HAL Id: jpa-00229982
https://hal.archives-ouvertes.fr/jpa-00229982
Submitted on 1 Jan 1989
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
DEVELOPMENT OF THE FIELD ION-SCANNING TUNNELING MICROSCOPE AND ITS
APPLICATIONS
I. Kamiya, T. Sakurai
To cite this version:
I. Kamiya, T. Sakurai. DEVELOPMENT OF THE FIELD ION-SCANNING TUNNELING MICRO-
SCOPE AND ITS APPLICATIONS. Journal de Physique Colloques, 1989, 50 (C8), pp.C8-491-C8-
495. �10.1051/jphyscol:1989884�. �jpa-00229982�
COLLOQUE DE PHYSIQUE
Colloque C8, Supplement au nail, Tome 50, novembre 1989
DEVELOPMENT OF THE FIELD ION-SCANNING TUNNELING MICROSCOPE AND ITS APPLICATIONS
I. KAMIYA and T. SAKURAI
Institute for Solid State Physics, The University of Tokyo, 7-22-1 Roppongi, Minato-ku, Tokyo 106, Japan
Abstract
A scanning tunneling microscope (STM) equipped with a field ion microscope (FIM) which operates at room
temperature has been constructed and operated successfully to obtain atomically resolved STM images with 100%
reproducibility. This instrument, which we call the FI-STM, has been employed for the quantitative study of the Si(100)
"2xn" phase.
1. Introduction
The importance of the role of the shape of the tip playing in the
resolution pf STM has been well recognized and studied theoretically. Tersoff and Hamann, for example, treated the problem by a simple model where the electrons tunnel between a planar metallic surface and a spherical tip. And they came up with the conclusion, "the sharper the tip, the better the
resolution." This result seemed to agree with our intuitive understanding, and thus it has been widely accepted as a standard theory. However, there have been very few experimental works actually performed to prove this theory. Kuk and silverman2 were the first to carry it out using the FIM.3c4 By
scanning the same surface with tips of different shapes, they confirmed the relation between the tip size and the observed resolution obtained by Tersoff and Hamann. FIM is the only technique available to observe the tip geometry on an atomic scale, and is the most suited for this study owing to its capability of manipulating the shape of the tip by field evaporation. Their apparatus, however, had some difficulties in the actual operation. One of the most critical was the thermal drift. Thus we have designed our apparatus to overcome these problems.
2. Instrumentation
Our UHV STM unit is a lever type, where the sample to tip approachment is accomplished by pushing the edge of the lever by a micrometer which can be manipulated externally.(Fig.l) As soon as the sample hits the stopper, the pivot point would switch for fine approach, and finally an electrostriction element which has a maximum displacement of 1 2 p enables the tip to reach the sample vicinity where the tunneling current would flow between the tip and the sample. This unit sits on five stage viton damper and is suspended by three 40cm long springs for vibration isolation. The details of our instrument are described elsewhere.5
The FIM setup consists of a negative electrode and a chevron channel plate-screen assembly with which fine FI images could be obtained even at room temperature without applying high voltage to the tip. This results in a great advantage to the actual operation of the instrument since the unwanted thermal drift and breakdown of the piezo elements could be avoided. With this FIM setup, the geometry of the probe tip could be monitored and modified, in-situ.
Owing to this tip characterization procedure, where we can prepare clean and locally sharp W<111> oriented tips, we have so far been successful in obtaining atomically resolved STM images of graphite(0001), Si(ll1)-7x7 and Si(100)-2x1 without failure. To our knowledge, this performance of being able to observe atomic resolution STM images with 100% reproducibility and reliability, has never been achieved by anyone else.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1989884
The whole instrument is housed in a ultra-high vacuum chamber mounted on an air suspended vibration isolator. The STM unit is fixed on the 8" O.D. flange located in the left bottom, and the FIM unit is fixed on the 10" O.D. flange located fn the left edge, and so are the sample preparation media, such as the heating electrodes, alkali dispenser and/or.the hydrogen effusion source.
Evacuation of the imaging gas used for FIM, and from atmospheric pressure, is carried out by a magnetic levitated turbo molecular pump. An during the STM experiment, the base pressure of the system is kept to 4x10- 19 T o m by ion pump and Ti sublimation pump, which is necessary for the studies of
semiconductor surfaces.(Fig.2) 3. Experimental
In the present work, we have employed this FI-STM for the studies of the clean and modified Si(100) surfaces. Following a thorough degassing f both the sample and the sample holder, our Si(100) sample, cut into 4x18nunq from commercially available As-doped n-type silicon wafer(R=O.l ohm-cm), was flashed up to 1100'~ for several times, and showed the 2x1 structure due to the
formation of dimers. The annealing of the sample was performed by resistive heating.
4. Results and Discussion
After this appropriate preparation procedure, the defect density on the clean Si(100)-2x1 surface was less than a few percent, which is far less than previous reports.6 In the images, the dimer rows are observed as stripes in the filled state and thebifurcatedstructure of the dimers are visible in the empty state. (Fig.3)
The clean Si(100)-2x1 surface has been known to transform into "2x11"
structure upon high temperature annealing, and this structure has been studied intensively using LEED (Low Energy Electron Diffraction) 7-1° and ABD(At0mic Beam
if fraction)^^.
However, the details of the structure have still not been elucidated. First of all, the periodicity "n" has been reported to be a variable. Mu ller et a1.7 and Aruga and ~ u r a t a ~ reported "n"to be
91
Martin et al.8 and Kato et a1.10 reported 6<n<10 and Rohlfing et al.
reported n=8,10, although they all agree that the structure is due to missing dimer defects. And this ructure seems to differ from the ordered defect model proposed by Panday.?$ Some of the factors that determine this
structure that have been discussed are the existence of Ni impurities and the quenching rate. Kato et a1.10 attempted a thorough investigation concerned with these factors. By measuring the AES(Auger Electron Spectroscopy) intensity Ni LMM(847eV) and Si LW(92eV), they correlated thz periodicity n versus the surface Ni concentration, and concluded that the 2xn" phase is formed when the Ni concentration is above 0.35%, and the value of n decreases from10 to 6 as the Ni concentration increases from 0.35% to 1%. And they claimed that the quenching rate and the peak temperature determine the degree of the surface segregation of Ni. However, ~ u e l l e r ~ and ~ r u ~ a ~ claim that their samples were free of Ni, and below 0.03% with Martin's casee9 This is still under conflict, as Kato et al. believe that this is only a matter of the sensitivity of the AES, especially as based on their LEED result, pairs of Ni atoms are located at the interstitial sites of Si crystal near the missing dimer defects, but not substituting the missing dimer atoms themselves. Aruga and Murata do not argue with the Ni induced reconstruction model either, as the the slight difference in the lattice constants between Si(5.431A) and
NiSi2(5.406A) might be large enough to cause the lattice distortion and
change the surface structure. But in any case, they attribute the annealing as the activation process to transfer the surface Si atoms, quenching as the suppresion of evaporation and the surface strain energy as the determining factor for the periodicity n.8
Another intriguing subje t in this phase was the structure of the defect itself. While Martin etlyf
- 5
concluded that they are metastable single d i m e defects, Rohlfing et al. based on their ABD result found that the defect width is 2.6ao where a, is the surface lattice constant 3.84A, with adepth of Si interplanar spacing. This indicates the pairing of missing dimers, and they claim that the single missing dimer is metastable being stabilized by impurities.
These kind of arguments are never settled without performing a precise quantitative study on an atomic scale. STM, owing to its power to provide us
with a three dimensional information in atomi~~scale, has recently been employed to study this system. Niehus et al. were the first to carry it out. By adding Ni impurities of 5% or less on purpose prior to high
temperature annealing, they were able to observe the "2x8" reconstruction due to missing dimer defects. They have found that both single and double dimer defects were present, and even a structure which they called 'split off dimer' was observed.
Although we do not argue with the Ni induced "2xnW, it seemed that
observation of the Ni free "2xn" was also warranted. We have thus studied this surface with our instrument. The images in Fig.4 were taken after repetitive annealing upto 1200°C. We see the missing dimer defects forming semi-periodic array of defect cluster bands, even under the condition where Ni was not detected with AES. The resolution here is high enough for carrying out a precise quantitative analysis. First, we have Fourier Transformation of one of our ima and obtained a pattern showing the 2x8 structure similar to Niehus' result.qsS'This 2x8 pattern, however, was not obtained from all of our
images. We attributed this to the fact that the structure is not perfectly ordered but rather semi-periodic, and hence we measured the inter-band spacing D in units of 3.84 A for detailed analysis. Fig.5 is the histogram obtained.
The average spacing is 8.3 with a peak at D=8 and a sub-peak at D=ll. These seem to explain the discripancy between the previous LEED and ABD reports.
By studying our data carefully, for instance by taking the cross section, we observe the 'split off dimers', and also the single, double and even multiple missing dimers forming clusters. The variety of the detailed structures seem to be due to the slight difference in the experimental conditions or the nature of the sampl~ gmployed. Just by comparing the two LEED patterns both catagorized as 2x8 c in detail, it is easy to recognize that the two are exhibiting different structures in real space. Also,
recently, Hoeven et a1.14 reported that similar structure was formed upon MBE growth of Si on Si(100) where the sample must be totally free from
contamination. Strain in the substrate or subsurface may be responsible, but this need not be Ni contamination. In any case further investigations seem warranted to give a conclusive remark.
4. Conclusion
We have constructed the field ion-scanning tunneling microscope, and the performance where atomically resolved STM images are obtained with 100%
efficiency has been achieved. Using this instrument, the "2x11" phase of Si(100) induced by high temperature annealing has been studied. We have found that
1) the structure is made from missing dimer defects forming cluster bands,
2) these cluster bands run perpendicular to the dimer rows to form semi-periodic arrays of which the inter-band spacing D has a wide distribution with a peak at D=8 and a sub-peak at D=ll, which seem to explain the results of previous diffraction studies.
Acknowledgements
The authors thank Dr,T.Hashizume and Mr.Y.Hasegawa for their collaboration.
References
1. J-Tersoff and D.R.Hamam, Phys.Rev.Lett.a(1983)1998, Phys.Rev.B31(1985)805.
2. Y.Kuk and P-J-Silverman, Appl.Phys.Lett.g(1986)1597.
3. E.W.Mueller and T.T.Tsong, "Field Ion Microscopy, Principle and Applications," Elsevier Publishing House, New York, 1969.
4. T-Sakurai, A.Sakai and H.W.Pickering, Adv.Electronics and Electron Phys., Suppl.20, Academic Press, 1988, "Atom-Probe Field Ion Microscopy and Its Applications. "
5. T.Sakurai, T.Hashizume, I.Kamiya, Y.Hasegawa, T.Ide, M.Miyao, 1-Sumita, A.Sakai and S.Hyodo, J.Vac.Sci.Technol.A7(1989)1684.
6. R.J.Hamers, R.M.Tromp and J.E.Demuth, Phys.Rev.B34(1986)5343.
7. K.Mueller, E-Lang, L.Hammer, W-Grim, P-Heilman and K-Heinz, in
"Determination of Surface Structure by LEED", ed. P.M. Marcus and F-Jona, Plenum, New York, 1984, p.483.
8. T-Aruga and Y-Murata, Phys.Rev.a(1986)5654.
9. J.A.Martin, D.E.Savage, W.Moritz and M.G.Lagally, Phys.Rev.Lett.56(1986) 1936.
10. K.Kato, T-Ide, S-Miura, A-Tamura and T-fchinokawa, Surf-Sci.l~(1988)t87, 11. D.M.Rohlfing, J-Ellis, B.J.Hinch, W.Allison and R.F.Willis, Surf.Sci.2&7
(19891L955.
12. K.C.Pandey, in "Proceedings of the Seventeenth International Conference On the Physics of Semiconductors," ed. D.J.Chadi and W.A.Harrison, Springer-
~erlag, New York, 1985, p.55.
13. H.Niehus, U.K.Koehler, M.Copel and J.E.Demuth, J.Microscopy &2(1988)735.
14. A.J.Hoeven, J.M.Lenssinck, E.J.van Loenen, D.Dijkkamp and J-Dieleman, Presented at STM189, July 1989, Oarai, Japan.
1
o SCREEN CHANNELPLATEFig.1 Schematic of the operation of the FI-STM.
HORIZONTAL
MANIPULATOR OF SAIAPLE
Fig.2 Schematio of the FI-STM system.
v s = +1.2v vs = -1.2v
EMPTYSTATES FILLED STATES
Fig.3 STM image of clean Si(100)-2x1.
Fig. 4 Si(100)-"2xn" obtained after repetitive annealing at 1200°~.
Fig.5 Distribution of inter-band spacing.
