PRACTICAL GUIDE FOR MEASURING HD UNDERCOAT THICKNEsis

Fuji0 Matsub', Teresita E.Guiala*, David J.Liston' and Yasuyuki Yamamoto"

C. Uyemura & Co. Ltd., Tokyo, Japan Uyemura International Corporation, Ontario, California

'*

ABSTRACT

This paper will discuss and review various methods of thickness measurement of Ni-P undercoat on aluminum magnetic hard disk (HD). Accurate measurement of these films is of considerable importance due to the trend of the hard disk industry towards smaller size disks as well as thinner Ni-P undercoat. Focus will be on the X-ray Fluorescence (XRF) method, pointing out the different variables effecting the accuracy of this method.

Jntroduct ion

The magnetic disk drive industry is still experiencing market growth evidenced by a smooth spiral in electroless nickel ( EN ) sales. There has been major shifts during recent years. Industry publications have expressed that the market has boomed for the 3.5-inch drives while the larger format 8-inch and 14-inch drives is almost gone . The personal notebook computers have moved to a 2.5-inch and 3.5- inch format disk drives.' Keeping up with technology advances, price competiveness and changing conditions remain as barriers for the magnetic disk industry's success.

Not only has the size of the disk decreased, the high phosphorous (12%) EN coating thickness has also decreased. We are seeing that where 30-40pm thicknesses were common, now 6-1 Opm has become common. Cost increases in proportion to the plating thickness, so we must be able to provide a sufficient film thickness, corrosion resistance, wear resistant and a durable non-magnetic surface.

The process of selecting a measurement method requires you to consider accuracy, the cost of the measuring device, repeatability and confidence levels. Accurately measuring the thickness of the undercoat poses a significant challenge as films become thinner.

Four basic thickness measurement techniques are gravimetric, microscopic cross sectioning, beta backscatter and, X-ray fluorescence (XRF). The trend over the past years has been away from

1

3

the first three stated above and toward XRF.2 XRF is favored mainly because of the accuracy and repeatability that can be achieved with minimal measurement times. XRF is capable of better than -15 Yo accuracy, especially where extra precautions are taken to eliminate errors. Acquiring significantly greater accuracy would be difficult because of the inherent roughness of the substrate and the coating surface themselves.

Comparison of the different thickness measurement methods for high phosphorous content electroless nickel (HPEN) hard

disk undercoat.

Gravimetry (two-types, non-destructive [weight gain] and destructive.): Non-destructive: Accuracy using this method causes the result to be affected by the loss in weight during etching due to the dissolution of the AI substrate material. Destructive method: Ni-PIA1 plated parts are dissolved in 10% wlv NaOH allowing nickel to peel-off from the AI substrate. The dry Ni foil can then be weighed. Thickness is calculated as equal to weight of Ni divided by the product of the density and the area.3 This destructive method is considered the most accurate and a good reference.

Microscopic cross section (destructive, optical): Good reference. The coating thickness is measured in a magnified image of a cross section of the coating. used for mounting, polishing, and etching the specimens. A filar micro- meter ocular or image-splitting eyepiece is used to measure the thickness. In some instances, the image is projected onto a ground glass plate, the measurement is made with a ruler at an approp- riate magnification.4 This is accurate for films of aboutlo-100 pm * 5%. Thickness capability is dependent on microscopic equipment used and skilled technician.

Normal metallographic procedures are

Beta backscatter ( non-destructive, mass per unit area method): The principle of beta backscatter is that when beta particles impinge upon a material, a certain portion are backscattered. The intensity of the backscatter can be used for the measurement of mass per unit area of the coating, which is directly proportional to

2

4

the thickness. Measure thicknesses ranging from 0.5 - 40 pm. For best accuracy with the backscatter method, appropriate isotope for the target coating and to have standards established for each turnover to adjust for phosphorous increase. Two disadvantages of beta backscatter are that frequent calibration is necessary to compensate for decay of the isotope ater rial.^ Second is that parts to be tested come in contact with the isotope probe and this will leave surface marks that may affect the integrity of the HD undercoat deposit, especially after polishing.

&Ray Fluorescence (non-destructive, mass per unit area method): A contactless measurement that indicates thickness by directing an X-ray beam through the coating . The beam is absorbed and the amount of absorption yields an instanteous readout of thicknesses between 0.5pm-30pm. Advantages are the ability to measure every part, with very high speed and without contacting and destroying the test piece. Areas as small as 2 mm in diameter can be measured.6 Disadvantages are the one-time cost of the special equipment and the upper limitation to measuring thickness.

Table 1 Compares the different methods.

I XRF Gravimetric 1 Cross-section Beta Backscatter 1 Non-destructive Destructive) Destructive Nondestructive Plating operator Skilled technician Plating operalor Less than 60 sec. More than 1 hour Less than 60 seconds

Common laboratory equipment Speciality equipment 0.4 Fm-unlimited 1 10-100 pm 0.5-40 pm I 0.5-30 pm

Based on the above data , we chose to explore to explore applic- ations of XRF for measuring Hard-disk electroless nickel undercoating.

XRF ExDerimental Details

The following experiments were carried out using a Seiko SFT-I57 Fluorescent X-ray Coating Thickness Gauge. The phosphorous content was set at 12% unless otherwise specified. X-ray intensity was 10 for collimator size 0.1 and 0.3 mm and X-ray intensity 4 was used for collimator 1.0 mm. The instrument was calibrated with 4 . 9 3 ~ ~ and 21.1pm Ni standards on AI; and with Ni, AI and 10.6 wt.%P Ni-P infinite standards .

3

5

Effect o f measurement time The thickness of a Ni-P foil (5, IO, and 15pm) on aluminum was measured 30 times using different count times (2,5,10,20,30,60 and 100 sec.).

Effect of dif ferent col l imator sizes The thickness of Ni-P foils on AI of various thicknesses were measurc using different collimator sizes ( 0.1, 0.3 and 1.0 mm). The count time for each measurement (IO trials) was 60 seconds.

Effect of 2-stage height The thickness of Ni-P /AI magnetic disk (approx. 13 pm) was measured for 30 sec. (10 trials) using a collimator size of 1.0 mm wit the sample stage at different heights, ranging from 5 mm below to 3 mm above the focal position'. This experiment was also performed using collimator sizes of 0.1 and 0.3 mm.

The point where the sample parts appears clear with defined sharp image when viewed through the C.R.T. monitor .

Effect of thickness of the AI substrate The thickness of Ni-P foil was measured with increasing number of AI foil pieces (0.02 mm per piece) using a measurement time of 20 sec. (10 trials). The procedure was repeated until the thickness of the AI was 0.3 mm.

Effect o f Phosphorous content The thickness of a 22.1pm Ni-PIA1 disk with a phosphorous content 01 12.3% was measured using a collimator size of 1.0 mm using differen % P content in the calibration.

X-ray dr i f t The thickness of a Ni-PIA1 standard disk was measured over time to check for any shift .7

4

Results and Discussion

We found these six factors which influence the measurement of thickness of Ni-P films on AI.

1. Measurement time 2. Collimator size 3. Position of stage 4. Thickness of substrate 5. Phosphorous content 6. Instrumental drift over time

1. Effect o f Measurement Time

Measurement time is the length each measurement lasted. The results in figure 1 show that the accuracy and precision of the measured Ni-P thickness can be increased with an increase in the measurement time, until a maximum point where any time increase had no significant effect.

C 0

a > 0 U

- e -

z

ti

a Q E a

15.0 I I , I I I 0.02 0 20 40 60 EO 100 120

Measurement time (see.)

Figure 1 Effect of measurement time on the Thickness of Ni-P film on aluminum.

Similar trends were observed using .a 5pm and a 10pm Ni-P foils on AI. The standard deviation decreased with measured thickness. These results suggest reliable thickness with 95% confidence level for both 0.3 mm and 1.0 mm collimator size can be obtained using a

5

7

measurement time of 10 seconds. However, even better results were obtained with slightly longer times. For example using the 1.0 mm collimator most reproducible results to measure a Ni-P film of about 15pm in shortest time was when measurement time of 20 seconds was used .

However, your production situation will dictate whether the time required obtaining the extra precision is cost efficient.

2. Collimator size

XRF Coating Thickness Gauges are designed to measure film thickness on small areas by focusing the X-ray beams from the X-ray tube into a concentrated ray. A collimator is used for this purpose.

Cross section is a good reference to counter-check XRF resul One direct way of comparing accuracy is to graph the measure- ments from one method against the other. Thus, the results in figure 2 show that the best overall correlation of the thickness obtained using these two methods was found using a 1.0 mm coll imator.

For Ni-P films with thicknesses below 20pm , the deviation of the thickness obtained using the 2 methods was less than 2.5%, usins collimator sizes of either 0.3 mm or 1.0 mm; however, for the 0.1 mi collimator, the thickness difference was up to 8%.

The error in the measurement of Ni-PIA1 thicknesses outside t range of calibration, that is in the present study below 5pm and abo\ 21pm showed a tendency to increase with the thickness of the Ni-P coating. Thus, the small deviation of less than 0.5% for the 58.4pm Ni-P film using the 1.0 collimator is misleading as the XRF measure ment above 40pm-50pm is generally less than that obtained by cross-sectioning. This is in contrast to the situation using the smal collimators (0.1 mm or 0.3 mm) where the XRF measured thickness showed an increase, after 25pm compared to the cross-sectioned thickness measurement.

6

I I I I I

0 1 0 20 . 30 4 0 5 0 I 0

Thickness measured by microscopic cross-section (pm)

Effect of collimator size on the thickness measurement of Ni-P lilms on alluminum alloy. Rgure 2

The collimator size also has a large effect on the reproduc- ibility of the measurement as the smaller the collimator the larger the measurement's standard deviation for a given measurement time and coating thickness.

3. Effect of the Position of the Sample

The sample is positioned on the stage and its height is adjusted so that the image of the sample on the C.R.T. monitor is in focus. The results in figure 2 show the effect of changing the position of the sample in the 2- direction on the thickness measurement.

In some of our experiments the position that corresponded to the highest thickness position was not the focus position ,in fact, the image was just out of focus at the highest thickness reading. This was also confirmed with a different collimator size and by testing on another XRF instrument.

The position of the Z-stage can be changed by about 3.5 mm and still obtain a measurement with a 95 % confidence level.

7

9

25

Combined thickness of upper and lower Ni-P films

Collimator size: 1.0 mm Measurement time: 20 sec.

I .I

Thickness of upper Ni-P film

0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 .0

Thickness of Alumlnim substrate between films (mm)

Figure 5 XRF measurements with AI substrate size changes.

35

30 Collimator size: 1.0 mm Measurement time: 20 sec.

25

20

15

10 0 5 1 0 1 5 2 0 2 5

Phosphorous content entered in callbration ( %P )

Figure 6 Effect of Phosphorous content on Ni-P film thickness

1 0

12

6. instrumental Dr i i t w i th Time

Note that another source of 'inaccurate readings is the uncorrected drift of the X-ray detector andlor tube, You will need to determine for yourself, the correction interval that will work best for your operation. Drift was tracked by plotting the thickness of a Ni-P film (10.8pm) on AI over time when no corrections were made (Figure 7). Over a period of 8 hours, the thickness drifted downward by about 0.03pm. The coefficent of variation was 0.3 %, hence no correction factor was necessary in our case, to maintain a 95% confidence for 8 hours.

1

10.64 I I

0 100 200 300 400 500

Time (min)

Figure 7 X-ray drifl (average lhickness vs. time)

3umrnary

Ensuring precision and accuracy of Hard disk Ni-P undercoat thickness measurement using X-ray fluorescence method requires:

1. Measurement time: The longer the measurement time, the more precise and accurate the result . Production application of 10 seconds produced relatively good results.

11

I

2.

3.

4.

5.

6.

Position : Focus distance between the sample and the collimator determines the confidence level of the measurement. This varies from instrument to instrument, i f absolute rather than relative data is desired , you will need to experiment with ideal positioning for your application. Collimator size: Exposing the maximum area, especially for thicker deposits gave a more precise measurement. Larger measurement area lessens any localization effect. We would recommend 1.0 mm collimator for use in the hard disk industry. Substrate thickness: As disks have become smaller they have also become thinner. As this trend progresses to even thinner aluminum alloy substrates, the thickness measurement techniques for the EN undercoating are sure to be effected. %P: Take extra care to enter the proper phosphorous content in the calibration to maximize the accuracy of your measurement using XRF equipment. Instrument drift: Protecting against natural drift in X-ray test equipment is the responsibility of you, the user. To maintain the XRF instrument at its ideal calibration before testing a sample , read the pages relating to calib- ration procedure in the USER GUIDE. The time spent doing this will prove worthwhile.

References 1. Arensman,R., Flectronic Business As ia, Nov. 1990, p. 77. 2. Hoffman,P.L., "Thickness Measurement on Electronic

Connectors: The Key to Plating Process Control", Platina and Surface Finishing, Sept. 1985, p. 20.

3. Sajdera,N., Thickness Testing, Meta I Finishing, Vol. 88, No. 1A (1990) p. 552.

4. Ogburn,F. , Methods of Testing; F. Lowenheim, Ed.; In Modern Electroplating, 3rd ed. N. J.,1974 ; p. 686.

5. ASTM Designation B 567-91, Measurement of Coating Thickness by Beta Backscatter Method, American Society for Testing and Materials, 1991.

12

14

6. ANSVASTM A 754-79, Coating Thickness by X-Ray Fluorescence, American Society for Testing and Materials, 1979. . Coating Thickness Gauge Operation Manual, Stability Check, p. 153.

Special Ed., pp. 1-6.

7. Seiko Instruments & Electronics Ltd., Fluorescent X-Ray

8. Koga, T. and Tomoda, K. ; Kinzoku ( Jap.), July ,1982,

13

10

Collimator size: 1.0 mm Measurement time: 30 sec.

9 0 5 10 1 5

Sample stage height ( mm )

Figure 3 Effect of changing stage height on the mickness measurement

4. Substrate Thickness

Below are data from current industry producers. As you can see, not only is the physical size of the disk being reduced, but the disk thickness itself is being forced down by the demand for compactness.

Table 2 Thicknesses of AI substrate currently being manufactured for HD.

~

Outer diameter (mm) Thickness (mm) 210 &inch 1.8 130 5.25-inch 1.8 95 3.5-inch 1.2 9 5 0.88 6 5 2.5-inch 0.86 5 0 2-inch 0.58' 3 6 1.5-inch 0.38'

The thickness of the 36 mm and 50 mm disk is not yet a the market place as these are not in wide commercial use

8

function of yet.

In regards to the substrate thickness, it is known that silver coatings8 for example, are influenced by the thickness of the underside film, especially if the substrate is thin. This phenomena can be understood from the following diagrams - one where the substrate is regular and the other where the thin substrate allows for the secondary fluorescent X-ray is from the underside coating as well as the base material.

X-RAY

\ RUORESCENT X-RAY .X-RAY

X-RAY FROM BASE

I SUBSTRATE ~~ 1 . . ~ ..:. ~ i_ .... <*...

Figure 4 Theoretical effect of decreasing the substrate thickness.

The results in figure 5 (next page) show that a thickness of the Ni-P film can be obtained with 95% confidence if the substrate is over 0.12 mm. Thus, for the present, this effect is negligible, however, care should be exercised in the future to take this phenomena into account.

5. PhosDhorous content

XRF measurements are based on conversion of mass per unit area to thickness, the coating density must be accounted for when using this method. The phosphorous content in a Ni-P film affects the density of the deposit and thus at least a close approximation of the deposit's phosphorous 6 show that the phosphorous content entered at the time could only be changed by about k 1% to obtain a thickness with a 95% confidence level.

must be entered into the calibration. The results in figure

9

F P / P / o c; 5 c The Proceedings of the 79th AESF Annual Technical Conference

SlR4N SUWF~NB '92

June! 22-25' 1-92 Atlanta, Georgia

The American Electroplaters and Surface Finishers Society, Inc. (AESF) is an international, individual- membership, professional, technical and educational society for the advancement of electroplating and surface finishing. AESF fosters this advancement through a broad research program and comprehensive educational programs, which benefit its members and all persons involved in this widely diversified industry, as well as govemment agencies and the general public. AESF dissemi- nates technical and practical information through its monthly joumal, Plating and Surface Finishing, and through reports and other publications, meetings, symposia and conferences. Membership in AESF isopen to allsurface finishing professionals as well as to those who providesetvices, supplies, equipment, and support to the industry.

According to the guidelines established by AESF's Meetings and Symposia Committee, all authors of papers to be presented at SUWFIN@have been requested to avoid commercialism of any kind, which includes references to company names (except in the title page of the paper), proprietary processes or equipment.

Statements of fact or opinion in these papers are those of the contributors, and the AESF assumes no responsibility for them.

All acknowledgments and references in the papers are the responsibility of the authors.

Published by the American Electroplatets and Surface Finishers Society, Inc. 12644 Research Parkway Orlando, FL 32826-3298 Telephone: 407/2814441 Fax:407/281W

CoPyriaht 1992 bv American EiectmDiaten and Surface Finishers Sociehr. Inc. All riahts resewed. Printed in the United States of ' Amen& This wbiicat~on may not be reproduced. stored in a retnevai sysiem. or trani" in whde or part. in any form or by an1 , Dhotccopying, recotding. or omemise mthout the prior wnnen permission of AESF, 12644 Research

Printed by AESF Press

SUWFIN'is a regisbered trademark of the American Electinplaters and Surface Finishers Society. Inc