RU2466417C1 - Method of selecting cmos/soi transistor structures resistant to effect of full absorbed dose of ionising radiation - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: in the method of selecting CMOS/SOI transistor structures which are resistant to the effect of full absorbed dose of ionising radiation, for each process solution, an additional pair of n- and p-channel transistor channels made from the same technology is formed on the chip along with the main LSI circuit; for limited sampling, LSI specimens are removed discretely with a step of about 100 krad(Si) in the TID range from zero to the level of requirements for resistance of the LSI to the effect of maximum TID plus 20% of the relationship I_D=f(V_Gs); data are obtained therefrom on change in shift of threshold voltage ΔV_TH from the TID value through a dose of not less than about 1000 rad(Si)·s^-1 taking into account manufacturing tolerance and possible boundaries of variation of the relationship V_TH(TID); lower

and upper

allowable boundaries of relationships V_TH(TID) of transistors of both conductivity types are approximated; the dispersion of measured values

is determined for each fixed TID value; the value ΔV_TH(↑↓) without exposure is determined for the minimum on the curve

for an n-channel transistor and for the virtually achieved maximum level of exposure; ΔV_th(↑↓) is determined for the same conditions for all possible combinations of differences "V_Th(TID)-Up" and "V_th(TID)-Down" of n-MOS and p-MOS transistors, ΔV_th(↑↓) is determined for three fixed LSI TID values; the LSI are ranked on four gradations of the observed values of ΔV_th(↑↓): "high" ("1") - maximum change; "significant" ("2") - strong change; "average" ("3") - moderate changes; "low" ("4") - weak changes, which on the average value are sorted as "1"%:"2"%:"3"%:"4"%, where the "low" gradation is taken as 100%; ΔV_th(↑↓) is measured for non-exposed industrial LSI specimens and the results of ranking exposed specimens are used to determine the possibility of using in exposure conditions those specimens for which the gradation of change in the value of ΔV_th(↑↓) corresponds to categories "4" and "3".

EFFECT: possibility of selecting industrial LSI which are resistant to TID effects without conducting an exposure experiment to measure radiation drift of threshold voltage ΔV_th.

7 dwg

Description

The invention relates to the field of measurement technology, in particular to assessing the resistance level of large integrated circuits (LSI) based on metal-dielectric-semiconductor (MIS) structures to the effects of dose effects from pulsed ionizing radiation when implementing varying technological processes.

In modern electronics, complementary, i.e. complementary metal-semiconductor (CMOS) structures formed in a thin silicon instrument layer on a sapphire substrate (CMOS / CNS) or in heteroepitaxial structures with a hidden silicon dioxide insulator (CMOS / SOI) layer. Such structures have the common name "silicon-on-dielectric" (KND).

When conducting pulsed radiation studies of integrated circuits manufactured by the KND technology, the urgent task is to control the level of resistance to dose effects (Total Integrated Dose (TID) effects caused by the generation of radiation-induced charges Q _ot , into the gate insulator and to the levels structural defects at the semiconductor - insulator interface - Q _it . This is especially important for p-type conductivity layers (“p-pockets”) in which n-channel field effect transistors of the MOS / KND structure (hereinafter referred to as KND) are formed.

The statistical spread of the radiation change in the basic electrophysical characteristics (EPC) of the transistor structures of CMOS technology, as the main elements of the LSI, is directly related to the statistical nature of the interaction of photon radiation with electronic materials (photoelectric effect, Compton scattering effect, electron-hole pair generation), and the non-uniformity of the EPC of the initial KND heterostructures according to their surface area, technological features of production, which generally determines the differences in the level e durability individual samples LSI chip fabricated using specific process features.

A combination of experimental studies and physical ideas about the mechanism of interaction of photon radiation with electronic equipment materials are used to select TID-resistant LSI samples of CMOS / KND technology. So, in / 1-3 / after irradiation with X-ray or gamma-quanta of LSI samples, the mobility µ of the main charge carriers is controlled as a criterion for assessing resistance to TID effects. However, when using X-ray sources with a low energy (up to 100 keV) and a low exposure dose rate (up to 10 ³ P s ^-1 ) as a source of ionizing radiation (II), practically no structural defects that affect the charge generation Q _it and charge transfer in a semiconductor material. Such a “subthreshold” mechanism is associated only with the recharging of the initial impurity-defective composition (PDS) of the heterostructure and, therefore, cannot be reduced only to the mobility μ of charge carriers in the semiconductor.

.There is also a method of selecting resistant LSI to the effects of TID, which is selected as a prototype / 4 /. To take into account the nonequivalence of the low-energy exposure to x-rays and the reference spectrum, we use the results of processing the experimental data on a simulating AI device with an _X-ray dose rate of P _X-Ray = 1800 rad (Si) · s ^-1 and a reference nuclide source of Co ⁶⁰ gamma quanta with an average quantum energy

.

, при этом мощность дозы источника Со⁶⁰ Р_γ=Р_X-Rау=Const, Р_γ≥10³ paд(Si)·c^-l /5, 6/; 2) снимают зависимость тока утечки межтранзисторных структур I_L для условий облучения на эталонном источнике Со⁶⁰ от поглощенной дозы TID рентгеновских и гамма-квантов

, при этом P_γ=10⁶ рад(SiO₂)·с^-1, то же производят для условий облучения на источнике X-Ray; 3) строят корреляционную зависимость тока стока при условиях нормального функционирования электрической схемы БИС (напряжение «сток-исток» V_DD=5 В) для условий облучения на источнике Со⁶⁰ (Р_γ=10⁶ paд(SiO₂)·c^-l) от тока утечки, снятой для условий облучения на источнике X-Ray (P_γ=1800 рад(SiO₂)·с^-1)

; 4) затем строится зависимость тока стока I_DD от TID в диапазоне 10^-2…10 Мрад (SiO₂) для трех модификаций технологического маршрута «А», «В» и «С», например, на моделирующей установке (ускоритель протонов с энергией 230 МэВ) критерий отбора составляет I_DD=3 мА; Р_X-Ray=10⁶ paд(SiO₂)·c^-l); 5) контролируют изменение сдвига порогового напряжения V_TH, зависимость I_DD от TID источника X-Ray и Со⁶⁰ и корреляционную зависимость тока I_DD с

;

;

при TID=Const,

;

,

при TID=Const; 6) выбирают критерии отказов по I_DD до и V_TH для выбранного уровня по TID для источника X-Ray для разных технологий

, a критерий отказа задают по требуемому значению I_DD.The selection procedure includes the following stages: 1) first, the KND test structure, designed and manufactured using 3 technological modes "A", "B" and "C", is irradiated with the AI of each source in order to remove the dependences of the drain current I _DD transistors of the KND test structure located on the LSI chip for laboratory (X-Ray) and reference (Co ⁶⁰ ) sources from TID≡D

while the dose rate of the source Co ⁶⁰ P _γ = P _X-Rau = Const, P _γ ≥10 ³ rad (Si) · s ^-l / 5, 6 /; 2) remove the dependence of the leakage current of the transistor structures I _L for the irradiation conditions at a reference source of ⁶⁰ from the absorbed dose of TID x-ray and gamma rays

while P _γ = 10 ⁶ rad (SiO ₂ ) · s ^-1 , the same is done for the exposure conditions at the source of X-Ray; 3) build the correlation dependence of the drain current under conditions of normal functioning of the LSI circuit (voltage "drain-source" V _DD = 5 V) for the conditions of irradiation at a source of Co ⁶⁰ (P _γ = 10 ⁶ rad (SiO ₂ ) · s ^-l ) from the leakage current measured for the exposure conditions at the X-Ray source (P _γ = 1800 rad (SiO ₂ ) · s ^-1 )

; 4) then, the dependence of the drain current I _DD on TID is constructed in the range 10 ^-2 ... 10 Mrad (SiO ₂ ) for three modifications of the technological route “A”, “B” and “C”, for example, on a simulator (proton accelerator with energy 230 MeV) selection criterion is I _DD = 3 mA; P _X-Ray = 10 ⁶ rad (SiO ₂ ) s- ^l ); 5) control the change in the shift of the threshold voltage V _TH , the dependence of I _DD on the TID of the X-Ray and Co ⁶⁰ source and the correlation dependence of the current I _DD s

;

;

with TID = Const,

;

,

at TID = Const; 6) select failure criteria according to I _DD to and V _TH for the selected level by TID for the X-Ray source for different technologies

, a failure criterion is set according to the required value of I _DD .

The disadvantage of this method is the need to perform a significant amount of irradiation experiments with the entire batch of LSI to achieve the goal - the implementation of the method of selection of CMOS / KND technology resistant to the effects of the full integral dose of AI transistor structures.

An alternative to this method is the proposed method, based on the measurement of the radiation shift ("drift") of the threshold voltage ΔV _TH (D) of MOS transistor structures with n- and p-channels embedded in a chip with LSI-controlled CMOS / KND technology.

The technical result of the proposed method is the selection procedure for TID-resistant LSI CMOS / KND technologies based on preliminary irradiation of a limited selection of LSIs with built-in test structures based on a pair of CMOS / KND transistors on a simulating x-ray source and a reference nuclide source With ⁶⁰ and the possibility of further selection of stable TID effects of industrial LSIs without conducting an irradiation experiment to measure the radiation drift of the threshold voltage ΔV _TH (D).

и верхних

допустимых границ зависимостей V_TH(TID) транзисторов обоего типа проводимости, определяют для каждого фиксированного значения TID разброс измеренных значений

, определяют значения ΔV_TH(↑↓) в отсутствие облучения, для минимума на графике зависимости

для n-канального транзистора и для практически достигнутого максимального уровня облучения, определяют для тех же условий ΔV_TH(↑↓) для верхнего значения «V_TH(TID)-Up» транзистора n-МОП и нижнего значения «V_TH(TID)-Down» транзистора p-МОП, ΔV_TH(↑↓) для верхнего значения «V_TH(TID)-Up» транзистора n-МОП и верхнего значения «V_TH(TID)-Vp» транзистора p-МОП, ΔV_TH(↑↓) для нижнего значения «V_TH(TID)-Down» транзистора n-МОП и нижнего значения «V_TH(TID)-Down» транзистора p-МОП, ΔV_TH(↑↓) для нижнего значения «V_TH(TID)-Down» транзистора n-МОП и верхнего значения «V_TH(TID)-Up» транзистора p-МОП, ранжируют облученную выборку образцов по полученным максимальным значениям величин радиационного дрейфа зависимостей вида ΔV_TH(↑↓)=f(TID) для транзисторов n-МОП и p-МОП для ряда предельных случаев при трех фиксированных значениях величины TID БИС на четыре градации наблюдаемых величин ΔV_TH(↑↓): «высшую» («1») - максимальное изменение; «значимую» («2») - сильное изменение; «среднюю» «3» - умеренные изменения; «малую» («4») - слабые изменения, которые по средней величине соотносятся как «1»%:«2»%:«3»%:«4»%, где за 100% принята «малая» градация, производят измерения ΔV_TH(↑↓) у необлученных промышленных образцов БИС и по результатам ранжирования облученных образцов судят о возможностях использования в условиях облучения таких БИС, у которых градация изменения величины ΔV_TH(↑↓) соответствует категориям «4» и «3».The technical result is achieved by the fact that in the method of selection of LSIs resistant to the effects of the total absorbed dose of ionizing radiation based on transistor structures of CMOS / KND technology by statistical processing of experimental data on the change in the threshold voltage shift ΔV _TH on the output characteristic "drain current from voltage between the gate and source ”I _D = f (V _GS ) of n- and p-channel transistors of a CMOS pair depending on the total integrated dose of ionizing radiation (TID), for which, for each technological solution, Along with the main LSI circuit, an additional pair of n- and p-channel transistor structures manufactured using the same technology is chopped on a chip, for a limited sample of LSI samples, they are discretely taken with a step of ~ 100 crad (Si) in the TID range from zero to the level of resistance requirements LSI for the effect of maximum TID plus 20% of the dependence I _D = f (V _GS ) for n- and p-channel transistor structures; data on them are obtained on the change in the threshold voltage shift ΔV _TH on the value of TID with a dose rate of at least ~ 1000 rad ( Si) · s ^-1 taking into account the technological spread and possible gr anits changes in the dependence V _TH (TID), approximate the lower

and upper

permissible dependencies of the dependences V _TH (TID) of transistors of both types of conductivity; for each fixed value of TID, the scatter of the measured values

, determine the values ΔV _TH (↑ ↓) in the absence of irradiation, for a minimum in the dependence graph

for the n-channel transistor and for the practically reached maximum level of exposure, determine for the same conditions ΔV _TH (↑ ↓) for the upper value "V _TH (TID) -Up" of the n-MOS transistor and the lower value "V _TH (TID) - Down "of the p-MOS transistor, ΔV _TH (↑ ↓) for the upper value of" V _TH (TID) -Up "of the n-MOS transistor and the upper value of" V _TH (TID) -Vp "of the p-MOS transistor, ΔV _TH (↑ ↓) for the lower value of “V _TH (TID) -Down” of the n-MOS transistor and the lower value of “V _TH (TID) -Down” of the p-MOS transistor, ΔV _TH (↑ ↓) for the lower value of “V _TH (TID) -Down "of the n-MOS transistor and the upper value of" V _TH (TID) -Up "of the p-M transistor OD, rank the irradiated sample of samples according to the obtained maximum values of the radiation drift of dependences of the form ΔV _TH (↑ ↓) = f (TID) for n-MOS and p-MOS transistors for a number of limiting cases with three fixed values of the LSI TID value for four gradations of the observed quantities ΔV _TH (↑ ↓): "highest"("1") - maximum change; “Significant” (“2”) - a strong change; “Average” “3” - moderate changes; “Small” (“4”) - weak changes, which are correlated by the average value as “1”%: “2”%: “3”%: “4”%, where “small” gradation is taken as 100%, measure ΔV _TH (↑ ↓) for non-irradiated industrial LSI samples and according to the results of the ranking of irradiated samples judge the possibilities of using such LSIs under irradiation conditions for which the gradation of the change in ΔV _TH (↑ ↓) corresponds to categories "4" and "3".

Figure 1 shows a typical dependence of the drain current I _{DD of} the MOS transistor structure on the value of the total integrated dose TID (SiO ₂ ) absorbed in the gate dielectric SiO ₂ .

Figure 2 shows the radiation shift of the interface component ΔV _it and the volume component ΔV _{ot of the} threshold voltage at the output current-voltage characteristics I _DD = f (V _GS ) of the transistors of the p-MOS and n-MOS structures.

Figure 3 shows the radiation shift of the threshold voltage ΔV _TH and its volume ΔV _ot and interface ΔV _it components in relative units of the total integral dose in relative units.

Figure 4 shows the zone of dominant influence on the shift of the threshold voltage ΔV _{TH of} volumetric and interface radiation-induced charges in the structure of n-MOS and the zone of their superposition in the structure of p-MOS.

Figure 5 shows the experimental dependence on the total integral dose (along the X axis in relative units) of ionizing radiation of threshold voltage V _{TH of the} MOS test transistors (along the Y axis in B): a) n-channel; b) p-channel.

Figure 6 shows the approximation of the lower (a) and upper (b) allowable limits of the dependences of V _TH on the total integrated dose (TID) of gamma radiation threshold voltage of the n-MOS test transistor.

и нижними значениями

n- и p-канального транзисторов МОП.7 shows possible combinations of the radiation shift of the threshold voltage ΔV _TH (↑ ↓) between the upper values

and lower values

n- and p-channel MOS transistors.

The proposed method is implemented as follows.

When the MOS transistors are irradiated, the radiation shift of the threshold voltage at the output I – V characteristic I _DD = f (V _GS ) in FIG. 2 is caused by two competing processes — the radiation-induced shift ΔV _it corresponding to the charge accumulation Q _{it to} the Si / SiO ₂ interface and radiation -induced shift ΔV _ot corresponding to the capture of charge Q _ot in the volume of the gate dielectric, as shown in Fig.3. In the transistor of the n-MOS structure, the first kind process is dominant at high levels of irradiation, and ΔV _it has a positive value, and therefore the resulting shift ΔV _TH after a certain maximum begins to decrease again (Figure 4). In the transistor of the p-MOSFET structure, the ΔV _it value has a negative value and, together with a similar change in ΔV _ot, causes a steady increase in the negative value ΔV _TH / 7 / with increasing TID (Figure 4). The resulting effect is presented in Fig.5 / 8 /. The gate assembly is represented as a MOS capacitor, for which equality (1) / 6 /

where C is the capacity of the node; ε is the relative dielectric constant of the gate dielectric material; ε ₀ is the absolute dielectric constant of the vacuum; S is the area of the shutter; t _OX is the thickness of the dielectric oxide. The capacitance C and the charge Q in the capacitor are connected by a simple relation

where V is the applied voltage.

Since the radiation-induced charge increases with increasing TID, and the applied external voltage decreases due to the compensation of this charge by the electric field, it follows from (2) that the equivalent node capacitance should also increase, which causes a deterioration in the frequency properties. In accordance with (1), while maintaining all topological parameters, this can be attributed only to the equivalent increase in the thickness of the gate oxide t _OX , or to the SPICE parameter “TOX”.

For modern LSI CMOS / SOI technology, a relationship is established that establishes a relationship between the shift of the threshold voltage ΔV _TH and the component ΔV _ot / 7 /:

The radiation-induced charge increment in the volume of a gate dielectric in the relation (3) can be represented as follows:

where D is the absorbed dose of AI [rad (SiO ₂ )]; k _g - carrier generation constant [rad (SiO ₂ ) ^-1 ]; f _it is the fraction (fraction) of the number of charge carriers that have not recombined in the Si / SiO ₂ interface; f _ot is the fraction (fraction) of the number of charge carriers that have not recombined in the volume of the dielectric.

In the method for selecting transistor structures that are resistant to the full absorbed dose of ionizing radiation, the CMOS / KND technologies on a chip with the main LSI of this technology form a test transistor CMOS / KND structure in a single technological cycle, which includes both n-channel and p-channel MOS transistors. The experimental data on the change in the threshold voltage shift ΔV _{TH are} statistically processed on the output characteristic “drain current versus voltage between the gate and source” I _D = f (V _GS ) of n- and p-channel transistors of a CMOS pair for each technological solution, depending on the full integral dose of ionizing radiation (TID).

и верхних

допустимых границ зависимостей V_TH(TID) транзисторов обоего типа проводимости, определяют для каждого фиксированного значения TID разброс измеренных значений

, определяют значения ΔV_TH(↑↓) в отсутствие облучения, для минимума на графике зависимости

, для n-канального транзистора и для практически достигнутого максимального уровня облучения. Определяют для тех же условий ΔV_TH(↑↓) для верхнего значения «V_TH(TID)-Up» транзистора n-МОП и нижнего значения «V_TH(TID)-Down» транзистора p-МОП, ΔV_TH(↑↓) для верхнего значения «V_TH(TID)-Up» транзистора n-МОП и верхнего значения «V_TH(TID)-Up» транзистора p-МОП, ΔV_TH(↑↓) для нижнего значения «V_TH(TID)-Down» транзистора n-МОП и нижнего значения «V_TH(TID)-Down» транзистора p-МОП, ΔV_TH(↑↓) для нижнего значения «V_TH(TID)-Down» транзистора n-МОП и верхнего значения «V_TH(TID)-Up» транзистора p-МОП, в соответствии со схемой Фиг.7. Затем ранжируют облученную выборку образцов БИС по полученным максимальным значениям величин радиационного сдвига зависимостей вида ΔV_TH(↑↓)=f(TID) для транзисторов n-МОП и p-МОП для ряда предельных случаев при трех фиксированных значениях величины TID (1 - без облучения; 2 - для минимума

для структуры n-МОП; 3 - для TID=1000 у.e.) на четыре градации наблюдаемых величин ΔV_TH(↑↓): «высшую» («1») - максимальное изменение; «значимую» («2») - сильное изменение; «среднюю» «3» - умеренные изменения; «малую» («4») - слабые изменения, которые по средней величине соотносятся как «1»%:«2»%:«3»%:«4»%, где за 100% принята «малая» градация, производят измерения ΔV_TH(↑↓) у необлученных промышленных образцов БИС и по результатам ранжирования облученных образцов судят о возможностях использования в условиях облучения таких БИС, у которых градация изменения величины ΔV_TH(↑↓) соответствует категориям «4» и «3».The I _D = f (V _GS ) dependences for n- and p-channel transistor structures in the TID range from zero to the level of requirements for the LSI resistance to the effects of the maximum TID plus 20% are taken discretely in increments of ~ 100 crad (Si) to account for the dosimetric error , they obtain data on the change in the shift of the threshold voltage ΔV _TH from the TID value with a dose rate of at least 1000 rad (Si) · s ^-1 , taking into account the technological spread and possible limits of variation of the dependence V _TH (TID), approximate the lower

and upper

permissible dependencies of the dependences V _TH (TID) of transistors of both types of conductivity; for each fixed value of TID, the scatter of the measured values

, determine the values ΔV _TH (↑ ↓) in the absence of irradiation, for a minimum in the dependence graph

, for an n-channel transistor and for a practically reached maximum level of exposure. For the same conditions, ΔV _TH (↑ ↓) is determined for the upper value “V _TH (TID) -Up” of the n-MOS transistor and the lower value “V _TH (TID) -Down” of the p-MOS transistor, ΔV _TH (↑ ↓) for the upper value “V _TH (TID) -Up” of the n-MOS transistor and the upper value “V _TH (TID) -Up” of the p-MOS transistor, ΔV _TH (↑ ↓) for the lower value “V _TH (TID) -Down ”Of the n-MOS transistor and the lower value of“ V _TH (TID) -Down ”of the p-MOS transistor, ΔV _TH (↑ ↓) for the lower value of“ V _TH (TID) -Down ”of the n-MOS transistor and the upper value of“ V _TH (TID) -Up "p-MOS transistor, in accordance with the circuit of Fig.7. Then, the irradiated sample of LSI samples is ranked according to the obtained maximum values of the radiation shift of the dependences of the form ΔV _TH (↑ ↓) = f (TID) for n-MOS and p-MOS transistors for a number of limiting cases with three fixed values of TID (1 - without irradiation ; 2 - for a minimum

for the structure of n-MOSFET; 3 - for TID = 1000 у.e.) by four gradations of the observed values ΔV _TH (↑ ↓): “highest” (“1”) - maximum change; “Significant” (“2”) - a strong change; “Average” “3” - moderate changes; “Small” (“4”) - weak changes, which are correlated by the average value as “1”%: “2”%: “3”%: “4”%, where “small” gradation is taken as 100%, measure ΔV _TH (↑ ↓) for non-irradiated industrial LSI samples and according to the results of the ranking of irradiated samples judge the possibilities of using such LSIs under irradiation conditions for which the gradation of the change in ΔV _TH (↑ ↓) corresponds to categories "4" and "3".

In the claimed method, firstly, they use the radiation-critical parameter ΔV _TH , which fully reflects the physical processes in the structure of the MOSFET under the influence of TID effects, secondly, the volume of preliminary measurements is reduced to obtain the selection criterion ΔV _TH (↑ ↓), thirdly, for this technology for the fabrication of MOS structures, this allows us to draw a conclusion about the resistance of LSIs to TID effects without conducting a radiation experiment based on data on ΔV _TH (↑ ↓) values obtained by measuring without irradiation of reading.

An example implementation of the proposed method

The I – V characteristics of the n- and p-channel MOS / KND transistors were measured on the plates D-01-09, D-01-11, D-012 similarly to Figure 2. According to the results of measuring the threshold voltages V _{TH of the} test transistors of n-MOS and p-MOS, it was found that the threshold voltages change significantly during the dose collection process. When samples are irradiated to a dose level of 1200 rel. units the V _TH values of n-MOS transistors varied on average from 1.3 V to 0.9 V (by 30–40%), and p-MOS transistors varied from 1.3 V to 2.1 V (by 60–70%).

To make comparisons in the values of the radiation shift of the threshold voltages ΔV _TH at the output I – V characteristics, the upper (Fig. 6-b) and lower (Fig. 6-a) allowable limits of the dependence on the total integrated dose (TID) of gamma radiation threshold voltage of the test transistors were approximated (V _TH ) of an n-MOS transistor.

For the n-MOS transistor, the following relationships are obtained:

V _TH = 2.0389 · 10 ^-12 D ⁴ -6.0860 · 10 ^-9 D ³ + 6.4506 · 10 ^-6 D ² -2.7599 · 10 ^-3 D + 1.1683 (5)

for the lower permissible boundary obtained from measurements with a linear regression coefficient R ² = 0.9904;

V _TH = 2.0 · 10 ^-12 D ⁴ -7.0 · 10 ^-9 D ³ + 8.0 · 10 ^-6 D ² -3.5 · 10 ^-3 D + 1.3958 (6)

= 428 отн. ед., для нижней границы (Фиг.6-б)) -

отн.ед. Тогда данные по разбросу измеренных значенийfor the upper permissible limit obtained from measurements with a linear regression coefficient R ² = 0.9918. From the data of Fig.6-a), there is a minimum in the graphs of the dependence V _TH (D) for the n-MOS transistor. For the upper boundary (Fig.6-a)), this minimum corresponds to

= 428 rel. units, for the lower boundary (Fig.6-b)) -

rel. Then the data on the scatter of the measured values

for fixed dose values are presented in table 1.

Table 1 The spread of the measured values of the threshold voltage of the transistors of the n-MOS structure from the value of TID Designations TID value, rel. units 0

1000 Parameter ΔV _TH (↑ ↓), V +0.230 +0.103 -0.0084

From the data obtained in table 1 it follows that

- ΔV _TH (↑ ↓) values in the absence of irradiation reach 0.23 V;

составляют до 45% от исходного значения;- ΔV _TH values (↑ ↓) for the minimum in the dependence graph

make up to 45% of the initial value;

- ΔV _TH values (↑ ↓) for the practically reached maximum level of exposure D = 1200 rel. units almost equal to zero.

The latter circumstance testifies to the fact that, at very high levels of radiation load, technological differences in the test samples become insignificant, since the number of structural defects causing changes in the I – V characteristics of transistors exceeds their initial value in the heterostructure and is decisive.

For a p-MOS transistor, the approximating dependence V _TH (D) for the upper and lower experimental dependences is the ratio:

- текущее значение величины V_TH при фиксированном значении D;Where

- the current value of V _TH at a fixed value of D;

- максимальное по абсолютной величине измеренное значение порогового напряжения на графике зависимости V_TH(D), В;

- the maximum absolute value of the measured value of the threshold voltage in the graph of the dependence V _TH (D), V;

α is the radiation degradation constant V _TH , rad (Si) ^-1 ;

D is the absorbed dose of AI, rad (Si).

The results of approximation of the dependences V _TH (D) for the upper curve of Fig.6-b) (Up) and the lower curve of the same dependence (Down) are given in Table 2.

table 2 The results of approximating the dependences V _TH (D) for the upper curve of Fig. 6-b) (Up) and the lower curve of the same dependence (Down) of the p-MOS transistor The value of α in equal to (8)

Value

Parameter

0.8295 -1.5429

0.881 -1.938

Values for D = 1000 rel. units

Table 3 shows the scatter of the measured values of the threshold voltage of the transistors of the p-MOS structure from the value of TID. It can be seen here that for the value D = 1000 rel. units ΔV _TH (↑ ↓) is maintained at 64% of the initial value.

Table 3 The spread of the measured values of the threshold voltage of the transistors of the p-MOS structure from the value of TID TID value, rel. units Designations 0

1000 Parameter ΔV _TH (↑ ↓), V -0,440 -0.391 -0.280

Table 4 shows the maximum values of the radiation drift of the dependences of the form V _TH (D) for transistors n-MOS and p-MOS for three fixed values of TID: D ₁ = 0; D ₂ = 428 rel. units; D ₃ = 1000 rel. units for the following extreme cases:

1) ΔV _TH (↑ ↓) for the graph of “V _TH (D) -Up” of the n-MOS transistor and “V _TH (D) -Down” of the p-MOS transistor;

2) ΔV _TH (↑ ↓) for the graph of “V _TH (D) -Up” of the n-MOS transistor and “V _TH (D) -Up” of the p-MOS transistor;

3) ΔV _TH (↑ ↓) for the graph of “V _TH (D) -Down” of the n-MOS transistor and “V _TH (D) -Down” of the p-MOS transistor;

4) ΔV _TH (↑ ↓) for the graph of “V _TH (D) -Down” of the n-MOS transistor and “V _TH (D) -Up” of the p-MOS transistor.

. В табл.4 введены четыре градации наблюдаемых величин ΔV_TH(↑↓): «высшая» («1») - максимальное изменение; «значимая» («2») - сильное изменение; «средняя» - («3») умеренные изменения; «малая» («4») - слабые изменения. По средней величине они соотносятся как 127%:118%:109%:100%, где за 100% принята «малая» градация. Таким образом, максимальное изменение, связанное с технологическим разбросом с учетом радиационной реакции структур МОП обоего типа проводимости, составляет 27%.The last column of Table 4 shows the total values of possible changes in threshold voltages

. In table 4, four gradations of the observed quantities ΔV _TH (↑ ↓) are introduced: “highest” (“1”) - maximum change; “Significant” (“2”) - a strong change; “Medium” - (“3”) moderate changes; “Small” (“4”) - weak changes. According to the average value, they are correlated as 127%: 118%: 109%: 100%, where the “small” gradation is taken as 100%. Thus, the maximum change associated with the technological spread, taking into account the radiation reaction of MOS structures of both types of conductivity, is 27%.

Table 4 Maximum values of radiation drift of dependences of the form V _TH (D) for transistors n-MOS and p-MOS for a number of limiting cases for three fixed values of TID D rel units

Category 0 +1.3958 (Up) -1.54 (Down) 2.94 "one" +0.670 +1.3958 (Up) -1.1 (Up) 2,50 "3" +1.1683 (Down) -1.54 (Down) 2.71 "2" +1.1683 (Down) -1.1 (Up) 2.27 "four" 428 +0.88 (Up) -1.938 (Down) 2,818 "one" +0.494 +0.88 (Up) -1.543 (Up) 2.42 "3" +0.78 (Down) -1.938 (Down) 2,718 "2" +0.78 (Down) -1.543 (Up) 2,323 "3" 1000 +0.8958 (Up) -2.1 (Down) 2,912 "one" +0.272 +0.8958 (Up) -1.82 (Up) 2,632 "3" +0.8119 (Down) -2.1 (Down) 2,996 "one" +0.8119 (Down) -1.82 (Up) 2,716 "2"

From the analysis of the data in Table 4, it follows that the gradations “1” and “3” in the absence of irradiation are maintained at an irradiation level of D = 1000 rel. units, “2” transforms into “1”, and “4” - into “2”. On this principle, the principle of sorting non-irradiated structures can be based, which coincides with the ideas about the conservative behavior of the left wing of the distribution of radiation-critical parameters (RCP) from the level of radiation load. For the given example, samples with a radiation shift of ΔV _TH (↑ ↓) in the range of 18-27% are removed from the irradiated sample.

Since the radiation shift of the threshold voltage refers to degradation processes, weaker samples, i.e. having an initially elevated level of structural defects, respond less to the effects of AI and, therefore, are less susceptible to radiation exposure.

при D=0 до величины

при D=1000 отн. ед. составил ≈42%, а радиационный дрейф нижней границы

при D=0 до величины

, составил ≈23%. Эти величины для n- и p-МОП транзисторов приведены в табл.5.From an analysis of the data in Table 4, it follows that the radiation drift of the upper boundary

at D = 0 to the value

at D = 1000 rel. units amounted to ≈42%, and the radiation drift of the lower boundary

at D = 0 to the value

amounted to ≈23%. These values for n- and p-MOS transistors are given in Table 5.

Table 5 Threshold voltage radiation drift at D = 1000 rel. units D rel units Radiation drift ΔV _TH ,%

1000 42 23 49 43

From the analysis of the results obtained, it follows that, based on previously obtained experimental data on the radiation shift of the threshold voltage ΔV _TH n- and p-MOS transistors, it is possible to implement the procedure for selecting CMOS / KND technology that is resistant to the effects of the full absorbed dose of LSI without additional radiation experiments.

Thus, the use of the proposed method allows you to quite simply and accurately select, in industrial production, CMOS / KND technologies that are resistant to the effects of TUBIS effects.

Literature

1. Petrosyants K.O., Kharitonov I.A., Orekhov E.V., Sambursky L.M., Yatmanov A.P. Instrument-technological modeling of the element base of CMOS KNI BIS taking into account factors of radiation exposure // M .: IPPM, 2008, Sat. scientific works of the 3rd All-Russian scientific and technical conference "Problems of developing promising microelectronic systems - 2008". - S.266-271.

2. Benedetto J.M., Boesch H.E. MOSFET and MOS capacitor responses to ionizing radiation // IEEE Transacnions on Nucltar Science, 1984. - V. NS-31. - No. 6. - pp. 1461-1466.

3. Petrosyants K.O., Kharitonov I.A., Orekhov E.V. Modeling the effects of single charged particles on the submicron SOI CMOS cell ROM // EMN 2007, Sat. scientific works, under the editorship of V.Ya. Stenina. - M .: MEPhI, 2007 .-- P.38-41.

4. Spratt J.P., Nickel V.V., McCollin J.L. Proposed Test Strategy For Radiation Hardened Custom LSI / VLSI // IEEE Transactions on Nuclear Science, 1981.- V.NS-28. - No.6. - pp. 427-428.

5. Millard D.C. et all. Time Dependence Dose Enhancement Effects on Integrated Circuits Transients Response Mechanisms // IEEE Transactions on Nuclear cience, 1985.- V. NS-32 / - No.6 / - pp.4376-4381 /.

6. Pozire C. M., Brown D.B., Sandelin J.W. Dose and Dose Rate Dependence of 8080A Microprocessor // IEEE Nransactions on Nuclear Science, 1980 .-- V. NS-27. - No.4. - pp. 1299-1304.

7. CERN Training / Aptil 11/2000 / Radiation Effects on Electronics Components and Circuits, part 1 of 2 / Martin DENTON.

8. Reference amateur radio. Under the general ed. A.A. Kulikovsky. / M., L .: SEI, 1955 .-- 256 p.

Claims (1)

A method for selecting large integrated circuits (LSI) that are resistant to the full absorbed dose of ionizing radiation based on transistor structures of CMOS / KND technology by statistical processing of experimental data on the change in the threshold voltage shift ΔV _TH on the output characteristic "drain current from voltage between the gate and source" _{I D = f (V GS)} n- and p-channel transistors in the CMOS pair, depending on the total integrated dose of ionizing radiation (TID), for which for each process solution is formed on the chip outfit with the main LSI circuit, an additional pair of n- and p-channel transistor structures manufactured using the same technology, characterized in that for a limited sample of LSI samples, discrete samples are taken discretely with a step of ~ 100 crad (Si) in the TID range from zero to the level of resistance requirements LSI for the effects of maximum TID plus 20% of the dependence I _D = f (V _GS ) for n- and p-channel transistor structures; data on them are obtained on the change in the threshold voltage shift ΔV _TH on the value of TID with a dose rate of at least ~ 1000 rad ( Si) · c ^-1 subject process may scatter and boundary changes depending V _TH (TID), an approximation of the lower

and upper

permissible dependencies of the dependences V _TH (TID) of transistors of both types of conductivity; for each fixed value of TID, the scatter of the measured values

, determine the values ΔV _TH (↑ ↓) in the absence of irradiation, for a minimum in the dependence graph

for the n-channel transistor and for the practically reached maximum level of exposure, determine for the same conditions ΔV _TH (↑ ↓) for the upper value of V _ТН (TID) -Up of the n-MOS transistor and the lower value of V _TH (TID) -Down of the transistor p -MOS, ΔV _TH (↑ ↓) for the upper value of V _TH (TID) -Up of the n-MOS transistor and the upper value of V _TH (TID) -Up of the p-MOS transistor, ΔV _TH (↑ ↓) for the lower value of V _TH ( TID) -Down of the n-MOS transistor and the lower value of V _TH (TID) -Down of the p-MOS transistor, ΔV _TH (↑ ↓) for the lower value of V _TH (TID) -Down of the n-MOS transistor and the upper value of V _TH (TID ) -Up p-MOS transistor, rank bluchennuyu take samples of obtained maximum values of the magnitudes of radiation drift dependency type _{ΔV TH (↑ ↓) = f} (TID) for the transistors n-MOS and p-MOS for several limiting cases at three fixed values of the magnitude TID LSI four gradations observed values ΔV _TH (↑ ↓): “highest” (“1”) - maximum change; “Significant” (“2”) - a strong change; “Medium” (“3”) - moderate changes; “Small” (“4”) - weak changes, which are correlated by the average value as “1”%: “2”%: “3”%: “4”%, where “small” gradation is taken as 100%, measure ΔV _TH (↑ ↓) for non-irradiated industrial LSI samples and according to the results of the ranking of irradiated samples judge the possibilities of using such LSIs under irradiation conditions for which the gradation of the change in ΔV _TH (↑ ↓) corresponds to categories "4" and "3".

Images