ML20198F931
| ML20198F931 | |
| Person / Time | |
|---|---|
| Site: | 07109033 |
| Issue date: | 12/22/1997 |
| From: | Charles Brown NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Chappell R NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| References | |
| NUDOCS 9801120212 | |
| Download: ML20198F931 (24) | |
Text
- _ _ _ _ _ _ - _ _ - _ _ _ - _ _ _ _ _ - _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _
- December 23, 1997 MEMORANDUM FOR: Ross Chappell, Chief Pcckage Certification Section SFP0 FROM: Christopher Brown, Materials Engineer Package Certification Section, SFP0
SUBJECT:
MEETING
SUMMARY
CONCERNING AMERSHAM'S TEST PLANS FOR THE MODEL 660 RADIOGRANlY DEVICE
&ljendees:
BRG Amersham NAC International David Tiktinsky Khaja Afeef Kris Kunert Nancy Osgood Cathleen Roughan Bernie Waite Steven Grenier
.Li Yang William McDaniel Chris Brown Greg field Andrew Gaunt Ross Chappell Andrew Barto Daniel Huang Introduction A meeting was held at the request of Amersham Corporation at Rockville, Maryland, on December 9, 1997,.concerning failure of the Model 660 radiography device. Amersham discussed several corrective measures to improve and prevent failure of the package during hypothetical accident conditions, in particular, the use of stain'iess steel end bolts.
Discussion
- 1. Discussed briefly the failure analysis conducted on the Modei 660,
- 2. Discussed Amersham's corrective measure to prevent failure of the N package by using high strength stainless steel end boltr instead of carbon steel end bolts, in addition, drop and puncture test orientations were discussed with emphasis on causing damage to the l }
stainless steel end bolts.
- 3. Discussed Amersham's package drawings and revisions to the Certificate of Compliance.
Docket No.: 71-9033 Attachments: Meeting Handouts Y Distribution: 'NRC FC NMSS r/f NRC FDk SfP0 r/f Attendees w/o Attachments CJHaughney SFShankman-To receive a copy of this document, indicate in the box: *C' Copy without attachment / enclosure
'E' - copt with attachment=/ enclosure *N* - No copy =
- . -. w
NAME CLBro[nk NL0sdhli ERJdgler CRChad DATL 9901120212 471222" 12[d[97 2))[97 [2/ /97 1242f/-97 POR ADOCK 071 g w J
[Gh4 o1. Luc 'll
=
AC .
EC --
45ao i'acinc iiich> r t=t tcom . W :hington 08424 2G33
- ck ging echnology,Inc. O (253) 9221450 D (253) D221445 s
, f;\\
November 25,1997 d u=nr un acages ox fit 97012
[vla Fax (617) 273 2216)
Ms. Cathleen Roughan Regulatory Affairs and Safety Manager Amersham Corporation 40 North Avenue Burlington, MA 01803
SUBJECT:
Failure Evaluation Report of Model 660 Series Projector
Dear Ms. Roughan:
Please find enclosed the PacTec's Failure Evaluation Report of the Sentinel Model 660 Series Projector for your information and use. This evaluation was based on the information which I received and observations which I made during my visit to the Amersham offices on November 10 - 11,1997.
Should you have any questions or require additional information regarding this report, please contact me.
Veiy Truly Yours, Packaging Technology, Inc.
h5$
' Gary L. Clark, P.E.
Vice President ,
encl: As stated
F. 4
- 1 .
Amersham Corporation November 25,1997 ,
Failure Evaluation of Sentinel Model 660 Projector introduction The Model 660 Series Projector is an US Nuclear Regulatory Commission (NRC) certified Type B(U) package in accordance with Title 10, Code of Federal Regulations, Part 71 (10 CFR 71). The package consists of a source tube enclosed within a depleted uranium (DU) shield, a rear endplate with a locking assembly, a front endplate with a storage plug assembly, four connecting rods, a stainless steel st. ell, and polyurethano foam. The DU shleid is totally enclosed by the foam, which is in tum surrounded by the stainless steel shell and the endplates. The endplates are fastened via the connecting rods using eight (0) %-Inch flathead machine screws. At the top of the package, a cast aluminum handle, which is fastened by two (2) 10-32 machine screws through each endplate, spans between the two endplates for lifting and moving the package. The Model 660 Series Projector has a maximum gross weight of 56 pounds and is illustrated in Figure 1.
countctuc noos [^5' v "$'y>
no m ,vu aca"*c
" mu s'o g s NL, st~ sit e > g ou s,o
\ / j c=gt,s o,
\ / -ss, scc sm a somem-
'O' f Q W V
.- ? [
~
""\_1 p 'kk' D m
8 8s Eft s'W ^"' ~
sc ' r ~ ~, 3 0d \
M
/ '
/ / \
ESEE/'" [' " ' ' " ' " '
/ Ewan-t ItEE'4t s sE"/4"C""
. / @dP eat 5sioato sit sHT s ut sui s / =switto - - 9l rosmou)
/ ul
=;,w Figure 1 Model 660 Series Projector
- p. n ACesasEC
~mw am 1 (W5ent660 dx
. A'mersham Corporation November 25,1997 in response to a Confirmatory Action Letter (CAL), the NRC has required Amersham to re-test this package to demonstrate compliance with 10 CFR 71. Amersham prepared a Test Plan (Test Plan 70) delineating the details of the tests to be performed. The Test Plan included the normal conditions (NC) of transport tests per 10 CFR S71.71 as well as the hypothetical accident conditinns (HAC) of transport per 10 CFR 971.73. As required by the cal., Amersham submitted the Test Plan for review and approval by the NRC.
The Test Plan mandated four test articles (A, B, C, and D) for four specific drop orientations. The testing was performed with each test specimen chilled to less than
-40 'C to simulate worst-caso initial test conditions. Following the drop and puncture tests, each specimen was to be placed in a furnace to comply with the thermal test requirement per 10 CFR $71.73(c)(4). After the thermal testing, each specimen was then to be evaluated to demonstrate that radiation level did not exceed 1,000 mrom/hr at a distance of 1 m from the external surface of tha package [10 CFR S71.51(a)(2)).
Description of Failure Of the four test articles, one article (Specimen D) sustained a failure of the two upper M inch flathead machine screws on the rear end plate. The 30-foot drop orientation which caused this screw failure was the CG-over the bottom edge with the package oriented at approximately 45' from the horizontal plane. Upon impact, this orientation imparted a deceleration which drove the DU shield via the source tube into the rear endplate. The endplate was plastically deformed from this force and moved axially and rotated about the upper and lower ends, it is also clear from the photographs and package design that the deformed condition of the rear endplate is a deflection-limited rather than a load-limited condition. The failed screws failed across the root diameter of the threads and showed signs of bending-plus tensile load. Thus, the rear endplate was attached to the pahage body by the two lower %-inch flathead screws and the aluminum handle. lhe front endplate and its fasteners were not damaged by this drop and subsequent puncture tests.
Following the puncture test directly on the front endplate end of the handleiattempting to drive the rear endplate off), Specimen D was oriented at an incline with the rear endplate down and plcced into a fumace. Shop air was supplied via a tube into the furnace at a rate of 10 CFM during the test. The test article was then thermally soaked to a minimum of 800 'C prior to initiating the 30-minute regulatory thermal event. When the specimen was removed from the furnace, the rear endplate was only attached by the two lower % inch flathead machine screws. Because the handle material was aluminum, the handle melted and thus allowed the DU shleid to further move the endplate and expose the DU directly to the radiation heat transfer of the furnace's heating elements. With this thermal heat transfer mode, the DU shleid was heated te a temperature which provided a sustaining oxidation reaction for approximately four dhys following removal from the furnace. Once the specimen had cooled to ambient temperatures, is was found that a large portion of the DU shield had been oxidized and
[DACTEC
.L a.,4 % + 2 rae sentsso e
. =. __.
Amersham Corporation November 25,1997
, removed from the solid DU. With this DU removed, the rear portion of the source tube 1 was exposod Based on an X-ray examination of the test specimen, an apparent unshielded path to the source resulted from the loss of the DU.
Discussion of itotential Failure Factors i in reviewing this test documentation (video, photographs, radiographs) and the failed !
flathead screws, a number of factors can bt, identified which may have had a minor or major role in the overall failure. These factors include the following:
. Thermal test being too conservative ,
. -40 'C initial temperature condition e Gaps between the connecting rods and endplate e Commercial grade flatnead screws e Foam density
- Damaged threads on the ends of the failed screws Each of these factors are individually discussed below.
Ihermal Test The thermal test which was implemented for the Model 660 testirig required that the test specimen be heated uniformly to 800 'C prior to initiating the regulatory 30-m.nute test. This preheating test approach was to address the forced convection requirement mandated by 10 CFR 971.73(c)(4). The preheating of the test specimen to a uniform tempcrature of 800 'C effectively exceed the amount of heat input which is required by the regulations. At first glance, this excessive amount of heat input may have had a significant contribution to the loss of the shielding. An indication of the effect of this excess can be determined by an examination of the heat-up rate of the intemal DU shield thermocouple. -
From the video tape, temperature readings were obtained as a function of time. These readings were then plotted and are shown in Figure 2. As shown by this data, the intemal temperature rose at an approximately linear rate of 53 'C/ min through 400 *C and 24 'C/m2n from 400 *C through 750 *C, Within 15 minutes, the internal temperature ,
- exceeded 600 'C. If the specimen had been exposed to an actual fire test, a similar temperature rise would be expected. Test Specimens A and B were also exposed to this preheating approach and successfully passed the post-test radiation survey. The difference between these specimens and Specimen D was that Specimens A and B
- retained the DU shield within the cavity formed by the stainless steel shell and endplates while Specimen D did not. With the DU directly exposed to the heating elements, the surface temperature of the DU would actually be higher than that recorded by the internal thermocouple.
IDACTEC
'1 % , L ,,w 3 swsontsooooc
- 1
- a. -
Amersham Corporation November 25,1997 F,or there reasons, it is concluded that the excessive heat input due to preheating was not a major contributor to the failure. The primary cause of the loss of DU was the direct exposure of the material to the heating elements of the furnace, i i
e00 l :
600 s ---
700-600 600 -
I 100 -
200 100 i !
p
, 4 ,
OJ 00 0 07.12 0 14740 21 36 i 20400 M 00 0 .3120 60240 67.M 0 Elasped T6me (min:sec)
Figure 2 Model 660 Specimen D Thermal Test - !nternal Temparature loltlal Tampatature Condillen The test specimen was chilled to less than -40 'C prior to each drop test. T,his pre-chilling was to address the potential brittle fracture effects on the carbon steel materials used in the Model 660 design. The structural components which are fabricated of carbon steel are the endplates (0,120 inch thick), the connecting rods (0.375 inch diameter), and the flathead machine screws (0.25 inch diameter) A secondary area of concern was the behavior of the DU at low temperatures, To assess the importance of this loa temperature, a review of Recommendations for Protecting Against Failure by Bn'ttle Fracture in Ferritic Steel Shipping Containers IJp to Four inches Thick (NUREG/CR-1815) was performed. Per this document, the Model 660 package is classified as a Category lli structure (i.e., special form, less than 30A, and lecs than 30,000 Cl), For Category lli structures, no brittle fracture requirements are specified for mat:: rials less than 0.4 inches thick, Because all of the Model 660 carbon steel materials' thicknesses are below this value, brittle fracture is not a AC EC
'adasinsfahnes aw 4- rde:Sent660 doc l
}. A'mersham Corporation November 25,1997 c,oncern. For this reason, the initial temperature condition is judged to have no significant effect on the flathead screw's failure.
One effect of the -40 *C initial temperature is the stiffening of thre polyurethane foam.
The foam utilized in the Model 660 package has a nominal density of 8 lbs/ft' at room temperature. At the lower temperature, the foam will respond similar to a higher density foam. An illustration of this stiffening is provided in Figure 3.
30000 2600 0 <
- 000 0 -
$ I g 16J0 0 -
~
j-y 20 peg'eetF [
- *
- koorn Temteresom o
- 000 0 7
[ --
l 600 0 ---
,c A*****
j ..... ....y.....
0 00 0 10 0 20 0 30 0 40 0 60 0 60 0 70 0 80 Strain (inAn)
Figure 3 Polyurethane Foam (8 lbs/ft') Compressive Stress Strain Behavior
(
Reference:
General Plastics Last A Foam FR 3708)
At a -40 C temperature, additional stiffening would be expected than what is presented here. The primary effect of foam stiffening on the Test Spec.imen at the lower temperature would be to possibly generate higher impact forces. In turn, the higher terultant impact foice may have resulted in a reaction load on the flathead screws which contributed to their failure. However, based on the amount of foam and the steel construction of the package, the contribution of increased foam stiffness due to low temperatures to the screw failure is probably not significant.
l,DACT_EC
.La.,,,<_.L-o.n m S Ne sentsso coc
t Amersham Corporation N:vember 25,1997
- . Gaps Balwagn_Connacting Rods _and Endplaig The presence of a gap between the connecting rods and rear endplate is a possibility for several reasons. First, the fabrication toleran:es of the individual components can result in a maximum gap of %-inch. Second, the tightening torque of 30 in lbs may not have been adequate to maintain a tight joint through the normal condition tests which preceded the test that failed the fiathead screws. The effect of a gap between the endplate and the connecting iod would be to allow a benalng moment being applied di Xtly to the threaded portion of the screw. If the joint is r#a'ained in metal-to-metal corrtact (rod to endplate), then no bending moment would e wetted on the screw shank.
Examination of the failed screw heads provided evidence that the screws where loaded by both a bending moment and a tensile force when the DU shield pushed on the endplato. Additionally, it.e screw failure surface extended beyond the rear face of the er dplate. From this evidence, it can be concluded that a gap did exist between the sorows and the upper connecting sods.
Bt.ited jointe fallinto two categories: tencile or tension joints and shear joints. Bolts and fasteners shw , never ta t ubjected to loads or forces which are not part of the ir, tended design load conditi)n nor in uending. The design oNhe Model 660 projector is such that these flathead screws P intended to be a tension joint and therefore, only subjected to a tensile forces. , c. a , msile joint, the fasteners should generate a sufficient clamping force in prevent separation of the jolnt members (i.e., connecting rod and endplate). For the current material, the tensile strength of the screws (which is futher discussed below) is questionable in being capable to develop this sufficient clamping force to maintain a tight joint. Given the small root diameter cross sectional area of the % inch screws (0.0269 in') coupled with the stress concentration of the threads (approximately an increase of 2.3), a small rotational deflection of the endplate will result in a bending moment which will fall the flathead screws.
Therefore, it can be concluded that a gap between the connecting rods and the rear endplate may have been a significant coninoutor to the failure of the flathead screws.
CommonlaLGrade_Eastenern The %-inch flathead screws which have been utilized in the fabrication of the Model 660 Series projectors are procumd as commercial grade fasteners. This type of fastener is t,upplied without material test reports (MTRs), either certified or typical, and as such, the actual mechanical properties are unknown. If it is assumed that these fasteners are equivalent to the lowest SAE grade which specifies properties (Grade 1), their minimum ultimate tensile strength would then be 55,000 psi. Using the tensile stress area of the screws, tensile failure can be estimated as follows:
Estimated Failure load = S,, x (tensile area) = SS,000(0.0318 in') = 1,7491bs/ screw AC EC ewin a s.a.is ii. 6 fue'Senf660 dx
- r .
t-Amersham Corporation November 25,1997 Although bending of the endplate was observed, an estimated impact deceleration can be computed which would result in failure of the flathead screws ifit is conservatively assumed to neglect this bendirm as well as assuming that only the failed upper two screws reaci all of the DU shield load. For these conservative assumptions,'(e failure impact would then be: -
~ 2 (1,749) '3,49F' impact g's = ~
40 ,
if the minor diametrical area is substituted for the tensile area, the impact deceleration '
falls from 87.45g to less than 74g. Based on previous drop test data for r 'ckages (e.g.,
125B TRUPACT-il, RTG, etc.), it is conservatively estimated that Specin 1 D l
experienced a minimum vertical deceleration of 200g For the test inclination of approximately 45', the resultant impact deceleration parallel to the axis of the screws would be in excess of 1409, .<hich is significantly in excess of the current screws' capability.- When the bending moment is added into the loadir.g condition, the actual strength of the joint is greatly reduced from that estimated here.
- From this rough analysis, it has been demonstrated that the quality and strength of the flathead screws played a pivotal role in the failure of the test specimen.
Foam Density in discussions with Amersham personnel, it was noted that over the life of the Model 660, the type polyurethane foam has been changed from one manufacturer to another.
There has been some noticeable change in the " hardness" oithe foam between these two manufacturers. As noted above in the discussion of the low temperature effects, a harder foam would generally result in higher impact loads. However, because of the relative higher stiffness of the steel structure, minor changes in the foam density or hardness (on the order of 5% or less) are judged to have only a very minor, if any, effect on the failure of the fiathead screws.
h Damaged Thr.eads on Failed Screws During the disassembly / assembly of Spe:Imen D to determine the CG location, it was noted that the ends c% Ici ad screws had damaged threads. Even though these screws could have beco W m..fi prior to the testing, Amersham personnel intentionally uti!! zed them in Specimen D to simulate an worst cace condition. Although damaged threads can affect the strength cf the fastener, the fact that this condition occurred on
- the ends of the screws had no bearing on the screws' strength. This conclusion can be supported by t' s fact the screws failed closer to the head rather than near the damaged threads in addition, the strength of any fastener is developed fully within a length equivalent one diameter,P. this case 0.25 ir.ches. Based on these data, it is concluded that damaged threads att te ends of the screws played no role in the failure.
l, ACTEC ,
m usine .L =had o im I Se:Sant660 doc
m + .. -
. - n< y: ' Amersham Corporation ' Nov:mber 25,1997- l j
,. Lflocommendations to Preclude Failure : ,,
o : Based on the examination of the failed flathead screws, the previous discussion, and
- the design of the endplate-connecting rod joint, there are several possible retrofits for1 ;
1 existing packages which would result in a high confidence to preclude the failure that j 1was experienced by Specimen D< Additionally, there are possible' design- j
- enhancements which could also be implemented to increase the overall design margin:
- of future fabricated Model 660 Projector units.1 Although recommeNiations'are being
! made, they are only recommendations. It should be noted thal. thme recommendations -i are not all inclusive l other options are possible to preclude tM obseved failure. The - ?
. _ final design modifications and/or design enhancements which a re lo be implemented l l should be determined by knowledgeable Amersham personnel considering a!I ' j appropriate factors: cost, manufacturing, operational, and reguatory.
n These recommendations are listed in the order of importance and ease of j
. Implementation to reduce or prevent the observed failure. _ At a minimum, the strength L E of the endplate screws should be increased from the current commercial grade quality
- screws.
- 1. Change Flathead Screw Material it is apparent through a rough analysis of the failed flathead screws that the most
- effective prevention of failure of the fasteners is to increase the strength of the joint. .
b Since the current screws are the weakest structural member, they should be replaced with a higher strength screw which are supplied with MTRs. With a higher strength -
E screw, the tightening torque at installation should be increased to the maximum
, possible value, based on both the internal threads of the connecting rod and the new screw material strength. The objective would be to increase the strength and tensioning of the joint so that the joint does not separate due to the impact forces, if the joint does not experience an impact load exceeding the preload, then it will not fail crhe any condition.
Ne e aslest method to increase the preload in the screw is to increase the tightening m:c;e during assembly. For example, if the screw was an SAE Grade 5 ot8 flathead '
E screw, the maximum torque which could be applied is 10 or 14 ft-lbs respectively (based on a minimum proof strength of 85,000 psi [Gr.5) or 120,000 psi [Gr. 8]). These
- - values compare with a maximum of 54 ft-lbs for the shear tearout of the internal threads.
In the 12L14 connecting rod (assuming a 0.5 inch thread engagement).- Therefore, the
- controlling element is the screw itself with a tightening torque value of 10 or 14 ft-lbs.
- . At the 10 ft-lbs value, the initial tension had would be 50% greater than the current
- value developed by tightening torque of 80 in-lbs (6.7 ft-lbs).
Additionally, since the screw failure was a deflection-limited condition, another design - 1
- approach to preclude failure.would be to ensure that the screw's strength and ductility is
- sufficient to withstand the imparted deflection. Without actual deflection measurements i
of the test specimen's endplate, material selection for the screw will need to be based on test rather than analytical methods.
t IDACTEC a a.m =-- + 8 m:sontseon i L
4 v,' . , - ..
~.O - - , , . . . - - , , - -.~i..i...~r ,,..,,e .,,, E .-v,,, ,,,n, --h- , - - -, y
m ,.
I
..-. Amersham CCrporation - November 25,1997 Without impact accelerations, it can not be specifically determined that increasing the -
- screw strength alone would be sufficient to preclude failure. However, it is apparent from the observed failure mechanism that the flathead screws require additional strength to increase the capability of the joint. Re-testing of the Model 660_ with stronger screws in this specific drop orientation will provide the objective evidence of whether the strength a!or.a is sufficient to preclude the observed failure.
2.: Stiffen Endplate-Constectina Rod Jelnt in addition to increasing the screw strength, a most positive method of eliminating bending of tk.a flathead screws would be to add a close-fitting machined boss around the connecting rod. This boss would also be chamfered to mate up with the screw -
recess in the endplate. An illustration of this boss is shown in Figure 4.
When the DU shield impacts the endplate, the deflection of the endplate would result in a bending moment at this joint. However, rather than the screw reacting the bending moment, the boss would then react the moment as a couple between the connecting rod and the endplate. As long as the screw remains tight, this joint configuration would prevent failure of the flathead machine screws.
.g fr/7/mW '~
% 3=3 AWi i 3 4 '
Figure 4 Proposed Joint Reinforcement Configuration
- 3. Change Handle Material it was apparent from the testing of Specimens A and B that if the DU shield is maintained within the cavity created by the endplates and the stainless shell, the DU shield will not experience significant loss of shielding material. If the handle had not melted and allowed the endplate to be further dislocated by the DU shield weight, then the performance of Specimen D to the thermal test would have been similar to the other AC EC unaema w % ir 9 me:sessaoooc
- - - . _- - - _ _ _ _ _ _ _ ~- -. . _
- .E :,wp , _
, l. >9'~ Amersh:m Corporati:n November 25, igg 7 ,
,. test' specimens, even with the failure of ths two upper flathead screws. Therefore, a : -
' handle fabricated of a material which has a molting temperature above the test --
. environment would have maintained the DU shield within the surrounding cavity. .
i4h Stiffen Endplata j
- Another method of eliminating bending of the endplate and on the screws would be to 2 l increase the stiffness of the endplate. In reviewing the video and photographs of the
' test specimen, it was noted that the locking mechanisrn provided some stiffening,- but did not cover sufficient plate area to eliminate bowing of the endplate.LThe most _ !
e effective way to achieve this would be to weld longitudinal stiffeners on the back side of each endplate (l.e., increasing the moment of inertia of the endplate).L The result of -
stiffening the endplate would be to react the impact load as direct tension and distribute 3 the DU shield impact load over more fasteners and thus reduce the potential for failure. - 1
[ <
The following recommendations fall within'the category of design enhancements and as
- such, should only be considered in the fabrication of future units.
- 1. Attach the DU Shield Directly to the Endplates
.it is apparent from the esults of the Specimen D testing that if the DU shield was
.- restrained by the endplates, then is highly likely that the shield would have remained .
I Twithin the steel cavity, Currently, both ends of the DU shield are attached to the .
endplates via slip connections on the titanium S-tube which provide little axial restraint.
l Connecting the DU shield to both endplates could be accomplished by several
- - methods. Possible methods include: 1) directly bolting through the endplate and into thresded holes in the DU; 2) adding brackets to the connecting rods which capture the shield; or 3) a combination of these two methods.
_2. Coating of the DU Shield o
in discussions with a Starmet representative (manufacturer of the DU shields).-it was' e disclosed that DU will oxidize at all temperatures. The only variation of this oxidation is L the rate at which it occurs.- At room temperature, the oxidation rate is low. However, at -
- elevated temperatures, the oxidation rate is very high and can be self-sustaining in the
[ right geometry. This behavior is what occurred with the DU shield of Specimen D.-
- To preclude the oxidation through all temperatures required by 10 CFR 71, the application of a coating or cladding which can withstand elevated temperatures (i.e., > ;
' 800 *C) would be required. 'A coating may be metallic which is applied via flame =
spraying or organic which provides a oxygen barrier upon combustion.- _ Cladding would u require a seal welded, thin steel sheet which surrounds the DU material. Other possible 7 protective DU coatings for this application may exist which could be investigated with - :
the DU shield supplier.
i, ,
INACTEC
- .n..a . .n.-s e 10 me:semesom
NA\
The informotion you need...when you need it
- November 24,1997 Mr. Steve Grenier Amersham Sentinal 40 North Avenue Burlington,MA 01803 Report No: 33631 P.O. No: Pl726 PURPOSE OF ANALYSIS:
To perform a fracture analysis of a screw and to determine the composition of the screw, SAMPLES:
The top part of one (1) fractured screw.
METHOD OF ANALYSIS:
Scanning Electron Microscopy (SEM)
Energy Dispersive X-Ray Spectroscopy (EDS)
CONCLUSIONS:
The primary fracture feature on the failed surface of the screw was tensile overload, but there was some evidence of quasi-cleavage due to high impact loading on the fracture face of the failure.
The screw is made of a carbon steel.
RECOMMENDATIONS:
To determine the precise alloy used here requires a more sensitive and accurate technique, such as Inductively Couple Plasma Spectroscopy (ICP).
RESULTS:
The screw was photographed as-received (see photographs M1 and M2).
After optical examination the screw sample was mounted on a Scanning Electron Microscope
_ (SEM) specimen mount and placed in the SEM to examine the fracture surface. Micrographs 4 Arrow Drive,Wobum,Massachuse'ts 01801 A Telephone:(617)938-0300 A Fax:(617) 935-5087
c n
were taken at the top of the sample away from the screw head (see photographs M3 and M4), at middle of the fracture (see photographs MS and M6) and near the base of the screw head (see photographs M7 and M8). In each location that was photographed, the cause of the fracture was tensile overload.
Energy Dispersive X Ray Spectroscopy (EDS) analysis was also performed to check the composition of the screw. This was done at a location near the top of the fracture 'Ihe composition was almost all iron (Fe), with small amounts of silicon (Si) and potassium (K) detected as well Carbon steels typically contain a fraction of a percent of silicon, but some contain 1 2 pucent. The potassium is probably an extemal contaminant.
The enclosed data present the results of the analysis.
The enclosed data sheets further describe the Scanning Electron Microscopy (SEM) and Energy
- Dispersive X Ray Spectroscopy (EDS) analyses.
N Charles F.Tuson Failure Analyst / Microscopist CFT/dmh
Enclosures:
Samples: 1 Micrographs: 16 Spectra: 1 Data Sheet: 2 Evaluation: 1
> Analytical Answers,Inc.
.?".
7, p # '
', ,' N.~ -j J' [ s is an optical micrograph of G
ig"Q~ Q,k 4 p 5-y, the fractured screw shown x
- mY,!, h .T6 'i as received. The arrows
'[ ). t ' - indicate the areas of the fracture !
I m.:. ,, [k r[k '
.I .
i r. .: that were photographed in the ll f,); g
\~ V ; . SEM.
- y i y "' ij. ; r m-)Kh' '"
l 1
?f kfhlklW f' g .
pH!f
+4W,Ign. .
r.f -
, .g i
M2 10X
- g .f ' . .
.v,.
This is an optical micrograph of the screw fracture face. The
'h
,, p 4 arrows indicate the areas of the
\ .y .- g .- .,
fracture that were photographed g
- 'O4 s i .. . :
in the SEM.
l.4 { ' ' ' 1
, , .t
?
j h ?
i
'"~
- . m;
.V , h , . h.f '
- ~.
i j
e Analytical Answers.Inc.
m ..; _.___x., ,s,,.
1 -
i e
y.
~. " ' -
M3 55X I
- {. . ]i sg'* '5N htgep#g.. ,[
- 8 -
'r - F ' ,,' ! Scanning electron micrograph
~
' J '. 3 ,!
i l (g c' (SEM) of the fracture at the top
- . Yf."ag,, g ig. 4(J p ' ;'k'D.
.E. ? *
' (away from the screw head).
0*
i i
I$t3% ,,, sg 35x, 'ssu % eiras n-ai-si
- M4 1000X s,
, f . ,. ; j't i
Scanmng electron micrograph
/- '
'! ' (SEM) of the top of the sample.
Note the tensile overload
% g~ pattern on the sample. The
.4 '
arrowindicates a single grain A ff ,
w' l that has evidence ofcleavage l
- ' C . --
, due to high impact load.
However, tensile overload l '7- l. O ' [ micro void coalescence is the
'p I,
l primary fracture feature.
- g. '
ms t
i ', O '
la I
i
,' ~
> Analytical Answers,Inc.
i i , .-
l_
I
'L' 8' % % ,>. hi., r * % n electre 3 micrograph
~ lc dk.' ..t9 ' # ','Y,
, s i (SEM) of the fracture at the M ,i . 0 3 -
middle of the sample.
Wik :
l l , ghj ,
=
i l
.' . pf 'gf n -
'e 1 f; l r; . , x%
~
- s' l $ '
M,
- 1. -
. '[% .f,, W-S4, .
~%
i ; , ,,. . . . ,
. . ~,.
\ . J' -
- r. .
~ ,. . o - - t ,
' Y "
l ,
. ., ; ,e l
. <' ;>, w - _-
.3
. 2%h,=
~
l :
. 4 E k ::.
. amo u asuv ssa sei . earn it-ai or i !
M6 500X
- Scanning electron micrograph 7 l t of the middle of the sample.
l i l . ! Note the tensile overload i pattem on the sample. The
- . t arrow mdicates a single grain j .
that has evidence of cleavage due to high impact load.
i
, However, tensile overload
~
I' micro void coalescence is the
. . . primary fracture feature.
4
! i
- . 2
{
4
- i h
> Analytical Answers,Inc.
! l ~. , . . .. ; M7 60X et Y ' ' sh ,, "4
, .rVr ; , Scanning electron micrograph s? - (SEM) of the fracture at the l
. :. N bottom (near the screw head).
l
- r. .
1 .
j .
A,f
. t
+
- l 1
! .' 't
, ;- Y
~* -I
! l
@, ..- ~.
- .. . , 7
. + . . ..
. e
- 3. ...r3 . . , .,. 1
.I: = l & d h $, . ;[ [ g, j
s
'2mo st asse asu N ' e1174 an-pi-gr I 4
- l- - . - !
l ' '
l .
l M8 Scanning 1000X electron micrograph
( (SEM) of the bottom of the
. . , sample. Note the tensile i
' overload pattern on the sample.
l ! The arrows indicate single grains that have evidence of a j l cleavage due to high impact t .
, load. However, tensile
- l , overload micro void
- ', coalescence is the primary l ,' fracture feature.
4 .
- . i 4 .
l l .:
a .
-_... _ _ __ _ e i
l <
t
- > Analytical Answers.Inc.
i
_ S
_ _ - i p
e m
c i
t g
r P
i u -
F i
m e . i .
j 3 i 3
. 6 i 3 i
1 -
g A -
i a
S i S i i i i e
i i
i i
i i
i R
a i
n .
,. K i g
e
,g i
i 4
e 0
i
.g S i
k C
i
_ i e
. R i V g E ;
i N W i i
C i M O i j
M i
. P i O
w S r I l
T I
i 5
O i T i o g N l t
>a
> l
- t C
9 o en u i n i t i s
=
i i 2 i 0 .
O i 9 Q i 8 i
3 i 7 i .
' L gl i i 0 i n ,
i e i a i
i r i
e i A i u i t i o i -
i V i S i =
- 1 2 3 I 3
0 0 0 i 6 0 0 0 .
7 J. _0 0 -0 ' 3
,MA\
Analpical Annt" ine.
SCANNING ELECTRON M1CROSCOPY (SEM)
Scanning Electron Microscopy (SEM) is a high resolution, great depth of field imaging technique. It shows topographical, structural and some elemental information at magnifications of 10X to 300,000X.
SEM Applicationsinclude:
- 1. Materials Evaluation:
Grain size distribution Surface roughness and porosity Particle sizing Materials homogeneity Intermetallic distribution Characterization of elemental diffusion
- 2. Failure Analysis:
Contamination location Examination for mechanical damage Electrostatic discharge determination Microcrack detection
- 3. Quality Control Screening:
Comparison of good to bad samples Material thickness determination Dimension verification MIL-standard screening Principle of Operation:
A finely focused electron beam is scanned across the surface of the sample generating, secondtry electrons, backscattered electrons and x ray signals. These signals are collected by specific detectors and displayed on a viewing cathode ray tube, lhe raster on the cathode ray tube corresponds to the raster on the sample, while the brightness on the cathode ray tube corresponds to the amount of signal generated at each point on the sample, 4 Anow Drive,Wobum,Massachusens 01801 A Telephone:(617)938-0300 A Fax: (617) 935-5087
_ +
e :,
p a ,- . >
[; , ; --
~1 a L
l@ '
Secondity Electron Imaging (SEI) shows the topography of surface features as small as 6 nm. a The production of the SEI signal is primarily dependent on surfaa roughness.
p High.Resolutien Secondary Electron Imaging (HRSEI) shows the topography of features as -
small as 3 nm. HRSEI can also image films and stains as thin as a few atomic monolayers. An ,
HRSEI equipped SEM can evaluate electron beam sensitive and charging sensitive materials at - :
magnifications up to 300,000X, often without the need for sample coating and without sample damage.
Cryogenic Secoedary Electron Imaging (CSEI) shows the sim,' structure, and shape _of wet :
.,_ ; materials such as hydrated polymers, slurries, oils, biological materials and food products. - an *
- SEM equipped with a cryo-preparation system will allow all SEM imaging and analysis.
capabilities without the need for drying the sample or extensive extraction procedures. ;
p .
. Backscattered Electron Imaging- (BEI) shows the : lateral distribution- of elements or -
compounds within the top micron of the sample.. An SEM equipped with a high resolution Robinson -type' detector (RBEI) can analym features as small as 10 nm and composition L variations of as little as 0.2 percent. The production of the RBEI signal is primarily dependent on surface composition. The Robinson Backscattered Electron Signal is sorted by intensity to produce images which show the distribution of elements and compounds within the top 0.5 -
microns of the sample's surface.'
Electron Beam Induced Current (EBIC) Imaging shows the location of sub-surface opens or .
shorts in microelectronic devices. It is a useful failure analysis diagnostic tool.'
Voltage Contrast (VC) Imaging shows presence of applied bias on the surface of a circuit or device. It identifies opens or shorts as well as voltage drops across a circuit. >
Electron Channeling Patterns (ECP) show localized crystallinity in a 3 micron area. It can analyze the crystalline structure of a material on a microscale and locate defects within structures.
- f -
[ Data Output:
- The SEM. images'are viewed on a TV screen and photographed from a high resolution (2000 4
- lines per inch) cathode ray tube with positive or positive / negative Polaroid film. ,
Sample Constraints:
4 De sample can be up to 15 cm x 10 cm x 7.5 cm in size.- The sample must be compatible with a
[ -
4
= 10 torr vacuum; i.e., non-volatile and not susceptible to electron beam induced damage, ,
= .- - - -. . - - - .. - -. . -. - .
e AJA\
The informatic.n you need...when you need it
- ENERGY DISPERSIVE X-RAY SPECTROSCOPY (EDS)
Energy Dispersive X Ray Spectroscopy (EDS) is an analytical technique that qualitatively and quantitatively identifies the elemental composition of materials analyzed in an SEM. EDS analyzes the top two microns of the sample with a spatial resolution of one micron.
Beryllium windowed EDS detects all elements with atomic numbers greater than oxygen at concentrations greater than 0.1% " Windowless" El>S detectors can also detect carbon, nitrogen and oxygen at concentrations greater than 1.0%.
EDS displays the distribution of elements as either dot maps or line proCles with a spatial resolution of one micron.
EDS Applications Include:
- 1. Materials Evaluation Contaminant location and identification Alloy and intermetallic identification Materiel composition verification Discrimination between electroless and electroplated nickel Elemental diffusion profiles Multiple spot analysis of areas from 1 micron to 10 centimeters
- 2. Failure Analysis Contaminut identification Identification and quantification of unknown materials Stringer location Cosmetic stain identification
- 3. Quality Control Screening Material verification Alloy identification Certifying platings to specification 4 Anow Drive.Wobum.Massachusens 01801 A Telephone:(617)938-0300 A Fax:(617) 935-5087
2 . .
v.
q , :e j
, y t
Principle Of Operation: J f
When the electron beam of the SEM is scanned across the sample, it generates x-rays from the -
atoms in the top two microns. The energy of each x-ray is characteristic of the atom from which +
it escaped. The EDS system collects the x-rays, sorts them by energy and displsys the number of -
x-rays versus their energy. This qualitative EDS spectrum can be either photographed or plotted.
-i This data can then be furser analyzed to produce either an area elemental analysis (displayed as'~
- a dot raap) or a linear elemental analysis (displayed as a line scan) showing the distribution of a -
- particular element within tne top two microns of the surface of the sample. The EDS data can be compared to either known standard materials or computer generated theoretical standants to produce either a full " quantitative" or a " semi-quantitative" analysis.
~ Data Output: -
EDS dot maps and line scans may be smoothed, background corrected and overlaid to show the
. distributions of several elements together. EDS systems also produce color dot maps which show each element's distribution in a different color. These systems also compute concentration
' line profiles displaying exact composition in steps as small as 1 micron across the sample.
, . Qualitative EDS data is typically presented as color photographs or as full-page spectral plots
, while quantitative EDS data is typically presented as tables.
Sample Constraints:-
l The sample can be up to 15 cm x 10 cm x 75 cm in size. The sample must be compatible with a 104 torr vacuum, i.e., non volatile and not susceptible to electron beam induced damage, p
i i.
I:
4 'p.
t l
r e4 ,. *, , , y-- + o---- , , . - - ,--m.- - .,, ,- - - , , w , ~ , - - , . * ,, + . ,
tdp wIururtrivet v;t<
\
A Assown,lat.
m inknr.ation m need alas m need tr
- w w g,t997 Mr. Steve Grenier Amersham Sentinal 40 North Avenue Bwlington,MA 01803
Dear Steve:
Below ytru will find a brief explanation of what a maallurgical fmeture analysis is:
METALLURGICAL FRACnJRE ANALYSIS When the fracture festmes of failed metallurgical components are examined the fractures can often be categorized by the fentwes dnt are discovered. There are many different fracture features, but only four primary failure modos generate the faaeture featwes i.e. tensile oveeload, fatigue,decohesive rupture, mod cleavage. he samples in Report 33631 exhibited two of the few failure modes. Tenvile overload and quasi cleavage. Quasi cleavage is differentiated imm cleavage by ordet of magnitude. Clasale cicava8e forms detta patterns that are not ateociated with quasi-cleavage. Quasi-cleavage is characterimod by isolated single grain cleaving. Cloevage is genecated by high impact loading or high triaxial strus.
He failure history of this bolt indicated that the isolated cleaved grains found on the free surface were the result of high impacticading. Because the evidence of the quasi-cleavage only ocuppies a small area relative to the total fracrure area, the fracture can not be catagorhed as a telt crimpact loadios. The primary fracture featum across b entire fradure surface was microvoid coaleocwx et dimptos. Microvold coaloocence is due to tensite overload, The load this bolt was subjected to, when it was rg.vn.JJy dmpped, exceeded the strength of the material.
Ifyou have any fwther questions or concerns, please feel free to contact me at any time, h
Charles F.Tuson Failure Analyst / Microscopist CFT/jinh 4 AnowDrive Wdurn,Massachuseus 01801 A Teleybooe:(617)9384300 A Fax:(617) 935-3087
- _ _ - _ _ _ _ -