The substrate is an aluminum annulus. The specifications for each product demand flatness, maximum thickness, and the outside diameter. The flatness of the disk is very critical as this rotating annulus undergoes deformation when rotated at high speeds―as it does in the disk drive, over 10,000 rpm. The substrates are typically less than 0.1 cm, and gauging the thickness is an important process for the subsequent operation of polishing and testing. At one point in time the outside diamter was 5 inches then shrunk to 3.5 inches, and an inner diameter of 2.5 cm. The spindle of the disk drive fit through the central whole.

Nickel is deposited onto the aluminum disk by electroless deposition. The thickness of the nickel coating is less than 12.5 m m (500 m -inches). A characteristic of the electroless nickel plating is the incorporation of phosphorus in the nickel deposit; and, it is not uncommon to refer to that layer as nickel-phosphorus. The nickel layer acts as a metallurgical buffer and adhesive layer between the subsequent chromium layer to be deposited and the aluminum substrate.

Due to the nature of the electroless nickel plating residual stresses are built into the nickel coating. Residual stresses are defined as those stresses which are internal to the material without an effect from an external source. These stresses may induce curvature into the substrate. The baking process serves two purposes: (1) to relieve the residual stresses, and (2) annihilate any magnetic properties of the nickel coating. If the residual stresses are not relieved the disk may warp, and if the coating is thin the nickel layer may crack. The complete cycle of baking takes about 5 hours, with soaking at a peak temperature at slightly less than 300 C for 1 hour.

The electroless nickel plating may result in rough surfaces and unacceptable uniformity. In the course of polishing a thin layer of the nickel is removed leaving behind a smooth surface. Following polishing, sample disks are tested for surface smoothness, i.e. to check for scratches and for the uniformity of thickness of the nickel coating.

The rate of nickel removal depends on the polishing pads, specific gravity of the slurry, variations in the type and condition (life) of the slurry and its pH, and the load applied during polishing. The polishing media are particles of industrial grade diamond, aluminum oxide, or any non-metallic hard particles with a predetermined particle size distribution. The chemistry of the slurry includes defloculants to keep the particles from settling; and, it may be necessary to continuously agitate the slurry. The quality control parameters for polishing are typically the remaining nickel thickness on each surface of the substrate, the average roughness (Ra), and the peak-to-valley length. The thickness of the remaining nickel is determined by X-ray Fluorescence (XRF). The metrology tools used for detecting scratches and roughness are based on light scattering techniques.

The texturing of the surfaces of the disk is a process to make the surfaces rough. When two extremely flat and smooth surfaces are brought close to one another they tend to stick. If the surface of the disk is extremely smooth then it will attract the head and that will cause damage to the disk and head. The purpose of texturing, thus, is to roughen the surface of the disk in a controlled way to avoid the head sticking to the surface of the disk. There are, notwithstanding reliability specifications and tests to determine wear or tribological characteristics of the disk in operation, and these include Fixed Track Contact-Start-Stop (CSS), Sweep CSS, Sweep Test, Fly Stiction (static friction). These topics are out of the scope of this presentation.

Texture roughness is controlled by the size of the particles in the slurry. For some products, a slurry with a single particle distribution might be used, and on others, two distinct particle size distributions might be used. In the latter case, the resulting texture has two overlapping textures, a fine texture and a coarse texture.

After texturing the disks are placed in a tank with di-ionized water with a special soap. The purpose of this post-texture clean or wash is to remove any particles or surface contaminations that might have adhered to the surface during the polishing and texturing operations.

There are certain specifications for the roughness of the surfaces of the disks of the different products. Roughness is determined by directing a fine laser beam onto the surface of the disk, and the beam is scanned a pre-determined length. The reflected beam at every scanned point goes through a set of sensors that translate the changes in intensity to numbers. If the resulting roughness numbers fall within the control limits for the product the disks of the lot pass to the next step.

The sputtering machine is designed with several stations each configured to deposit a single layer. The entire film-stack composed of the underlayer, the magnetic film stack, and the protective top-layer, are deposited in one run without breaking vacuum. In general, there are three distinct functional layers deposited on each surface of a disk, the first is the chromium film, the second is the hard-magnetic film of cobalt or its alloys, and lastly the top diamond-like carbon film. The role of the chromium layer is to help the adherence of the magnetic cobalt-alloy film to the nickel layer and to separate and prevent the cobalt layer from interacting magnetically (interfering with the orientation of the magnetic domains of the cobalt film) with the nickel. The top carbon film is reduce wear between the disk and the head.

The following tests check the quality of the sputtered magnetic film stack and the diamond-like carbon film. These are:

- the thicknesses of the different films

- Coercivity

- The product of the remnance and thickness. The product of remnance and the magnetic film(s) thickness is referred to as Mrt or Brt.

- Raman spectroscopy to evaluate the extent of bonding in the carbon film

Lubrication of disks is the process of applying a thin coating of an organic material – a few angstroms. The coating function is essentially to decrease friction between the head and the disk in case of contact. Baking helps adherence of the organic coating to the disk surface, as well as promotes the polymerization of the lubricant. Some lubricants require baking to polymerize and others polymerize by exposure to air and ambient temperature.

The surface of the disk may not be very smooth at the microscopic level after sputtering. Occasionally, the sputtered film may have ‘growths’ (as hillock growths) that may damage the heads on the testers. The burnishing processes utilizes a tape with embedded fine particles which smooth out any growths which may have happened during sputtering. The force applied onto the disk surface is a light force (about 1-2 pounds). Burnishing removes only the peaks of the aspirates on the carbon layer as well as a very thin layer of lube.

Back to Top of Page    Back to Part I    Go to Books

Good disks are the zenith of good process control throughout the manufacturing steps. Failure analysis labs usually have procedures mapping test failures to processes that were out of control at the time of manufacturing that disk. Further, certification tests are a testament to the exquisiteness of the disk-drive technology.

There are three consecutive tests to a test program. First, is the burnish test—a non-contact test—and is done first. Next, the glide test is to test for surface irregularities. And finally, the certification and modulation tests for the quality of the magnetic media, and these two are done simultaneously. Certification testing is to qualify the magnetic medium to the effectiveness of detecting magnetic transitions, that is reading an electrical pulse.

The disk is placed on a tester with different types of heads. The test would proceed as follows: the burnish and glide heads load at loading zone of the disk. The burnish head sweeps the surface and returns back to where it started. Next, the glide head sweeps the surface. The glide head is a piezoelectric or acoustic emission sensor to detect any impact between the head and any irregularity on the surface of the disk. Thus, glide defects, if any, are mechanical in their origin. Finally, the certification and modulation tests follow the glide test. The certification head tests for the magnetic film parametrics (changing timing, variables, or levels thereof), and these include: amplitude of the written and read-back bits, and over-written bits. The results obtained are indicative of the functionality of the disk to store data and to retrieve it. The general objective of the certification and modulation tests is to qualify the magnetic qualities of the magnetic layer through reading signals and reading them back. Magnetic defects do not allow the writing head to write clean bits on the magnetic medium for the read-head to legibly read those signal (or actually the change, or transition, in the magnetic fields). Thus, the certification and modulation defects are magnetic in their origin.

Here is a list of some of the certification and modulation tests that are done:

A. The Track Average Amplitude (TAA) test is to determine the strength of the written and read signals

B. Resolution for the clarity of the signal

C. Over-write is an indication of the quality of the magnetic layer

D. AC and DC Noise, Phase Margin, and Pulse Width, Carrier-to-noise ratio (CNR), and Bit Shift indicates the how well the disk will function when in operation

E. Missing Pulse (MP) Errors: In this test, the electronics searches for bits in the read-back signal that fall below a predetermined (voltage) threshold. The number of bits that fall below the pre-programmed value of the threshold voltage are counted as missing pulse errors. The length of an error is measured in bits. The byte location which the error belongs to is recorded relative to a physical mark made on the disk, the radius, sector and surface.

F. Extra Pulse (EP) Errors: In contrast to the missing pulse test, here the test begins with an erased disk. A fixed threshold is set, and voltage signals detected with values above the threshold are counted as defects.

G. Correctable and Non-correctable Determination: The classification of defects as correctable or non-correctable depends on the number of defects causing either MP or EP that pass through an error window of pre-determined duration

H. Positive Modulation and Negative Modulation: Modulation results are related to the uniformity in thickness of the magnetic layer. Positive modulation would be attributed to thicker regions of the chromium film; and, conversely, negative modulation to thinner parts of the film.

I. Phase Margin Analysis (Bit Shift) is to measure the timing accuracy of a recording system; this is done through a programmable frequency separator (on the order of 1 nsec) and a Phase Margin Detector (on the order of 500 psec).

Back to Top of Page

Finn Jorgensen, The Complete Handbook of Magnetic Recording, Fourth Ed., McGraw-Hill, New York, NY USA 1996.

Eric D. Daniel and C. Dennis Mee, Magnetic Recording Technology, Second Ed., McGraw-Hill, New York, NY, USA 1996.

Bharat Bhushan, Tribology and Mechanics of Magnetic Storage Devices, Second Ed., Springer-Verlag, New York, NY, USA, 1996.

A. M Taratorin, Characterization of Magnetic Recording Systems: A practical approach. San Jose, CA, USA, 1996.