KINETICS OF FIELD EVAPORATION OF GALLIUM PHOSPHIDE IN THE PRESENCE OF HYDROGEN

HAL Id: jpa-00229923

https://hal.archives-ouvertes.fr/jpa-00229923

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

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.

KINETICS OF FIELD EVAPORATION OF GALLIUM PHOSPHIDE IN THE PRESENCE OF HYDROGEN

A. Gaussmann, W. Drachsel, J. Block

To cite this version:

A. Gaussmann, W. Drachsel, J. Block. KINETICS OF FIELD EVAPORATION OF GALLIUM

PHOSPHIDE IN THE PRESENCE OF HYDROGEN. Journal de Physique Colloques, 1989, 50 (C8),

pp.C8-141-C8-146. �10.1051/jphyscol:1989825�. �jpa-00229923�

COLLOQUE DE PHYSIQUE

Colloque C8, supplbment au n o 11, Tome 50, novembre 1989

KINETICS O F F I E L D EVAPORATION O F GALLIUM PHOSPHIDE I N THE PRESENCE O F HYDROGEN

A. GAUSSMANN' , W. DRACHSEL and J. H. BLOCK

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, 0-1000 Berlin 33, F.R.G.

Abstract - Field evaporation of gallium phosphide and field desorption of hydrides from different gallium phosphide surfaces were studied in the temperature range of 26 to 340K.

Field evaporation rates spanning up to 3 orders of magnitude were measured as a function of temperature and at different partial pressures of hydrogen. The temperature dependence of the evaporation rate showed a pronounced minimum at 200K. Below 200K the evaporation rate was found to be proportional to the hydrogen gas pressure. The evaporation rate was proportional to the hydrogen coverage. The temperature dependence of the evaporation rate was fitted by Langmuir- and Volmer-isobars. From these fits, a binding energy of 45 meV was evaluated. Above 200K the activation energy for field evaporation was determined to be 662meV. Two models describe the occurrence of different activation energies. In the first case field desorption of gallium phosphide hydrides are influenced by an increased hydrogen supply by surface diffusion of hydrogen from the shank to the apex of the tip or by impingement from the gas phase. Conversely, the second model neglects the influence of hydrogen and regards the onset of field evaporation of gallium phosphide, as usually observed for metals and semiconductors.

1. Introduction

Field evaporation and field desorption are thermally activated processes.

These processes are visible in a field ion microscope and can be used to study chemical reactions on solid surfaces. The interaction of hydrogen with unsaturated dangling bonds at semiconductor surfaces is of great interest.

FIM studies of silicon and gallium phosphide have shown a strong reduction of the evaporation field strength due to the presence of hydrogen2y3. The rate determining step in the evaporation process was found to be associated with surface hydride formation as found by Sakata and ~ l o c k ~ l ~ . In this study, field-induced dissociation of field-adsorbed molecular hydrogen at the semiconductor surface was assumed to lead to the observed hydride formation. Thus, field desorption of hydrides would be responsible for the observed reduction in field strength. This mechanism was experimentally confirmed by Kellogg, using a pulsed-laser atom-probe for studying silicon6.

In the case of the compound semiconductor gallium ptiosphide, the temperature dependence of field evaporation showed two different regimes3.

Below room temperature the rate of removal of layers was found to decrease with increasing temperature, while at higher temperatures the evaporation rate increased as the temperature increased. The evaporation rate was proportional to the hydrogen pressure in the investigated temperature range of 80 to 330K.

( 1 )present address:

Eidgenlissische Technische Hochschule, Universitatsstrasse 6, 8092 Zurich, Switzerland Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1989825

The aim of this investigation was to characterize the low temperature behavior of the field evaporation process. The accessible range of tip temperatures was extended down to 26K. Influence and nature of the hydrogen supply were studied in detail. In particular, hydrogen may be supplied from the gas phase, or alternatively, via diffusion from the shank to the tip of the emitter. Accordingly, experimental data were fitted by two different kinetic models. The evaporation rates from different surfaces with variable concentrations of gallium and phosphorus were measured and information about surface reactivities could be obtained.

2. Experimental

This investigation was performed in a conventional ultra-high vacuum field ion microscope. The apparatus is equipped with a closed cycle helium refrigerator so that tip temperatures can be adjusted within 26 to 500K.

Gallium phosphide samples were prepared from a n-doped wafer with a sulfur concentration of approx. 1017 ~ r n - ~ . The specific resistivity of the water was about l o - * Ohm.cm. In a first step of tip preparation the wafer was cut in thin (110)-oriented rods. The thin rods were fixed by conductive epoxy glue to a platinum support loop which was used for sample heating. Chemical etching was performed at approx. 350K in a freshly prepared so!ution of 1:3 H N 03:HCI. After the sample was introduced into the chamber and UHV-conditions were established, it was further cleaned by field-induced chemical etching in hydrogen gas at 30K 2.

The field evaporation rate was determined by measuring the time of removal of subsequent surface planes. This technique was described elsewhere in detai14s5. To maintain a constant local field strength, the applied voltage was corrected, because of the increase of the tip radius during continuous field evaporation (see figure 1 .).

.06.

.05-

.

. 0 4 - - 0

-

Y

.03.

.02-

x .

. O l ,

.

.

.

.

.

.

0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 4 5 5 0 5 5 6 0 6 5 7 0 number of evaporated layers

F~gure 1. Callbration curve for the evaporation rate k (layers per second) of the gallium phosphide (1 1 1 ) plane. Temperature T=28K, gas pressure ~ = 4 x 1 0 - ~ hPa, Voltages U=8.03 kV and U=8.25 kV. To maintain a constant evaporation rate k one must increase the voltage U for each evaporated surface plane by 6.5 V.

3. Results and Discussion

T h e reduction of field evaporation voltage U at constant rate k could be observed down to 26K. Below 80K the field evaporation decreased more sharply. As shown in figure 2., a change in temperature of 60 degree from 86 to 26K corresponds to a reduction in field evaporation voltage of 2.0kV or 15%.

Figure 2. Temperature dependence of the evaporation voltage at constant k of the gallium phosphide (111)-surface in the temperature range of 28 to 150K. Hydrogen pressure Pt4x10-5h~a, evaporation rate k=0.1 planeslsec.

The evaporation rate in the temperature range of 30 to lOOK was determined for several surface planes using two different hydrogen pressures. Figure 3.

shows the results.

A I n k ( 1 1 1 )

I

I n k ( 1 l l ) P = l x l 0 - 2 p a H2

v

In k (001) In k (001)

- 6

Figure 3. Temperature dependence of the evaporation rate k for different gallium phosphide surface planes in the temperature range from 30 to 100K. The voltage was kept constant at U=14.00 kV.

-

-

-

Ty,,

^.:

^{. :}

1000 J T [K]

No difference in k could be discerned for the close-packed gallium (1 11) and the phosphorus (1 1 T) planes. However, the less densely packed (001) plane, with a concentration of 50% phosphorus atoms and 50% gallium atoms, showed an evaporation rate lowered by a factor of 0.7 as compared to both other planes. This factor reflects exactly the difference in the number of surface sites per unit cell of the (111) and the (001) planes. The evaporation rate is proportional to the density of surface sites. If the number of atoms evaporating per second would be constant, a reverse relation would be expected. Obviously, the rate k is proportional to the number of surface sites which can adsorb hydrogen and then the hydride formation on (111) planes is 0.7 times slower than on (001) planes.

Furthermore, a reduction of the gas pressure from 1.0x10-~ hPa hydrogen to 1 .Ox1 o - ~ hPa hydrogen decreases the evaporation rate k by a factor of 10.

Under the experimental conditions chosen ( at temperatures below 200K and at hydrogen pressures in the 10-5 hPa range), k is proportional to the hydrogen pressure. The behavior of the evaporation rate k is shown in figure 4.

for the (111) and the (001) planes at constant field strength and at temperatures between 32 and 336K.

In the temperature ranges from 30 to 120K and from 270 to 340K the hydrogen gas pressure was p=5x1W6 hPa, while- in the range from 120 to 270K, measurements at a 4 times higher pressure ~ = 2 x 1 0 - ~ hPa were performed. This procedure was found necessafy because of the dim images at these temperatures. By using the proportionality between the evaporation rate and the hydrogen pressure, the image intensity could be amplified. Up to temperatures of 200K, a decrease of the evaporation rate k was found, while at higher temperatures, a strong increase of k was observed. The well developed minimum of k at constant field strength was determined to 200+30K for different tip radii.

Figure 4. Temperature dependence of the evaporation rate k of the gallium phosphide (1 11) and (001)surfaces in the temperature range of 3 2 to 336K. k is plotted relative to the constant hydrogen pressure P=2x10-~ hPa. The field strength was kept constant (relative to U=10.57 kV by adding a correction of 6.5 V for each evaporated plane).

Under the assumption that the evaporation rate k is proportional to the relative hydrogen coverage ⁸ of the specimen, experimental data can be fitted by two different kinetic models, i.e. the Langmuir isobar and the Volmer i ~ o b a r ~ ~ ~ . Langmuir's model assumes identical adsorption sites and no interaction between adsorbed particles. Volmer's model includes surface diffusion of adsorbed atoms or molecules. The constant value of the evaporation rate k at low temperatures was chosen as unity for the relative hydrogen coverage of the (1 11) plane. The optimized Volmer-isobar for k results in an adsorption energy of H=45+5 meV, the lower Langmuir value is less plausible due to the bad fit. Using a similar method, Sakata and Block found a value of about 33 meV for the adsorption energy for molecular hydrogen on a silicon (1 11) plane4.

The Volmer model gives a closer fit to the data at low temperatures or high hydrogen coverage. Diffusion of hydrogen seems to be evident. This effect was also found on metal surfacesg. Unless the surface temperature is too low, hydrogen impinging on a surface plane would migrate to its peripher , where hydride formation is assumed to take place above kink site atoms3v4*

!i! .

10 1 5 2 0 2 5 3 0 3 5 40

1000 I T [K]

Figure 5. Langmuir- and Volmer-plot of the evaporation rate k of the gallium phosphide (111) surface in the temperature range of 30 to 80K. k is plotted relative to the constant hydrogen pressure P=4x1 o - ~ hPa. The field strength was kept constant (relative to U=12.15 kV by adding a correction of 6.5 V for each evaporated plane).

Below 200K, the influence of the hydrogen pressure and the difference in hydride formation on the (1 11) and (001) surface planes can be interpreted on the basis of an adsorption/desorption equilibrium of molecular hydrogen on the emitter. The adsortion time T (determined by the Frenkel equation using the energy H=45 meV and ~ 0 = 1 0 - l ~ sec) is so short that the coverage (even at 30K) is far below the monolayer limit. The proportionality between the evaporation rate k and the hydrogen pressure, as well as the difference in k between the close-packed (1 11) and the less densely packed (001) planes, are expected, because the impingement rate j of hydrogen is directly proportional to the gas pressure. The higher the gas supply, the more the hydrogen can adsorb and the faster the rate k is.

For T>200K, the evaporation rate k increases drastically with temperature.

The Arrhenius-plots in the temperature range of 200 to 360K were measured three times and each plot resulted in a activation energy Q=662f240 meV. The occurrence of different activation energies may be explained by a reduced influence of hydrogen on the process, because for T>200K the surface concentration of field adsorbed hydrogen is very small. In this way, the gallium phosphide field evaporates as atoms or clusters. On the other hand, Yamamoto et al. detected phosphorus hydrides also at room temperature and observed a layer by layer evaporation of gallium and phosphorus from the (1 11) planes1 O. In particular, gallium evaporated much faster than phosphorus, which can be understood by the difference in the ionisation potentials V~,=6.06 eV and Vp =10.57 evl l, respectively. On the basis of these findings, one may interpret Q=662 meV as the activation energy for the field evaporation of phosphorus.

As compared to the binding energy of 45 meV for molecular hydrogen on a gallium phosphide (111) plane, field evaporation of phophorus needs a nearly fifteen-times higher activation energy. However, this finding depends strongly on the field strength and the emitter material under study. Kellogg, for example, found an activation energy of 260 meV for field evaporation of tungsten12. Small amounts of hydrogen in the vacuum system reduced the activation energy to only 60 meV or 23% of its former value. The question as to whether hydrogen influences the observed increase in the evaporation rate should stimulate further investigations on this subject.

Acknowledgements

The authors wish to acknowledge the helpful discussion with Dr. N. Ernst and Dr. W.A. Schmidt. Gratefully, this work was supported by the Deutsche Forschungsgemeinschaft Bad Godesberg (SFB 6/81).

References

1 E.W. ^Mijller and T.T. Tsong, Field Ion Microscopy, Principles and Applications (Elsevier, ^New York,

1969)

2 A.J. Melmed and R.J. Stein, Surf. Sci., 49, 645 (1975)

3 T. Sakata, J.H. Block, M. Naschitzki and W.A. Schmidt, in: Proc. 34th Intern. Field Emission Symp.,

Osaka, J. Physique, 48, C6-239 (1 987)

4 T. Sakata and J.H. Block, Surf. Sci..

m,

m,

8 A. Clark, The Theory of Adsorption and Catalysis (Academic Press, New York and London, 1970) 9 N. ^{Ernst and} J.H. Block, Surf. Sci., X25, 397 (1983)

10 M. Yamamoto, D. N. Seidman and S. Nakamura, Surf. Sci., LtB, 555 (1982)

11 D.D. Wagman et al., The NBS tables of chemical thermodynamic properties. Selected values for inorganic and C1 and C2 organic substances in SI units, J. Phys. Chem. Ref. Data, U, Suool.a,

(1 982)

12 G.L. Kellogg, Physical Rev. B, 29, 4304 (1984)