IPC-4556 印制板化学镍钯浸金（ENEPIG）规范ENG - 第37页

T able A4-1 Summary of XRF Conﬁguration Solutions Of fered for Measurement of ENEPIG Plating on PCB’s with Advantages and Disadvantages ENEPIG Layer Proportional Counter XRF Solid State Detector XRF V acuum & Solid S…

Solid state based XRF instruments offer some advantage for Pd measurement as well.

Although the varying background scatter issue is still present, the overall background level tends to be lower compared with

proportional counter systems, again allowing a clearer view of the Pd peak above the background. This translates into the

ability to measure lower thicknesses of Pd using solid state detectors relative to proportional counters. The disadvantage in

terms of Pd measurement is that solid state detectors are not as efficient as proportional counters in detecting the high energy

Pd K-line x-rays. As such, overall x-ray intensity for the same Pd thickness will be less when using a solid state detector than

a proportional counter, given all other factors being equal. This disadvantage is partially negated by the overall lower back-

ground levels observed with solid state detectors.

Other considerations noted for the proportional counter based XRF systems must be applied as well with solid state detector

based systems. These include corrections for phosphorous content in both the nickel and again, to a lesser extent, the Pd

layer. The advantages offered by the solid state detector when measuring ENEPIG come with other disadvantages, as well.

First, the cost for solid state detector based systems is higher than proportional counter based systems. Second, the size of the

detection area is smaller than proportional counters. As a result, less x-rays are detected per unit time under the same exci-

tation conditions. Generally, larger x-ray collimators must be used to compensate for the smaller detection area. This means

that measurement of sample areas less than 12 mils wide or in diameter can be problematic with solid state detector based

XRFs. Typically longer measurement times must be used to achieve the same level of measurement repeatability as a pro-

portional counter based XRF. However, if collimators that are 20 mils or larger can be used, then measurement time for both

types of XRF’s can be comparable to achieve similar repeatability.

XRF Instruments Using Cr Target X-ray Source Some proportional counter based XRF systems have been provided by at

least one manufacturer using a Cr target x-ray source. The principle advantage of the Cr target source is its ability to excite

the Pd L series peaks (~ 3 keV). Once these peaks are sufficiently excited, measurement precision is signiﬁcantly improved

for the Pd layer since the sensitivity of Pd L line x-ray intensity, in terms of unit thickness change, is excellent. This high

sensitivity to Pd thickness change is especially advantageous in the working range for immersion Pd thicknesses used in

typical ENEPIG applications. The improved measurement repeatability and shorter measurement times that may be used,

allow for better process control. Furthermore, by shifting the Pd analysis from the high energy K line typically used to the

low energy L line, the issue of background scatter variations from the substrate is eliminated. This is because low energy

scatter will not vary in this case.

Thus the measurement of Pd is less dependent again on operator know-how and vigilance. By using the Cr target source with

a solid state detector, one may have the same advantages the Cr target offers for the Pd layer and achieve the advantages for

the Au layer offered by the solid state detector. Again there is a down side to this strategy as well. Very small areas (<12

mils) are difficult, if not, impractical to measure (at least long measuring times are required). Secondly, Cr target sources

typically do not have as long an operating life as the common W or Mo target sources. Therefore, cost of ownership is

higher.

Vacuum Based XRF Systems Even higher cost XRF systems are available that use only solid state detectors and offer the

ability to evacuate the X-ray chamber of air. By evacuating the X-ray chamber, one gains the ability to directly detect phos-

phorous x-rays. Detection and counting phosphorous x-rays allows for the direct measurement of phosphorous content in

electroless nickel. This provides the user with direct, accurate composition measurement and allows the XRF to simultane-

ously calculate electroless nickel thickness based on the actual phosphorous content. This represents the most accurate way

to measure electroless nickel layers.

The same advantages and disadvantages described above for measurement of Au and Pd layers using solid state detectors

applies with one exception. If one chooses to conﬁgure the vacuum XRF with a Cr target source to analyze the Pd L peaks

intensity or alternatively uses x-ray optics to produce microbeams to measure very small sample areas, the use of an evacu-

ated environment reduces background substantially for the Pd L peak. When measuring Pd L intensity in atmosphere, argon

ﬂuorescence interferes with Pd L peaks (Ar is 1% of the atmosphere). By evacuating the sample chamber, this effect is elimi-

nated, making Pd L peak intensity determination more precise.

To summarize the various effects one must address when calibrating and measuring ENEPIG coatings by XRF, and the vari-

ous solutions offered by different XRF conﬁgurations, the following Table A4-1 is offered. In addition, the supporting section

to this article includes some example spectra that illustrate some of the effects discussed.

IPC-4556 January 2013

Table A4-1 Summary of XRF Conﬁguration Solutions Offered for Measurement

of ENEPIG Plating on PCB’s with Advantages and Disadvantages

ENEPIG

Layer Proportional Counter XRF Solid State Detector XRF Vacuum & Solid State Detector XRF

Au 1. Use peak deconvolution to correct

for Br interference

2. Check calibration at least twice a

day - drift correct or recalibrate as

needed

3. Difficult to measure <51 µm

4. Measure time 60 - 90 sec

depending on collimator size

5. Can measure areas as narrow as

76 µm wide or 102 µm diameters

with longer measurement times

(120 sec)

1. No deconvolution needed - can

measure <25 µm - better overall

accuracy

2. Good long term stability

3. Measure areas <305 µm wide -

impractical unless instrument

conﬁgured with x-ray optic

4. Measure time 60 - 90 sec; longer

for layers <25 µm. Less time if

collimator is 508 µm or larger

or if x-ray optic used.

1. No deconvolution needed - can

measure <25 µm - better overall

accuracy

2. Good long term stability

3. Measure areas <305 µm wide -

impractical unless instrument

conﬁgured with x-ray optic

4. Measure time 60 - 90 sec; longer

for layers <25 µm. Less time if

collimator is 508 µm or larger

or if x-ray optic used

Pd 1. Must use background correction

methods unless Cr target source

used

2. Measurement of Pd <51 µm is

problematic unless Cr target

source used

1. Must use background correction

methods unless Cr source or x-ray

optics used

2. Measurement of Pd <51 µm is

possible with longer measurement

times

3. Measure areas <305 µm wide

impractical unless instrument

conﬁgured with x-ray optic

4. Use of Cr target source or x-ray

optics improves precision, reduces

time

1. Must use background correction

methods unless Cr source or

x-ray optics used

2. Measurement of Pd <51 µm is

possible with longer measurement

times

3. Measure areas <305 µm wide

impractical unless instrument

conﬁgured with x-ray optic

4. Use of Cr target source or x-ray

optics improves precision, reduces

time and allows for measurement

of minimum of 7.6 µm

Ni-P Correct thickness measurement for

difference in % P between calibration

Most accurate - measures % P

and corrects thickness of Ni layer

Correct thickness measurement for

difference in % P between calibration

standard and sample

automatically

January 2013 IPC-4556

Supporting Section

Above Figure A4-SS-1 displays overlay of one spectrum collected on half ounce Cu/epoxy and spectrum of one ounce

Cu/epoxy. Blue trace is from half ounce Cu/epoxy, red trace is from one ounce Cu/epoxy. Note the signiﬁcant reduction in

Br peak when Cu thickness increases from half ounce to one ounce. Also note the difference in background levels in the

indicated region of the spectrum where Pd intensity will be counted. Again, background ‘‘noise’’ decreases as Cu thickness

increases.

Figure A4-SS1 XRF Spectra of 1/2 oz Cu/Epoxy vs. 1 oz Cu/Epoxy

Smoothed, Comparison of Cu epoxy half ounce.cts(blue) vs. Cu epoxy one ounce.cts(red)

0.9

795

Br Pd

636

477

Br from epoxy

Background noise

in Pd region

318

159

8.4 15.8

KeV

Counts

23.3 30.7 38.2

IPC-4556 January 2013

Spectral Analysis Mode- WPHA

Fle Edt Yew Display Ogerations ROI Help

BLUE

Dx 2109.2

Peak(Ch.,KeV,Cts)

１９１，８．０，１２５４８．６

Ready

讼

箱物箱箱着

NM[