R. Brennan now at ARL

1
Ultrasonic Nondestructive Evaluation of Armor
Ceramic Materials
Rutgers University Piscataway, NJ
R. Brennan (now at ARL) S. Bottiglieri R.
Haber D. Niesz

• Army Research Laboratory
• Aberdeen Proving Grounds, MD

J. McCauley W. Green E. Chin
2
Thesis Objectives and Tasks

• Primary Objectives
• Determine the feasibility of utilizing ultrasound
  C-scan imaging and quantitative analysis for
  nondestructive detection and evaluation of
  micron-range and larger defects and features in
  high density bulk armor ceramics
• Develop a method for using ultrasound to
  establish a representative materials
  fingerprint that will describe microstructural
  features, including defect input data that can be
  quantified and applied to armor ceramic property,
  design, and performance optimization

• Primary Tasks
• Determine the limits and optimize the conditions
  necessary for nondestructively detecting
  micron-range and larger defects in bulk armor
  ceramics
• 2. Develop a method for quantitative analysis of
  ultrasound data for evaluation and comparison of
  bulk armor ceramic material integrity
• 3. Develop a technique to analyze properties such
  as size, shape, and proximity of individual
  features detected using ultrasound and to
  determine the defect size distribution of these
  features

3
Hot Pressed SiC (Sample B) High Frequency
Ultrasound Fingerprint
- Hot pressed SiC test specimen from set of
three samples manufactured under same conditions
Average Density 3.18 g/cm3
Average Thickness 12.76 mm
Average Z 38.7x105 g/cm2s
TOF C-Scan Image 75 MHz
AMP C-Scan Image 125 MHz
TOF Scan Average 2.080 ?s TOF Scan
Standard Dev 0.0041 ?s AMP Scan
Average 19.0 mV AMP Scan Standard Dev
4.60 mV TOF AUTC 0.42 TOF
FWHM
0.018 Velocity AUTC
1,912 Velocity FWHM
18 AMP AUTC
1,262 AMP FWHM
3 Longitudinal Velocity Avg 12,175
m/s Shear Velocity Average 7,632
m/s Elastic Modulus Average 436
GPa Shear Modulus Average 186
GPa Bulk Modulus Average 225 GPa
TOF
AMP
VEL
4
Sintered SiC S3 High Frequency Ultrasound
Fingerprint
- Sintered SiC test specimen with TiB2 additives
from set of eight samples manufactured under same
conditions
Average Density 3.221 g/cm3
Average Thickness 7.71 mm
Average Z 36.6x105 g/cm2s
TOF C-Scan Image 75 MHz
AMP C-Scan Image 125 MHz
TOF Scan Average 1.339 ?s TOF Scan
Standard Dev 0.0235 ?s AMP Scan
Average 41.0 mV AMP Scan Standard Dev
3.68 mV TOF AUTC 2.76 TOF
FWHM
0.013 Velocity AUTC
23,750 Velocity FWHM
92 AMP AUTC 1,495 AMP FWHM
14 Longitudinal Velocity Avg
11,370 m/s Shear Velocity Average
7,240 m/s Elastic Modulus Average
390 GPa Shear Modulus Average 170
GPa Bulk Modulus Average 210 GPa
AMP
TOF
VEL
5
Reflected Signal Amplitude Visualization Methods
S2
3D Visualizations
Schematic Representations
Regional Mapping
S3
6
Amplitude Distribution Analysis of Various SiC
Materials
Sample B
0-16 mV (17.6)
17-22 mV (33.7)
23-27 mV (43.9)
28-47 mV (4.8)
Sample C
0-16 mV (18.2)
17-23 mV (41.1)
24-30 mV (40.3)
31-42 mV (0.4)
Sample S2
0-40 mV (19.1)
41-100 mV (16.4)
101-120 mV (50.8)
121-150 mV (10.0)

• Amplitude distribution analysis based on
  selected range to show percentages and regional
  locations

7
Peak Deconvolution Analysis of Sintered SiC
Material
Lorentzian Two-Curve Fit R2 0.921
Area 2662 100 Area
Area 597 22.4 Area
Area 2065 77.6 Area
Normalized Amplitude Range

• Peak deconvolution of AMP histogram curves used
  to separate and compare tail regions separately

8
Peak Deconvolution and Distribution Analysis of
Sintered SiC Materials
Tail Peak 30 - 74 mV (16.2)
Main Peak 75 - 228 mV (67.4)
Sample 30 - 228 mV (83.6)

• Peak deconvolution results used to show
  distribution percentage different results for
  sintered SiC

9
Image Conversion of Hot Pressed SiC Non-Defect
Regions
20 mm
125 MHz C-Scan Image of Hot Pressed SiC Sample
High Resolution Gray Scale Image

• Hot pressed sample homogeneous enough to achieve
  thresholding in single step without sectioning
• Step size of 0.05 used to construct image with
  more than 3.6 million data points for evaluation

10
Image Conversion and Separation of Sintered SiC
Non-Defect Regions
20 mm
125 MHz C-Scan Image of Sintered SiC Sample
High Resolution Grayscale Image
20 mm
Region 2
Region 3
Region 1
Full C-Scan Image
11
Low and High Amplitude Regional Feature
Thresholds of Sintered SiC Sample
Thresh120
Thresh150
Thresh180
Amplitude Spectrum

• Distinct features isolated by applying threshold
  to lower (low AMP) and higher tail regions of AMP
  spectrum
• Sample divided into specific regions to extract
  individual features without saturating or washing
  out large areas

Region 1
Region 2
Region 3
Low AMP Features
Thresh 147
Thresh 137
Thresh 130
20 mm
High AMP Features
Thresh 150
Thresh 185
Thresh 158
12
Low and High AMP Feature Imaging of Hot Pressed
SiC Sample
High AMP Feature Color Changes
20 mm
High AMP Features
Combined Features
20 mm
Low AMP Features
13
Low and High AMP Feature Imaging of Sintered SiC
Sample
High AMP Feature Color Changes
20 mm
High AMP Features
Combined Features
20 mm
Low AMP Features
14
Low and High AMP Labeling and Measurements of Hot
Pressed SiC Sample
Low AMP Labeling
All Feature Labeling
High AMP Labeling
Magnification of Numbers
Magnification of Labels
20 mm
Area 5.98E5 ?m2 Perimeter 3708
?m Eq. Diameter 872 ?m Length 1154
?m Width 791 ?m Aspect Ratio
1.46 Symmetry 0.93 Roundness 0.57
15
Low and High AMP Labeling and Measurements of
Sintered SiC Sample
Low AMP Labeling
All Feature Labeling
High AMP Labeling
Magnification of Numbers
Magnification of Labels
20 mm
Area 6.99E5 ?m2 Perimeter 3393
?m Eq. Diameter 943 ?m Length 1113
?m Width 886 ?m Aspect Ratio
1.26 Symmetry 0.97 Roundness 0.72
16
Measured Low and High AMP Parameters of Hot
Pressed SiC Sample
Average Feature Parameters
Additional Average Feature Parameters
Minimum and Maximum Selected Parameters
17
Measured Low and High AMP Parameters of Sintered
SiC Sample
Average Feature Parameters
Additional Average Feature Parameters
Minimum and Maximum Selected Parameters
18
Equivalent Diameter Size Range Images of Hot
Pressed SiC Sample
20 mm
50-99?m Features (62)
100-199?m Features (43)
200-299?m Features (22)
300-399?m Features (7)
400-499 ?m Features (6)
500-550 ?m Features (3)
19
Equivalent Diameter Size Range Images of Sintered
SiC Sample
20 mm
99-150?m Features (276)
151-250?m Features (254)
251-350?m Features (79)
351-500?m Features (57)
500-999 ?m Features (134)
1-6 mm Features (34)
20
Equivalent Diameter Curve Fit of Hot Pressed SiC
Sample
Origin 7.0 Curve Fit Function
Model FreundlichEXT Equation y
ax(bx(-c)) Chi2/DoF 17.73389 R2
0.96878 a 3.0868E-17
2.9355-16 b 27.97883 5.80017 c 0.24445
0.0055
Area8,363
21
Equivalent Diameter Curve Fit of Sintered SiC
Sample
Origin 7.0 Curve Fit Function
Model FreundlichEXT Equation y
ax(bx(-c)) Chi2/DoF 590.08468 R2
0.96264 a 5.5764E-49
6.96E-48 b 59.52138 6.4149 c 0.18656
0.00119
Area177, 840
22
Defect Size Distribution Curve Comparison to
Bakas et al. Data
Ballistic Fragment Data
Ballistic Fragment Data
Cumulative Feature Size Distribution Data
and Curve Fit from Hot Pressed and Sintered SiC
Samples
Ultrasound Defect Distribution Data
23
Summary and Conclusions

• Comprehensive ultrasound fingerprint allowed for
  various acoustic
• parameters to be directly compared among armor
  ceramic samples
• Visual representation methods created to show
  ultrasound C-scan image
• data in several different formats including
  schematic overlay, regional
• mapping, and three-dimensional mapping
• Amplitude distribution post-processing technique
  developed to describe
• reflected signal amplitude range percentages
  and distribution maps
• Peak deconvolution post-processing technique
  developed for two-phase
• histogram curves to compare critical tail
  region of various test specimens
• Image analysis protocol developed using
  grayscale thresholding for
• detecting distinct isolated features and
  determining size distributions
• Size distributions of distinct material
  inhomogeneities from C-scan images
• fit to power law function for hot pressed and
  sintered SiC and compared to
• defect size distributions from ballistic
  studies

24
Future Plans

• Q4-2007
• Evaluate the use of higher frequency broad range
  transducers (in conjunction with TRS
  Technologies) to more precisely map defect
  distributions
• Q1 and Q2-2008
• Work with ARL and CCMC Members to classify a
  larger population of dense parts into separate
  acoustic populations for subsequent ballistic
  testing.
• Conduct microfocus x-ray on selected samples to
  corroborate UT analyses.