JPH0590375A - Evaluation method of semiconductor insulating film - Google Patents

Abstract

(57) [Abstract] [Purpose] To improve the long-term reliability of an insulating film by providing a method for evaluating the semiconductor insulating film and an insulating film evaluation method capable of physically clarifying the breakdown voltage failure mechanism. [Construction] Leakage generated by applying constant current stress to a MOS semiconductor to move injected carriers in the direction of electrode 3 and then applying constant current stress of opposite polarity to move the carriers in the direction of electrode 6. The time series data of the current is converted into a frequency spectrum, the breakdown voltage failure mechanism of the insulating film is clarified by the shape, and the insulating film is evaluated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to reliability evaluation of an insulating film in a semiconductor integrated circuit.

[0002]

2. Description of the Related Art As the density of LSIs has increased, the insulating film has become thinner, and it has been required to form a highly reliable insulating film. One of the most important parameters in evaluating the reliability of such an insulating film is the evaluation of the insulating film withstand voltage. As a method for evaluating the withstand voltage of the gate oxide film, which is an insulating film, a method of applying an electric field stepwise to a MOS capacitor and measuring the dielectric breakdown electric field strength of the MOS capacitor, and a method of applying a constant electric field to the MOS There is a method of measuring the rate at which a capacitor is destroyed by changing the magnitude of the electric field (TDDB method).

FIG. 1 shows the former stepwise application of an electric field,
It is a histogram showing the relationship between the breakdown electric field and the breakdown frequency by the method of measuring the breakdown electric field strength. The A mode shown in FIG. 1 is a mode in which dielectric breakdown occurs due to an initial short circuit, the B mode is a mode in which dielectric breakdown occurs in a field strength range of 1 MV / cm or more and 8 MV / cm or less, and the C mode is a leak current without breakdown. Is a mode in which a predetermined judgment current is reached.

Further, it has been found that the above-mentioned TDDB measurement of a MOS capacitor which is destroyed in the B mode causes dielectric breakdown of the MOS capacitor by applying a voltage for a long time even with a low electric field strength. From this, it can be said that the B mode is a defective MOS capacitor related to long-term reliability, but the main cause of the failure cannot be clarified from the above-described evaluation method.

[0005]

As described above, the conventional insulation film withstand voltage evaluation method is a simple method for evaluating an insulation film, but it is not possible to elucidate the breakdown voltage failure mechanism of the insulation film. The present invention has been made in view of such circumstances, and it is possible to improve the long-term reliability of an insulating film by providing an insulating film evaluation method capable of physically clarifying a breakdown voltage failure mechanism together with the evaluation of the insulating film. To aim.

[0006]

A method for evaluating a semiconductor insulating film according to the present invention comprises a step of applying a unipolar constant current stress to a MOS semiconductor in a method for evaluating the reliability of a MOS semiconductor against dielectric breakdown, and Then, a process of applying a constant current stress of reverse polarity, a process of measuring leak current flowing when a constant current stress of reverse polarity is applied to obtain its time series data, and obtaining the frequency spectrum of the leak current from the time series data. And a process.

[0007]

The semiconductor insulating film evaluation method of the present invention is performed by applying a constant current stress to a MOS semiconductor and measuring the leak current that has flowed. The mechanism until a constant current stress is applied to a MOS semiconductor and a leak current is generated will be described with reference to FIG. A MOS capacitor, which is a MOS semiconductor, has an insulating film 4 on a substrate 5, and electrodes 3, 6 are provided on the insulating film 4 side and the substrate 5 side.
have.

When the substrate 5 of the MOS capacitor is an n (p) type, a positive (negative) polarity constant current stress is applied to the electrode 3 as shown in FIG. 2 (A) (FIG. 2 (B)). At this time, the carriers are injected into the interface between the substrate 5 and the insulating film 4 and into the insulating film 4, and move toward the electrode 3 on the insulating film side by an electric attraction. Next, a constant current stress having a polarity opposite to that of the constant current stress is applied to the MOS capacitor. At this time, the carrier starts moving in the opposite direction due to the electric attraction with the electrode 6 on the substrate side. At the same time, carriers of opposite polarity existing in the substrate move toward the electrode on the insulating film side due to the electric attraction.

A leak current is generated by such carrier movement. This carrier and the opposite polarity carrier,
There is a variation in the leak current that occurs due to movement in the insulating film while repeating coupling and separation. By expressing the time series change of the leakage current as a frequency spectrum, the evaluation of the insulating film and the breakdown voltage failure mechanism can be clarified.

That is, the frequency spectrum of the insulating film having a good withstand voltage extends over a wide range, and the frequency spectrum of the insulating film having a poor withstand voltage has few high frequency components and is biased to the low frequency side. This means that carriers slowly move in the insulating film having a poor breakdown voltage, and suggests that carrier generation-recombination centers (gr centers) exist in the insulating film. The withstand voltage of the insulating film can be evaluated by comparing with the frequency spectrum of the desired withstand voltage performance.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings showing its embodiments. FIG. 3 is an explanatory diagram of the test circuit of the present invention. The MOS capacitor to be inspected has an insulating film 4 formed on a substrate 5, an electrode 3 formed on the upper surface thereof, and an electrode 6 formed on the lower surface of the substrate 5. The pulse generator 2 is connected to the electrode 3. The pulse generator 2 is connected to the pulse voltage control unit 1 and generates a pulse voltage based on the polarity current value, the voltage value, and the application time set by the pulse voltage control unit 1, and the pulse voltage is generated between the electrodes 3 and 6. Give stress.

The leakage current generated by this is the electrode 6
Is detected by the ammeter 7 connected to the leak current measuring unit 8
Measured by Based on the measured current value, the pulse voltage control unit 1 performs voltage control so that the current is constant. The time-series data of the leak current measured by the leak current measuring unit 8 is Fourier-transformed by a computer (not shown) connected to the leak current measuring unit 8 and expressed as a frequency spectrum. The distribution state of this frequency spectrum enables the evaluation of the insulating film and the breakdown voltage failure mechanism to be clarified.

The present invention will be described below with reference to specific examples. The substrate is p-type Si manufactured by the Czochralski method (CZ method)
(<100>, specific resistance about 10 Ω · cm) was used. Gate oxidation is 9
A dry oxygen method was performed at 50 ° C. to form a 20 nm thick SiO ₂ oxide film. After depositing 400 nm of polysilicon by the CVD method, phosphorus was diffused and a polysilicon electrode having an area of 0.5 mm ² was formed by photolithography.

Under the same conditions as above, different wafers are used.
The oxide film breakdown voltage of various types of MOS capacitors, Samples 1, 2, 3 and 4, was evaluated. Sample 1 is a p / p ⁺ epitaxial wafer, Sample 2 is a wafer in which the gate oxide film has been etched and then gate oxidized again, and Sample 3 is p
Type Si surface wafer with root mean square roughness of 0.8 nm,
Sample 4 has a root mean square roughness of 1.7 nm on the p-type Si surface.
It is composed of a wafer.

First, using the test circuit shown in FIG.
Negative constant-current stress (-stress) is applied to sample 1 Constant-current stress: -900nA Stress pulse: period; 1000μs, pulse width; 500μs Application time: 300sec Injection charge: -0.027 C / cm ² Number of injected carriers: It was applied under the condition of 1.69 × 10 ¹⁷ / cm ² , and the leak current flowing at this time was measured by the leak current measuring unit 8.

FIG. 4 is a chart showing the change with time of the leak current. All measurements are performed in a vacuum chamber at 24 ℃,
The degree of vacuum was set to 0.01 Torr or less. Samples 2, 3, 4
About-The chart showing the change with time of the leakage current generated when stress was applied showed almost the same signal as in FIG. From this, it can be seen that the leak current generated by applying stress has no correlation with the quality of the insulating film.

Subsequently, a positive constant-current stress (+ stress) was applied under the conditions of constant-current stress: + 900nA stress pulse: period: 1000 μs, pulse width: 500 μs, application time: 300 sec, and the leak current flowing at this time was measured. .. Similarly, for Samples 2, 3 and 4, + stress was applied to measure the leak current.

FIG. 5 is a chart showing the change with time of the leak current. As shown in FIG. 5, the leak current generated by applying + stress, which is a constant current stress of opposite polarity, was clearly different from that of Samples 1, 2, 3, and 4. FIG. 6 shows the result of Fourier transforming these time series data and representing them by frequency spectrum.

As shown in FIG. 6, sample 1 shows a 1 / f spectrum shape over a wide frequency band, whereas samples 2, 3 and 4 have high frequency components decreasing in that order.
It can be seen that the shape of the 1 / f ² spectrum is shown. The leak current showing such a 1 / f ² spectrum indicates that the carriers have moved slowly, which means that the carrier generation-recombination center (gr center) exists in the insulating film. From these results, it can be said that Samples 1, 2, 3 and 4 have good insulation film withstand voltage in this order.

Sample 1 is an epitaxial wafer
Therefore, it is considered that there are few crystal defects and the breakdown voltage is good.
Sample 2 is that the gate oxide film was etched
Since the defects near the Si surface are removed more than the CZ wafer
It is thought that the pressure resistance is good. Also, the surface roughness of Si is large.
In this case, when this Si surface is gate-oxidized, Si / SiO ₂interface
Stress increases, and₂Bond or sio₂The structure of
I get stressed out. This stress causes SiO₂Conclusion
It becomes easier to cut the bond, and the cut site becomes the gr center.
So sample 4 is sample 3
It can be said that the breakdown voltage is poor.

With respect to the wafers constituting these samples 1, 2, 3 and 4, the electric field strength was 8 MV / cm with an area of 5 mm ^2.
The ratio of the capacitors showing the above insulation resistance is
There are data of 98%, sample 2 80%, sample 3 45%, sample 4 5%, and the above-mentioned results are in agreement with this data.

Although it is possible to make a comparative evaluation of the dielectric breakdown voltage only from the time-series data of the generated leak current, the difference between the data becomes clear by obtaining the frequency spectrum.

[0023]

As described above, according to the method for evaluating an insulating film of the present invention, the insulating film is evaluated, and it is clear that the existence of the g-r center is involved in the defective mechanism of the insulating film breakdown voltage. Therefore, the present invention has excellent effects such as obtaining an index for improving the long-term reliability of the insulating film.

[Brief description of drawings]

FIG. 1 is a histogram showing a typical frequency of a dielectric breakdown electric field of an oxide film.

FIG. 2 is a schematic diagram showing a mechanism in which a leak current is generated when a constant current stress is applied.

FIG. 3 is an explanatory diagram of a test circuit used in the present invention.

FIG. 4 is time-series data of leak current when a constant current stress is applied.

FIG. 5 is time-series data of a leak current when a constant current stress and a constant current stress having an opposite polarity to the constant current stress are continuously applied.

FIG. 6 is a frequency spectrum of time series data of leakage current.

[Explanation of symbols]

 1 pulse voltage control unit 2 pulse generator 3, 6 electrode 4 insulating film 5 substrate 7 ammeter 8 leak current measuring unit

Claims (1)

[Claims] 1. A method for evaluating the reliability of a MOS semiconductor against dielectric breakdown, comprising the steps of applying a constant-polarity constant current stress to the MOS semiconductor, subsequently applying a reverse-polarity constant current stress to the MOS semiconductor, and reverse polarity. The method for evaluating a semiconductor insulating film, comprising: a step of measuring a leak current flowing when a constant current stress is applied to obtain time series data thereof; and a step of obtaining a frequency spectrum of the leak current from the time series data. ..