, Glasuritstrasse 1, 48165 Münster, Germany. |
This journal is © the Owner Societies 2014 |