%

i'^Jh

X NBS TECHNICAL NOTE 1032

*^fAU 0* *
 .

NATIONAL BUREAU OF STANDARDS

The National Bureau of Standards' was established by an act of Congress on March 3, 1901
The Bureau's overall goal is to strengthen and advance the Nation's science and technology
and facilitate their effective application for public benefit. To this end, the Bureau conducts
research and provides: (1) a basis for the Nation's physical measurement system, (2) scientific
and technological services for industry and government, (3) a technical basis for equity in
trade, and (4) technical services to promote public safety. The Bureau's technical work is per-
formed by the National Measurement Laboratory, the National Engineering Laboratory, and
the Institute for Computer Sciences and Technology.

THE NATIONAL MEASUREMENT LABORATORY provides the national system of

physical and chemical and materials measurement; coordinates the system with measurement
systems of other nations and furnishes essential services leading to accurate and uniform
physical and chemical measurement throughout the Nation's scientific community, industry,
and commerce; conducts materials research leading to improved methods of measurement,
standards, and data on the properties of materials needed by industry, commerce, educational
institutions, and Government; provides advisory and research services to other Government
agencies; develops, produces, and distributes Standard Reference Materials; and provides
calibration services. The Laboratory consists of the following centers:

Absolute Physical Quantities — Radiation Research — Thermodynamics

2
and
Molecular Science — Analytical Chemistry — Materials Science.
THE NATIONAL ENGINEERING LABORATORY provides technology and technical ser-
vices to the public and private sectors to address national needs and to solve national
problems; conducts research in engineering and applied science in support of these efforts:
builds and maintains competence in the necessary disciplines required to carry out this
research and technical service; develops engineering data and measurement capabilities;
provides engineering measurement traceability services; develops test methods and proposes
engineering standards and code changes; develops and proposes new engineering practices;
and develops and improves mechanisms to transfer results of its research to the ultimate user.
The Laboratory consists of the following centers:

Applied Mathematics — Electronics and Electrical Engineering- — Mechanical

Engineering and Process Technology 2
— Building Technology — Fire Research —
Consumer Product Technology — Field Methods.

THE INSTITUTE FOR COMPUTER SCIENCES AND TECHNOLOGY conducts

research and provides scientific and technical services to aid Federal agencies in the selection,
acquisition, and use of computer technology to improve effectiveness and
application,
economy in Government operations in accordance with Public Law 89-306 (40 U.S.C. 759),

relevant Executive Orders, and other directives; carries out this mission by managing the
Federal Information Processing Standards Program, developing Federal ADP standards
guidelines, and managing Federal participation in ADP voluntary standardization activities;
provides scientific and technological advisory services and assistance to Federal agencies; and
provides the technical foundation for computer-related policies of the Federal Government.
The Institute consists of the following centers:

Programming Science and Technology — Computer Systems Engineering.

'Headquarters and Laboratories at Gaithersburg, MD, unless otherwise noted: '

mailing address Washington, DC 20234.

Some divisions within the center are located at Boulder, CO 80303.
 ATlOKAl. BtJMAO
or stakdabm
UBltABY
Measurement of Far-Field and JUN 1 5 1981

Near-Field Radiation Patterns

from Optical Fibers no. (o>^.-

, *>

Ernest M. Kim
Douglas L. Franzen

Electromagnetic Technology Division

National Engineering Laboratory
National Bureau of Standards
Boulder, Colorado 80303

***

U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary

NATIONAL BUREAU OF STANDARDS, Ernest Ambler, Director

Issued February 1981

 NATIONAL BUREAU OF STANDARDS TECHNICAL NOTE 1032
Nat. Bur. Stand. (U.S.), Tech. Note 1032, 48 pages (Feb. 1981)
CODEN: NBTNAE

U.S. GOVERNMENT PRINTING OFFICE

WASHINGTON 1981
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402

Price $2.50 (Add 25 percent additional for other than U.S. mailing)
 . 1

CONTENTS

Page
1 INTRODUCTION 1

2. RADIATION PATTERN CONSIDERATIONS IN OPTICAL FIBERS 1

2.1 Measurements Utilizing Far-Field Patterns 5

2.1.1 Radiation Angle (Numerical Aperture) 5
2.1.2 Attenuation with Restricted Launch via a Mode Filter 6
2.2 Measurements Utilizing Near-Field Patterns 6
2.2.1 Index Profile.'. 6
2.2.2 Core Di ameter 7
2.3 Mode Volume Transfer Function--A Measurement Utilizinq
Both Far- and Near-Field Patterns 7

3. DESCRIPTION OF FAR-FIELD MEASUREMENT SYSTEM 8

3.1 Source 8
3.2 Detector 10

4. PERFORMANCE OF FAR-FIELD SYSTEM 11

4.1 Precision 11
4.2 Accuracy 11
4.3 Dynamic Range IS

5. TYPICAL RESULTS FROM FAR-FIELD MEASUREMENTS IS

5.1 Radiation Angle (Numerical Aperture) 15
5.2 Attenuation with Restricted Launch via a Mode Filter 15
5.3 Symmetry of Radiation Patterns 18
5.4 Mode Dependence of OH" Absorption 18

6. DESCRIPTION OF NEAR-FIELD MEASUREMENT SYSTEM 21

7. PERFORMANCE OF NEAR-FIELD SYSTEM 23

7 . Prec i s i on 23
7.2 Resolution and Accuracy 23
7.3 Dynamic Range 31

8. TYPICAL RESULTS FROM NEAR-FIELD MEASUREMENTS 31

8.1 Claddinq Liqht 31
8.2 Step-Index Fibers 31
8.3 Symmetry of Radiation Patterns 31
8.4 Index Dip Pathologies 31

9. CONCLUSION 38

10. REFERENCES 38

m
 MEASUREMENT OF FAR-FIELD AND NEAR-FIELD
RADIATION PATTERNS FROM OPTICAL FIBERS

Ernest. M Kim* and Douglas L. Franzen

Systems are described for measuring the far- and near-field radiation pat-
terns from optical fibers. Parameters which affect measurement precision, accu-
racy, resolution, and signal-to-noise ratio are discussed. Measurements utiliz-
ing radiation patterns are covered; this includes radiation angle (numerical
aperture), attenuation using mode filters, index profile, core diameter, and mode
volume transfer function. Experimental examples are given in many instances.

Key words: Attenuation: core diameter; far field; index profile; mode filter;
numerical aperture; radiation angle; radiation patterns.

1. INTRODUCTION
Radiation patterns exiting optical fibers yield important information used in fiber
specification. For example, in some instances core diameter and numerical aperture are
obtained from radiation patterns. Measurement practices with regard to these parameters
continue to evolve as accumulated practical experience and the efforts of standards groups
lead toward uniformly accepted procedures. This Technical Note is one of a series intended
to describe the present design and capabilities of fiber measurement systems used at the

National Bureau of Standards. These systems are perhaps representative of current practice
in tne industry ana many of the present techniques and methods will be relevant to future

systems.

2. RADIATION PATTERN CONSIDERATIONS IN OPTICAL FIBERS

The radiation propagating in an optical fiber can take on various spatial and angular
distributions depending upon the relative excitation of fiber modes. Useful information
about the fiber can be ootained oy analyzing the radiation exiting an end. This analysis
can take place in either the far- or near-field region of trie end. A far-field measurement

determines the angular dependence of intensity (

irradiarice) sufficiently far from the end of
tne fiber (fig. 2-1); wnereas, a near-field measurement determines the spatial intensity
distribution in the plane of the end face.
In the far-field, tne normal i zed, angular intensity distribution no longer depends on
distance from the fiber. While tne intensity at any given angular coordinate decreases as
the square of the distance from the fiber end, trie normalized pattern remains constant. If

the far-field pattern is circularly symmetric about the fiber axis, then a single angular
coordinate is sufficient to describe the pattern.
The far-field, which is synonymous with the Fraunnoffer diffraction reqion, is usually
said to start at a distance

e
NBS Coop graduate student on leave from New Mexico State University, Las Cruces, NM.
 INTENSITY, I *-

(a) FAR FIELD

FIBER
(b) NEAR FIELD

igure 2-1. Radiation pattern measurements from optical fibers, (a) far field is qiven by
1(8) at z>>z (b) near field is given by I(r) in xy plane.
,

L
AXIS INDEX

RADIUS

igure 2-2. The meridional ray in a step index fiber makinq the maximum possible angle, e^,
with the fiber axis.
 2a)
(1)

from the fiber end, where X is the wavelength of liqht and a the core radius [1]. When
z » z
Q , the amplitude (electric field) distribution in the far-field is closely related to
the spatial Fourier transform of the amplitude distribution existing at z = 0, [1]. In

practice, far-field measurements are made at distances greater than 10 z from the fiber
end. For the standard graded index core diameter of 50 \m and a wavelenqth of 1 ym, 10 z
Q
is 2.5 cm. For the standard step index core diameter of 100 iim, the distance becomes

10 cm. Both of these lenqths represent convenient laboratory workinq distances.

Radiation patterns are measured on either short or lonq lenqths of fiber, dependinq
upon the application. Frequently the short lenqth is a 2 m cutback lenqth left over from an

attenuation or bandwidth measurement. In this Technical Note, a lonq lenqth refers to the
entire lenqth of fiber under test and, for telecommunications fibers, qenerally exceeds
1 km.

Radiation patterns depend upon launchinq conditions and contain contributions from both
quided and tunneling leaky modes of the fiber [2,3]. In short lenqths, patterns result
mainly from guided modes with a significant contribution from leaky modes. The relative
importance of the leaky mode contribution depends upon the index profile and launchinq con-
ditions. In lonq lenqths, leaky modes have qenerally decayed and the patterns result from
guided modes modified by differential mode attenuation and mode coupling. Far- and near-
field patterns resultinq from overfilled launching conditions on graded index fibers gener-

ally become restricted with further propagation along the fiber. This occurs because dif-
ferential mode attenuation is generally higher for higher order modes. Eventually an equi-
librium is reached between mode couplinq and differential mode attenuation to yield a steady
state pattern. However, steady state is seldom observed in practice since mode couplinq
lenqths in good fibers are several kilometers T41 if not tens of kilometers [5].
The remainder of this section briefly discusses the measurement applications of far-
and near-field patterns. Various applications are summarized in table I. For some of these
measurements, procedures are pending before standards groups [61. In general, more atten-
tion has been given to standardizing procedures for graded index fibers than for step index
fibers.
 Table I
Measurement Applications of Radiation Patterns from Multimode Fibers

Appl ication Fiber length Radiation Launching General

employed pattern condition comments
Short Long Far Near
(2 m) field field

Radiation Angle Overfill Meridional ly defined NA

(Numerical aperture) approximated for near
parabolic graded index
fibers. Special consi-
derations necessary for
step index.

2 Attenuation using Overfill Used to qualify mode

restricted launch filters; applies to
via a mode filter graded index fibers over
1 km in length.

3 Index profile Overfill Little or no leaky mode

correction reguired for
near parabolic qraded
index fibers; large cor-
rection for step index.
Limited resolution,
information on guiding
regions of fiber only.

4 Core diameter Overfill Possible uncertainty due

to low index "barrier"
layer, poor resolution
near core-cladding
boundary, may have
problems conforming to an
index profile definition
of core diameter, is easy
to implement.

Mode volume Restricted Applies to graded index

transfer function fibers, measurement sys-
tem must have ability to
control launch mode vol-
ume and therefore beam
optics systems capable of
independently variable
launch spot and NA dre
usually required.
 2.1 Measurements Utilizing Far-Field Patterns
2.1.1 Radiation Anqle (Numerical Aperture)
The cone of liqht rays accepted by a fiber is of practical importance. Launching effi-
ciency of sources like LEDs and loss performance of joined sections of dissimilar fiber de-
pend upon this parameter. "Radiation anqle" is a measure of the cone anqle transmitted by
the fiber [7]. This measurement is usually made in the far field of a short, 2 m lenqth ex-

cited at a specified wavelenqth using overfilled launchinq conditions; i.e., constant irra-

diance over the core with a launch anqle greatly exceedinq the fiber acceptance anqle.

In early fiber work, the cone anqle containinq 90 percent of the transmitted power was
determined by either inteqratinq a far-field intensity pattern or by translatinq an aperture
in the far field. The sine of the half anqle of this cone was frequently termed "90 percent
numerical aperture."
More recent procedures pendinq before standards qroups define radiation anqle directly
from a far-field intensity pattern [8]. In this instance, radiation anqle is defined as the
half anqle where the far-field intensity has decreased to 5 percent of the peak value. The
sine of this angle, for near parabolic index multimode fibers, is close to the index defined
numerical aperture.
Historically, the term "numerical aperture" (NA) has been used to describe the larqest

angle meridional ray accepted by a fiber, figure 2-2, with

NA = sin 9 = AV - n 22 (2)

n !-n 2
= u x
/2E with A «- (3)

where m is the core index of refraction, m the cladding index, and A is defined by the
above equations. While figure 2-2 applies to step fibers, analysis shows that the numerical
aperture of graded index fibers is also given by eg (2) with ni being the on-axis index of
refraction [9].
The sine of the radiation angle based on the 5 percent intensity points is close to the
index defined NA of eg (2) for near oarabolic index fibers. Theoretical work shows the ac-
ceptance angle of skew rays is always less than that of meridional rays in near parabolic
index fibers [10]. Adams, et al., show that leaky modes in a near parabolic index fiber are
all contained within the meridional ly defined numerical aperture [31. Thus, when all modes
are excited, the largest observed angle in a far-field pattern would be given by 6 in eq
Q

(2). At the 5 percent level, patterns are close to the maximum observable angle.
For step fibers, the situation is quite different. Here the acceptance anqles for skew
rays can exceed 6 . When all modes are excited, radiation would appear at anqles qreater
than 8 . Thus, to obtain the numerical aperture based on eq (2) for short step index
fibers, special launchinq conditions would be necessary. This would probably mean launchinq
only meridional rays. This could be accomplished by small spot excitation at core center.
 2.1.2 Attenuation with Restricted Launch Via a Mode Filter
In measuring the attenuation of lonq telecommunications type fibers, it is useful to

launch with a distribution which qives good concatenation predictions [111. This is pre-
sently accomplished by restricting the launch, to avoid the excitation of certain high loss
modes which do not propagate far in the fiber. In general, this can be accomplished by one
of two approaches. The beam optics approach avoids the initial excitation of certain high
loss modes by restricting the launch spot size and numerical aperture at the fiber input
end. (70 percent of the core diameter at 70 percent of the fiber NA is the choice pending

before standards groups.) In the mode filter approach, the fiber is initially overfilled
and a mode filter is applied to strip out certain high loss modes.
Far-field radiation pattern measurements on lonq and short fiber lenqths are used to
qualify a mode filter for an attenuation measurement. Current proposals call for overfil-
ling the input of the fiber under test in both spot size and numerical aperture, then mea-
suring the radiation angle (5 percent intensity) exiting the long length. An appropriate
mode filter is applied so the radiation angle exiting a 2 m length is less than that of the

long length by 3 + 3 percent. Now the mode filter has stripped away enough of the hi qh
angle-high loss modes to simulate, in a 2 m cutback length, the radiation pattern resulting

from a kilometer or more of propagation. Common mode filters include dummy fibers and
mandrel wraps [12], [13].

2.2 Measurements Utilizinq Near-Field Patterns

2.2.1 Index Profile

When a multimode optical fiber havinq a large number of modes is uniformly illuminated
to equally excite all bound modes, the near-field pattern approximates the refractive index
profile of the core [14]. In short lengths where differential mode attenuation and mode
coupling are negligible, the deviation from the index profile is due mainly to the presence
of tunneling leaky modes. Sladen, et a! .
, have derived a leaky mode correction factor
allowing one to obtain the index profile from transmitted near-field data [2]. The radial
index profile difference, n(r) - n->, is given by

n(o)-n
n(r) - n2 *
2
W-
D{ r \

(4)
C(r,z) P(o)

where n(r) is the index in the core, no the cladding index, P(r) the near-field intensity as

a function of radius, and C(r,z) the correction factor. C(r,z) depends on the nominal pro-
file shape, fiber lenqth, and radial position. C(r,z) is zero at the core center and grad-
ually rises to a maximum f\ear 0.9a, where a is the core radius. As an example [14], for a

1 m length of fiber havinq a near parabolic index profile, a core diameter of 80 pm, an NA
of 0.18, and at a wavelenqth of 0.9 vm, the correction factor is 8 percent at 0.6 a, rising
to 20 percent at 0.85 a.

Whether a leaky mode correction should be made for near parabolic profile, graded index
fibers is still open to debate. Calculations by Petermann [15] indicate a slight core el-
lipticity can cause leaky modes to attenuate more rapidly than indicated in references [2]
and [14] and corrections would not be necessary for practical near parabolic fibers. Leaky
mode correction factors have been derived for elliptical core fibers [16].
Regardless of the appropriateness of a correction, the near field is a close approxima-
tion to the index profile for near parabolic fibers. Also, corrections mainly affect the

0.5a to 0.9a region and the apparent width of the pattern near the baseline is not affected.
This is of importance in determining core diameter from near-field measurements.
For step index fibers, leaky modes are definitely significant and for short lengths the
near field profile differs substantially from the index profile (see sect. 8.2).
It should be pointed out that the above comments apply only to multimode fibers with
large numbers of modes. In fibers with single or few propagating modes, the near-field in-
tensity distribution is given by the mode patterns themselves which can differ substantially
from the index profile shape.
The near-field intensity distribution gives information on the spatial location of
power in the fiber. Thus, one is measuring only the guidinq regions of the f iber. What
happens at the core cladding interface or in the cladding is not indicated.

2.2.2 Core Diameter

Core diameter is usually defined from the fiber refractive index profile. In one re-
commended definition, the core diameter is that diameter on the index profile where the

refractive index of the core exceeds that of the cladding by k times the difference between
the maximum refractive index in the core and the minimum index in the cladding, where k is a

specified constant (0 < k < 1) [7]. Near-field measurements can be used to determine core
diameter in some circumstances because of the close relationship between near field and re-
fractive index profile. Leaky mode corrections, if they are even necessary for near para-
bolic index fibers, do not affect the measured width of the near-field pattern near the
baseline. A core diameter measurement based on the full width at the 2 to 3 percent inten-
sity points may be appropriate. A more serious uncertainty in determining core diameter

results from the use of low index "barrier" layers between the core and cladding. How the
near field is affected by such fluctuations near the cladding has not been established.
Also, in graded index fibers, there is a loss of resolution near the core claddinq boundary
due to the decrease in local numerical aperture [17]. Some of these problems may be alle-
viated in a recently described "modified near-field" technique [18]. Despite some limita-
tions, near-field measurements do indicate where power is spatially located in the fiber
and, in some practical situations, this may be as important as the core diameter determined
from an index profile.

2.3 Mode Volume Transfer Function--A Measurement

Utilizing Both Far- and Near-Field Patterns
Holmes, et al .
, have introduced a concept describing the effect of propagation on
spatial and angular intensity distributions in graded index fibers [4]. They define an
"effective mode volume", EMV, based on far- and near-field intensity distributions. Speci-
fically, the EMV at a qiven point along the fiber is the square of the product of the FWHM
(Full Width at Half Maximum) of the near field and the sine of the HWHM (Half Width at Half
Maximum) of the far field.
EMV is related to the number of modes excited in the fiber. This can be intuitively
seen when applied to the expression for the total number of quided modes, N, in a fully
excited qraded index fiber having a power law profile a [9].

2 ™i a
aA
N
2+a

where a is the core radius, X the wavelenqth, and other terms have been previously

defined. After arranging terms,

N =
2+a
—X
2
) C(NA) • a]
2
(6)

2
a / 2tt
^ EMV (7)
2
2+a X

Of interest is the EMV transfer characteristic for a fiber. This describes how various
EMVs propagate in the fiber and is a plot of output EMV versus input EMV. By controlling

launch spot and NA, various input EMVs can be generated. Input EMVs are determined from

radiation patterns exiting a 2 m length at the input. The corresponding output EMVs are
obtained from radiation patterns exiting the lonq lenqth.
The EMV transfer characteristic qives information on mode couplinq, differential mode
attenuation, and quasi-steady state EMV. A fiber without mode couplinq or differential mode
attenuation appears as a 45° line on the transfer characteristic plot; i.e., the output EMV

is equal to the input EMV.

A principle use of the EMV concept is in the prediction of concatenation. EMV transfer

characteristics and attenuation versus EMV for all the fibers in the link allows one to pre-
dict accurately the concatenated attenuation for any input EMV [41.

3. DESCRIPTION OF FAR-FIELD MEASUREMENT SYSTEM

3.1 Source
The radiation source used with both far- and near-field systems is capable of producing
independently variable spot sizes and launch numerical apertures, fiqure 3-1. Aperture 1,

which controls launch spot size, is illuminated by either a strip filament lamp or a flat,

close-coiled filament, quartz haloqen lamp havinq an etched envelope. The lamp should have
qood spatial uniformity and, from this viewpoint, the strip lamp is preferred.
A broadband interference filter selects the nominal wavelenqth. Since these measure-
ments are relatively insensitive to wavelenqth [191, a broadband filter with 80 nm linewidth
gives improved signal-to-noise over narrowband filters. Angles of incidence on the filter
are so small that angular dependent transmission effects are avoided.
The launch lens produces a demagnified image of aperture 1 on the fiber end; a demagni-
fication ratio of 22 is typical and spot sizes down to 18 wn have been obtained. Aperture 2

controls the launch numerical aperture from a maximum of 0.36 to a minimum of 0.03.
 CHOPPER
WHEEL

LAUNCH
LENS

RADIATION
FROM
INCANDESCEN-
BROAD BAND
INTERFERENCE
FILTER
A vU TRANSLATION

?
LAMP 1 FIBER

APERTURE
NO. 1
APERTURE
NO. 2

Figure 3-1. Launch optics used with far- and near-field systems.

PIN SILICON
DETECTOR TO LOCK-IN
AMPLIFIER

APERTURE
CLADDING
MODE STRIPPER

FIBER

STEPPER MOTOR
CONTROLLED ROTARY TABLE

Fiqure 3-2. Far-field measurement system usinq a fixed fiber end and a rotating detector.
 Overfilled launchinq conditions are used for most of the measurements, Table I. There-
fore, additional optics to optically center the aperture image on the fiber end are not
usually employed. Peak power transmitted throuqh the fiber has been the aliqnment criterion

when overfilled launchinq conditions are used.

A chopper wheel is located in the beam to modulate the liqht at 900 Hz. This provides
a reference siqnal for narrowband, phase sensitive detection with a lock-in amplifier.

3.2 Detector
The far field may be scanned by a number of different techniques: ('1) fixed fiber end,
rotating detector, (2) fixed detector, rotatinq fiber end, (3) fixed fiber end, fixed detec-
tor, rotating mirror, and (4) fixed fiber end, detector array.
The system described in this Technical Note uses a fixed fiber end and a rotatinq
detector, fiqure 3-2. This approach represents perhaps the simplest construction of
available choices.
The fiber first passes throuqh a claddinq mode stripper consistinq of two 10 cm lonq
felt pads wetted with index matchinq fluid. Buffer coatinqs are removed from the fiber
where it contacts the mode stripper. Near-field scans show this type of mode stripper
effectively removes light from the cladding.
A vee groove positions the fiber so the end is coincident with the axis of rotation.
Some measurements, especially near fields, are affected by the amount of force used to hold

the fiber in the vee groove. These distortion effects are eliminated when a small, felt
padded weiqht holds the fiber in the qroove. Before makinq a measurement, the fiber end is

visually inspected for flatness and perpendicularity. The vee qroove position can be
slightly adjusted by a two dimensional translation stage to assure that the detector scan
passes through the maximum intensity part of the pattern.
Scanning is accomplished with a stepper motor controlled rotary table which swings the
detector through a 12 cm radius arc. At this distance the far-field criteria is satisfied
for core diameters of 100 m or less. Angular motion (V? arc min. per step) is fine enouqh
to qive smooth far-field curves.

Detector aperture size is chosen with a trade-off between resolution and signal-to-
noise ratio. Presently, a 0.8 mm diameter aperture is used qivinq a resolution of 0.38°.
Reducing aperture size to improve resolution, and hence accuracy, would not be impor-
tant for most measurements which are relative comparisons (Table I). Radiation anqle (NA)
is the only far-field measurement which is absolute. A simple one-dimensional model based
on a parabolic shaped far-field pattern with an NA of 0.2 predicts an error of 1.6 percent
in the determination of radiation angle when the curve is acquired with 0.38° resolution.

Measurement precision for determining radiation angle is in the ranqe of 1 to 2 percent when

a new output end is prepared and realigned (Sect. 4.1); therefore, a significant reduction
in aperture size would not result in much improvement.
A silicon PIN diode operating in the photovoltaic mode is used as the detector [201.
This detector with an active area of 5.1 mrrr has a built-in operational amplifier and is
mounted directly behind the aperture. To improve signal-to-noise ratio, a time constant of

10
0.4 s is used on the lock-in amplifier. The scanning rate is chosen so a resolution element
(0.38°) is scanned in approximately three time constants. At this rate a fiber far-field
pattern is obtained in three to four minutes.

Data may be acquired by several methods. For curve fitting or numerical analysis,
digital acquisition is necessary; however, for most applications, a good, wide paper strip

chart recorder is adequate.

4. PERFORMANCE OF FAR-FIELO SYSTEM

4.1 Precision
Precision is a statement of measurement repeatability. It says nothinq about a "sys-

tematic" error which is the offset of the averaqe of a larqe number of measurements from the

true value. Best precision is obtained when a measurement is repeated without disturbinq
any parameter. A typical example of this precision is qiven in fiqure 4-1. Here a far-

field measurement was repeated seven times on graded index fiber A usinq overfilled launch-
inq conditions at a wavelength of 860 nm. Standard deviations for determining the full
widths at the 80, 50, 20, 10, and 5 percent intensity points are shown. Error bars repre-
sent + one standard deviation (2 standard deviations in lenqth) and are qiven as a percent-

age of the full width at that particular point. A typical value is in the 0.5 percent

ranqe. It should be noted that the width of an ink line is approximately 0.3 percent of the
curve FWHM and therefore represents a limitation in determininq actual precision. The main
conclusion drawn from fiqure 4-1 is that the system has no appreciable dc drift problems.

For many measurements, the above precision is not applicable. In practice, new ends
must be prepared and the system aliqned. To determine this precision, far-field measure-
ments were repeated seven times, for each measurement a new output end was prepared and the
system slightly realigned to the maximum intensity point. Three different fibers represent-
ing two different manufacturers were measured, figures 4-2 and 4-3 [21]. Standard devia-
tions for graded index fiber B (not shown) were about 1 percent on lower parts of the curve.
For graded index fiber A the correspondinq values are a little hiqher--in the 1 to 2 percent
ranqe. Fiber C, with a step index, has a much flatter top in qeneral aqreement with theory
[22]. Here the precision is also in the 1 to 2 percent ranqe. From these tests, a preci-
sion of 1 to 2 percent is indicated for a radiation angle measurement.

4.2 Accuracy
Accuracy is determined by the absolute value of the angular calibration factor, angular
resolution, and system linearity. The angular calibration factor was determined by reflect-
ing a low divergence He-Ne laser beam off a mirror attached to the rotary table and onto the
laboratory wall. The table was rotated through 360° and the beam returned, within the
nearest step, to the original position. This procedure determines the angular rotation per
step more accurately than is practically needed.
As previously mentioned, the finite angular resolution introduces an error when deter-
mininq the radiation angle. This error is estimated to be less than 1.6 percent.

11
 " >
-o
S_
CD "C
JO S_
i- (C

C
ra

on

CJ O)
c c:
'

ra o
o +i

0) .*
h E
c
<T3

ui c:
i_ ^c
<u co
+->
a) «
E
rO
<
Zo
• c
-r-
S- o
"3 CO - D.
Q- C\J CV
4->
O O -C
•

O «3

•-+J
c E m +J
-^
CD
.c
S
^
ce
C •
«
. * JC
C cc -i->

O LO
-
c -c
•

l/l »
•i- x: +->
o -4-> "3 5_
O) CT rt3

s_ c O) r—
Q. cd
ro cr
4-> C C
c E i- "3
CD -^
E .— S- i—
o i—
QJ
l- •
3 CTM-
r

C
O M-
<D X r— O
E CD ro
<D
-o C -O D-.
O) res

•i- -o re c
4- a> O CD
r- O
s_ <a T3 i-
c: CD
u_ ex •i- Q.

QJ
S-
3
ex

12
 o
4- -O
CD
+-> J->
C re

C ID

re c
CD O
S_ •!-
4-)
"O re
c >i-
T3 >
CD
OC X3
CD TT
S_
+j re
3 -O
Q- C
4_> fO

o tyi

T3
co
> o
cm
CD

o
re
CD +i o
o .»-u
CD <C c
s_ re

r (D^
+j xi
s_

•r-
2 4-
O-r-
•

C X CM
O CD •

f- T3
C
l/l
O
O LO
CD T3
S-
O.T3
CD O
re •>

+-> s- oo
c cs
CD
E
O •

CD • J->
S- 4J TO

c/) CD CD
re E +->
CD CD <T3

E S- C
3 T-
T3 CO T3
r- re s_
CD CD O

S- u o
re re r—
u. cd re

CD
S-

cn

13
 S_ <DO
c
vo oin
o
+-> +i •

c -o •
(1)

E E "O
c c c
CD re
r-
I—
O
<Xi "
re cc i—
cu
s- -c
E
re DC •

-o c
0> LC
+->
13
-E
-t->
O
Q- CT «
4-> C 00
3 <D
o -— o •

T3 E •*->
re
o> -2^
>
re c\j <u
a> • -m
,— i— re
o «-•>—
c
a;
s_ x -a
a> s_
.c "O o
+-> c
i- •.- CT
3 C
D. C
C <U <—
o
•i—
+j
en
re

• .—
<_>
O)
O" OJ
+->
re
S_ S- <J
a. a) •-
xi -o
+-> -r- C
C U- •>-

E c
a> • o
•<-
S- +j
1/1 re
re e >-
<V 03 >
E i- 0)
3 T3
T3 </>
i re TJ
a> a> s- E
3
.,- e re
m- -o E
i_ u re x
re re 4-> re
u. <u </i E

14
 Detector non-linearity could affect the shape of the measured pattern. A calibrated

neutral density filter was used to confirm detector linearity up to the highest power levels
used with the system.

4.3 Dynamic Ranqe

Signal -to-noise ratio was determined by attenuating the source until the noise level

was observed. Figure 4-4(A) shows the far field from a fiber having a loss of 5 dB when

excited with overfilled launch conditions. In figure 4-4(B), the source is attenuated by
approximately 17 dB to display the noise level. Also shown is the noise level on the base-
line and at the peak value; in both cases the noise is about 4 pV oeak-to-peak. Signal-to-
noise ratio for a short length of fiber, having negligible attenuation, is 34 dB at 860 nm
using an 80 nm bandpass interference filter, overfilled launching conditions, and a time
constant of 0.4 s on the lock-in amplifier. In figure 4-4(B) the siqnal-to-noise ratio is

12 dB and represents the lowest level for making a reasonable measurement. Based on this

criteria, the system has a dynamic range of 22 dB. There are measurements, especially with
very restricted launching conditions, where more signal-to-noise would be desirable. The
easiest way to increase signal is to increase source linewidth. For some measurements this

may be acceptable. An additional 7 dB improvement in signal can be obtained by replacing

the broadband interference filter with a long pass filter having a cut-on wavelength of 0.78
urn. In this case the long wavelength cutoff, 1.1 un, is determined by the spectral response
of the silicon detector.

5. TYPICAL RESULTS FROM FAR-FIELD MEASUREMENTS

5.1 Radiation Angle (Numerical Aperture)
Radiation angle was determined for a number of commercially available multimode fibers.
Measurement precision was consistent with the results of section 4.1. Fiqure 5-l(A) is a

typical result from a graded index fiber with a rather low NA of 0.16; while figure 5-1 (
B)

shows a fiber with a rather high NA of 0.23. Results of figure 5-1 were obtained using
overfilled launching conditions on a 2 m length of fiber. In almost all cases, far-field
patterns from commercial graded index fibers have been smooth, bell shaped, and exhibit a

single central maximum.

5.2 Attenuation with Restricted Launch via a Mode Filter

Radiation pattern measurements on both long (1 km or more) and short (2 m) lengths of
the same fiber are necessary to qualify mode filters for restricted launch attenuation mea-

surements. A successful mode filter produces an output radiation anqle from a 2 m length
equivalent to the output radiation angle from the full test length excited by overfilled
launching conditions. Restrictions in radiation angle, due to fiber propagation, are a

function of differential mode attenuation and mode coupling. For many typical graded index
fibers, the radiation anqle from an overfilled launch restricts by 4 to 8 percent after a

kilometer of propagation.

15
 _1
o CD -
z -O CD
<
IT?

_o 4-
l+-
o
O
<u
i/l 1/1
(/) r—
O o
<— c
re o
+->
CD
C 1

r— re
> c
nl a
-C • r—
1/1

S-
cu ff

.o ce
r- -o
M-
r^
X ,

O)
T3 >)
Ci
-C

-o
a
0)
QJ +j
-o n3
r« 3
S- C
cr QJ
+J
E +J
O «
5-
14- <u
o
"O 5-
r— 3
<u O
(/I
<+-
-o Q < s-
.c
-M
LU rC •i-
_]
o
z ^
< <;

s-

cr>

16
 C
CD
(_>
—
CO

s- -a
<d
Q.
c
T3

LD .—

<!J (/)
i CD
CT >
C i-
<n =3

e
O E
•i

4->
o
s_
fO t-
"C
ro
o<u
S- c
3 T5

'
O 4-J

o
T3
C a>
13
<T3
-H
CT OO
•(— cv
sz
sz
o
+->
• r-
ac
3 <v

C/l tc
s_ 1

cu
jO o
l*- 4-
o
X
a> i/>
T3 —
C <c
,,_ z:

XJ
CU
"C
n3 _£r
l_ +J
cn-o
•i >)
4- 3
O 01
>> >
V) +J r—
QJ *->
i i/l <J
o_ c a:
E QJ a
r5 4J i/i
X cz a>
LU f— s_

17
 One mode filter possibility is a series of serpentine, macroscopic bends, fiqure 5-2.
This geometry in a step index fiber has been previously used as a mode-scrambler for band-
width measurements [23] [24]. The specific geometry used here consists of eleven, 9 mm
diameter nylon posts--six fixed and ffve' moveable. A fiber is placed in the mode filter and
the moveable posts are translated to produce sinusoidal bends having a period of 22 mm with
variable amplitude. Maximum far-field restrictions are obtained when the posts are evenly
aliqned (9 mm peak-to-peak amplitude).
The effect of this mode filter on the far-field pattern from a short length of graded
index fiber is shown in figure 5-3. The solid curve was obtained without the mode filter
using overfilled launching conditions to the 2 m length. The dotted curve is the same but
with the mode filter set to maximum amplitude. A restriction of 16 percent in radiation
angle is achieved at maximum amplitude. This particular fiber, when excited by overfilled
launching conditions, exhibits a 6 percent restriction in radiation angle after 0.9 km of
propagation. Thus, the serpentine mode filter in this case would be more than adequate to
achieve the necessary restriction (9 percent) for an attenuation measurement.
It should be emphasized that mode filters affect different fibers in different ways.
The exact nature of the mode reduction is a function of: mode filter, buffer coating, clad-
ding diameter, core diameter, numerical aperture, etc. Fibers have been found where the
serpentine mode filter has produced significantly different results than figure 5-3. It is

therefore necessary to "qualify" a mode filter for the particular class of fibers beinq mea-
sured.

5.3 Symmetry of Radiation Patterns

In general, far-field measurements on sinqle core fibers qive smooth, symmetric, bell
shaped curves havinq a single maximum. Fine structure from the core would be diffracted
into relatively large angles. For example, a plane wave illuminating a circular aperture of
diameter d has a diffraction angle of 1.22 X/d. This is 1.4° when applied to a 50 urn

diameter core at a wavelenqth of 1 vm.

Pattern symmetry is indicated in fiqure 5-4. Here far-field patterns are folded about

a vertical axis passinq throuqh the midpoint of the width at half maximum. These examples
were selected from measurements on a large number of commercial fibers and are typical of
good and poor symmetries. Even in the poor case, the symmetry is fairly good.

5.4 Mode Dependance of OH" Absorption

The OH" radical causes spectrally localized absorption in optical fibers. One of the
stronger lines occurs near 950 nm. Previous backscatter measurements show that OH" is not
necessarily distributed uniformly along the fiber length [25]. This section describes
far-field measurements to qualitatively determine the mode dependence of OH" absorption in a

particular fiber.
For these measurements, a 0.9 km fiber was chosen with a large OH" absorption peak and
little mode coupling. Spectral attenuation measurements indicate a loss of 5 dB/km at
850 nm which increases to 40 dB/km at 950 nm. By analyzing far-field patterns at 850 nm

18
 NYLON POST
9mm DIA.

Figure 5-2. One type of mode filter consistinq of a series of serpentine, macroscopic
bends. Moveable posts can be translated into alignment with fixed adjacent
posts.

WITHOUT MODE FILTER

ANGLE, DEGREES

Figure 5-3. Effect of mode filter on 2 m length of graded index fiber. Solid curve is
without mode filter, dotted curve is with mode filter set at maximum amplitude.

19
 o ti
.C "C
+->
i- li- 1)
re O •

CL 4-
+-> 1

+-> C S-
M- •
r— re
<v o u-
_j Q.
T3
</> E a>
S- c
<D .c: •
r-
jQ CJ '

•f— 3
4- o -o
i_ • i—
X :C
CL! 4-J O
T3 tfi

C 1/1

X c
T3 re S-
0) a>
•
-o i
4->
re rC +-> -
S- u <T3
Q--
fit
cr.
JJ
E w 4- -o
o <D o
s- > +J
a 1

u- r-
rC s_ H-
<S) ro 1

c (J CL S-
s_ 3 ro
CD O 4-> 4-
4-> -O -C
+-> rO
•r—
cr C
ro
ro
Q-T3 S_ jr
^ +j
T3 -a x:
i ( i-J >>
0) o r— s-
4- 3 4J
4- O)
1 i/> CD E
S- ••"
rO rO 1,
4- „ — a. (/>

E a>
4- c C S-
O •p— o a>
i -i-j

>> o j-j
S_ T3 -(-> aj
+J aj -Q
QJ +J x:
E -!-> 4-J (/)

E o
>>-c
-c 4->
•i

l/l - 5 -O
c E -C
re s- 3 X
<_ a E QJ
4- •

Q.+. X —
i

.»

>, a: re co
1— Q- E

a)

cn

20
 ;

from overfilled and restricted mode volume launches, this fiber was found to have little
mode coupling.
Qualitatively, if more OH" were located near the core-cladding interface, higher order
(large radius-large angle) modes would be attenuated more and a narrower pattern would
result at the output from an overfilled input; the converse would be true if the OH" were
concentrated near the axis.
Figure 5-5 is the superposition of normalized far-field patterns at 857 nm and 957 nm
using overfilled launching conditions and 70 nm bandpass filters. Precision in determining
these differences should be quite good since only interference filters were changed between
the two measurements. As shown, the presence of OH" absorption at 957 nm does little to

alter the shape from the pattern at 857 nm where no OH" absorption is present. Therefore,
the OH" in this fiber appears to be uniformly distributed with a slight indication of more
OH" near the axis (957 nm curve is wider). The distribution of OH" would depend upon the
preform fabrication process so the above results apply only to the particular fiber.

6. DESCRIPTION OF NEAR-FIELO MEASUREMENT SYSTEM

The near-field system produces a magnified image of the fiber end face. A single

microscope objective gives a magnification of about 50, fiqure 6-1. By scanning a detector

across a diameter of the image, the near-field intensity distribution is obtained. In

figure 6-1, the detector aperture is located at a fixed distance from the objective so the
magnification ratio remains constant independent of other system adjustments.
The detector is an apertured silicon PIN photodiode identical to the one used with the
far-field system. This detector is scanned across the image, which is nominally a few mm in
diameter, by a stepper motor controlled translation stage having a movement of 0.4 pm per
step. The detector aperture size, 80 un diameter, was chosen to be consistent with theo-
retical resolution limits. A smaller size does not improve resolution but results in lower
signal-to-noise ratio. An 80 urn aperture gives a resolution of 1.5 urn when referenced to
the fiber endface.
Care should be used in selection of the fiber holder, which is positioned just after
the cladding mode stripper. The present system uses a vee groove with a small, felt padded
weight to hold the fiber in the groove. If the fiber is held with too much force, stress

will distort the near-field intensity distribution and scatter light into the cladding. The
vee groove is mounted on an xyz translation stage so the image can be brouqht into focus on
the detector aperture.

Alignment is accomplished using the movable mirror and eyepiece shown in figure 6-1.
First, a fiber is imaged in the system at the measurement wavelength (typically 850 nm)
good sensitivity is obtained by imaging the on-axis index dip. Then, using deep red visible
light, the movable mirror is positioned to illuminate the eyepiece. The eyepiece is trans-
lated until the image appears in sharp focus. Now, if a fiber end is brouqht into focus
using the eyepiece, it will be in proper focus at the detector aperture when the mirror is

noved out of position. This method of alignment was checked using a number of fibers.

21
 ANGLE. DEGREES

Figure 5-5. Experiment to qualitatively determine mode dependence of OH" absorption,

Little mode dependence is indicated for this particular fiber.

FIXED DISTANCE

CLADDING

FIBER
MODE STRIPPER IU MOVEABLE
<\ MIRROR
PIN SILICON
DETECTOR;

Q--0V
VEE .4NA
GROOVE MICROSCOPE L APERTURE
HOLDER OBJECTIVE

M EYEPIECE STEPPER
MOTOR
v CONTROLLED
TRANSLATION
STAGE

Figure 6-1. Near-field measurement system based on a radial scan of a magnified near-field
image.

22
Fibers were first aligned visibly with the eyepiece; no further improvement in the focus of
index dips could be obtained at the detector using 850 nm light.
The source used with the near-field system was previously described in section 3.1.
Scan rate and lock-in amplifier time constant were chosen with a trade-off between mechani-
cal drift and signal-to-noise ratio. Presently, the system scans a 50 urn core in 5 minutes
with a detection time constant of 0.13 s. These issues are addressed more fully in sections
7.2 and 7.3.

7. PERFORMANCE OF NEAR-FIELD SYSTEM

7.1 Precision
Best precision is obtained when the measurement is repeated without disturbing any
parameter except refocus. Figure 7-1 shows the results of four repeated measurements on a

1.1 km length of graded index fiber D. The standard deviations for determining the full
widths at the 80, 50, 20, 10, and 3 percent intensity points are indicated along the ordi-
nate in the figure. A typical value is in the 0.3 percent range for the lower part of the
curve. It should be noted that the width of an ink line is approximately 0.2 percent of the
FWHM and therefore represents a limitation in determining actual precision. The main con-

clusion from figure 7-1 is that the system has no appreciable dc drift during a 10 minute
period.
In practice, new ends must be prepared and the system refocused. To determine this

precision, near-field measurements were repeated four times with new end preparation and
alignment for each measurement. Three fibers, two graded and one step, 0, E, and C, were

used for this test; results on E and C are given in figures 7-2, and 7-3. For fibers E and

C, standard deviations for determining full-widths on the lower parts of the curves range
from 0.8 to 1 percent. For fiber D (not shown), the corresponding values are slightly
higher, 1 to 2 percent. This precision data contains effects of core ellipticity since
fibers were remeasured without regard to angular orientation. A precision of about 1.5

percent would apply to a core diameter measurement based on a determination of the near-
field full width close to the baseline on a short length of fiber.

7.2 Resolution and Accuracy

When all light emitted by the fiber is collected by the objective, the near-field
resolution is determined by fiber numerical aperture. This is a consequence of the expres-
sion for the resolution, Y, of a microscope objective with

Y ~M* . (8)
NA

Where x is the wavelength and NA refers to the collected numerical aperture [26]. At a \ of
0.85 un and an NA of 0.2, Y is 2.6 ym.

23
 —

Q >
o>

(1)

•r- S_ <C
|
i re
E -O
C 3
re E
-o -.

+->
CO in X
O" <T3

c d)E
re c

s_ .-o
re E
CO
.

to
i- c -c
+J CO <C

re ec
s- z: O •

Q-ro »
C\J cv
o
coo • •
•

c » * c
CD E LO -r-
O
x: -^
3
CD
^O •
Q-
CT3 •> +->
O CC "3
•i-
to
ID
O .E
+J
•

O -C 4-J 4->
0) !-> re re
S- CT
Q- C
co x:
00 -(-> +->
+-> r- re t:
c •-
00 E .- 2
E -* T3
1- r—
00
5- r- O i—
3 3 •

(/) CTM-
i

n3 C
CO " O 4- •

E X r— O
re 00
-O T3 co
i— C "O CT.
<D -i- co re
+-> +->
4- -o re c
I CO o oj
S- T3 •i- o
fO <C -o s-
CO S- C 00
Z CT •r- Q-

CO
S-
3
CT.

24
 O 4-
*+- QJ O
C
p on
+ia'
E O •

D1ET3
13
OJ <jr
c — p

S_ CO

T3 *
O
C
TO ~Zi
< CV r>

T3
C
co
cv
OJ rs

-M
O U"
•

3
Q- _C
* o
+-> +-> «*

3 croc
O c •

a i
QJ o
a> -M
> E TO
to
QJ LT) QJ
i +J
<_> w> TO
QJ X <—
C
i_ CD •

TD T3
.c: C S_
+-> •r- o
3 T3 CJl
<U c
c X3 o
o TO 1

•<
s_ TO
l/l cr
r— T3
(_> * QJ
QJ UJ +->
S- TO
O- S_ U
p—
QJ •

+-> JD T3
C •«— C
•p—
aj U_
E
QJ C
S_ O
3 +-> •p—
1/1 c +J
TO QJ TO
QJ E r—
E <u >
S- QJ
T3 3 T3
»— </)
QJ TO T3 •
-i— QJ E S-
4- E Z3 TO

fo CTO T-EX
1 T3
S-
TO
ts 4-> TO
^QJ QJ E </>

C\J

P»»

QJ
S-

25
 O OJ M-
<4- C O
o
c +i O
D1C-C
—CO O CCO
•r-
i

<£
CD CO .—

oc -*r
E «-
CO >-. CM
CG .

DXlO
QJ VD "
+J •
3
CL+->
.cr
»-

O
+-> cr «
3 c CC
O 01
X3
i— O •

4J E-l-1
> -*: <o
co
<u c\J QJ
i— • +->
O i— n3
QJ
«•-
C
S_
X T3
-C QJ S-
4-J "O O
•i- C
S -r- CT
c
C Q- O
O
r— +J
OJ i—
CO

•i- T7
O - QJ
QJ <_> +->
S_ CO
O- S_ o
QJ •.-
+1J3TJ
C -c- C
QJ Ll_ •>-
E
QJ C
S- • O
3 4-> •>-
CO
CO
C +J
CO
QJ
OJ E -r-
E QJ >
S- QJ

—
i (/)
co T3
QJ
•i- QJ S-
l»- E CO
I T3
i- .£= C •

CO O CO
QJ co -|->

2: QJ 00

26
 In a more detailed treatment where mode amplitudes are summed, Adams et al .
, show that

near-field resolution is related to the number of propagating modes and, for a near para-

bolic fiber is approximately 2 a/V, where V is the normalized frequency [27]. V is given by

[9] as

V =ll a (NA) . (9)

X

Thus, the resolution becomes X/ttNA, which is about a factor of two less than eq (8). Theo-

retical limits therefore are approximately 2 un for a 0.2 NA graded index fiber at a wave-

length of 0.85 un.

Experimentally, upper limits to resolution were obtained by observing abrupt features
in near-field patterns. This includes index dips and core-cladding boundaries. Upper
limits are obtained since the actual shapes of the features are unknown. Figure 7-4(A)
shows the narrowest index dip, 2.5 m FWHM, observed with the system. Figure 7-4(B) is a

1.9 urn 10 to 90 percent edge response to a core cladding boundary in a step index fiber

after l.l km of propagation (leaky modes have died out). Both of these cases suqqest a

resolution near 2 urn consistent with theoretical expectations.

The scanned dimension must be calibrated to make an absolute measurement. At present,

this is accomplished with a microscope stage micrometer having rulings every 10 urn. These
rulings are illuminated from behind the stage substrate with light, at the actual measure-

ment wavelength, from a large core, large NA fiber. A typical calibration scan is shown in

figure 7-5. Here the imaged rulings transmit more light than the background and appear as a

series of peaks separated by 10 urn. A close inspection of figure 7-5 shows the peaks are
not exactly evenly spaced. We believe this is caused by uneven illumination across the
widths of the rulings. Such behavior can be observed by eye using visible light. Presently
a centrally located pair of lines spaced by 150 un is used to determine the calibration fac-
tor. In this way, the uncertainty in calibration due to uneven ruling illumination is

reduced so resolution and stage micrometer accuracy become the limiting factors.
Preliminary results have been obtained to verify the calibration of commercial stage
micrometers. In this measurement the spacings between rulings are compared to transmission
fringes from a Fabry Perot interferometer, figure 7-5. Rulings which have higher trans-
mission are located by a HeNe laser beam focused to 2 un. A motorized differential micro-
meter drives the stage and translation is measured by counting fringes. With an HeNe laser,
transmission fringes occur every X/2 or 0.3164 un. The Fabry Perot free spectral range is
large enough so cavity modes from a common HeNe alignment laser are not observed. Figure
7-7 shows a typical result; in this case ruling centers are separated by 316 fringes (near-
est integer fringe) or 99.98 urn. The average of 9 measurements over various ruling pairs
was 100.4 un with a standard deviation of 0.5 )jm--in reasonable agreement with the stated

spacing of 100 un. Further improvement would be expected by focusing the HeNe beam to a

smaller spot and using a stage micrometer with sub-un ruling widths.

27
 \- 90%

1.9 MICROMETERS

CORE CLADDING

U 10%

(B)

(A)

Figure 7-4. Experimental upper limits to near-field resolution at 860 nm. (A) narrowest
on-axis index dip observed, (B) in to 90 percent edqe response to core cladding
boundary on Fiber C.

28
 3£

CL
O
O
(/I

O
s-
c_>

CT1

i*-^

a;

_!?=-

s_
o
4-
i/l
^/ U
n3 CD
CD +J
Cl a.'

F
C o
O i-
•»—
u
4-> • r-
re F
l
_o C1J
r- rr
1 ro

^-

29
 .

HeNe

DIFFERENTIAL *-
MICROMETER
STAGE
s ,'\ MICROMETER
FABRY
PEROT

MOTOR
DRIVE
DET y —A— HeNe

Figure 7-6. Apparatus to reference stage micrometer ruling spacings to transmission fringes
from a Fabry Perot interferometer.

FABRV PEROT FRINGES

nmrinrniwmiviiiimiii'.iiiMiiimirmiimffliii iiiiiiiiiimiiiimtiiwimiiimiifmmmwmmTOmmimvmi^^

-316

Figure 7-7. A comparison of ruling spacing (100 un nominal) to counted fringes

(316 x 0.3164 = 99.98 un) vjtti

30
 7.3 Dynamic Ranqe
The source was attenuated until noise was observed on the near-field pattern from a 2 m

length of graded index fiber, figure 7-8. In this instance the source (850 nm wavelength,

80 nm linewidth) was attenuated by 22 dB, thus indicating a signal-to-noise ratio of 39 dB

with an unattenuated source. Figure 7-8 with a signal-to-noise ratio of 17 dB represents
the lowest signal-to-noise ratio for extracting useful information; thus, fiber losses of up
to 22 dB can be accomodated by the system.

Further improvements in signal-to-noise ratio could be achieved by increased source

linewidth or lock-in amplifier time constant. Figure 7-8 was scanned in about 10 minutes,
substantially longer scan times would cause the system to drift out of focus. Mechanical
drift is minimized by mounting all components securely to a 5 cm thick aluminum plate. With
a 10 minute total scan time, the lock-in amplifier time constant could be increased about a

factor of five over figure 7-8 to pick up an additional 4 dB in signal-to-noise ratio.

8. TYPICAL RESULTS FROM NEAR-FIELD MEASUREMENTS

8.1 Cladding Light
Signal-to-noise ratio is sufficient to observe light scattered into the cladding due to
stress on the fiber, incomplete mode stripping, etc. Figure 8-1 shows a scan over an entire
fiber diameter to detect light in the cladding; the cladding-air boundary is clearly shown.
Cladding light can be measured with a signal-to-noise ratio consistent with section 7.3.

8.2 Step Index Fibers

Several examples have already been given in this Technical Note of near fields from
graded index fibers. In contrast, short step index fibers are greatly affected by the
presence of leaky modes. Figure 8-2 is the superposition of near-field patterns from the
same step index fiber at a length of 1 m and 1.1 km. In 1.1 km, the leaky modes have been
attenuated so the near field closely resembles the index profile of a step fiber.

8.3 Symmetry of Radiation Patterns

From the very nature of the preform fabrication process, fibers should exhibit good
symmetry across a diameter. However, at times, cores can be out of round and not neces-
sarily concentric with the outside cladding boundary. Figure 8-3 shows typical near-field
measurements from fibers. To illustrate radial symmetry the patterns have been folded about
a vertical axis passing through the midpoint of the width at half maximum. As shown, the
deviations Are slight and lines nearly overlap. The more asymmetric behavior that has been
observed has usually occurred within the index dip on-axis.

8.4 Index Dip Pathologies

Fibers made by chemical vapor deposition processes generally show an on-axis dip in the
index profile. This happens when the preform is either consolidated or collapsed and dop-
ants evaporate out of the axial region [9]. Attempts have been made to compensate this pro-
cess, giving rise to index peaks in cases of overcompensation [28]. This section is

31
 30 30
RADIAL POSITION, MICROMETERS

Figure 7-8. Near-field from a 2 m length of graded index fiber with source attenuated by
22 dB. Signal-to-noise ratio of this figure is 17 dB.

32
 60^ lio
RADIAL POSITION, MICROMETERS

Figure 8-1. Near-field scan across a fiber diameter including the core and cladding regions
to observe light in the cladding region.

33
 +J
s_
o
.c
l/l

re
c:
E O)
O CD
s- JO
CD
c >
CD -C
+->
-4-> in
13 CD
Q.T3
o
-O E
"cd >>
i- _*:
H_ re
I CD
i- i—
re
CD 4_>
c l^
o
CD E

CD
cO s-
CD
E -t->

J*
re
a> •
i— s-
CD

O f-

O X
CD CD

U- C
re -r-

CD Q.
jn CD
+-> 4->
en
u- -o
o m- a>
O +J
cd re
I— J= 3
a_+-> c
E CD CD
re c +->
X —CD +J
ll'i re

34
 (/) OJ
" l—
(0
Q.
CD E
c o
•T— o
'

"CJ
o
4->
CD
+J _c
J-J +J
o T3
— T3 r—
3
c E
S- 5
0) E
4-J i—
J-J X
<rs (O
Q_ E
u- 4-
O
rO
+J -d
s_
ra 4-
Cl o
+-> +->
4- c
0) r—
_J O
1

</l r— *^-N
c E 01
V c
JC r—
OJ
-M CJ r—
+J 3
Q.
fO O T3
s_ r-
_£T r—
•o
«
+-> o
l/l
CD ts>

4- X c
I r« 5.
i- CD
n) r— 4->
0) 03 +->
c o T3
^« Q.
4- +J
O l_ 4-
CD O
>i >
J- +-)
-t-J ro i_
OJ fO
E +j Q.
3
i, o +->
on _c sz
fO en
(O T3 s_
O CD
•i "D _c
Q. i (->
>) O 1—
1— 4- 5

35
 -o
1/1

o
cu
x:
+->

X!
CD

li-

re

ai
c
re
+->
.a
O

oc •

<s>
•r- <S>
CD
(/) O
•i- O
X S-
Q-
re
I

c c
o o
i— re
fO T3
to x
o
re
CD
4- 1/1

O "5
-C
CU Q-
D- S_
E o
re a.
x re
uj >

36
 -Q

T3
CD

er.

o
u-
-o
c
+->

en

o
cu

Q-
E
X
QJ

LT)
r

CO
!U
s-
ZS
CD
LL

37
intended to show the diverse behavior of index dips found in commercially available graded
index fibers.

The outside chemical vapor deposition process qenerally gives smaller dips than the
inside process. Figure 8-4 is an example of a fiber made by an outside process and exhibits
the smallest dip we have observed.
Figure 8-5 shows the largest dip observed with the system to date. The depth is about
half of the peak height of the curve. In many of the index dips measured, the widths are
close to the system resolution. Consequently, the dips are deeper than the near-field mea-
surements indicate. Other methods for determining index profile such as refracted near
field offer better resolution for these kinds of measurement.

9. CONCLUSION
The determination of radiation angle (numerical aperture) from far-field data at the 5

percent intensity points should result in good measurement precision and accuracy when
applied to near-parabolic, index fibers. In these fibers, leaky modes are contained within
the meridional ly defined numerical aperture so they should not affect the intensity at the
largest observed angles. However, in step index fibers the situation is different. Here
leaky modes cause uncertainty if meridional ly defined numerical aperture is desired. At

present, standards groups have not made recommendations for step index fibers.
More work is needed if core diameter is to be determined from a near-field measure-
ment. Correlation between the near field and index profile at the core-claddinq boundary
must be established. The use of a low index "barrier layer" may cause differences between
core diameter determined by near-field and index profile measurements. By judiciously
choosing definitions, it should be possible to reduce systematic errors between these two
measurements. Technically, near-field measurements are much easier to impliment than index
profile measurements which, for the most part, rely on interferometry over small dimensions.

10. REFERENCES

[1] Cathey, W. T. Optical information processing and holoqraphy. New York: John Wiley;
1974.

[2] Sladen, F. M. E., et al . Determination of optical fiber refractive index profiles by

a near-field scanning technique. A.P.L. 28(5) :255-258: 1976 March.

[3] Adams, M. J., et al . Leaky rays on optical fibers of arbitrary (circularly symmetric)
index profiles. Electronics Letters, 11(11) :239-240; 1975 May.
[4] Holmes, G. T. Propagation parameter measurements of optical waveguides. Proceedings
of the SPIE, Fiber Optics for Communication and Control, Vol. 224: 1980.

[5] Kitayama, K., et al . Impulse response prediction based on experimental mode couplinq
coefficient in a 10-km long graded-index fiber. IEEE J. Ouant. Elect.
QE-16(3):356-362; 1980 March.

38
 [6] Procedures for determining radiation angle (numerical aperture) and core diameter from

far and near-field patterns respectively are under consideration by Committee P6.6 of

the Electronic Industries Association.

[7] Hanson, A. G., et al . Optical waveguide communications glossary. NTIA Special

publication, NTIA-SP-79-4; 1979 September.

[8] Electronics Industries Association Committee P6.6 is currently recommending that
radiation angle be based on the far-field pattern half width at the 5 percent
intensity points.
[9] Miller, S. E.; Chynoweth, A. G. Optical Fiber Telecommunications. New York:Academic
Press; 1979.

[10] Matsumura, H. The light acceptance angle of a graded index fibre. Optical and

Ouantum Electronics 7:81-86; 1975.

[11] Eriksrud, M., et al . Comparison between measured and predicted transmission
characteristics of 12 km spliced graded-index fibres. Optical and Ouantum

Electronics. 11:517-523; 1979.

[12] Tateda, M. Optical loss measurements in graded-index fiber using a dummy fiber.
Applied Optics. 18(19) :3272-3275: 1979 October.

[13] Cherin, A. H.; Gardner, W. B. Standardization of optical fiber transmission

measurements. Laser Focus. 16(8):60-65; 1980 August.

[14] Adams, M. J., et al . Length-dependent effects due to leaky modes on multimode graded-
index optical fibres. Optics Communications. 17(2) :204-209; 1976 May.

[15] Petermann, K. Uncertainties of the leaky mode correction for near-square-law optical
fibres. Electronics Letters. 13(17) :513-514; 1977 August.

[16] Barrell, K. F.; Pask, C. Leaky ray correction factors for elliptical multimode
fibres. Electronics Letters. 16(14) :532-533; 1980 July.

[17] Hazan, J. P. Intensity profile distortion due to resolution limitation in fibre index
profile determination by near field. Electronics Letters. 14(5) : 158-160; 1978 March.

[18] Sabine, P. V. H., et al . Fibre refractive-index profiling by modified near-field

scanning. Electronics Letters. 16(23) :882-883; 1980 November.

[19] Sladen, F. M. E., et al . Measurement of profile dispersion in optical fibers: a

direct technique. Electronics Letters. 13(7) :212-213; 1977 February.

[20] PIN photodiodes with built-in operational amplifiers are available from a number of
sources. A 22 megohm feedback resistor is used with the operational amplifier.
[21] The alphabetical fiber designations are consistent within this Technical Note but do
not apply to any other NBS publications.
"22] Gloge, D.; Marcatili, E. A. J. Multimode theory of graded-core fibers. B. S.T.J.
52(9):1563-1578; 1973 November.
.23] Tokuda, M. , et al . Measurement of baseband frequency response of multimode fibre by
usinq a new type of mode scrambler. Electronic Letters. 13( 5) : 146-147; 1977 March.

[24] Franzen, D. L.; Day, G. W. Measurement of optical fiber bandwidth in the time
domain. Nat. Bur. Stand. (U.S.) Tech. Note 1019; 1980 February.

39
[25] Conduit, A. J., et al . Spectral and length-dependent losses in optical fibres
investigated by a two-channel backscatter technique. Electronics Letters. 16(3): 77-
78; 1980 January.

[26] Born, M.; Wolf, E. Principles of Optics. New YorkrPergamon Press; 1964.

[27] Adams, M. .]., et al . Resolution limit of the near-field scanning technique.

Proceedings of the Third European Conf. Opt. Comm.; 1977; Munich.
[28] Eickhoff, W.; Krumpholtz, 0. Pulse response and microbending losses of an optical
glass-fibre waveguide with a bump in its index profile. Electronics Letters.
13(9):256-257; 1977 April.

40
JBS-114A ' REV. 2-8C)

U.S. DEPT. OF COMM. 1. PUBLICATION OR 2. Performing Organ. Report No 3. Publication Date.

REPORT NO.
BIBLIOGRAPHIC DATA
SHEET (See instructions) NBS TN-1032 March 19>i
4. TITLE AND SUBTITLE
Measurement of Far-Field and Near-Field Radiation Patterns from
Optical Fiber s

5. AUTHOR(S)
Ernest M. Kim and Douglas L. Franzen
,'. PERFORMING ORGANIZATION (If joint or other than NBS, see instructions) 7. Contract/Grant No.

NATIONAL BUREAU OF STANDARDS

DEPARTMENT OF COMMERCE 8. Type of Report & Period Covered
WASHINGTON, D.C. 20234

). SPONSORING ORGANIZATION NAME AND COMPLETE ADDRESS (Street. City, State, ZIP)

Supported in part by the Department of Defense, Calibration

Coordination Group (CCG).

0. SUPPLEMENTARY NOTES

Document describes a computer program; SF-185, FlPS Software Summary, is attached.

1. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a si gnificant
bibliography or literature survey, mention it here)

Systems are described for measuring the far- and near-field radiation
patterns from optical fibers. Parameters which affect measurement precision,
accuracy, resolution, and signal-to-noise ratio are discussed. Measurements
utilizing radiation patterns are covered; this includes radiation angle
(numerical aperture), attenuation using mode filters, index profile, core
diameter, and mode volume transfer function. Experimental examples are given in
many instances.

' KEY WORDS (Six to twelve entries; alphabetical order; capitalize only proper names; and separate key words by semicolons)

Attenuation; core diameter; far field; index profile; mode filter;

numerical aperture; radiation angle; radiation patterns.
5. AVAILABILITY 14. NO.OF
PRINTED PAGES
X Unl imi ted
For Official Distribution. Do Not Release NTIS
to
48
X Order From Superintendent of Documents, U.S. Government Printing Office, Washington, D.C.
20402. 15. Price

Order From National Technical Information Service (NTIS), Springfield, VA. 22161 $2. 50
 .

NBS TECHNICAL PUBLICATIONS

PERIODICALS NOTE: The principal publication outlet for the foregoing data is

the Journal of Physical and Chemical Reference Data (JPCRD)

JOURNAL OF RESEARCH— The Journal of Research of the published quarterly for NBS by the American Chemical Society
National Bureau of Standards reports NBS research and develop- (ACS) and the American Institute of Physics (AIP). Subscriptions,
ment in those disciplines of the physical and engineering sciences in reprints, and supplements available from ACS, I 155 Sixteenth St.,
which the Bureau is active. These include physics, chemistry, NW, Washington, DC 20056.
engineering, mathematics, and computer sciences. Papers cover a
broad range of subjects, with major emphasis on measurement Building Science Series —
Disseminates technical- information
methodology and the basic technology underlying standardization. developed at the Bureau on building materials, components,
Also included from time to time are survey articles on topics systems, and whole structures. The series presents research results,
closeiy related to the Bureau's technical and scientific programs. test methods, and performance criteria related to the structural and

As a special service to subscribers each issue contains complete environmental functions and the durability and safety charac-
citations to all recent Bureau publications in both NBS and non- teristics of building elements and systems.

NBS media. Issued six times a year. Annual subscription: domestic

$13: foreign $16.25. Single copy. $3 domestic: $3 75 foreign.
Technical Notes — Studies or reports which are complete in them-
selves but restrictive in their treatment of a subject. Analogous to
NOTE: The journal was formerly published in two sections: Sec- monographs but not so comprehensive in scope or definitive in
tion A "Physics and Chemistry" and Section B "Mathematical
treatment of the subject area. Often serve as a vehicle for final
Sciences."
reports of work performed at NBS under the sponsorship of other
DIMENSIONS/NBS — This monthly magazine is published to in- government agencies.
form and industry leaders, leachers,
scientists, engineers, business
students, and consumers of the latest advances in science and Voluntary Product Standards — Developed under procedures
technology, with primary emphasis on work at N BS. The magazine published by the Department of Commerce in Part 10, Title 15, of
highlights and reviews such issues as energy research, fire protec- the Code of Federal Regulations. The standards establish
tion, building technology, metric conversion, pollution
abatement, nationally recognized requirements for products, and provide all

health and safety, and consumer product performance. In addi- concerned interests with a basis common
understanding of the
for

tion, it reports the results of Bureau programs in measurement characteristics of the products. NBS administers this program as a
standards and techniques, properties of matter and materials, supplement to the activities of the private sector standardizing
engineering standards and services, instrumentation, and organizations.
automatic data processing. Annual subscription: domestic $11;
foreign $13.75.
—
Consumer Information Series Practical information, based on
NBS research and experience, covering areas of interest to the con-
NONPERIODICALS sumer. Easily understandable language and illustrations provide
useful background knowledge for shopping in today's tech-
Monographs —
Major contributions to the technical literature on nological marketplace.
various subjects related to the Bureau's scientific and technical ac-
tivities.
Order the above NBS publications from: Superintendent of Docu-
ments, Government Printing Office, Washington, 20402. DC
Handbooks — Recommended codes of engineering and industrial
Order the following NBS publications — FIPS and NBSIR's —from
practice (including safety codes) developed in cooperation with in-
the National Technical Information Services, Springfield, VA 22161
terested industries, professional organizations, and regulatory
bodies. Federal Information Processing Standards Publications (FIPS
Special Publications — Include
proceedings of conferences spon- PUB) — Publications in this series collectively constitute the
sored by NBS, NBS annual reports, and other special publications Federal Information Processing Standards Register The Register
appropriate to this grouping such as wall charts, pocket cards, and Govern-
serves as the official source of information in the Federal
bibliographies. ment regarding standards issued by NBS pursuant to the Federal
Applied Mathematics Series — Mathematical tables, manuals, and Property and Administrative Services Act of 1949 as amended.
studies of special interest to physicists, engineers, chemists, Public Law 89-306 (79 Stat. 1127), and as implemented by Ex-
biologists, mathematicians, computer programmers, and others ecutive Order 11717(38 FR 12315, dated May II, 1973) and Part 6
engaged in scientific and technical work. of Title 15 CFR (Code of Federal Regulations).

National Standard Reference Data Series— Provides quantitative NBS Interagency Reports (NBSIR) —
A special series of interim or
data on the physical and chemical properties of materials, com- final reports on work performed by NBS for outside sponsors
piled from the world's literature and critically evaluated. (both government and non-government). In general, initial dis-
Developed under a worldwide program coordinated by NBS under tribution handled by the sponsor; public distribution is by the
is

the authority of the National Standard Data Act (Public Law National Technical Information Services, Springfield, VA 22161,
90-396). in paper copy or microfiche form.
U.S. DEPARTMENT OF COMMERCE
National Bureau of Standards
Washington. D.C. 20234
POSTAGE AND FEES PAID
US. DEPARTMENT OF COMMERCE
OFFICIAL BUSINESS
COM-21 5
Penalty for Private Use, $300

SPECIAL FOURTH-CLASS RATE

BOOK