1536 OPTICS LETTERS / Vol. 20, No. 14 / July 15, 1995

Nondestructive method for attenuation measurementsin optical hollow waveguides

R. Dahan, J. Dror, A. Inberg, and N. Croitoru

Department of Electrical Engineering–Physical Electronics, Faculty of Engineering,Tel Aviv University, Tel Aviv 69978, Israel

Received March 14, 1995

A new method for characterizing hollow waveguides has been developed in which the laser radiation is coupledinto the waveguide hollow bore through an optical f iber. By moving the distal end of the f iber along thewaveguide we achieved scanning of the incident radiation in the waveguide at various points on the internalwalls. This method can be employed for measuring attenuation without cutback or for detecting point defectson the waveguide’s guiding layers.

Hollow waveguides are used for transmitting infraredradiation from a source to a target in a curved trajec-tory.1 – 3 The fabrication process includes deposition ofmetal and dielectric layers on the inner surface of ahollow tube (as a specific case we used AgyAgI lay-ers4,5). The transmission characterization of the hol-low waveguide depends on the deposited layer qualityand the roughness of the tube’s internal surface.

The attenuation A and the attenuation per length aof a waveguide are defined by the following relations6:

A ­ 210 logpz2

pz1dB , (1)

a ­A

z2 2 z1dBym , (2)

where z1 and z2 are points along the waveguide sz2 .

z1d and pz1 and pz2 are the powers at z1 and z2,respectively.

There appear to be two sources contributing to theattenuation: (1) losses during the process of propa-gation along the waveguide by incident radiation onthe internal wall and (2) coupling losses. Since onlythe first contribution is an inherent characteristic ofthe waveguide and is connected to the process of prepa-ration of the waveguide, we studied mainly this typeof loss.

The cutback method is usually used to evaluatethe attenuation of the waveguide without the cou-pling losses’ being taken into account. In this methodthe powers at the waveguide’s outlet and inlet aremeasured and introduced into Eq. (1). The distancez2 2 z1 is the waveguide length. One finds the attenu-ation by removing successively several sections szidnear the outlet by ablation and measuring the corre-sponding output power from the remaining waveguide.By plotting the transmission as a function of zi 2 z1one obtains the attenuation and coupling losses fromthe slope and intercept for zi 2 z1 ­ 0, respectively.By use of this procedure the losses at coupling are ob-tained separately since their contribution to the attenu-ation is kept constant. This measurement, however, isdestructive. To avoid this difficulty, we developed an

0146-9592/95/141536-02$6.00/0

alternative nondestructive method for evaluating theattenuation of hollow waveguides and finding the loca-tion of defects.

The measurement system is schematically illus-trated in Fig. 1. He–Ne laser radiation is coupledinto a silica or glass optical fiber that is inserted intothe hollow waveguide’s bore. This fiber delivers thecoupled radiation to selected places inside the hollowwaveguide. A detector is coupled to the outlet of thehollow waveguide. By moving the hollow waveguideand maintaining the fiber fixed we achieved scanningof the He–Ne radiation from the tip of the fiber alongthe bore of the waveguide. Using this method of scan-ning, we measured the transmission through the hol-low waveguide by changing the distance between thefiber tip and the hollow waveguide outlet. This isequivalent to measurements of transmission for vari-ous waveguide lengths by the cutback method withoutchanging the coupling, but it is not destructive. Fromthese measured data we evaluated the attenuation ofthe hollow waveguide according to Eqs. (1) and (2) andcompared the results with the cutback data.

The transmission measurements for glass andplastic hollow waveguides as a function of distancefrom the tip of the fiber to the outlet of the waveguideare shown in Fig. 2. The corresponding attenuationsfor glass and plastic waveguides obtained with ourmethod are 0.32 and 0.83 dBycm, respectively. Thevalues obtained from cutback measurements were0.11 and 0.74 dBycm for glass and plastic waveguides,respectively. The measured results were comparedwith the results from a numerical calculation basedon the ray-tracing method.7 This calculation takes

Fig. 1. Schematic illustration of the measurement setup.

1995 Optical Society of America

July 15, 1995 / Vol. 20, No. 14 / OPTICS LETTERS 1537

Fig. 2. Transmission versus the distance from the fibertip to the waveguide outlet for glass (filled circles) andplastic (open circles) waveguides. The correspondingattenuations are 0.32 and 0.82 dBycm for the glass andplastic waveguides, respectively, with negligible couplingloss (T ­ 100% for zi 2 z1 ­ 0d.

Fig. 3. Comparison between the calculated (solid curves)and measured data for glass (filled curves) and plastic(open circles) waveguides.

into account the refractive index and thickness of themetal and dielectric layers, e.g., AgyAgI, the surfaceroughness (calculated according to Ref. 8), and thewaveguide length. The data shown in Fig. 3 indicategood agreement between the experimental and theo-retical values for distances as long as 8 and 6 cm forglass and plastic waveguides, respectively. For longerdistances, small deviations of transmission betweenthe experimental and calculated values are shown.It can be theorized that these deviations are due totechnological problems related to the homogeneity ofthe deposited films along the hollow waveguide. Thesame experimental procedure can also be employed forobtaining information about roughness at specificpoints along the internal wall of a hollow waveguide.

Fig. 4. Transmission versus the distance from the fiber tipto the waveguide outlet for a plastic waveguide. A pointdefect at 10 cm from the waveguide distal end was detectedthat reduced the overall attenuation by 1.07 dB.

Figure 4 shows a plot of the power versus the dis-tance between the tip of the fiber and the hollowwaveguide outlet that we obtained while scanning aplastic waveguide. This figure clearly shows that thewaveguide has a point defect (larger scattering) at10 cm from the distal end, increasing the attenuationby 1.07 dB. Such a method can be used for qualitycontrol of hollow waveguides.

A method for internal scanning of hollow waveguideshas been developed. This method can be used for non-destructive measurements of attenuation as a substi-tute for the cutback method or to find the location ofpoint defects that reduce the transmission of radiationthrough the hollow waveguide.

References

1. N. Croitoru, J. Dror, and I. Gannot, Appl. Opt. 29, 1805(1990).

2. M. Miyagi, Y. Shimada, and S. Nishida, Opt. LaserTechnol. 17, 197 (1985).

3. J. Harrington and C. C. Gregory, Opt. Lett. 15, 541(1990).

4. N. Croitoru, J. Dror, E. Goldenberg, D. Mendelovic,and I. Gannot, ‘‘Hollow fiber waveguides and method ofmaking same,’’ U.S. patent 4,930,863 (June 5, 1990).

5. R. Dahan, J. Dror, and N. Croitoru, Mater. Res. Bull. 27,761 (1992).

6. C. Allard, ed., Fiber Optics Handbook for Engi-neers and Scientists (McGraw-Hill, New York, 1989),Chap. 4, pp. 4.7–4.18.

7. M. Alaluf, J. Dror, R. Dahan, and N. Croitoru, J. Appl.Phys. 72, 3878 (1992).

8. R. Dahan, J. Dror, A. Inberg, and N. Croitoru, ‘‘Scat-tering of IR and visible radiation from hollow wave-guides,’’ Proc. Soc. Photo-Opt. Instrum. Eng. 2396 (to bepublished).