WO1984004159A1

WO1984004159A1 - Method for the measurement of the diameter of a glass fiber using the resonant backscatter technique and measuring apparatus for carrying out the method

Info

Publication number: WO1984004159A1
Application number: PCT/HU1984/000025
Authority: WO
Inventors: Andras Podmaniczky; Lenk Attila Markus; Janos Giber; Tibor Behringer; Miklos Barabas
Original assignee: Budapesti Mueszaki Egyetem
Priority date: 1983-04-18
Filing date: 1984-04-18
Publication date: 1984-10-25
Also published as: HU187230B; EP0177496A1

Abstract

Method for the measurement of the diameter of a glass fibre using the resonant backscatter technique during the application of which the fiber to be measured is illuminated by a tunable wavelength laser and the extrema of the detected intensity of the rays backscattered from the fiber into a small solid angle are registered while tuning the laser wavelength. The paths of the illuminating and backscattered rays are separated. The illuminating laser beam of the measuring arrangement suitable for carrying out the method subtends a small angle with the normal of the fiber axis and the intensity detector that receives the backscattered rays is placed so that the principal direction of the rays accepted by it makes an angle of the same size but opposite sign with the said normal. Before the detector there is a slit of a size corresponding to the said small solid angle.

Description

METHOD FOR THE MEASUREMENT OF THE DIAMETER OF A GLASS FIBER USING THE RESONANT BACKSCATTER TECHNIQUE AND MEASURING APPARATUS FOR CARRYING OUT THE METHOD

The object of the invention is a method for the measurement of the diameter of a glass fiber using the resonant backscatter technique and a measuring apparatus for carrying out the method.

The present invention can be used primarily for high precision diameter-measurement of clad or unclad glass fiber of diameter of 10-150 μm.

Several methods of diameter measurements of thin glass fibers are known, L. S. Watkins describes in his paper entitled "Control of Fiber Manufacturing Processes" (Proc. of IEEE. Vol. 70, No. 6. June 1982 pp. 626-634) a method of diameter measurement where the fiber is illuminated by laser fiber axis and the angle distribution of the forward scattered rays is detected by a large number of photodiodes and analyzed by a computer.

This method provides a means for diameter measurement of an accuracy of 0,25 μm, but it can be used only with fibers of small diameter because the number of interference fringes increases rapidly with increasing diameter which circumstance makes the evaluation difficult. To attain good resolution a large number (approx. 1000) of photodiodes is necessary and this together with the use of a microprocessor renders the method expensive.

The paper cited describes also another method for diameter determination, in which the fiber to be measured is scanned perpendicularly to its axis by a focussed or unfocussed laser beam, which is detected behind the fiber by a detector, and the negative impulse caused by the shadow of the fiber is analyzed. For a given scanning velocity the duration of this impulse is characteristic of the fiber diameter. The accuracy of this measurement is about ± 0,5 μm and this precision deteriorates with decreasing diameter.

In the production and control of glass fibers a higher accuracy of measurement is required than that described above. An appropriate measuring method is presented in the paper of A. ASHKIN, J.M. DZIEDIC and R.H. STOLEN entitled "Outer diameter measurement of low birefringence optical fibers by a new resonant backscatter technique" (Applied Optics, Vol. 20. No.13. July 1981, pp. 2299-2303). This method called the near-field resonant backscatter technique is based on the detection of Fabry-Perot and surface-wave resonances of light backscattered from a fiber illuminated by a wavelength-tuned laser beam. The examination of Fabry-Perot resonances yielded an accuracy of 0,01 μm in the relative measurement of the diameter of communication fibers.

In the near-field resonant backscatter technique the unclad fiber is illuminated by laser light through a beam splitter in a direction perpendicular to its axis, backscattered rays are directed by the beam splitter to an imaging system (microscope) and the magnified image of the fiber is produced. Rays generating Fabry-Perot resonances and rays emerging from the surface wave proceeding as backscattered rays in the axial direction appear separately in the image plane and the rays to be analyzed are selected by a slit. The minima and maxima resulting from the interference of selected backscattered rays of the same type occurring when tuning the laser wavelength are detected and the diameter is computed from the values of these extrema.

Although the accuracy of the method is very large the following difficulties arise when applying it (especially in the case of small diameters). Due to the imaging that requires very large magnification even the smallest - and practically unavoidable - movement of the fiber results in a shift of the backscattered rays relative to the slit and thus the measurement cannot be accomplished or false minima are observed. This difficulty becomes more severe with decreasing diameter because very thin fibers are more prone to vibration and in the case of a smaller diameter the magnification must be increased.

Besides the sensitivity to movement the presence of the beam splitter is disadvantageous as well, because the fourfold attenuation caused by the beam splitter further decreases the intensity of the backscattered light which is already very much less intensive than the incident light.

In the case of singly or multiply clad fibers the applicability of the near-field resonant backscatter technique is diminished by a large extent because of the optical effects caused by the claddings. Image transmitting fibers a re usually doubly clad and their diameter is in the 10-20 iim range. The diameter measurement of such fibers with this method can be carried out only with great difficulty.

The task of the present invention is to provide a method for the measurement of fiber diameter which yields equally sufficient accuracy in the case of thin and thick and clad and unclad fibers respectively and by the use of which method there occur neither the difficulties arising from near-field imaging nor the other problems outlined above.

Therefore the object of the invention is a method for the measurement of the diameter of a glass fiber using the resonant backscatter technique, where the glass fiber is illuminated by laser light, backscattered rays lying in a small solid angle range around a specified direction are detected, the laser wavelength is tuned continuously within a given band, and the diameter is determined from the wavelength values corresponding to the extrema of the detected intensity, and according to the invention the paths of the illuminating beam and of the detected backscattered beam are separated by tilting the illuminating laser beam relative to the direction perpendicular to the fiber axis and/or by separating the said specified directions and the direction of the illuminating beam is the plane perpendicular to the fiber axis, and the total for field intensity of the backscattered rays is detected without the use of an imaging system.

The method according to the invention eliminates the necessity of the beam-splitter by spatially separating the illuminating beam from the detected light and this in itself causes an approximately fourfold increase in intensity. Detection of the resultant far field intensity eliminates the most inconvenient feature of the measurement i.e. the imaging system, and although the detect.ion of the resultant ray pattern causes difficulties as well especially in the case of clad fibers, these difficulties can be reduced to a minimum tuning evaluation. From the point of view of the reduction of the effect of backscattered rays that influence Fabry-Perot resonances it is advantageous if the direction of polarization of the illuminating light makes an angle with the fiber axis and this angle is advantageously a right angle. Separation of the two said ray paths can be advantageously attained if the plane determined by the illuminating laser beam and the principal direction of the backscattered rays is parallel to the fiber axis and the illuminating laser beaβ is tilted with respect to the normal of incidence. This tilt is of a very small angle the value of which is determined only by the smallest angle needed to separate the paths of illuminating and detected rays, and this angle is between apprσx. 0,5° and 2°. The accuracy of the evaluation can be increased if several wavelengths corresponding to the different minima and maxina obtained in the tuning range are taken into account when determining the unknown diameter.

From the point of view of the elimination of measurement errors arising from undesired interferences it is advantageous if from the wavelength data of all the minima and/or maxima lying in the given tuning range an average order of interference is determined corresponding to a selected extremum, and the fiber diameter is calculated from the measured wavelength of the selected extremum and this average order of interference.

The reliability of the result obtained this way may be checked if the wavelength of the selected extremum, the average order of interference and the resonance condition are used to compute the wavelengths of all the extreβa taken into account in the calculation of the average order of interference, and the value of the measured diameter is accepted as a correct result only if the average of the differences between the calculated and measured values of the wavelengths corresponding to the extrema is smaller than a prescribed threshold value, else the measurement is to be repeated.

The introduction of the average order of interference derived from several extrema drastically reduces the effect of interferences disturbing exact determination of the wavelength of the extrema and the influence of other undesired effects, and consequently the accuracy of the far-field resonant backscatter technique according to the invention is satisfactory (better than 0,1 μm) also in the case of small fiber diameters. With the invention we created also a measuring arrangement to carry out the method, which arrangement comprises a tunable wavelength laser light source illuminating the glass fiber to be measured and an intensity detector placed in the path of the backscatted rays lying in a given narrow solid angle range around a given direction, and according to the invention the laser beam of the said laser light source illuminating the filter subtends a small angle with the normal to the fiber axis and the intensity detector is placed in the path of the backscattered laser rays that lie in the plane of the illuminating beam and subtend an angle of the some size but of opposite sign with the said fiber normal, and in the path of the backscattered rays still before the detector there is a slit the size of which corresponds to the given small solid angle range. Between the laser light source and the glass fiber there is preferably a beam contractor and in the path of the backscattered rays still before the slit there is a beam deviating element separated from the path of the illuminating beam, which beam deviating element is advantageously a mirror. The laser light source consists preferably of a pumping laser, for example an argon-ion laser and a dye laser connected to it.

In the following the invention is described in detail in connection with examples of embodiments with reference to the figures.

Fig, 1 is a magnified image of the cross-section of the glass fiber and it shows the illuminating laser beam and the ray components backscattered in its direction, Fig. 2 is a sketch in a plane perpendicular to the plane of Fig. 1 and it illustrates the reflection of the sligh t ly t ilt ed illumina t ing rays wher e the size of the angles is exaggerated for the sake of demonstration, Fig. 3 shows the far-field intensity distribution arising in the case of illumination sketched in Fig. 1 as a function of the scattering angle for perpendicular and parallel polarization, Fig. 4 is a sketch demonstrating the generation of interference fringes arising at different scattering angles for different fiber tilt angles. Figs. 5 a, b and c are photographs of the interference patterns obtained with different polarizations and fiber diameters by scanning the fiber tilt angle according to Fig. 4 ,

Fig. 6 is the sketch of an embodiment of the measuring arrangement according to the invention and

Fig. 7 shows the diagram of the far-field intensity as a function of the wavelength as registered by the measuring arrangement according to Fig. 6 .

Fig. 1 illustrates the rays scattered by an angle of 180° from the doubly-clad fiber 10 illuminated by laser light. The illuminating laser rays L , perpendicular to the axis of glass fiber 10 are denoted by parallel arrows in Fig. 1 .

Rays I of the illuminating beam are incident at a sight angle on the central portion of the surface of glass fiber 10 and one part of these rays is reflected from the outside, front surface and the other part from the opposite inside back surface of fiber 10 thus producing rays II. Rays III of the illuminating beam are also reflected at an angle of 180º after having suffered multiple refraction at the boundaries of the claddings and having been reflected at the inner back surface.

Finally rays IV incident tangentially on glass fiber 10 propagate as surface optical waves along the fiber surface and leave fiber 10 also on its diametrically opposite sides.

Among the backscattered rays I-IV rays I and II are of the largest intensity, but rays III and IV cannot be ignored either. Interference between rays I and II gives rise to Fabry-Perot resonances because the central portion of the glass fiber may be regarded as a parallel plate. Fig. 2 is a sketch obtained by projection onto a plane perpendicular to the plane of Fig. 1, and here the illuminating laser beaα L makes an angle α with the normal of incidence. The outer cladding of the glass fiber 10 has an index of refraction n₁ and its thickness is t₁ , tne index of refraction of the inner cladding is n₂ and its thickness t₂ and the central core is of diameter 2r and of index of refraction n₃. The outer diameter of glass fiber 10 is d. The laser beam L incident at an angle α is refracted at the claddings and its direction of prepagation in the central core subtends an angle β with the normal of incidence. This angle β is approximately valid also for the claddings because these layers are thin. Examining the resultant intensity of the backscattered rays I-IV as a function of the scattering angle θ shown in Fig. 1 we obtain the function shown in Fig. 3. The full line demonstrates the intensity in the case when the polarization of the rays L is parallel to the axis of fibre 10 and the broken curve shows the intensity for perpendicular polarization. Ripples on the curve are due to the interference of rays III and IV with rays I and II and it is thus seen that the effect of rays III and IV is much weaker if the direction of polarization is perpendicular to the fiber axis. The following consideration concern the backscattering angle θ = 0 and its small neighbourhood.

Referring again to Fig. 2 we seek the condition under which the resultant intensity of rays I and II is a minimum. For this purpuse we introduce the concept of the average index of refraction defined by

In the case of θ = 0º rays I and II yield minimum intensity if the condition

is fulfilled. Here λ denotes the wavelength of the illuminating laser beaa L and m is an odd integer giving the order of interference. If we examine the resultant intensity of the backscattered rays as a function of both angles α and θ we obtain the peculiar interference pattern depicted in Fig. 4. Fig. 5a illustrates such an interference pattern obtained by scanning the angle of incidence α in the case of a doubly clad glass fiber of diaβeter

12,2 μm where the incident laser rays L are polarized perpendicularly to the fiber axis. Fig. 5b is a picture similar to Fig. 5a but the plane of polarization is parallel to the fiber axis. Fig. 5c shows the interference pattern obtained during the measurement of a communication fiber having a diameter of 120 μm. From the inspection of the values of the angle α corresponding to successive minima along to line θ = 0ºthe diaaeter d of the glass fiber 10 may be determined:

Equation (3) may be used for the measurement of the diameter d but because of the uncertainty inherent in the determination of the wavelength corresponding to a minimum (which is still enhanced by the effect of rays III and IV. and due to the error of the angle measurement this procedure yields the value of the unknown fiber diameter with an accuracy of only about 2-3 μm.

According to the invention the measurement of the unknown diameter d is performed in the neighbourhood of θ = 0° and by a very small value of the angle α or by the value 0º of the angle α and a very small value of the angle θ on the basis of the detection of the extrema of the backscattered radiation during the continuous tuning of the wavelength. Now we refer to Fig. 6 which shows the skecth of a measuring arrangement suitable to carry out the method according to the invention.

The arrangement comprises an argon-ion laser 11 to which there is connected a dye laser 12 whose wavelength is continuously tunable. The intensity as well as the wavelength of the output beam of dye laser 12 are known. In Fig. 6 a first detector 13 is used to measure the intensity and a spectrometer 14 together with a third detector 15 is used to measure the wavelength. The appropriate deviations of the laser beam are accomplished by beam-splitters 16 and 17 as well as a mirror 18. The output beaa of dye laser 12 propagates toward the measured fibre 10 through a beaa contractor 19 which performs an approximately five-fold contraction of the beam diameter. This contracted beaβ is the illuminating beaβ denoted by L. In the embodiment sketched in Fig. 6 the axis of the glass fiber 10 is not exactly perpendicular to the beam L, but it is tilted by an angle of about 1 (α≈1º) with respect to the perpendicular position. Thus the laser beam L subtends an angle with the normal of incidence and we measure the far-field intensity of the backscattered rays BS reflected at an angle 2α with respect to the direction of the laser beam L. In the path of the backscattered rays there is a mirror 20 and in the path of the rays proceeding therefrom there is a slit 21 that selects the backscattered rays lying in the prescribed solid angle range. 8ehind the slit 21 there is a second detector 22 which detects the intensity of the backscattered rays BS. Between the slit 21 and the second detector 22 there may be placed a focussing lens 23. Concerning the arrangement sketched in Fig. 6 it. is essential that the ray path between the glass fiber 10 and the second detector 22 is separated from path of the illuainating beaa L leading to the glass fiber 10, although this separation is realized by a very small value of the angle α. Because of this separation (of the ray paths) we need not place a beam-splitter or any other device in the path of the backscattered rays BS that would further reduce the already very sβall intensity of the rays BS, It is also essential that the second detector 22 is placed into the far-field of the backscattered rays BS and that it receives the backscattered rays BS passing through the slit 21 without the use of an imaging device. In the case of small fiber diameters the absence of imaging means a very essential structural simplification which manifests itself in earlier application of the instrument because thus the second detector 22 is less sensitive to possible small displacements of the fiber 10 which are always to be expected in practice.

In the arrangement sketched in Fig. 6 the size of the angle a, is approximately between 0,5 and 2º . The paths of the laser beam L and of the backscattered rays BS are separated similarly if the glass fiber 10 is not tilted (α=0°) but the mirror 20, the slit 21 and the second detector 22 are displaced from the plane of the drawing and situated along a line making an angle of for example 1º with the above mentioned plane. Figure 5 shows clearly that the intensity is virtually unchanged in such a small angle range.

In the following we refer to Fig. 6 but it is readily seen that separation of the rays in a different plane leads to similar results. To determine the diameter d of the glass fiber 10 according to the invention we change continuously the wavelength of the laser beam L by tuning the dye laser 12 and with the second detector 22 we simultaneously record the variation of the intensity of the backscattered rays BS. Fig. 7 shows this recorded intensity as a function of the wavelength of the dye laser 12. In the case of Fig. 10 the diameter of the glass fiber 10 is d = 14,0 μm , the thicknesses t₁ and t₂ of the cladding layers are approximately equal to 0,7 μm and the indices of refraction are n₁ = 1,507 and n₃ = 1,62. From Fig. 7 it is seen that the value of the intensity at the minimum is greater than zero and that this value depends also on the wavelength. This effect is due partially to the difference between the intensities of rays I and II and partially to the influence of rays III and IV (see Fig. 1) which is also present in the output signal of the second detector 22. Fig. 7 illustrates the case in which the plane of polarization of the laser beam L is parallel to the axis of the glass fiber 10. In the case of perpendicular polarization the minima are sharper. The unknown diameter d may be determined from the differences of wavelength between adjacent minima or maxima on the basis of equation (2). Let us denote the wavelengths corresponding to the minima by λ _i and those corresponding to the maxima by . Then λ_l

denotes the first ainimum and the first maximum.

The orders of interference corresponding to two successive minima decrease by the value of 2 with increasing wavelength. We have

mi = m1 - 2(i-1) (4)

M_j = M_l - 2( j -1) ( 5) where m_i is the order of interference of the i-th minimum and M_j is the order of interference of the j-th maximum. If is smaller than λ

_l then m_l = M₁ + l .

Ignoring the cosine factor in equation (2) we have

In equation (6) the order of interference is unknown and therefore the diameter d canηot be determined. Using equation (6) together with the condition of resonance and with equations (4) and (5) the order of interference m₁ corresponding to the first

minimum can be determined from the value of λ_l and any of the values λ _i (≠ λ_l):

The knowledge of the value λ_l of the wavelength corresponding to the first minimum and of the value

corresponding to the j-th maximum provides further data of the order of interference m₁:

If the location of the first minimum or maximum and that of any other miniajm or maximum are known, the order of interference in ecuation (6) can be computed and then the unknown diameter d can be calculated as well.

From the diagram of Fig. 7 we see that the minima are slightly displaced from their ideal position and these shifts of wavelength may reach several tenth of a nanometer. These shifts are caused probably by a small phase jump suffered by ray II while traversing the claddings at certain wavelength values.

According to the invention the accuracy of the measurement can be substantially increased if the order of interference m_l in equation (6) is determined not only from the first mininurm or maximum and a single other extremum, b-t also from taking into account all the extrema lying in the wavelength range of the tunable dye laser 12, i.e. if equations (7) and (8) are evaluated for all possible values of i and j thus yielding a series of values for m₁.

After these values have been calculated their average is used instead of a, to determine the

diameter from equation (6). As a consequence of averaging the effect of the smaller shifts of wavelength is reduced to a negligible level. Therefore the use of wavelength data corresponding to several minima and maxima in the determination of the order of interference results in an essential increase of the accuracy of the measurement.

In order to check the reliability of the result obtained the following calculation may be carried out. From the average order of interference values

λ _{i c} ,

of the wavelength corresponding to each maximum and minimum are computed by

If each of these calculated wavelength values is substracted from the corresponding measured value then the errors between the calculated and measured values of the order of interference are expressed as wavelength differences. The average of the moduli of these differences may be regarded as an average wavelength measurement error present in the determination of the location of extrema. It may be proved that the value of this average wavelength error is very sensitive to measurement errors of any kind, i.e. it characterizes the coincidence of measured and calculated wavelength data in a suitable way. If in the case of i,j ≈ 5 the average wavelength error is smaller than 0,1 nm than the error of the diameter aeasurement is smaller than 0,09 μm. Therefore tne simple measuring arrangement sketched in Fig. 6 provides a means for the high-precision measurement of glass fibers of even the smallest diameters (10-15 μm) if several minima and maxima are taken into account. Control calculations showed that the raeasureaent error was almost always under the above mentioned threshold, and thus there was no need the repeat the measurement.

Although in the above description we give an example of the calculation of the order of interference corresponding to the first maximum orminimum, it may be readily seen that the computation might have been based on any other arbitrarily chosen i-th or j-th minimum or maximum with the use of slightly modified equations. The reliability and accuracy of the measurement increases with the number of minima and/or maxima taken into account in the calculation of the average order of interference used to determine the unknown fiber diameter.

Claims

1. Method for the measurement of the diameter of a glass fiber using the resonant backscatter technique during the application of which method the glass fiber is illuminated by laser light, rays backscattered from the fiber and lying in a small solid angle range around a given direction are detected, the wavelength of the laser light is changed continuously in a given band and the fiber diameter is determined from the values of the laser wavelength corresponding to the extrema of the detected intensity c h a r a ct e r i z e d by the separation of the path of the illuminating laser beam (L) from the path of the backscattered rays to be detected (BS), which (separation) is realized by tilting the illuminating leser beam (L) relative to the direction perpendicular to tne axis of the glass fiber (10) and/or by the separation of the said given direction from the direction of the illuminating beam (L) in a plane perpendicular to the axis of the fiber (10) and by the detection of the resultant far-field intensity being realized without the use of imaging.

2. The method as claimed in claim 1 c h a r a c t e r i z e d by the fact that the plane of polarization of the illuminating laser beam (L) is adjusted so as to aake an angle with the axis of the glass fiber (10).

3. The metnod as claimed in claim 2 c h a r a c t e r i z e d by that angle is a right angle.

4. The method as claimed in any of claims 1-3 c h a r a c t e r i z e d by the facts that the plane determined by the direction of the illuminating beam (L) and by the said given direction of the backscattered rays (BS) is parallel to the direction of the axis of the glass fiber (10) and that the illuminating laser beam (L) is tiltea relative to the normal of incidence.

5. The method as claimed in any of claims 1-4 c h a r a c t e r i z e d by the fact that wavelength values correspondint to several minima and/or maxima lying in the tuning range of the illuminating laser beam (L) are taken into account to determine the fiber diameter.

6. The method as claimed in any of claims 1-5 c h a r a c t e r i z e d by the fact that an average order of interference corresponding to a specified extremum is deterained by using the wavelength data of all the minima and/or maxima lying in the given tuning range and that the diameter is derived from this average order of interference and from the measured value of the wavelength corresponding to the specified extremum.

7. The method as claimed in claim 6 c h a r a c t e r i z e d by the fact that the wavelength of all the extrema taken into account in the determination of the average order of interference are calculated froa the wavelength corresponding to the specified extremua and from the condition of resonance, the moduli of the differences of the calculated and measured values of the wavelengths corresponding to each of the extreaa are averaged and the result obtained for the fiber disaster is accepted only if the said average wavelength difference lies under a prescribed threshold value, else the measurement is repeated.

8. A measuring arrangement for carrying out the method as claiaed in any of claims 1-7, comprising a tunable wavelength laser light source illuminating the measured glass fiber and a detector placed in the path of the rays backscattered from the fiber into a given small solid angle range around a given direction c h a r a c t e r i z e d by that

- the illuminating laser beam (L) subtends a small angle with the normal of the axis of the glass fibre (10),

- the intensity detector (22) is placed in the path of the backscattered rays (BS) lying in the same plane as the illuminating beam (L) and subtending an angle of same magnitude but different sign with the said normal and

- in the path of the backscattered rays (BS), still before the detector (22) there is a slit the size of which corresponds to the said small solid angle range.

9. The measuring arrangement as claimed in claim 8 c h a r a c t e r i z e d by the facts that between the laser light source and the glass fiber (10) there is a beam contractor (19) and in the path of the backscattered rays (BS) still before the slit (21) and out of the path of the illuminating laser beam (L) there is a light deviating device which is advantageously a mirror (20).

10. The measuring arrangement as claimed in claims 7 or 8 c h a r a c t e r i z e d in that the laser light source comprises a pumping laser (11) and a dye laser (12) combined with it.