This invention relates to a biaxially oriented copolyester film. More particularly, this invention relates to a biaxially oriented PENBB film suitable as a substrate for magnetic recording media with which a high quality picture can be obtained and which has excellent abrasion resistance.

As a substrate for magnetic recording media such as video tapes and audio tapes, biaxially oriented polyester films, especially PET films, are con¬ ventionally used. To give good slip properties to the recording media, the conventional substrates contain particles by which protrusions are formed on a surface of the film. The protrusions reduce the coefficient of friction of the film so as to give good slip properties to the media. It would however, still be desirable to improve the stiffness (tensile modulus) and mechanical strength of such film, which would allow the manufacturer to produce even thinner film, thereby increasing the recording capacity of cassettes manufac¬ tured therewith. At the same time, it would be desirable to still further improve dimensional stability.

United States Patent No. 4,720,412 discloses a biaxially oriented film suited as the substrate for magnetic recording media, which has protrusions of specific height and number. The film disclosed in this U.S. Patent has a drawback that the surface of the film is likely to be scratched by a roll or a guide which contacts the film in the process of applying the magnetic layer, in calendaring process or in the process of dubbing a finished video tape or the like to manufacture pre-recorded tapes. Since these processes are con- ducted at ever higher speeds, problems with scratching are becoming more serious. Further, conventional magnetic recording tapes have the disadvanta¬ ge in that when they are dubbed, the quality of the picture decreases, i.e., the S/N ratio (signal/noise ratio) of the dubbed tape is unsatisfactorily low. Accordingly, the object of the present invention is to provide a film suited as a substrate for magnetic recording media, which is more scratch resistant, even in a high speed process (hereinafter referred to as having good scratch resistance), and with which the degradation of the picture caused by

dubbing is reduced (hereinafter referred to as having good dubbing). Moreo¬ ver, the substrate film should be more dimensionally stable and should show improved mechanical strength and especially improved stiffness (tensile modulus) over conventional PET films.

This invention now provides a biaxially oriented PENBB film comprising PENBB and inert particles, which film has at least one surface with protru¬ sions formed by the existence of the inert particles, the surface having a parameter of surface topography (β/ ₀ ) of 0.1 to 0.7 and the average distance between adjacent protrusions being not more than 20 μm. Prefera- bly, the coefficient of friction of the surface and the parameter of surface

y ≤ 0.171 v (β7σ) + 0.25 (2)

topography satisfying the following equations (1 ) and (2): wherein y is the coefficient of friction, β is the mean flatness of the protrusions and σ is the standard deviation of height distribution of the protrusions.

Similar surface topographies are disclosed in EP-A 0 31 1 426 for PET films.

U.S. Patent No. 3,008,934 discloses copolyesters containing as acid derived units 4,4'-bibenzoate and a host of other dicarboxylates including

2,6-naphthalic dicarboxylate. It also discloses oriented fibers and films prepa- red from these copolyesters, however, biaxially oriented PENBB films are not disclosed or envisioned. In particular, those films with improved stiffness (tensile modulus) and tensile strength in both MD and TD as well as thermo- stability, UV stability, hydrophobicity, dimensional stability and impermeability toward gases in comparison to PET film are not disclosed in U.S. Patent No. 3,008,934.

According to the present invention, a novel biaxially oriented copoiye- ster film is provided, which is suited as a substrate for magnetic recording

media. By employing the PENBB film of the present invention as the sub¬ strate for a magnetic recording medium, the magnetic experiences reduced scratching, even in high speed processing, and the quality of the picture of the dubbed medium is higher than that of conventional media. Further, the PENBB film does not degrade the magnetic layer formed thereon and provides good slip characteristics to the recording medium. Moreover, the substrate film is more dimensionally stable and shows improved mechanical strength, especially improved modulus over conventional polyester films. The latter characteristics allow thinner film to be used, which in turn increases the recording capacity of cassettes. Thus, by employing the PENBB film of the present invention as a substrate, an excellent magnetic recording medium can be obtained.

The copolyester film of the present invention comprises PENBB and inert particles. PENBB as mentioned hereinbefore is a copolyester containing as acid-derived unit at least 5 mole-% of a radical of the formula

(bibenzoate, BB

In the case that more than 10 mole-% of terephthalic acid derived radicals are present in the copolymer, the content of bibenzoate derived units is at least 25 mole-%. Films of these copolyesters are mentioned in the unpublished German Patent Application P 4224161 .8, which is incorporated herein by reference. Preferably PENBB is a copolyester wherein at least 80 mole-% of the acid derived units (NBB) consist of bibenzoate (20 to 80 mole- %, preferably 40 to 60 mole-%) and naphthalate (80 to 20 mole-%, prefera¬ bly 60 to 40 mole-%). The remaining 20 or less mole-% may consist of other acid derived units, which e.g. affect the melting point or the crystallization kinetics. Preferably at least 80 mole-% of the diol-derived units consist of - O-(CH ₂ ) ₂ -O-units. The remaining 20 or less mole-% may consist of other diol- derived units, which e.g. may also affect the melting point or the crystalliza¬ tion kinetics. It may also be desirable to replace minor amounts of the acid-

and/or diol-derived units with hydroxycarboxylic-acid-derived units, e.g. such derived from p-hydroxybenzoic acid.

In order to achieve the desired mechanical properties in the biaxially oriented PENBB film it is recommended that the IV value (inherent viscosity) of the PENBB polymer after extrusion be > 0.5 dl/g and preferably > 0.55 dl/g.

Although not restricted, the preferred examples of the inert particles contained in the film of the present invention include essentially spherical silica particles originating from colloidal silica, synthetic calcium carbonate particles, titanium dioxide (rutile) particles, thermal type carbon black, and organic polymer particles such as crosslinked polymer particles, especially crosslinked styrene-divinyl benzene micro-spheres. Although the average diameter d {μm) and the content C (% by weight) of the particles are not restricted, if these satisfy the following equation (3), the scratch resistance is further improved and better dubbing may be obtained. Further, the mean distance between the adjacent protrusions required in the present invention may easily be obtained.

0.01/d ^. C ≤.0.45/d (3) Two or more kinds of particles may be incorporated in the film. In cases where a mixture of two or more kinds of particles with different aver¬ age diameters is employed, the average diameter d in the above-described equation (3) designates the overall average diameter of the particles, and the content C in equation (3) designates the overall total content of the particles. The preferred range of the average diameter of the particles is 0.2 - 0.7 μm for calcium carbonate particles and titanium dioxide particles.

"Non-external" particles may be used as the inert particles. " Non- external" particles herein means particles produced by precipitating at least one compound selected from the group consisting of calcium compounds, magnesium compounds and lithium compounds in situ during the polyconden- sation of the PENBB by adding a bonding agent. It should be noted that phosphorus and trace amounts of other metal components such as zinc, cobalt, antimony, germanium and titanium may be present in the non-external

particles in an amount not degrading the advantageous effects of the present invention. In cases where non-external particles are employed together with the added inert particles, the preferred content of the added inert particles ranges from 0.01 to 3.0 % by weight, since this amount promotes scratch resistant and good dubbing.

Although the film of the present invention mainly is composed of the above-described PENBB composition, as long as the advantageous effects of the present invention are not degraded, other polymers may be blended up to an amount of 20 % by weight. Further, the film of the present invention may contain inorganic and/or organic additives such as anti-oxidants, thermal stabilizers, lubricants, ultra-violet absorbers and crystallization nucleants in amounts conventionally employed in polyester substrates.

The PENBB film of the present invention is a biaxially oriented film. Non-oriented films and uniaxially oriented films are not preferred since the scratch resistance is poor, good dubbing cannot be achieved, and mechanical properties are insufficient for tape handling.

The film of the present invention has at least one surface on which protrusions are formed due to the existence of the inert particles. The para¬ meter of surface topography y/fβ/σ) of the at least one surface with protru- sions is 0.1 to 0.7, more preferably 0.15-0.65, still more preferably 0.25 -

0.62, wherein β is the mean flatness of the protrusions and σ is the standard deviation of the height distribution of the protrusions. The mean flatness of the protrusions and the standard deviation of the height distribution of the protrusions are determined as later described in detail. If the parameters of surface topography ^'(β/σ) of both surfaces of the film are smaller than the above range, good dubbing is not obtained. On the other hand, if they are larger than the above range, scratch resistance decreases.

Preferably, at least one surface of the film of the present invention has a relationship between the coefficient of friction y and the parameter of

surface topography _• which relationship satisfies the following equations ( 1 ) and (2), preferably (4) and (5), and more preferably (6) and (7):

y ≥ 0.171 v^(β/σ) + 0.16 (4)

y ≤ 0.171 /(P ⁺ 0.24 (5)

y ≥ 0.171 /(β/σ) + 0.17 (6)

y ≤ 0.171 /σ) + 0.23 (7)

If the coefficient of friction on both film surfaces is smaller than the above-described range defined in equation (1 ) good dubbing is not obtained.

On the other hand, if it is larger than the range defined in equation (2) on both surfaces, scratch resistance is reduced.

Preferably, on the surface satisfying the equations (1 ) and (2), the r-verage distance between adjacent protrusions is not more than 20 μm, preferably not more than 17 μm, and more preferably not more than 15 μm.

If the average distance is larger than 20 μm, the scratch resistance is redu¬ ced . Although the average distance between adjacent protrusions does not have a lower limit, about 1 μm may be the practical lower limit because it is difficult to form surfaces with an inter-protrusion distance of less than 1 μm. In view of the scratch resistance, the average distance between adja¬ cent protrusions with a height of less than 0.08 μm is preferably not more than 10 μm, more preferably not more than 8 μm, and the average distance between adjacent protrusion's with a height of 0.08-0.5 μm is preferably 15- 1 50 μm, more preferably 20-100 μm. It should be noted that the number of

protrusions with a height higher than 0.5 μm is preferably less than 5% , more preferably less than 2% of the total number of the protrusions.

Further, in view of the scratch resistance, the difference between the average height of protrusions with a diameter of less than 1 μm and the average height of protrusions with a diameter of 1 -8 μm is preferably 0.02- 0.42 μm, more preferably 0.05-0.30 μm. The number of protrusions with a diameter of more than 8 μm is preferably less than 5%, more preferably less than 2% based on the total number of the protrusions. In order to have good scratch resistance and improved dubbing, at least one surface of the film of the present invention preferably has an R ₂ (five point average surface roughness) of 60-190 nm, more preferably 70- 160 nm, still more preferably 80-1 50 nm.

In order to further improve scratch resistance and dubbing, at least one surface of the film of the present invention preferably has an R _p /R _a ratio

(average depth/center line average surface roughness, both R _p and R _a ex¬ pressed in nm) of 4-25, more preferably 6-20, still more preferably 9-1 5.

For further improvement of scratch resistance and dubbing, at least one surface of the film of the present invention preferably has an effective space volume ψ of the protrusions of 1 x 10 ³ -5 x 10 ⁵ , more preferably 5 x

10 ³ -5 x 10 ⁴ .

For even further improvement of scratch resistance and dubbing, at least one surface of the film of the present invention preferably has a ratio (H/d) of the average height (H) in nm of the protrusions to the average diame- ter (d) of the inert particles of 0.1 -0.5, more preferably 0.1 -0.4.

For still further improvement of scratch resistance and dubbing, at least one surface of the film of the present invention preferably has an average height of the protrusions of 40-130 nm, more preferably 50-120 nm, still more preferably 50-100 nm. Although the use of the film of the present invention is not restricted, the PENBB film of the present invention is useful as a substrate for magnetic recording media, where the use allows reduced scratching of the surface in

the processing steps to be achieved, especially as a substrate for video tapes which are now increasingly being dubbed, as pre-recorded video tapes beco¬ me more widespread. Further, when a film of the present invention in which only one surface has the above-mentioned parameters is used, the surface with the described parameters is preferably used as the running surface (i.e., in magnetic recording media, the side opposite to the magnetic layer, and in other uses, the surface on which a treatment such as printing and application of a layer is not performed).

The process of manufacturing the film of the present invention will now be described. It should be noted, however, that the manufacturing process is not restricted to the process hereinafter described.

The inert particles may be added before, during, or after the polymeri¬ zation to form the PENBB. To obtain the parameter of surface topography

•/(β/σ) and the relationship between the coefficient of friction y and the parameter of surface topography defined in the present invention, it is prefer¬ red that the inert particles be added to the diol component of the PENBB, such as ethylene glycol in the form of a slurry. In this case, it is preferred that the diol component slurry containing the inert particles be filtered through a filter with an absolute filtration precision of 0.5-4.5 μm, more preferably 1 .5-3 μm. In order to control the inert particle content so as to obtain the desired values for the parameter of surface topography ^(β/σ) the relationship between the coefficient of friction and the parameter of surface topography, and the average distance between the adjacent protru¬ sions defined in the present invention, it is preferred to prepare PENBB-based master pellets with a high content of the particles, preferably with a particle content of 1 -5% by weight and then to dilute the master pellets during the formation of the film. In cases where the inert particles are essentially spheri¬ cal silica particles originating from colloidal silica, to obtain the parameter of surface topography y/fβ/σ) ^" • the relationship between the coefficient of friction and the parameter of surface topography, and the average distance

between the adjacent protrusions defined in the present invention, it is prefer¬ red to heat the slurry of the diol component such as ethylene glycol at 140- 200°C, more preferably 180-200°C for 30 minutes to 5 hours, more prefera¬ bly 1 -3 hours. Further, in this case, to obtain the parameter of surface topography γ/(β/σ) _> the relationship between the coefficient of friction and the parameter of surface topography, and the average distance between the adjacent protrusions defined in the present invention, it is preferred to main¬ tain the sodium content of the slurry at not more than 0.5% by weight, more preferably not more than 0.2% by weight and to maintain the pH of the slurry at 7-10. In cases where the inert particles are particles other than silica, such as calcium carbonate and titanium dioxide, to obtain the parame¬ ter of surface topography /ζfϊjσ) . the relationship between the coefficient of friction and the parameter of surface topography, and the average distance between the adjacent protrusions defined in the present invention, it is prefer- red to add to the slurry ammonium phosphate in the amount of 0.5-2.0% by weight with respect to the weight of the particles.

In cases where non-external particles are utilized, the non-external particles may be generated by adding as bonding agent at least one com¬ pound soluble in glycol selected from the group consisting of calcium com- pounds, magnesium compounds, manganese compounds and lithium com¬ pounds, preferably together with an acid and/or an ester compound of phosphorus during the polycondensation of PENBB after direct esterification of the dicarboxylic acids and the diol, or during the polycondensation after ester interchange reaction between the alkyl esters of the dicarboxylic acid components and the diol. The compounds of calcium, magnesium, mangane¬ se and lithium which may preferably be herein used include inorganic acid salts such as halogenides, nitrates and sulfates; organic acid salts such as acetates, oxalates and benzoates; hydrides; and oxides, which are soluble in glycol, as well as mixtures thereof. Preferred examples of the phosphorus compound which may be herein used include phosphates, phosphorous acid and phosphonic acid, as well as esters and partial esters thereof.

After sufficiently drying the pellets containing the prescribed amount of inert particle^ (more particularly, a mixture of the master pellets containing large amount of particles and the PENBB pellets which do not substantially contain particles), the pellets are fed to an extruder and are extruded from a slot die at a temperature between the melting temperature (T _m ) and 60°C above T _m in the form of a sheet. The sheet is cooled and solidified on a casting roll to obtain a non-oriented, amorphous film. To obtain the parame¬ ter of surface topography ^(β/σ) , the relationship between the coefficient of friction and the parameter of surface topography, as well as the average distance between the adjacent protrusions defined in the present invention, the ratio of the width of the die slit to the thickness of the non-oriented film is preferably 5-30, more preferably 8-20.

Then the non-oriented film is biaxially oriented by biaxial stretching.

Either two or more sequential biaxial stretching steps or one simultaneous biaxial stretching may be employed. It should be noted, however, that to obtain the parameter of surface topography , the relationship between the coefficient of friction and the parameter of surface topography, and the average distance between the adjacent protrusions defined in the present invention, it is preferred to employ sequential biaxial stretching wherein the stretching in the longitudinal direction is carried out first and then the stret¬ ching in the transverse direction is carried out second. Although not always necessary, the stretching operations can be carried out such that the stret¬ ching in the longitudinal direction is conducted in not less than three steps, preferably not less than 4 steps to obtain a stretching ratio in the longitudinal direction of 2.5-5.5 times. Further, to obtain the parameter of surface topo¬ graphy _> the relationship between the coefficient of friction and the parameter of surface topography, and the average distance between the adjacent protrusions defined in the present invention, it is preferred to con¬ duct the first stretching in the longitudinal direction at a temperature between the glass transition temperature of the PENBB (T _g ) minus 20°C and the glass

transition temperature of the PENBB plus 20°C, and to conduct the sub¬ sequent stretching at a temperature higher than the first stretching. The stretching rate may preferably be 5, 000-50, 000%/min. As a method of stretching in the lateral direction, a tenter usually employed. The stretching ratio in the lateral direction may preferably be 2.5-5.0 times. The stretching rate in the lateral direction may preferably be 1 ,000-20, 000%/min. Biaxial drawing is performed such that the birefringeance is < 0.2, preferably < 0.1 to ensure adequately isotropic properties. Birefringeance as mentioned herein is the absolute value of the difference between the maximum and minimum refractive indices in the plane of the film, as measured on common instru¬ ments such as Abbe refractometer, optical bench or compensators.

Then the film is heat set. The heat setting may preferably be conduc¬ ted at a temperature of between the cold crystallization temperature (T _cc ) and the melt temperature (T _m ) of the PENBB copolyester composition for 0.5-60 seconds. It is useful for obtaining the average distance between the adjacent protrusions defined in the present invention to stretch the film in the lateral direction to 1 .05-1 .3 times, especially 1 .05-1 .2 times the original length during the heat setting.

Methods of Determining Physical Properties and Methods of Evaluating

Effectsl

The methods of determining the physical properties relating to the present invention and methods of evaluating the effects are as follows:

( 1 ) Average Diameter of the Particles The PENBB is removed from the film sample by subjecting the film to plasma treatment. The residue is dispersed in ethanol and the volume aver¬ age diameter of the particles are determined by centrifugal sedimentation.

(2) Content of Particles

A solvent, such as a 50/50 mixture of hexafluoroisopropanol and pentafluorophenol, which dissolves the PENBB but does not dissolve the particles is added to the PENBB film and the mixture is heated to dissolve the copolyester. The resultant mixture is centrifuged and the obtained particles

are washed with a volatile solvent, such as acetone, and dried in vacuum. The particles are subjected to differential scanning analysis using a DSC (differential scanning calorimeter). If a melting peak corresponding to the PENBB copolyester is observed, additional solvent is added to the particles and the resultant mixture is heated, followed by centrifugation. When the melting peak is no longer observed, the remaining particles are the separated particles. Repeating the centrifugation twice is usually sufficient. The content of the particles is defined as the ratio (% by weight) of the total weight of the thus separated particles to the total weight of the film.

(3) Glass transition temperature T _g and cold crystallization temperature T„.

The glass transition temperature T _g and the cold crystallization tempe¬ rature T _cc are determined using a DSC as follows:

Ten milligrams of a sample in a closed pan are placed in the DSC and then melted at a temperature 40 °C above the melting point of the PENBB copolyester for 5 minutes, followed by rapid quenching in liquid nitrogen. The rapidly quenched sample is heated at a rate of 10 ^° C/minute and the glass transition temperature T _g is determined. The heating is continued to determine the peak of the generation of crystallization heat from the glassy state. The thus determined peak is defined as the cold crystallization tempe¬ rature T _oc .

(4) Surface Roughnesses R _a , R _z and R _p

The surface roughnesses R _a , R _z and R _p are measured using a commerci- ally available surface roughness meter employing the following conditions: radius of stylus tip: 0.5 μm stylus load : 5 mg length of measured portion: 1 mm cutoff value: 0.08 mm R _a is determined according to DIN 4768.

R_ and R _p are calculated in accordance with the following equations:

R _z = (the third highest peak height) - (the third deepest valley depth) R _p = (the highest peak height) - (the center line value)

(5) Effective Space Volume φ of Protrusions Surface roughness curves are obtained in the same manner as in (4).

Peak counting levels parallel to each other and the center line, with an inter¬ val distance of 5 nm are overlaid on the surface roughness curve. For each peak counting level, the number of peaks reaching or crossing it in the mea¬ surement length is determined. For each level n, the number of peaks cros- sing it is PC(n). Level m is the first level where PC(m) = 0. The effective space volume ψ is defined as: m-1

(6) Average Height, Standard Deviation of Height Distribution, Flatness and Diameter of Surface Protrusions Using a scanning electron microscope the height of the protrusions is measured by scanning the surface of the film, while setting the flat region of the surface of the film to zero. The height of protrusions thus measured is transferred to an image processing system as gray value with 256 gradua¬ tions and the image of the surface protrusions are reconstructed on the image processing system. The circle radius is calculated from the area of each protrusion with not less than 10 graduations, which area is obtained by converting the graduations into two values. The thus calculated circle radius is defined as the mean diameter of protrusion, d (nm). The highest value in each two value-converted protrusion portions is defined as the height of the protrusion, and this height is determined for each protrusion. This measure¬ ment is repeated 500 times on different areas. For protrusions with a height of not less than 200 nm, the mean height H (nm) and the standard deviation σ (nm) of the height distribution are calculated by least square regression assuming the height distribution of the protrusions to be a Gaussian distribu-

tion (a Gaussian distribution of which center is the point of zero height). The flatness of protrusions β is defined as the ratio d/H of the mean diameter of protrusions (d) to the mean height of protrusions (H).

(7) Coefficient of Friction y

Using a commercially available tape running tester, a sample tape is run at 20 ° C, 60%RH, and the coefficient of friction y is calculated in accordance with the following equation: _y = π/iriiππyT,) wherein T., is the tension on the entering side and T ₂ is the tension on the exiting side. The diameter of the guide is 6 mm, the material of the guide has a surface roughness of 0.2S, the winding angle is π rad, and the running speed is 3.3 cm/min.

(8) Inherent Viscosity (IV) (expressed in dl/g)

The inherent viscosity is calculated from the solution viscosity in penta- fluorophenol/hexafluroisopropanol measured at 25 ^* C according to the follo¬ wing equation:

In (η

IV =

wherein η _ιel = (solution viscosity/solvent viscosity), C is the weight of dis¬ solved polymer per dl ( 100 ml) of solvent (g/dl, usually 0.1 or 0.2 g/dl). The solution viscosity and the solvent viscosity are measured using an Ubbelohde viscometer.

(9) Dubbing

A magnetic pigment coating with the following composition is applied to the film with a gravure roll coater, and the coated magnetic pigment is magnetically oriented and the coating dried. The magnetic film is subjected

to a calender treatment with a small test calender apparatus (steel roll/Nylon roll, five steps), at a temperature of 70 ^° C with a linear pressure of 200 kg/cm, followed by curing at 70 " C for 48 hours. The coated film is slit to 1 /2 inch (1 .25 cm) to prepare a pancake. From this pancake a tape with 250 m length is taken and is set in a video tape recorder (VTR) cassette to prepa¬ re a VTR tape.

(Composition of magnetic pigment coating in parts by weight) Co-containing F ₂ O ₃ (BET value 50 m ² /g) : 100

vinyl chloride/vinyl acetate copolymer 10

polyurethane elastomer : 10

polyisocyanate 5 lecithin : 1

methylethyl ketone 75

methylisobutyl ketone : 75 toluene 75

carbon black 2

lauric acid : 1 .5

On the thus prepared tape, 100% chroma signals from a commercially available television testing wave generator are recorded using a home VTR. Chroma S/N of the regenerated signals are measured using a commercially available color video noise-measuring apparatus, this value being designated as A. Also, a pancake of a tape on which the same signals as mentioned above are recorded is dubbed to another pancake of a tape (containing no previously recorded signals) of the same kind using a magnetic field tran¬ scription type high speed print system and the chroma S/N of the dubbed tape is measured as mentioned above, the value being designated as B. If the decrease in the chroma S/N (A-B) due to the dubbing is less than 4.0 dB, the tape is considered to have good dubbing, and if it is greater than or equal to 4.0 dB, the tape is considered to have bad dubbing.

( 10) Scratch Resistance

The above-described tape is repeatedly run on a commercially available high speed tape running tester (running speed of 800 m/min, 10 passes). After the test, the film is observed with a microscope to check if scratches are formed. If almost no scratches are formed, the scratch resistance of the film is considered high, and if at least 3 scratches per the width of the tape are formed, the scratch resistance of the film is considered low.

The following example, while not limiting its scope, serves to illustrate the present invention.

Example 1

An ethylene glycol slurry containing colloidal silica with an average diameter of 0.3 μm is prepared. The sodium content in the slurry is adjusted to 0.1 % by weight with respect to the weight of the silica particles. After heating the ethylene glycol slurry at 190 ^° C for 1 .5 hours, it is subjected to ester interchange reaction with dimethyl naphthalate/dimethyl 4,4'-bibenzoate (50 : 50 mole ratio). The reaction product is filtered through a filter with an absolute filtration precision of 3 μm. The filtered slurry is subjected to polycondensation reaction to prepare master pellets of PENBB. The polycon¬ densation reaction time is adjusted so as to obtain an IV of the polymer of 0.9 dl/g. In parallel, by a conventional process, pellets of PENBB with an IV of 1 .0 dl/g are prepared and then mixed with the above-described master pellets in such ratio as to attain the silica particle content of 0.5% by weight. The mixed pellets are then dried under reduced pressure (3 Torr) at 180°C for 3 hours. The resulting pellets are fed to an extruder and are melted at 310 ^° C. The melted polymer is extruded through a slot die and the extruded sheet is electrostatically pinned on a casting drum with a surface temperature of 20 ^° C to cool and solidify the sheet to obtain a non-oriented film. The ratio of the width of the clearance of the die slit to the thickness of the non- oriented film is 10. The non-oriented film is stretched in the longitudinal direction to 4.5 times the original length . This stretching is conducted in four

steps utilizing the difference of the circumferential velocity of two pairs of rolls. The stretching temperatures are in the range of 120-135 ^° C. The thus obtained uniaxially oriented film is then stretched in the lateral direction to 4.0 times the original length using a tenter at 130 ^° C with a stretching rate of 2, 000%/min. The resulting film is heat set at 250 ^° C for 5 seconds while stretching the film in the lateral direction to 1 .07 times the original length to obtain a biaxially oriented film with a thickness of 10 μm. The scratch resi¬ stance of the film is excellent and the film has excellent dubbing. Both surfaces of this film have the same parameters. The improved mechanical and dimensional stability over conventionally prepared PET films may be taken from Table 1 .

TABLE 1