Ballistic Electron Emission Microscopy Measurements Research Articles

Time dependent changes in Schottky barrier mapping of the W/Si(001) interface utilizing ballistic electron emission microscopy

The W/Si(001) Schottky barrier height is mapped to nanoscale dimensions using ballistic electron emission microscopy (BEEM) over a period of 21 days to observe changes in the interface electrostatics. Initially, the average spectrum is fit to a Schottky barrier height of 0.71 eV, and the map is uniform with 98% of the spectra able to be fit. After 21 days, the average spectrum is fit to a Schottky barrier height of 0.62 eV, and the spatial map changes dramatically with only 27% of the spectra able to be fit. Transmission electron microscopy shows the formation of an ultra-thin tungsten silicide at the interface, which increases in thickness over the 21 days. This increase is attributed to an increase in electron scattering and the changes are observed in the BEEM measurements. Interestingly, little to no change is observed in the I-V measurements throughout the 21 day period.

Direct Measurements of Lateral Variations of Schottky Barrier Height Across “End-On” Metal Contacts to Vertical Si Nanowires by Ballistic Electron Emission Microscopy

Ballistic electron emission microscopy measurements on individual "end-on" Au Schottky contacts to vertical Si nanowires (NWs) indicate that the local Schottky barrier height at the contact edge is 23 ± 3 meV lower than at the contact center. Finite-element electrostatic simulations suggest that this is due to a larger interface electric field at the contact edge resulting from (equilibrium) positive charge in Si/SiO(2) interface states near the Au/NW contact, induced by local band bending due to the high work function Au film.

Combined molecular beam epitaxy low temperature scanning tunneling microscopy system: Enabling atomic scale characterization of semiconductor surfaces and interfaces

A molecular beam epitaxy and low temperature scanning tunneling microscopy chamber have been integrated to characterize both compound and elemental semiconductor surfaces and interfaces. The integration of these two commercially available systems has been achieved using a custom designed sample transfer mechanism. The MBE growth chamber is equipped with electron diffraction and provides substrate temperature measurements and control by means of band-edge thermometry accurate to within ±0.5°C. In addition, the microscope can operate at temperatures as low as 4K and perform ballistic electron emission microscopy measurements. The chamber that houses the microscope includes a preparation chamber with an evaporation source for metals. The entire STM chamber also rests on an active vibration isolation table, while still maintaining an all ultrahigh vacuum connection to the MBE system.

Observation of a scanning tunneling microscopy induced photocurrent during ballistic electron emission microscopy

We report that a scanning tunneling microscope induced photocurrent (STM–PC) can occur during ballistic electron emission microscopy (BEEM) measurements on Schottky-barrier samples with typical efficiency of 10−4–10−6 and apparent BEEM threshold slightly above the substrate semiconductor band gap. This STM–PC resembles normal BEEM current, and hence, can interfere with BEEM measurements of structures with intrinsically low signal and/or intrinsic BEEM threshold voltage larger than the substrate band gap. We discuss a simple test for the existence of STM–PC during BEEM measurements.

Calculated potential profile near charged threading dislocations at metal/semiconductor interfaces

We have made finite element calculations of the expected potential profile around negatively charged threading dislocations (TDs) close to a metal–semiconductor interface, using a Pt contact on n-type GaN as a specific case. The potential was calculated as a function of the assumed linear density and energy level of TD-related acceptors. Our model shows good agreement with the model of Read [W. T. Read, Philos. Mag. 45, 775 (1954); 46, 111 (1954)] for an infinite dislocation, far from any interface. Assuming 1 acceptor/c-axis lattice spacing (c=0.52 nm), we found for our near-surface modeling that acceptors levels deeper than 1.3 eV below the conduction band minimum (CBM) should be charged all the way to the Pt/GaN interface. This should produce a significant local increase in the potential barrier and at the Pt/GaN interface and should be observable by ballistic electron emission microscopy (BEEM). In fact recent BEEM measurements by Im et al. on molecular beam epitaxy-grown GaN films [Phys. Rev. Lett. 87, 106802 (2001)] showed no significant increased barrier at TDs indicating an acceptor energy within 1.3 eV of CBM and/or an acceptor density less then 1 acceptor/c-axis lattice spacing.

Diamond tips in low temperature scanning tunneling microscopy

Scanning tunneling microscopy and ballistic electron emission microscopy measurements are presented using boron-doped diamond as tip material. The experiments were carried out at 77 K on CoSi 2/Si(1 1 1), a system previously investigated by conventional tungsten tips. Compared to tungsten tips, the measurement stability is improved, while atomic resolution is still easy to obtain in topography scans. A threshold shift in ballistic electron emission spectra is found to depend on the magnitude of the tunneling current. We explain this shift in terms of an ohmic series resistance in the tunneling circuit due to poor tip contacting or too low a carrier concentration in the tip material at low temperatures.

Interfacial scattering of hot electrons in ultrathin Au/Co films

We have used room-temperature, ballistic electron emission microscopy (BEEM) to measure hot-electron transport through ultrathin Au/Co multilayer structures deposited onto Si. The samples consist of Au/Co/Si or (Au/Co)n/Au/Si diodes, sputter deposited at 175 or 300 K, where n is the number of repeat layers. The thin-film Co attenuation length, λCo, is extracted from the BEEM spectra as a function of Co thickness, in single Co layer samples. Similarly, the interface attenuation number, or the number of Co/Au interfaces required for a 1/e attenuation, is determined from the multi-interface samples. BEEM barrier heights of Au/Co/Si decrease with increasing Co thickness (for thicknesses <1 nm), as the film becomes continuous and develops a Schottky barrier for Co or CoSi2(<0.7 eV). For these diodes, λCo, increases from 0.3 to 0.5 nm, each with an estimated uncertainty of 0.1 nm, when the deposition temperature is decreased from 300 to 175K. This result is associated with decreased silicide formation at the lower deposition temperature. When Co is isolated from the Si with a 2 nm Au layer, the barrier height is stable (0.82 eV), and λCo, increases further to 0.8±0.1 nm. The λCo values are independent of electron energy over the range measurable (1–1.8 eV). The interface attenuation number was 1.8 and 25 interfaces for 0.6 and 1.2 nm total Co thickness, respectively (in structures with a constant total Au thickness of 8 nm). Thus, significant interface scattering is observed in the thinner sample and negligible scattering when the Co thickness was doubled. Further BEEM measurements on similar superlattices are being carried out to understand these results.

Modification of GaN Schottky barrier interfaces probed by ballistic-electron-emission microscopy and spectroscopy

Ballistic-electron-emission microscopy (BEEM) and spectroscopy have been used to investigate the properties of Au/GaN interfaces. The effects of in situ and ex situ annealing on the starting GaN surface were examined, with the aim of increasing the surprisingly low value of interface electron transmission observed in previous BEEM measurements. BEEM imaging and spectroscopy have demonstrated that much higher, more uniform transmission across the Au/GaN interface can be achieved. However, while methods were identified that increase transmission by more than an order of magnitude, BEEM spectroscopy indicates that annealing can substantially alter the Schottky barrier height. These barrier height changes at moderate temperatures are attributed to vacancy diffusion.

Investigation of ultrathin SiO2 film thickness variations by ballistic electron emission microscopy

We investigate the feasibility of using ballistic electron emission microscopy (BEEM) to study possible thickness variations of ultrathin SiO2, which might exist at substrate defects, such as steps. We find that simple BEEM imaging of the oxide film sandwiched into a metal–oxide–semiconductor (MOS) structure does not reveal any features that could be related to the oxide film. We further present initial results suggesting that hot-electron resonance in the oxide conduction band could be observed by BEEM and could be sensitive to local film thickness. We also confirm the ability of oxide film to sustain injection of very high densities of hot electrons without any observable damage. In some cases we observe local damage of the MOS structure induced by BEEM measurements, but we conclude that it is most likely related to failure of the metal overlayer and probably not related to hot-electron breakdown of the oxide.

Metal/GaN Schottky barriers characterized by ballistic-electron-emission microscopy and spectroscopy

Ballistic-electron-emission microscopy (BEEM) and spectroscopy have been used to characterize the Pd/GaN and Au/GaN interfaces. BEEM spectra yield a Schottky barrier height for Au/GaN of ∼1.05 eV that agrees well with the highest values measured by conventional methods. For both Pd and Au, a second threshold is observed in the spectra at about 0.2–0.3 V above the first threshold. Imaging of these metal/GaN interfaces reveals transmission in nearly all areas, although the magnitude is small and spatially varies. Attempts to perform BEEM measurements on other GaN material have resulted in no detectable transmission in any areas, even at voltages as high as 3.5 V.

Ballistic-electron-emission microscopy and spectroscopy of metal/GaN interfaces

Ballistic-electron-emission microscopy (BEEM) spectroscopy and imaging have been applied to the Au/GaN interface. In contrast to previous BEEM measurements, spectra yield a Schottky barrier height of 1.04 eV that agrees well with the highest values measured by conventional methods. A second threshold is observed in the spectra at about 0.2 V above the first threshold. Imaging of the Au/GaN interface reveals transmission in nearly all areas, although the magnitude is small and varies by an order of magnitude. BEEM of other GaN material shows no transmission in any areas.

Ballistic electron emission microscopy measurements of epitaxially grown Pt/CaF2/Si(111) structures

The electronic transport properties and surface morphology of ultrathin Pt/CaF2/Si(111) metal-insulator-semiconductor structures have been characterized in-situ by ballistic electron emission microscopy (BEEM) and scanning tunneling microscopy (STM). Platinum thicknesses from 2–53Åand CaF2 thicknesses of 5 and 15Åhave been characterized. The STM images of the Pt/CaF2/Si(111) structures show the atomic steps of the underlying CaF2 morphology, as well as the formation of Pt nodules. The BEEM spectra from the thicker Pt coverages show an onset near the conduction band minimum (CBM) of the CaF2. The BEEM spectra from the thinner Pt coverages show an additional peak near 2 eV, which has not been observed in previous studies of metal/CaF2/Si(111) structures. This extra peak has been attributed to quantum mechanical resonant tunneling through the ultrathin MIS structure.

Design of a scanning tunneling microscope for in situ topographic and spectroscopic measurements within a commercial molecular beam epitaxy machine

A scanning tunneling microscope that performs scanning tunneling microscopy, scanning tunneling spectroscopy, and ballistic electron emission microscopy measurements on 2-in. wafers has been designed and constructed. The instrument is incorporated into the Si preparation chamber of a cryo-pumped Fisons V90H Si/III-V molecular beam epitaxy machine. Its design uses two commercial Burleigh inchworms: one for performing scanning tunneling microscopy measurements and a second for making a front contact which is necessary for the ballistic electron emission microscopy measurements. The substrate holder for the V90H system is designed to handle wafers up to 6 in. in diameter. Therefore, a custom 6-in.-diam. holder has been constructed which supports two 2-in. holders: one for performing reflection high energy electron diffraction measurements and a second which allows transfer of the wafer to the scanning tunneling microscope and also incorporates a removable shadow mask for growing metal/semiconductor Schottky diodes. Although the chamber vibrations generated by the cryopump and its compressor are quite severe, atomic-resolution images have been obtained with all of the system’s pumps in operation. The enhanced stability of our design is attributed to our unique support system of the inchworm and wafer which are both manufactured from machinable ceramic, the use of Be–Cu springs with an extension length of 16 in. for vibration isolation, and the use of silicone O-rings for vibration damping.

Comparison of Au contacts to Si, GaAs, In xGa 1 − xP, and ZnSe measured by ballistic electron emission microscopy

Ballistic electron emission microscopy (BEEM) measurements have been made on Au contacts to Si, GaAs, In x Ga 1 − x P, and ZnSe. For each semiconductor studied, a distribution of barrier heights (BHs) was observed. The distributions were compared with the current—voltage ( I/V) and capacitance—voltage ( C/V) characteristics of each metal/semiconductor contact. The BH distributions increase in width with increasing doping and/or complexity of the substrate. The effect of discrete, ionized dopants in the depletion region is considered as one source of BH distribution widening. Average BEEM measurements agree well with the BH computed from I/V and C/V measurements for all diodes if the effects of thermionic field emission and image force lowering are considered. These results demonstrate the unique application of BEEM for the nanoscale evaluation of semiconductor interface inhomogeneities critical to the control of macroscopic electronic properties.

Au/ZnSe contacts characterized by ballistic electron emission microscopy

Ballistic electron emission microscopy (BEEM) has been performed on Au/ZnSe (001) diodes prepared in ultra high vacuum. An average barrier height (BH) of 1.37 eV is found for Au/n-ZnSe in close agreement with previously published values for diodes measured by conventional techniques. The BH distribution is relatively narrow, from 1.32 to 1.43 eV, consistent with cross-sectional transmission electron microscopy which indicates that the interface is abrupt, and without reaction products. These results differ from those reported for BEEM measurements on chemically etched Au/ZnSe diodes. [R. Coratger et al., Phys. Rev. B. 51, 2357 (1995)].

Nanometer-resolved spatial variations in the Schottky barrier height of a Au/n-type GaAs diode.

Nanometer-resolved lateral variations in the Schottky barrier height (SBH) formed at a chemically prepared Au/n-type GaAs interface were measured using ballistic-electron-emission microscopy (BEEM). The spatial profile and the statistical distribution of the SBH's thus obtained were compared to current-voltage (IV) and capacitance-voltage (CV) characteristics of the same metal-semiconductor contact. This comparison showed that the macroscopic SBH obtained from the IV measurements can be successfully interpreted using the parallel conduction model applied to the BEEM-derived distribution, if the effect of thermionic field emission is included. The SBH obtained from the CV measurements is greater than the mean value obtained from BEEM measurements by nearly the image-force lowering expected for a Au/GaAs diode.

Gold–silicon interface modification studies

We have performed ballistic electron emission microscopy measurements on the Au–Si system with and without controlled monolayer impurities at the interface. At moderate sample to scanning tunneling microscopy tip biases (<2.5 V) we have observed, and at high biases (>3 V) modified the local ballistic transmittance (BT), i.e. the scaling factor of the collector current versus voltage spectra, of the interface. Spatially, the modification typically consists of a region of decreased BT a few hundred Å in diameter surrounded by a ring of increased BT. No change in Schottky barrier height is found. A model is presented which describes the decrease in terms of Au–Si interdiffusion, and the enhancement in terms of a thinning of an impurity layer between the Au and Si; connections are made to observations of the unstressed system.

Ballistic electron emission microscopy studies of the NiSi2/Si(111) interface

We have performed ballistic electron emission microscopy measurements on the NiSi2/Si(111) system under UHV conditions. Schottky barrier heights have been measured for types A and B interfaces and are found to be in agreement with other techniques. Through correlations between structural defects in the silicide and local increases in collector currents, we find that elastic scattering in the film can increase ballistic electron transmission rates across the interface.