Introduction to X-ray imaging for

industrial applications

Markus Tarin President & CEO

MoviMED

Agenda

Agenda

• What exactly are “X-rays” ?

• The X-ray tube

• Typical applications

• Methods of detection

• Detector types

• Challenges in x-ray imaging

• Detector selection

• Example application

• Conclusion

• Q & A

Head x-ray – “Homer Simpson”

The Electromagnetic Spectrum

Historical Background

• German Physicist

• Discovered x-rays or “Roentgen rays” on Nov 8, 1895

• During experimentation with vacuum tubes he noticed a faint glow (fluorescent) on a cardboard covered with “Barium Platinocyanide”

• Named the radiation “x-rays” as in “unknown” rays

• Wilhelm Roentgen received the Nobel Prize in 1901 for this discovery

Wilhelm Conrad Röntgen

*03/27/1845 † 02/10/1923

First X-ray image

• First X-ray images taken

by Wilhelm Roentgen

• 22nd December, 1895

• Shows the hand of

Wilhelm’s wife with ring

The X-ray tube

X-rays

Cathode

Filament

Glass body

Anode with Tungsten target

Electron beam kV

(keV)

The filament provides an

“electron cloud” which is

accelerated by the high

voltage potential

between the anode and

cathode.

The electrons collide and rapidly decelerate on the

high-density Tungsten anode. The energy decay

causes a photon emission. The emitted wavelength is in

relation to the energy loss of the electron.

X-ray Energy

Definition of X-ray energy:

The term xxx kV (kilo Volt or 1 x 1,000 Volt) refers to the high voltage supplied to

the x-ray tube. With other words the potential between the anode and cathode. A higher “kV” setting results in a higher x-ray energy output and a shorter wavelength.

The current (mA – milli-Ampere - 1/1000 A) is the selected current allowed to flow through the filament of the x-ray tube at the selected voltage. A higher current causes a higher x-ray flux.

The term “eV” or more commonly “keV” or “MeV” is the “electron Volt” or the energy given to an electron by accelerating it through 1 Volt. When considering x-rays, the keV or MeV is referring to the output energy of the x-ray photons generated by the x-ray tube.

Examples of Energy (eV)

The following are examples of eV energies:

• Visible light photons: 1.5 – 3.5 eV

• Approximate energy of an electron striking a color

television screen (CRT tube): 20,000 eV (20 keV)

• High energy medical, diagnostic x-ray: 200 keV

• 100W light bulb burning for one hour: 2.2 Trillion

TeV !!! (2.2 Trillion Trillion eV)

• Kinetic energy of an 1,900lb race car traveling at

230 mph: 28 x 10^24 eV

Typical Applications

• Airport/Homeland Security

• Electronics

• General Inspection

• Petro-Chemical

• Automotive

• Aerospace

• Non-Destructive Testing

• Medical/Diagnostic Imaging

• X-ray Fluoroscopy

X-ray detection methods

• Image Intensifier

• Scintillator (Gamma detector)

• Amorphous Silicon

Panel

• X-ray film

• Phosphor plate

scanner (CR)

• others

X-ray image intensifier

Image Intensifier

• Commonly found

in industrial x-ray

applications

• Converts x-rays

into photons using

phosphor

Scintillator

• User for gamma ray detection

• Could be coupled to a photon multiplier counter

• Uses inorganic material to convert gamma rays into photons in the visible waveband

Conceptual overview – Scintillator with detector

Amorphous Silicon Panel

• Energy range from

10 to 160 keV

• Resolution about 48

µm (~10lp/mm)

• Standard frame

grabber interface

• Easy to integrate

• “X-ray camera”

• Compact form factor

Challenges in X-ray imaging

• X-ray applications are generally “light starved”

• Signal to noise ratio is usually not very favorable

• Achieving acceptable image quality requires

careful selection of optics and camera

• X-ray imaging requires a considerable amount of

“domain knowledge”

• Image contrast depends on many factors (x-ray

energy, absorption bands in specimen, type of

materials, focus quality of x-ray beam, selection of

detector, lens and imaging sensor)

• Feature definition may be “fuzzy” or “faint”

Importance of Detector Selection

Why using a standard machine vision

camera for X-ray will not work…

• The light output produced by an image

intensifier is typically very low

• Increasing the x-ray energy to achieve

more light output will not necessarily

improve the contrast ratio

• Using a small pixel size sensor (<7.4um

with 8-bit ADC) results in a limited dynamic

range

• Light starved x-ray imaging competes with

detector noise of the camera

Importance of Dynamic Range

Sensor A Sensor B

Pixel Size 7.4um x 7.4um 16um x 16um

Full well capacity 20,000 e 150,000 e

Read noise 16 e 15 e

Dynamic Range 20,000e/16e = 1,250 150,000e/15e = 10,000

Dynamic Range 62 dB 80 dB

Gray levels ADC @ 10-bit - 1024 ADC @ 14-bit = 16,384

(Effective #of bits: 13.3)

• Sensor B has a dynamic range 8 x higher than sensor A

• X-ray images need to be processed in 16-bit format

Lens Selection

• Due to the large pixel size

requirement, the sensor size of a

camera is also large.

• F-mount lenses are often required

• The field of view and working

distance needs to be matched to

the output port of the image

intensifier

• The lens needs a large aperture

(low f#) to capture all available

light. Ideally f# < 1.0

Example application - BGA

Ball grid array inspection • Higher density integrated circuits also come

with a higher pin count

• The “BGA” or ball grid array IC package type

has little balls as connection leads

• The connections are underneath the chip

and are no longer visible, after assembly

• The solder paste is being placed onto the

circuit board, prior to chip placement

• The chip is then placed onto the PCB and

run through a reflow oven, where the solder

paste melts and forms an electric connection

between the PCB and the ball contacts

• X-ray inspection is the only method to “see

through the circuit board and verify that all

balls are in contact and that there are no

accidental short circuits

BGA – X-ray image

• The solder paste is very opaque

to the x-rays and provides a

very favorable contrast ratio.

• The electronics industry

frequently uses radio isotopes,

emitting gamma rays (MeV

energy)

• Automatic detection algorithms

can be used in this example for

verification.

Image Processing – X-ray

Step 1: Raw x-ray image

BGA – Threshold

Step 2: Threshold type “Moments”

Creates a binary image. Only black

areas are being considered.

BGA - Dilate

Step 2: Dilate

Dilates binary pixels in the object

BGA - Close

Step 3: Close

Closes holes in the blobs

BGA – Particle analysis

Step 4: Particle analysis

Counts objects, measures size and

location. 144 pins found – pass!

BGA defect

• This images show multiple

defects after the soldering

operation of a BGA IC.

• Multiple BGA contacts are

accidentally soldering

together.

• The bridged pins or ball

contacts are creating a

shortcut

• The circuit will not function

properly

BGA defect – raw image

BGA defect - Threshold

BGA defect - Erode

BGA defect – Remove particles

BGA defect - result

• The final image

processing step

reveals 7 shorts

• The blobs are

much larger than

the acceptance

criteria

• Their center of

gravity does not

coincide with the

IC package

definition

Conclusion

• X-ray imaging is a very useful, non-visible imaging tool

• A considerable amount of domain knowledge is required to successfully apply this technology

• Budgetary estimate for X-ray imaging systems range from $60k to $500k and up.

• Exposure to x-ray radiation is potentially dangerous to ones health and safety.

• Automated image processing may not always be possible, due to sometime poor defect definition

• 16-bit image processing is necessary, due to the need for large dynamic range.

Markus Tarin President & CEO

MoviMED

15540 Rockfield Blvd., Suite C110

Irvine, CA 92618

USA

Phone: 949-699-6600

email: m.tarin@movimed.com

www.movimed.com

Q & A