Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

860 Charged Particle and Photon Interactions with Matter

and Baze, 2001; Benedetto et al., 2004; Gadlage et al., 2004; Uemura et al., 2006). For example,

Figure 30.22 shows the individual contributions from an inverter and a latch to the overall error

rate (Buchner et al., 1997). At low frequencies, the SEU rate is dominated by those originating in

the latch and at high frequency by those in the inverter.

In this section, I will shift the emphasis away from transient current in single MOSFET

(See Figure 30.21) to transient voltage pulse in an inverter chain. The inverter chain consists of

n-MOSFETs and p-MOSFETs as shown in Figure 30.23a (Kobayashi et al., 2007a). In princi-

ple, the transient current in single n-MOSFET differs from that in n-MOSFET of inverter. Figure

30.24 shows an example of numerically calculated transient currents in single and integrated SOI

MOSFETs (Ferlet-Cavrois et al., 2006). The difference between these is due to the bias condi-

tion. During transient current measurements on single n-MOSFET, the drain voltage is articially

forced. On the contrary, the drain of the n-MOSFET in inverter chain is oating, and therefore the

drain voltage is not forced to a xed value. The look up table (LUT) technique, proposed by D.

Kobayashi, acts as a bridge between them (Kobayashi et al., 2007a,b). Figure 30.23b and c show

how to estimate the transient voltage pulse by using the transient currents on the single MOSFET.

The transient currents, I

, are measured under various constant drain voltage conditions. The mea-

sured data are replotted as a function of drain voltage, V

, for any time. The I

at 4 ps is represented

in the bottom left of Figure 30.23b. On the other hand, the recovery current, I

, versus drain voltage

are calculated for each time. The recovery current is the sum of p-MOSFET drain current, I

, and

1.0

0.8

0.6

0.4

0.2

0.0

1.6

1.2

0.8

0.4

0.0

–3

–4

–5

–1 0 1

Pad

Edge

Center

260 MeV Ne

520 MeV Ar

Diffusion

Bipolar

Ambipolar

2 3

Time (ns)

(a)

(b)

(c)

Current (A) Current (mA) Current (mA)

4 5 6 7 8

Figure 30.20 Typical transient currents at a reverse bias of 20V induced by 260MeV Ne (a) and

520MeV Ar (b), when the ion strikes at the edge, the center, and the bonding pad. The transient currents obtained

from the center of electrode on a log scale (c). The ambipolar, the bipolar, and the diffusion phases are denoted

in this gure. (Reprinted from Hirao, T. et al., Nucl. Instrum. Methods. B, 267, 2216, 2009. With permission.)

Radiation Effects on Semiconductors and Polymers for Space Applications 861

150

100

–50

350

300

250

200

150

100

–50

–100

–3 –2 –1 0 1

Time (ns)(a)

(b)

Source and drain current (μA)

Collected charge (fC)

Transient current

Bias condition

Drain

Source

Collected charge

L =0.25 μm

2 3 4 5 6

–3 –2 –1 0 1

Time (ns)

2 3 4 5 6

OFF

Bias condition

L =0.25 μm

OFF

Figure 30.21 Typical transient currents obtained from the source and drain electrodes of bulk n-MOSFETs

biased in the OFF-state or TG-state during irradiation (a). The charge collection on the source and drain, which

is calculated by integrating the transient currents is represented in (b). (Reprinted from Ferlet-Cavrois, V. et al.,

IEEE Trans. Nucl. Sci., 51, 3255, 2004. With permission.)

100

Sum

Combinatorial logic

Sequential logic

0.1

1 10

Frequency (arb. units)

Error rate (arb. units)

100

Figure 30.22 Error rate as a function of frequency for combinational and sequential logic elements as well

their sum. (Reprinted from Buchner, S. et al., IEEE Trans. Nucl. Sci., 44, 2209, 1997. With permission.)

862 Charged Particle and Photon Interactions with Matter

capacitance discharge current, I

. The inverter’s operating voltage is determined by the intersection of

-V

curve and I

-V

curve. Figure 30.23c shows the transient voltage pulse in the inverter chain

by using the LUT technique. A simulation result by TCAD is also shown as a broken line. The LUT

estimation agrees rather well with the simulated result. Thus, this indicates that the transient voltage

pulse in inverter chain can be described using this modeling method.

Nonirradiated parts

= Analytical models (SPICE)

0.8

0.4

(c)

4 ps

Operating

point

(at 4 ps)

Irradiated MOS

= LUT

2.0

1.5

1.0

0.5

0.0

–0.5

–12

–11

SOI

Gate-hit

LUT estimation

Mixed-mode

–10

Time: t (s)

–9

1 2

(V)

OUT

(V)

= V

OUT

I (mA)

= I

+ I

Heavy ion

OUT

OFF

Heavy ion

(a)

(b)

Figure 30.23 SET-pulse construction. A test inverter (a) is divided into two circuit blocks: irradiated

n-MOSFET and nonirradiated parts (b). The inverter’s operating point is determined by the intersection of

-V

curve and I

-V

curve. SET pulse is constructed from the operating points (c). A mixed-mode simula-

tion result with the same device model and circuit conguration is superimposed for reference. (Reprinted

from

Kobayashi, D. et al., IEEE Trans. Nucl. Sci., 54, 2347, 2007a. With permission.)

Radiation Effects on Semiconductors and Polymers for Space Applications 863

30.4 polymers

Polymer materials have been widely used for structure and thermal control in spacecraft because

of their light weight, workability, exibility, and electrical insulation properties. Recently, polymer

materials have been used for their enhanced and sophisticated performance due to their high specic

strength, high specic stiffness, high thermal stability, and high dimensional stability. Meanwhile,

these materials are exposed to severe and complex space-environment conditions. As a result, some

material properties—mechanical, thermal, and optical—are degraded. One of the polymer material

degrading factors is space radiation particles. Another degrading factor is intense solar ultraviolet

(UV) rays, atomic oxygen (AO) in low earth orbit (LEO), and thermal cycle. There have been cases

where volatile compounds from polymers in vacuum contaminated a high-prole sensor. A syner-

gistic

effect may be caused by a combination of these factors.

One

approach to solving these problems is the evaluation of the material properties on the ground

in the required mission period. For such evaluations, we utilize our combined space effects test

facility, which accommodates the irradiation of independent or coincidental electron beams, ultra-

violet

(UV) rays, and atomic oxygen.

Furthermore,

the material property data related to the effects of the natural space environment

is necessary for performing the evaluations on the ground. Space materials exposure experiments

have

been done not only for this purpose but also for demonstration for space use.

The

radiation effects on polymers, especially on aromatic polyimides which are typically used

in spacecraft material, are summarized briey, and then recent ground evaluation facilities are

described. Finally, space materials exposure experiment projects and results concerning radiation

effects are summarized.

30.4.1 general radiation effectS on polyMerS

Several authors have described degradation in polymers by means of bar charts showing the stages

of degradation versus dose (Holmes-Siedle and Adams, 2000). This type of chart is a plot of the

growth of damage versus dose, or a “radiation index” (RI) such as the logarithm of the dose in

Gy at which elongation at break, E, a well-dened mechanical property, is halved. Figure 30.25a

and b are classications of materials in decreasing order of their radiation resistance (Tavlet et al.,

1998). Figure30.25a and b list rigid thermoplastics, thermoset resins and composites with respect

1.0

0.8

0.6

0.4

0.2

0.0

0.0 0.1 0.2

Stuck

node

Stuck NMOS current:

Device alone

Device in inverter chain

inv8

Time (ns)

Node voltage (V)

Drain current (mA)

0.3 0.4

1.2

0.6

0.0

–0.6

–1.2

Figure 30.24 Simulation of a 130nm SOI n-MOSFET struck by a 5 MeV cm

/mg ion in the center of the

gate. The struck transistor is either simulated as a device alone in the OFF-state or integrated in an inverter

chain with mixed-mode simulation. The current driven by the struck device is shown on the left axis. The

output inverter voltage are indicated on the right axis for the struck node and the eighth inverter. (Reprinted

from

Ferlet-Cavrois, V. et al., IEEE Trans. Nucl. Sci., 53, 3242, 2006. With permission.)

864 Charged Particle and Photon Interactions with Matter

to their radiation resistance. This classication gives the order of magnitude of the maximum dose

of usability of the materials; it corresponds to long-term irradiations. In particular, thermoplastics

(Figure

30.25a) may be very sensitive to oxidation and hence to dose-rate effects; during long-term

irradiation in the presence of oxygen, their degradation starts at a much lower dose (Tavlet et al.,

1998). This datum is acquired in a nuclear reactor with a high dose-rate irradiation in air, corre-

sponding to the popular threshold dose limit. It is necessary to consider the “real radiation led”

where

polymers are used and screened.

Polyimide (PI)

Liquid Crystal Polymer (LCP)

Polyetherimide (PEI)

Polyamideimide (PAI)

Polyphenylsulfide (PPS)

Polyetheretherketone (PEEK)

Polystyrene (PS)

Copolymer PI + siloxane

Polyarylate (PAr)

Polyarlamide (PAA)

Polyethersulfide (PES)

Polysulfone (PSU)

Polyamide 4.6

Polyphenyloxyde (PPO)

Acylonitride-butadiene-styrene (ABS)

Polyethylene (PE)

Polyethyleneterephtalate (PETP)

Polycarbonate (PC)

Polyamide 6.6 (PA)

Cellulose acetate

Polypropylene (PP)

Polymethylmethacrylate (PMMA)

Polyoxymethylene (POM)

Polytetrafluoroethylene (PTFE)

Mild to moderate damage, utility is often satisfactory

Moderate to severe damage, use not recommended

Epoxy, glass laminate

Phenolic, glass laminate

Phenolic, mineral filled

Aromatic cured epoxy (special formulation)

Silicone. glass-filled

Silicone, mineral-filled

Polyester, glass filled

Polyurethane (PUR)

Polyester, mineral filled

Silicone (unfilled)

Epoxy (EP)

Phenolic (unfilled)

Melamine-formaldehyde (MF)

Urea-formaldehyde (UF)

Polyester (unfilled)

Anilline-formaldehyde (AF)

(a)

(b)

Figure 30.25 General classication of radiation resistance. (a) Rigid thermoplastics. (b) Thermoset resins

and composites. (Reprinted from Tavlet, M. et al., Compilation of radiation damage test data, Part II, 2nd

edn, Thermoset and thermoplastic regins; composite materials, Report No. CERN-98-01, CERN, Geneva

Switzerland.

With permission.)

Radiation Effects on Semiconductors and Polymers for Space Applications 865

In the aromatic polyimides, the imide rings and the linking groups between the aromatic groups

within the polymer chains, principally determine the chemical, thermal, and radiation stability

(Inoue et al., 1987). In addition, the concentration of the aromatic groups inuences the sensitiv-

ity because these groups offer protection against structural damage due to exposure to electron or

gamma irradiations (Inoue et al., 1987). This aromatic protective effect is believed to occur via the

ability of aromatic groups to dissipate absorbed energy to heat through their manifold vibrational

states (O’Donnell and Sangster, 1970; Megusar, 1997). So, polyimide is widely used for thermal

control

lms and lubricants for space use because it has a high radiation tolerance.

The

mechanical properties have been examined by means of tensile tests after high dose (5kGy/s)

electron irradiation, which is similar to a vacuum environment, and a gamma ray irradiation test under

pressure with 0.7MPa oxygen has been performed for 12 kinds of polyimides such as Kapton, polyether-

etherketones (PEEK); the results are reported in Sasuga and Hagiwara (1985, 1987). Figure 30.26 shows

stress–strain diagrams after electron irradiation of Kapton H, which is a typical example of polymer that

expands without constriction, and amorphous PEEK which is a typical example of polymer that expands

with constriction. The property value of breakdown strength and ultimate elongation is greatly inuenced

by the irradiation compared with the property value of yield strength and Young’s modulus. Figure 30.27

200

150

100

PEEK

10 MGy

100 200

(a)

(b)

300

0 10 20

Strain (%)

KAPTON H

Stress (MPa)Stress (MPa)

90%

2200 kg/cm

30 MGy

120

Figure 30.26 The stress–strain curves of (a) polyimide “Kapton” and (b) non-crystalline PEEK irradiated

various doses. (Reprinted from Sasuga, T. and Hagiwara, M., Polymer, 26, 1039, 1985. With permission.)

866 Charged Particle and Photon Interactions with Matter

25 MGy

γ rays in 0.7 MPa O

20 16 12

Dose (MGy) Dose (MGy)

8 4 0

0 20 40 60

: 50% of initial value

m/n =80/20

PPO

m/n =70/30

m/n =50/50

Electron beam

: 20% of initial value

80 100 120

Figure 30.27 Relation between radiation resistance and chemical structure of aromatic polymers under oxidative and non-oxidative irradiation conditions. 50%

of the initial value and

▭ 20% of the initial value. (Reprinted from Hayakawa, J., Useful Life and Its Prediction for Polymer Material [in Japanese], IPC, Japan, 1989.

With

permission.)

Radiation Effects on Semiconductors and Polymers for Space Applications 867

shows the dose at 50% and 20% elongation compared to the initial values for polymers under oxidative

and non-oxidative irradiation conditions. When even the aromatic polymers are irradiated under oxygen

environment, they are degraded at the dose from one-tenth to one-fth compared to a non-oxygen envi-

ronment (Sasuga, 1988). The radiation tolerance for the aromatic polymers is very different based on the

chemical structure. Figure 30.28 shows the residual elongation of 50% and 20% of the initial value as

a function of dose for each polymer. The radiation stability of the units that make up the main chain is

estimated by comparison of the chemical structure and the residual elongation as follows (Sasuga and

Hagiwara, 1985):

1. Polyimides such as Kapton and UPILEX show extensive radiation resistivity, indicating

that the aromatic imide ring has excellent radiation stability.

2. The two aromatic polyimides, A-Film and APH-50, show relatively high stability for radia-

tion,

so the aromatic amide is a highly resistive structure for radiation.

O O

N R

Kapton

UPILEX

ULTEM

APH-50

A-Film

PEEK

U-Polymer

Udel-Polysulfone

Crystallized PEEK

(× % –15)

PES

NORYL

0 10 20 30 40

Dose (MGy)

50 60 70

Figure 30.28 Relation between chemical structures and residual elongation. (Reprinted from Sasuga, T.

and

Hagiwara, M., Polymer, 28, 1915, 1987. With permission.)

868 Charged Particle and Photon Interactions with Matter

3. Non-crystalline PEEK also shows very high radiation resistivity, indicating that the

aromatic ether and aromatic ketone are stable under irradiation.

4. The polyetherimide ULTEM shows radiation resistivity of only several MGy, although it

consists of a radiation resistive imide ring and aromatic ether. Similarly, the polyarylate

U-Polymer, is spite of containing a stable aromatic ether and ester, shows low radiation

resistivity. These two polymers contain bis-phenol A group in the main chain. It is consid-

ered

that the bis-phenol A group is a low radiation stable structure.

5. Since the two polysulfones and the aromatic sulfones show very low resistivity, the aromatic

sulfone group does not have high radiation stability.

6. The modied poly (phenylene oxide) NORYL shows low radiation resistivity. The phen-

ylene oxide unit cannot be considered a weak structure for radiation on the analogy with

the case of aromatic ketone and ether. The modication, by blending with polystyrene for

example,

may weaken the radiation stability of this polymer.

The

mechanical properties of high performance polyimide, which has optically transparent char-

acteristics, are studied (Devasahayam et al., 2005). The mechanical properties of four optically

transparent polyimides prepared from the dianhydrides ODPA and 6FDA and the diamines ODA

and DAB were assessed. The four polyimides were synthesized with the objective of obtaining opti-

mum transparency for space applications where high-energy radiation doses of 15–20MGy may be

expected in geosynchronous orbits over a life span of 20 years. The polyimides were shown to main-

tain good optical and tensile properties at temperatures up to about 450K for a dose of 18.5MGy,

but above this temperature the module of the polymers began to deteriorate and there was a small

decrease

in the transmittance of the exposed polymer lms.

30.4.2 ground evaluation facility

In ground simulation tests, single, sequential, and simultaneous irradiation tests using AO, UV, and

radiation sources have been carried out to simulate the space-environmental degradation of polymer

materials (Kanazawa et al., 1992; Marco and Remaury, 2004; Miyazaki and Shimamura, 2007;

Shimamura

and Miyazaki, 2009).

Equipped

with AO, vacuum ultraviolet (VUV), and electron beam (EB) sources, the Combined

Space Effects Test Facility in JAXA can irradiate these beams simultaneously into an irradia-

tion chamber with a high vacuum of about 10

−5

Pa (Miyazaki and Shimamura, 2007; Shimamura

and Miyazaki, 2009). The standard specications of the facility are listed in Table 30.1; its sche-

matic view and the visual appearance of the facility are shown in Figure 30.29. The linear motion

feedthrough is installed in the irradiation chamber, which is located over the sample holder and

which can move parallel to the sample holder’s plane. A photodiode or a quartz crystal microbal-

ance (QCM) can be attached on the linear motion feedthrough. The AO generation in the facility is

based on the laser detonation phenomenon invented by Physical Science, Inc. The AO beam source

is composed of a pulsed valve and a pulsed CO

laser at a wavelength of 10.6μm with a pulse

energy of about 10 J. The translational energy of the hyperthermal AO produced by the facility

is controlled at approximately 5 eV to replicate the LEO environment. The VUV sources, a total

of 48 deuterium lamps (L2581, Hamamatsu Photonics K.K.), are installed in the lamp chamber,

which has been purged using high-purity Ar gas. The VUV beam from the lamps is concentrated

by the CaF

lens and a collection mirror coated with Al and MgF

. It is then injected into the irra-

diation chamber through an MgF

plate separating the lamp chamber and the irradiation chamber.

The facility named SEMIRAMIS, which was designed at ONERA in France for the evaluation of

thermal control coatings in a simulated space environment, is often used to simulate LEO or GEO

orbits (Marco and Remaury, 2004). Its main characteristics are cleanliness (very low organic resid-

ual partial pressures in vacuum) and reliability (samples in vacuum for several months). Cleanliness

is achieved by the use of stainless steel, the use of a majority of metallic seals, and the exclusive

Radiation Effects on Semiconductors and Polymers for Space Applications 869

use of cryogenic pumping units, each equipped with a gate valve. The reliability is increased by the

implementation of each function in a different vacuum chamber separated by its own gate valve and

redundant pumping units whose associated gate valve is automatically closed upon failure. The main

facility is composed of four vacuum chambers. The lower part is the irradiation chamber containing

three irradiation ports, one for protons normal to the samples, and the others for electrons and UV

in two symmetrical directions at an incidence angle of 30°. At the upper stage, is a chamber used for

measurements and the ultimate safeguard of the vacuum in which the sample holder can be introduced

by a vertical translation from the irradiation chamber. On one side of the upper chamber, an optical

measurement device storage chamber is mounted, and on one side of the lower chamber there is a con-

tamination device storage chamber. The proton and electron beams are supplied by 2.5 and 2.7MeV

Van de Graaff accelerators, respectively. The protons are obtained from plasma of pure hydrogen

and separated from other charged species by a magnetic mass analysis after acceleration. In order to

irradiate the samples, the protons are swept across the sample holder surface. In the case of electrons,

the beam is diffused through a thin aluminum window. The solar UV generator is based on short arc

xenon sources (4000 or 6500W) whose spectral distribution in UV is close to that of the sun. The

source emission is ltered in order to reject all visible and infrared light above 400nm and reproduce

only the UV outer atmosphere emission of the Sun. Multiple solar constants calculated in the range

of 200–380nm can be applied over the sample holder surface (1 Thekaekara solar constant is 9.46

mW/cm

). The number of laboratory irradiation hours multiplied by the number of used solar constants

gives the equivalent amount of sun hours (ESH) of space. Figure 30.30 shows the visual appearance of

the SEMIRAMIS (ONERA, http://www.onera.fr/desp-en/technical-resources/semiramis.php).

table 30.1

standard

pecications

of the Combined s

pace

ffects

est

Facility

items specication

Vacuum 10

−5

Pa (10

−3

–10

−2

Pa during AO irradiation)

Sample Size:

diam. × max 3

thick

Exposure area: 20

diam.

Quantity: 16 samples per holder

AO Method: laser detonation

Laser: pulsed CO

laser

Laser

wavelength: 10.6

μm

Pulse

rate: 12

Laser

output power: ∼10

J/pulse

Beam

velocity: 8

km/s

Translational

energy: 5

Flux:

∼5 × 10

atoms/cm

VUV Source:

deuterium lamps

Number of lamps: 48

Lamp current: 250–350

Tube

voltage: 70–90

Flux:

0.3–0.5

mW/cm

(Integral intensity at a wavelength of 120–200nm)

EB Accelerating

voltage: 100–500

Source

current: 0.1–2.0

Sample temperature −150°C–80°C

Source: Shimamura, H. and Miyazaki, E., J. Spacecr. and Rockets, 46,

241–247,

2009.