JPS59130B2 - Method for measuring electrical characteristics of defects at the interface between insulating films and semiconductors - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention describes the electrical characteristics such as the density, energy level, and trapping cross section of trap centers at the interface between an insulating film and a semiconductor to determine the square wave response of a metal-insulating film-semiconductor (MIS) diode. The present invention relates to a method for easily measuring temperature dependence with high sensitivity and high accuracy.

The electrical properties of the insulating film-semiconductor interface have a fundamental influence on the operation of semiconductor devices such as bipolar transistors and field effect transistors, and integrated circuits. It is extremely important to easily and sensitively measure not only static properties, such as the concentration, but also dynamic properties, such as the capture cross section. Conventionally, as a method for measuring the electrical characteristics of an insulating film-semiconductor interface, various methods have been known in which the electrical characteristics are determined from the bias voltage and frequency dependence of the admittance of a MIS diode. However, in this method, it is essentially necessary to obtain the semiconductor surface potential with high precision; however, it is necessary to obtain the surface potential due to fixed charges in the insulating film or two-dimensional dispersion of the surface potential due to irregularities at the insulating film-semiconductor interface, and the shallow impurity density in the semiconductor. Errors occur due to non-uniformity in the depth direction, etc., and both measurement accuracy and sensitivity cannot be said to be sufficient. Further, although static interface characteristics can be measured in a relatively short time, it generally takes a long time to determine dynamic characteristics. By the way, the square wave transient response of the junction capacitance of a semiconductor p-n junction or Schottky junction can be expressed as 2
Difference in capacitance ΔC between two sampling times 0 and 2
DLTS measures the electrical properties of deep impurity levels existing in the junction, that is, the spatial distribution of energy levels, trapping cross sections, and density, from the temperature sweep of this capacitance difference.
(Deep-LevelTransientSpect
There is a method called ``roscopy'', which has high sensitivity and
It has been confirmed that it has high accuracy, and it is also applied to the measurement of the interface state at the insulating film-semiconductor interface of MIS diodes.

However, in the conventional method of keeping the square wave applied voltage constant during temperature sweep, for example, the capacitance c(
tl) changes with temperature due to the temperature dependence of the Fermi level. Therefore, when applied to the measurement of the interface state of a MIS diode, it is necessary to correct each change in C(tl), and the spectroscopic method of DLTS It has the disadvantage that its characteristics are not fully utilized. In contrast, the method of the present invention is an improved version of the conventional DLTS for measuring the interface state at the insulating film-semiconductor interface of a MIS diode, and its purpose is to measure the sampling time tl during temperature sweep. Or, a feedback system is provided for the applied square wave voltage so as to keep the high frequency capacitance at 2 constant, and the energy distribution of the interface state density is directly recorded on the X-Y recorder without complicated computer processing, and The objective is to provide a measurement method that allows the captured cross-sectional area to be determined very easily from the drawing on the recorder.

Hereinafter, the method of the present invention will be explained in detail with reference to examples.

FIG. 1 shows an outline of the configuration of an example of an interface level measuring device for carrying out the measuring method according to the present invention. The MIS diode 2 to be measured is placed inside the cryostat 1, which can be heated from about 80K to about 400K during a temperature sweep for about 30 minutes to 1 hour, and its high frequency capacitance C(t) and the capacitance value of the standard capacitance 3 are measured. The difference C(t)-CO with CO is measured with a high-speed response differential measurement type capacitance meter 4. The measurement frequency of this high frequency capacitance can be, for example, 1 MIIz. Further, as the standard capacitance CO, a capacitance value in a resting state when a bias voltage is applied or a value close to the capacitance value can be selected. At this time,
A square wave-like repetitive voltage as shown in FIG. 2a is applied to the MIS diode 2 by a square wave generator 5 and a feedback DC power source 6 via a capacitance meter 4. However, FIG. 2 illustrates the case of the MIS diode 2 having an n-type semiconductor substrate. The pulse width and period of this square wave voltage can be set arbitrarily, and the peak value voltage Vb is set to be a voltage more positive than the flat band voltage, so that an electron accumulation layer is formed on the semiconductor surface and electrons are captured in the interface state. state. Here, MI
When the bias voltage applied to the S diode 2 is changed to a voltage Va at which a depletion layer is formed on the semiconductor surface, for example -3, electrons are emitted from the interface state to the conduction band, and along with this, the MIS The high frequency capacitance C(t) of the diode 2 changes as illustrated in FIG. 2b. This transient response is divided into two sampling times T,, t2, for example t1=5ms. ．．
Capacitance difference ΔC (=-C(T2) at t2=10ms
-C(T2)). This capacitance difference is measured using a two-input type box scan averager 7. Now, in the conventional DLTS, this capacitance difference ΔC is plotted against the temperature of the MIS diode 2, but in the method of the present invention, during temperature sweep, the high frequency capacitance C(t1) at one sampling time t1 is the standard capacitance. A feedback path is provided for the bias voltage Va so that it is equal to CO. That is, as shown in FIG. 1, the difference C(T,)-CO between the high frequency capacitance C(T,) measured by the capacitance meter 4 and the standard capacitance G is sent to the feedback type DC power supply 6 via the box filter and averager 7. supply In this case, b changes at the same time, but there is no problem as long as an electron storage layer is formed on the semiconductor surface, so the peak value of the square wave voltage should be set sufficiently large. Now, the temperature of the MIS diode 2 is measured by, for example, a thermocouple 8, and the output is further converted to an absolute temperature T(k) by a linearizer 9.
Convert to an output proportional to . Then, the divider 10 outputs the ratio ΔC/T between the capacitance difference ΔC and the absolute temperature T.
When C/T is recorded on the X-Y recorder 11 against the absolute temperature T, ΔC/T is proportional to the interface state density and T is approximately proportional to the energy level, so it is recorded directly on the X-Y recorder. The energy distribution of the interface state density is drawn. Furthermore, the capture cross section is measured once again by changing the sampling times Tl and t2, and is easily determined from the changes in the two curves. Next, the above matters will be explained below using theoretical formulas.

Now, taking MIS diode 2 of an n-type semiconductor substrate as an example, the interface state density Nss(E) at energy level E
is expressed as.

However, εs is the dielectric constant of the semiconductor, ND is the shallow donor density,
k is Boltzmann's constant, COx is oxide film capacitance, A is metal electrode area, and other symbols have already been defined. Also, the energy level E is expressed as.

However, Ec is the energy at the bottom of the semiconductor conduction band, Nc is the effective level density of electrons in the conduction band, Vn is the thermal velocity of electrons, σn
(E) is the capture cross section for electrons at that energy level. Equation (1) shows that by keeping C(t1) constant, ΔC/T is proportional to Nss. Equation (2) shows that if the capture cross section does not strongly depend on the energy level, then the absolute temperature T It is clear that Ec is proportional to the energy level measured, and it can be seen that the energy distribution of the interface state density is directly recorded on the XY recorder 11. Now, the capture cross section and the relationship between temperature and energy level can be determined using equation (2).

The specific method is shown in Figure 3. That is, the ratio T2/t1 of the sampling times Tl and t2 is kept constant and changed,
Let Tl,t'2. At this time, curves A and B representing ΔC/T versus T shift relative to each other as shown in FIG. Incidentally, since the interface state densities corresponding to the same energy level are naturally equal, they have the same ΔC/T value according to equation (1). Conversely, the temperature corresponding to any equal ΔC/T is the third
As shown in the figure, T and T' correspond to the same energy level, and the capture cross section σn is given by using equation (2).

However, considering the temperature dependence of T2 of VnNc, VZ
, V〃 means the value at 300K. given by.

In this way, the capture cross section and energy level can be determined by simple calculations. Now, as a specific example,
FIG. 4 shows the results of measurement of the interface level at the interface between SiO2 formed on the silicon surface by high-temperature oxidation in dry oxygen and the substrate silicon. Curve a in Figure 4 is t1=5m
ΔC measured with a sampling time of s, t2 - 10ms
/T vs. T is converted into a scale of interface state density Nss vs. energy level E. It can be seen that measurement over a wide range from approximately Ec-0.4 eV to Ec-0.5 eV is possible. Figure 4b shows the energy dependence of the capture cross section. The average capture cross section remained constant at approximately 10-16 cIL. Although it is a matter of course, it is not necessarily necessary to perform the temperature sweep more than once; if you use a device that can sample at three times, Tl, t2, and t3, and take out the two outputs of A sweep is sufficient and the measurement time is cut in half.

In this case, the capacitance value at any one sampling time may be made equal to the standard capacitance CO, but it is easier to make the measurement process by making the capacitance value C (T2) at the intermediate T2 equal to CO. It is advantageous because As described above, the measurement method of the present invention is an improved version of DLTS, which is a type of junction capacitance transient response, for measuring the interface state at the insulating film-semiconductor interface of a MIS diode. By applying feedback to the applied voltage so that t1) does not change during the temperature sweep, the energy distribution of the interface state density can be directly recorded on the X-Y recorder, and the capture cross section can also be determined at the same time. It is a highly sensitive measurement method.

Furthermore, it is a highly accurate measurement method because it is not a measurement method that uses the semiconductor surface potential, which generally has an in-plane distribution, and there is no need to make special assumptions about the shallow depth distribution of impurity density.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of the configuration of an example of an apparatus for carrying out the measurement method of the present invention, and FIGS. 2a and 2b illustrate the applied voltage and M
Conceptual diagram of time change of high frequency capacitance of IS diode, Part 3
The figure is an explanatory diagram of the procedure for determining the energy level and capture cross section, and FIG. 4 is an example of measurement for a SiO2/Si-based MIS diode. 1... Cryostat, 2... MI
S diode, 3...Standard capacity, 4...
・Differential measurement type capacitance meter, 5...Square wave generator, 6
... Feedback type DC power supply, 7 ... Boxcar averager, 8 ... Thermocouple, 9 ...
... Linearizer 10 ... Divider, 11.
...X-Y recorder.

Claims (1)

[Claims] 1. Evaluate the square wave transient response of the high-frequency capacitance of a metal-insulating film-semiconductor (MIS) diode by the difference in capacitance at at least two sampling times t_1 and t_2, and use the temperature sweep of this capacitance difference to determine the interface of the insulating film-semiconductor interface. When measuring the level, sampling time t_1 or t
An interface state measurement method characterized by applying feedback to a bias voltage applied to a metal electrode during capacitance measurement so as to keep the high frequency capacitance at _2 constant during a temperature sweep.