US 3869356 A
A ferromagnetic layer having high information density capabilities is deposited on a flexible substrate in a thin, continuous, uniform and coherent manner utilizing a process which involves coating the substrate with a smooth layer of gold and a copolymer of vinylidine chloride and acrylonitrile, chemically depositing a layer of copper thereover and electrodepositing the ferromagnetic layer of cobalt no greater than 20 microinches, on top of the copper. A chemical plating bath of copper formate is used to plate the copper while the cobalt is deposited in the form of hexagonal crystals from a bath of cobalt chloride buffered with a low molecular weight carboxylic acid in a pH range of between about 3.6 and 5.8. The cobalt layer possesses a nominal coercivity within the range 200-500 oersteds which, upon exposure to a 60 cycle per second alternating current field having a strength of 1000 oersteds, exhibits a coercivity tolerance not greater than about 40 per cent of the nominal coercivity. This results in an extremely rapid switching time, thereby permitting theretofore unattainable high bit density capabilities.
Description (OCR text may contain errors)
finite @tates [1 1 Long et al.
METHOD OF MAKING A THIN, FLEXIBLE MAGNET1C MEMORY LAYER Inventors: Kenneth E. Long, Cleveland Heights, Ohio; David W. Taylor, Edgemont, Pa.
Nico Magnetics lnc., Wilmington,
Filed: July 5, 1973 Appl. No.: 376,855
US. Cl. 204/38 R, 204/38 B, 204/48, 117/130 E. 106/1 Int. Cl. C23f 17/00, C23b 5/08 Field of Search.-. 204/38 B, 38 R, 48; 3 117/130 E; 106/1  References Cited UNITED STATES PATENTS 3,360,397 12/1967 Koretzky ..'..l17/54 3,423,214 l/l969 Koretzky 117/54 Primary Examiner-R L. Andrews Attorney, Agent, or Firm- -.lam'es A. Lucas, Esq.
' 1 Mar. 4, 1975 57 ABSTRACT A ferromagnetic layer having high information density capabilities isdeposited on a flexible substrate in athin, continuous, uniform and coherent manner utiliz ing a process which involves coating the substrate with a smooth layer of gold and a copolymer of vinylidine chloride and acrylonitrile, chemically depositing a layer of copper thereover and electrodepositing the ferromagnetic layer of cobalt no greater than 20 microinches, on top of the copper. A chemical plating bath of copper formate is used to plate the copper while the cobalt is deposited in the form of hexagonal crystals from a bath of cobalt chloride buffered with a low molecular weight carboxylic acid in a pH range of between about 3.6and 5.8. The cobalt layer possesses a nominal coercivity within the range 200'500 oersteds which, upon exposure to a 60 cycle per second 1 14 Claims, 8 Drawing Figures Ill Almlll I1 Ill-- I11 SIGNAL vows Ill] 1000 FIELD STRENGTH (OERSTEDS) +1000 LAJ llf'rllll sum 2 0F 4 llllllll PATENIEDHAR 4:975
-!OOO FIELD STRENGTH (OERST EDS) +IOOO II III! IIIII lll Tl llH llll -IOOO FIELD STRENGTH (OERSTEDS) +IOOO PATENTEU 4 I975 sum 3 pr 4 Illl Iljl FIELD STRENGTH (OERSTEDS/ +1000 -IOOO (PRIOR ART) l||| Illl llll yll llll III'I Ill! HIT ill] II -ICOO FIELD STRENGTH (OERSTEDS) +|OOO (PRIOR ART) PAIENIEBH R 191s W VI Summ ng FIELD STRENGTH (OERSTEDS) +|OOO (PRIOR ART) fin IIII III] llll Illl lllllyfll :000 FIELD STRENGTH (OERSTEDS) +1000 (PRIOR ART) I. I I METHOD-OF MAKING A THIN, FLEXIBLE vIVIAGIIWE'lllIQ MEMORY LAYER BACKGROUND OF THE INVENTION Prior to the present invention, ferro-magnetic memory layers of several types have been available commercially, and countless others have been postulated by enterprising inventors in paper patents. The commercially available layers are formed on rigid substrates such as discs and drums as well as on flexible substrates such as'tapes, cards, and the like, depending upon the intended end use. In general, ferro-magnetic layers comprising particles of metal oxides dispersed in organic binders applied to such substrates constitute a large percentage of'cornmercia-l production to date, but thin layers of mixtures or alloys of metals without any binder overcome many of the problems connected with the oxide coatings, and these are beginning to find commercial favor for avariety of reasons.
The problems connected with oxidecoated substrates preventbit densities in excess of about 3,000 hits per linear inch (B.P.I.). First, the relatively low volume of magnetic material capable of being dispersed in an organic binder limits the total magnetic energy in any given thickness of deposition, and thus puts a low finite limit on the desirably high signal voltage. Secondly, the low permeability of such coatings, and the relatively low saturation induction prevents the utilization of coatings thinner than about 100 micro-inches, and the popular commercial types are between'l25 and 400 micro-inches in thickness. The consequence of suchthicknessjis an undesirably long flux penetration time causing low bit density limitations.
Even more important, the broad range of magnetic coercivities measured in oersteds between essentially fully magnetized states, or coercivity tolerance, due
inlarge part to broad particle size and density distribution in such layers, is directly responsible for poor switching time from a first fully mangetized state through complete de-rnagnetization to a second fully magnetized state. This limitation makes higher bit density capabilities virtually impossible, as explained in more detail below. Other problems connected with the oxide coated substrates are the tendency of the particles undesirably to abradethe write-read transducers, and the gradual disintegration of the binder causing undesirable fault counts.
The ferro-magnetic layers formed by depositing metal alloys on a substrate do not make use of organic binders, and hence there can be no faults arising from binder disintegration. However, metal depositions are in some cases thinner than oxide coatings and are susceptible to mechanical damage, such as abrasion, either of the deposition itself, or the transducers associated therewith which may cause loss of stored information. It is thus advisable to protect the ferro-magnetic layer with a coating formulated to be low in abrasion and sufficiently hard to absorb impacts without damage to the underlying ferro-magnetic layer.
Such coatings create a significant loss in signal, how- Metal'alloy depositions presently in use, such as nickel-cobalt are somewhat higher in permeability, and thus can be made thinner to attain potentially lower flux penetration time and higher bit density capability. But the nickel and cobalt are frequently alloyed with other materials such as phosphorous or sulfur in order to con- I trol the desired magnetic parameters. The addition of ever, due to the separation distance between the translimitations.
these materials and others which have similar properties, inevitably results in an increase in the range of coercivities and a reduction in the magnetic permeability. The latter requires a thicker deposition for equivalent signal, and the broadening of the band of coercivities indicates that it takes a relatively-long time to switch the magnetic energy from one state of magnetization to another again causing low bit density limitations.
The use of nickel is considered disadvantageous because nickel has alow saturation induction as comthat error signals can readily occur. Thus, the presence of secondary coercivities is equivalent to long switching time with additional hazard of adding error signals to the recorded magnetic layer. In other cases, alloys of magnetic metals will create a non-symmetrical hysteresis loop and a large coercivity tolerance indicating poor switching time, although the nominal coercivity and magnetic permeability may be adequate for present day bit density capability requirements, I
' Considerable research and development has been conducted over a period of years to solve the foregoing along recognized problems and to produce a ferromagnetic layer with bit density capabilities suitable for use with future generations of digital computers. This is substantiated, for instance, by Moline US. Pat. No. 2,730,491 (January 1956) which discloses the production of a ferro-magnetic, layer consisting of a nickelcobalt alloy plus sulfonamides having a nominal coercivity of about 230 oersteds, but the layer is 400 microinches thick which explains the extremely slow switching time indicated by the large coercivity tolerance evident from FIG. 2. I I
The Koretsky US. Pat. No. 3,360,397 (December 1967) discloses the chemical deposition of a ferromagnetic layer of cobalt from a bath maintained at a high pH. A citrate and/or malonate ion is used in the bath as a complexing agent. The layer is 250-5000 angstroms thick with a nominalcoercivity of 600-700 oersteds, but theicoercivity tolerance evident from FIG. 2 is over percent thereof thus indicating a slow switch ing time.
In a later patent, US. Pat. No. 3,423,214, Koretzky discloses the further use of an electroless cobalt bath to deposit a ferromagneticlayer on to a'catalytic'surface. The bath is maintained at a high pH of 8-9 by use of ammonium salts as a buffer. A short chain dicarboxylic acid is used as a co'mplexing agent. The product has relatively low coercivities in the range of 20-200 oersteds.
BRIEF DESCl Q IPTION OF INVENTION The present invention relates broadly to a method of forming a thin ferromagneticmemory layer on a flexible substrate whereby many of the aforementioned problems are overcome. More specifically, it covers a method of forming on thin, flexible film or substrate, aferromagnetic layer having a high bit density, short switching time and good resistance to abrasion.
The method comprises applying to a thin film a smooth continuous polymeric coating containing gold from a solution of LiAuC1 and a copolymer of vinyli dine chloride and acrylonitrile, depositing an electroless coating of copper from a plating bath containing copper formate and thereafter electrodepositing a layer of cobalt from a plating bath buffered in a pH of 3.6 to
5.8 with a substituted or unsubstituted mono or dicarboxylic acid.
The smooth continuous film ofa copolymer of vinylidine chloride-acrylonitrile and gold is roll coated on to the substrate, such as Mylar polyester film, to a thick ness of between and 150 millionths of an inch. The gold is typically present in an amount of less than 1.5percent by weight of the dry film.
Next, the copper is deposited from an aquous plating bath containing copper formate formaldehyde and ethylene diamine tetra acetic acid and maintained at a pH of between 12 and 13 with sodium hydroxide.
Asecond layer of copper may be electrodeposited on top of the copper layer if desired. The cobalt layer is then electrodeposited on the copper to a thickness of between about 4 and microinches from a bath of CoCl buffered with anacid such as monochloroacetic acid. A current density of at least 150 asf is typically used for plating on to the film moving through the plat ing bath at a speed of at least about 8 feet per minute. The cobalt is then covered with a protective layer of graphite.
BRIEF DESCRIPTION OF THE DRAWINGS Numerous advantages of the present invention will become apparent to one having ordinary skill in the art from a reading of the detailed description in conjunction with the accompanying drawings, wherein similar reference characters refer to similar parts, and in which:
FIG. 1 through FIG. 8 are illustrations of the oscilloscope display of a conventional B-H meter. FIG. 1 depicts the unintegrated hysteresis loop showing the magnetic characteristicsof a preferred embodiment of the ferro-magnetic layer of the present invention;
FIG. 2 depicts the hysteresis loop showing other magnetic characteristics of the ferro-magnetic layer of FIG.
FIG. 3 depicts the unintegrated hysteresis loop showing the magnetic characteristics of a ferro-magnetic layer having a relatively low nominal coercivity within the scope of the present invention;
FIG. 4 depicts the unintegrated hysteresis loop showing the magnetic characteristics of a ferro-magnetic layer having a relatively high nominal coercivity within the scopeof the present invention;
FIG. 5 depicts the unintegrated hysteresis loop showing the magnetic characteristics of a ferro-magnetic layer composed of metal oxide particles dispersed in an organic binder as presently available commercially;
FIG. 6 depicts the hysteresis loop showing other magnetic characteristics ofthe ferro-magnetic layer of FIG.
FIG. 7 depicts the unintegrated hysteresis loop showing the magnetic characteristics of a ferro-magnetic layer composed of a mixture or alloy of metals deposited on a flexible tape substrate as presently available commercially; and
FIG. 8 depicts the hysteresis loop showing other magnetic characteristics of the ferro-magnetic layer of FIG. 7.
DETAILED DESCRIPTION The following detailed description is divided into three sections to facilitate understanding. The first section discloses the magnetic characteristics indicating higher bit density capabilities of the ferro-magnetic layer produced according to this invention, the second discloses the ferro-magnetic layer with reference to its physical characteristics, and the third discloses the novel methods and techniques for depositing the ferromagnetic layer on a flexible substrate such as tape.
1. Magnetic Characteristics It has heretofore been conventional to define bit density capabilities of ferro-magnetic layers by reference to the bit density requirements of the write" and read transducers with which the layer is used in a computer. but this by definition is inadequate for measuring bit density capabilities which are beyond the ca pability of the transducer in question. For instance. inherent losses in the magnetic materials which are an essential component of the transducers undermines their reliability in determining the bit density capability of a given magneticlayer. In addition, since bit density ca-.
pabilities are ordinarily defined as bits per linear inch,
this value introduces as a variable the component of relative speed between the transducer and the magnetic layer. Other variables, such as the frequency of the pulses in the transducer, the spacing between the transducer and the magnetic layer, and the like, render the bit density capability concept too vague to be useful for comparison purposes.
A more reliable manner of defining bit density capability is a determination of switching time which cen be accurately and reliably measured by direct reference to the unintegrated voltage curve of aspeciman which is cyclically magnetized under prescribed test conditions.- The unintegrated voltage curve may be used to compute switching time in fractions of a cycle or fractions of a second. These fractions are also indicated with less precision on the integrated curve or hysteresis loop which is'commonly used for measuring other magnetic characteristics, but the unintegrated voltage curve is considered'much more reliable in determining switching time under prescribed test conditions.
Switching time as used herein isthe length of time required to take a given area of a ferro-magnetie sample from a first essentially fully magnetized state through complete demagnetization to a second essentially. fully magnetized state. In terms of the unintegrated voltage curve, switching time is computed from that portion of acycle occupied by the voltage curve during its excursion between substantially horizontal portions indicating first and second states of essentially full magnetization. In terms of the. more conventional hysteresis loop.
I switching time may in many cases be computed from age peak, and on the hysteresis loop switching time is suggested by the angle of the slope as it passes through horizontal axis against signal in the sample, measured in volts, on the vertical axis. On the oscilloscope display of the signal curve, states of essentially full magnetization are indicated by substantially horizontal portions of the curve generally superimposed on the horizontal axis. The switching ofthe sample from a first fully magnetized state through complete de-magnetization to a second fully magnetized state is indicated as the curvedeparts from the horizontal axis and traces a curve forming a peak indicating complete de-magnetization and returns to the horizontal axis. The point at which the sample becomes completely de-magnetized is con-, ventionally referred to as the samples coercivity, or H expressed in oersteds, which indicates the field strength gauss, on the vertical axis. On the oscilloscope display of the hysteresis loop, states of substantially saturated induction are indicated by essentially horizontal portions of the curve parallel with but spaced from the horizontal axis. Switching time is indicated as the curve departs from the horizontal and traces a curve passing through the horizontal axis indicating complete de-magnetization and returns to the horizontal. The coercivity H, of the sample is indicated by the point at which the loop crosses the horizontal axis. The term nominal coercivity is used herein to designate H of the sample as measured in oersteds.
The term coercivity tolerance (W) is used herein to designate that portion of a cycle wherein the sample is less than essentially fully magnetized in either direction. In terms of the unintegrated voltage curve on the oscilloscope display, the coercivity tolerance can be expressed in oersteds by measuring the distance along the horizontal axis between the point where the voltage curve departs from the horizontal to the pont where it returns to the horizontal. This can-be spoken of as the width of the voltage peak at its base. In terms of the standard hysteresis loop, the coercivity tolerance in oersteds is indicated by the horizontal distance be-' tween the point where the curve departs from the horizontal to the point where it returns to the horizontal. This can be spoken of as the slope and shape of the loop.
Switching time (in seconds) is directly proportional to coercivity toleranceand represents the time required to go from zero field to a fully magnetized state and back to complete demagnetization. It is determined by dividing the coercivity tolerance (W) by twice the coercivity (HE) and by the frequency in cycles per second of the field strength. For instance, for a sample exposed to a 6 cycle per second alternating current field havind a strength of 1000 oersteds which shows on the 8-H meter oscilloscope display a coercivity tolerance of about 100 oersteds, the switching time is 100/(2 X 1000 X 6 0)- or 0.833 milliseconds. Thus the switching'time under prescribed test conditions for a given sample. is readily and reliably determined by reference to a standard B-H meter.
To illustrate this more graphically, reference is made to FIGS. 1-8. These are illustrations of the oscilloscope display of a conventional B-H meter, in this case a Model 65lB B-H Meter available-from Scientific Atlanta, Inc., Atlanta, Georgia, equipped with an oscilloscope which is sufficiently responsive to extremely rapid rise times such as a Tektronix RM503 Oscilloscope. The sample in every case has been exposed to a conventional and easily reproducible 60 cycle per second alternating current field having a strength of 1000 oersteds, prescribed test conditions in which inductance and capacitance do not substantially influence the sample being measured. I FIG. 1 depicts the-unintegrated hysterosis loop or signal curve'showing the magnetic characteristics of a preferred embodiment of the ferro-rnagnetic layer of the present invention, and FIG. 2 depicts the hysteresis loop showing other magnetic characteristics of the same layer. In FIG. 1, the nominal coercivity H is indicated by thehorizontal spread between-the two voltage peaks. With the oscilloscope trace properly centered with respect to the verticalaxis of the oscilloscope grid, the nominal coercivity is seen to be about 300 oersteds in both the plus and the minus direction. Similarly in FIG. 2, by properly centering the trace with the grid, the points at which the hysteresis loop crosses the horizontal axis indicates a nominal coercivity ofabout 300 oersteds in both the plus and the minus direction.
To determine the coercivity tolerance W of the sample, thewidth of the voltage peak at its base in FIG. I is measured in oersteds along'the horizontal axis, and is seen to be about 60 oersteds. Similarly, in-the hysteresis loop of FIG. 2, the coercivity-tolerance W of the sample is seen to be about60 oersteds. A coercivity tolerance of about 60 oersteds is only about 20 percent of the 300 oerstednominal coercivity ofthe sample. A coercivity tolerance of about 60 oersteds, representing 1 about 1/33 of the time required to trace a full cycle;
can be computed under the prescribed test conditions as a switching time of about. 0.0005 second. Such a switching time is extremely rapid for a sample having a coercivity as high as about 300 oersteds, indicating that the field strength will rise to a value of about 270 oersteds before any individual portions ofthe crystalline structure of the sample even begin to become de-' magnetic layer having a relatively low nominal coercivity within the scope of the present invention; the corresponding hysteresis'loop 'of the same layer is not illustrated. The nominal-coercivity H is seen to be about 200 oersteds, and the coercivity tolerance W is seen to be about oersteds. A coercivity tolerance of about 80 oersteds is about 40 percent of the 200 oersted nom-' inal coercivity. A coercivity tolerance of about 80 oersteds represents a switching time of about 0.00067 second, which is materially faster than heretofore possible in a ferro-magnetic layer having an II value of 200 oersteds.
FIG. 4 depictsa stylized unintegrated hysteresis loop showing the magnetic characteristics of a ferromagnetic layer having a relatively high nominal coercivity within the scope of the present invention; the corresponding hysteresis loop of the same layer is not illustrated. The nominal coercivity H is seen to be about 500 oersteds, and the coercivity tolerance W is seen to'be about 200 oersteds. A coercivity tolerance of about 200 oersteds is about 40 percent of the 500 oersted nominal coercivity. A coercivity tolerance of organic binder as presently available commercially,
and FIG. 6 depicts the hysteresis loop showing other mangetic characteristics of the same layer The voltage peak of FIG. 5 is extremely broad and curvalinear at the base before blending with the horizontal portions 600 oersteds represents a switching time of about 0.005
second, which is' extremely slow for a ferro-magnetic layer having an H value of 300 oersteds.
FIG. 7 depicts the unintegrated hysteresis loop show ing the magnetic characteristics of a fe'rro-magnetic layer composed of a mixture or alloy of metals deposited on a flexible tape substrate-as presently available commercially, and FIG. 8 depicts the integrated loop showing other magnetic characteristics of the same layer. The voltage peak of FIG. 7 is relatively broad at the base and ragged on one side before blending with the horizontal portions representing full magnetization, and the hysteresis loop of FIG. 8 crosses the horizontal axis at a low angle and curves gradually on one side and more abruptly on the other side of the horizontal axis before joining the horizontal portions representing saturated induction. The nominal coercivity H is seen to beabout 550 oersteds, and the coercivity tolerance W is seen to be about 360 oersteds. A coercivitytolerance of about 360 oersteds is about 65 percent of the 550 oersted nominal coercivity of the sample. A coercivity tolerance of about 360 oersteds represents a switching time of about 0.0029 second, which is insufficient even for a ferromagnetic layer having an H value of 550 oersteds.
The ratio of saturated induction (8,) to remenant induction (8,) as determined from a conventional hyster- 8 switching time, which is directly related to bit density capability since it is well known that the rate of change of magnetization, ratherthan absolute magnetization, is responsible for the useful signal in the read transducer of a digital write-read" system.
2.'The Ferro-Magnetic Layer The ferro-magnetic layer exhibiting the foregoing magnetic characteristics is comprised primarily ,of cobalt metal desposited in a uniform layer. The layer is as thin as possible consistent with the ability to create a flux reversal of sufficient magnitude to generate a signal voltage in the read transducer sufficiently large to discriminate from inherent electrical noise in the system.
As applied, to a flexible tape for digital computer use in which thefwrite-read" transducers normally contact the tape surface, the layeris no greater than about 15 micro-inches in thickness, and preferably of the order of 10 micro-inches or less-in thickness. The thinness requirement will also vary with the signal requirements at various frequencies employed in, and physical relationships encountered between, the magnetic layer and the read transducers in such applications as discs, tapes, drums, wire, flexible and rigid cards, static surfaces as in memory banks, and the like. I
A desirable requirement to obtain the physical and magnetic features of the magnetic layer is a surface which is smooth to. the point of being optically reflective, with an arithmetical average variation of less than about 2 micro-inches. A decrease in the smoothness of the layer appears to have an adverse effect upon its magnetic properties.
In addition to the extreme thinness, the ferromagnetic layer is continuous with no substantial abber ations, uniform within extremely close tolerances. physically coherentwith no gaps or voids, electrically conductive, free from undesirable mechanical stress and, particularly in the case of flexible substrates, sufficiently ductile to pass around small diameter spools. To this end, it is now believed that the layer should consist essentially of hexagonal crystal form cobalt metal. The hexagonal form, when deposited in the manner described below, will exhibit the magnetic properties obtained with the thin, continuous, coherent and conductive layer.
The magnetic layer consists of the deposition products of a plating solution containing prescribed amounts of a cobalt salt or salts in aqueous solution. This is now believed to produce, for instance when the layer is obtained by electrodeposition, close-packed hexagonal crystals of essentially pure cobalt free of undesirable impurities and containing withinthe grain boundaries of the crystals minor amounts of cobalt in a form or forms other than pure cobalt. These minor amounts are now believed to be essential to achieve the desired magnetic parameters, and they may be controlled within close limits in the manner described below by the operative factors of the presently preferred methods of practicing this invention. Poisoning agents such as phosporous, sulphur and the like must be avoided.
An important feature of this invention isto produce the thinnest possible magnetic layer to give a saturated induction (8,) consistent with maximum squareness of the hysteresis loop. This induction is conventionally measured as the vertical height between the horizontal portions of the hysteresis loop. To this end, elemental 9. cobalt is desirable because it is possible to obtain the highest possible saturated induction with the least possible thickness of'magnetic material. Alloys of cobalt and iron although exhibiting high magnetic induction are not believed useful because they do not lend themselves to control of coercivity comtemplated by this invention. Nickel is similarly not useful, either in its pure state or alloyed in more than minor amounts with cobalt, despite excellent squareness because it has low saturation induction and inherently low coercivity.
Furthermore, under many conditions of deposition, nickelcobalt alloys do not necessarily exhibit a true alloy condition, and the magnetic properties are far from ideal. In some cases, the magnetic layer will exhibit more than one major coercivity, with the .second and tertiary coercivities at amplitudes such that the coercivity tolerance is increased and error signals can readily occur. Thus, the presence of secondary coercivities is for all practical purposes equivalent to a long switching time, with the additional hazard of adding error signals to. the recorded magnetic layer. In other cases, alloys of magnetic metals will create a non-' symmetrical hysteresis loop, which will also exhibit poor switching time, although the magnetic properties such as the nominal coercivity and magnetic permeability may be adequate for bit density capabilitiesin common use today.
3. Deposited on a Flexible Substrate The tape comprises, very basically, an elongated flexible mechanical support for the ferromagnetic layer, such as a polyester base film of desired width and thickness. Tapes ofpolyamide or cellulose acetate, although not possessing all of the favorable characteristics of polyester, can likewise be utilizedas a substrate. The non-magnetic side or back of the tape is coated with a conductive layer of a carbon-loaded plastic to eliminate static charge during computer operation. On the front or magnetic side of the tape, there is adhered to the polyester film a layer of plastic containing activating materials to form catalytic nuclei suitable for the plating ofcopper and to achieve improved surface smoothness. A layer of copper is chemically plated on the activated tape and the copper layer provides a suitable base for the electroplating of the extremely thin ferromagnetic cobaltlayer which provides the superior magnetic properties of the tape according to this invention. Finally, a protective coating of graphite is providedon the magnetic cobalt layer to complete the magnetic side of the tape.
The magnetic tape is fabricated in a continuous process which deposits the necessary layers in sequence. The first step consists of the provision of a roll of polyester base film approximately 0.00l inch thick. Film three inches wide in rolls ofabout 3,500 linear feet are currently available from 'Celanese, FMC and others.
The roll of film may be unwound according to conventional means and drawn through suitable apparatus for for practical purposes be 0.1-1.5 percent of the weight of the dry Saran film. This dry film has a thickness of 10 feet per minute is immersed for about 20 seconds, and such cleaning operation is deemed satisfactory.
Third, a layer of a solution of 5-10 percent Saran F'- l20 or Fl30 (a trademark of the Dow Chemical Co., covering a copolymer of vinylidine chloride and acrylonitrile or other suitable polymer'or copolymer) with LiAuC1 '2l-l O in a high boiling solvent, such as methyl ethyl ketone, methyl isobutyl'ketone or mixtures thereof, is applied by roller coating with a steel roller. The LiAuCl,2H O is typically prepared by-adding lithium hydroxide to tetrachloroauric acid solution in stoichiometric amount and evaporating to dryness on a water bath. The amount of gold in the Saran'coating is notcritical, but does have an effect on the time required todeposit the initial coating of copper, discussed below. Small amounts of gold require a longer time than larger amounts. Since other factors such as the composition of the electroless copper bath, its pH and temperature also effect the time for the initial coating of copper, however, the amount of gold used should between 5 and 150 millionths of an inch.
Next a layer of copper is chemically plated on the smooth Saran coating from a solution prepared as follows:
16.7 g copper formate crystals 107 cc of tetrasodium salt of ethylene-diamene tetra-- acetic acid 30% by weight) cc of sodium hydroxide (25% by volume) are made to one liter with water. When the plating solution is ready for use, 20 cc of 37% of formaldehyde solution, 12 cc of 25%sodium hydroxide, 20 cc of isopropyl alcohol, and 1 cc of 1% Duponol WAQE (a wetting agent comprising laural alcohol sulfate) are added, after which the solution is heated to between 5560C with air agitation to prevent spontaneous decomposition of the bath. This bath is capable of chemically plating a nearly opaqueconductiv e copperlayer on the moving film in 30 to 60 seconds for a period of about, minutes. The addition of more formaldehyde and sodium hydroxide allows the bath to be used for a longer period of time while the addition of more copper formate solution serves to replace the cop- 'per that is plated out. A copper'film thickness of less than 5 millionths of an inch has been found to be satisfactory.
As an alternative, the following chemical copper This electroless copper plate obtained in 30 to 60 seconds with either of the two aforementioned baths is sufficiently conductive, to act as a base for the subsequent electrodeposit of cobalt at medium current densities.
However, a second layer of. copper can be electrodeposited as an optional step on top of the chemically deposited layer, prior to the deposition of the cobalt layer. This will increase the conductivity and allow the use of higher current desities duringv the cobalt deposition. A 30 second electrodeposit at e.g. 40 amps from a bright acid copper bath such as a sulfate or fluoborate bath has been found to be suitable for the deposit of this second'copper layer.
Preparatory to depositing the cobalt layer, it is necessary that the copper layer be rinsed and dried of surface liquids. A final rinse in alcohol which eliminates water droplets and evaporates rapidly has been found to prevent local dilution of the cobalt solution which is next applied.
Next, a layer of cobalt is applied by electroplating techniques. This is accomplished by providing a tank containing a cobalt chloride solution, such as a percent by weight solution of cobalt chloride hexhydrate (CoCl 6H O) in water. A solid slab of electrolytic cobalt metal immersed in the solution serves as the anode, and the copper layer on the film which is passed through the solution acts as the cathode. A current end ofthis range is preferable. More important than the particular level of the temperature, it now appears that the primary consideration is maintaining a constant temperature during the plating operation and avoiding a fluctuations uppor down from that temperature. A pH value of about 3.6 is considered a minimum'to obtain a closepacked hexagonal structure in the plated cobalt. A ph value higher than about 5.8, on the hand, has a tendency to deposit basic cobalt salts instead of cobalt on the copper as desired. lt-has been found that the cobalt layer in the finalproduct has a higher signal if the film is maintained under maximum recoverable longitudinal tension as it progresses through the plating bath, and is thereafter permitted to relax upon completion of the cobalt. plating operation.
The purity ofthe cobalt plating solution is of primary importanceQ Impurities generally found in commercially available cobalt, and particularly copper, have an undesirable influence on the magnetic qualities of the plated cobalt layer. The cobalt as plated must be relatively free of impurities normally native to cobalt such as iron, copper, zinc, lead, bismuth, tin, and gold, although small amounts, for example lpercent of nickel may be tolerated.
A graphite coating is applied to the cobalt from a graphite dispersion. This coating is appliedin a very thin layer and is tenaciously adherent to the cobalt. It may, if necessary, be burnished-to remove excess material to avoid contaminating the read" and write heads of the computer on which the tape is used. Such a layer provides protection against chipping, abrasion and the like which'give rise to fault counts in previously available tapes. The graphite layer protectsthe cobalt from direct contact with the guides, rollers, tensioning devices, cleaning knives, wiping pads, and similar equipment found in digital computers.
Lastly, a layer of carbon-loaded plastic is applied to the back side of the polyester base film to provide electrical conductivity in order to remove static charge from the tape during operation. Finally, the completed tape is wound up into a roll suitable for slitting to proper widths for various uses, such as use in digital computers.
While the above described embodiments constitute the presently preferred mode of practicing the invention, other embodiments and equvalents are within the scope of the actual invention, which is claimed as:
1. Method of producing an electromagnetic tape comprising:
a. coating a surface of a thin film'substrate with a smooth layer of a copolymer of acrylonitrile and vinylidine chloride and containing gold,
b. thereafter chemically depositing a layer of copper upon said smooth layer from an electroless plating bath containing copper formate as the source of copper ions,
c. electrodepositing a layer of cobalt on top of the copper from a plating bath containing cobalt ions and buffered in a pH range of between about 3.6
and 5.8 with a low molecular weight substituted or unsubstituted carboxylic or dicarboxylic acid. and
d. applying a protective coating over the layer of cobalt.
2. The method according to claim 1 wherein the cobalt is added to the cobalt plating baths as CoCl -6H O and monochloroacetic acid is used as the buffer.
3. The method of claim 2 wherein the cobalt is deposited to a thickness of between about 4 and about 20 microinches using a filtered rectified AC. curreht of 15 amps at 8 volts, and a cobalt anode.
4. The method of claim 3 further using an electically conductive cathode roller located above the liquid level of the plating bath with an anode to cathode spacing of about 0.25 inches.
5. The method ofclaim 1 wherein the layer of copper is deposited from a plating bath prepared by mixing to gether the following:
copper formate ethylene diamine tetraacetic acid formaldehyde sodium hydroxide, and
water and maintained at a pH of between about 12 and 13.
6. The method of claim 5 wherein the copper is maintained at a level of between about 0.8 and 5.0 g/l in the bath and the ethylene diamine tetraacetic acid is maintained at a level of between about 9 and about 45 g/l v in the bath. v
7. The method of claim 1 wherein the copolymer of acrylonitrile and vinylidine chloride is deposited on the film substrate from a ketone, solution to which the gold is added as lithium gold chloride.
8. The method of claim 1 wherein a layer of copper is electrodeposited over the chemically deposited layer of copper.
9, A method ofproducing a thin, continuous, uniform and coherent ferromagnetic layer upon a flexible substrate wherein the layer consists essentially of hexagonal cobalt crystals, has a thickness of less than about 15 microinches and has a nominal coercivity within the rangeof 200 to 500 oersteds, a switching time of less than lmillisecond and a coercivity tolerance no greater than 40percent of the nominal coercivity when exposed to a 60 cps field having a strength of 1,000 oersteds comprising:
13 14 about 1.5 weight percent of gold from a ketone sosisting of acetic acid, dig lycolic acid, chloroacetic lution containing the gold added as lithium gold acid and malonic acid and using a cobalt metal anchloride, ode. Chemically depositing a layer of pp p the l0.'The method of claim 9 wherein copper is deposfimooth activated layer from electrolfis Plating ited to a thickness of less than about 5 microinches. bath maintained at a temperature of between about ll. The method of claim 10 whemin a secofid layer 50 and about 70C and composed primarily of copper formate, formaldehyde, ethylene diamine tetraacetic acid and sufficient sodium hydroxide to maintain the bath at a pH in the range of about 12 10 to 13;
12. The method of claim 9 wherein the flexible substrate consists of a polyester filml 13. The method of claim 9 wherein the coblat is electrodepositing to a depth of between about 4 plated at a curfenhdensilty of at least amp/cmz and microinches a ferromagnetic layer of cobalt from a bath maintained at a temperature of between on to the copper surface from a plating bath using about, 180 and I v a filtered, rectified A.C. current, said bath contain- 15 Thie method of Clalm 9 further Including pp y I of copper is electrodeposited on top of the first layer.'
ing cobalt ion dd d as C C1 -6H O d b ff d a thin protective layer over the cobalt comprising a disto a pH of between about 3.6 and 5.8 with a mono persion of graphite in sodium silicate or dicarboxylic acid selected from the group con-