812 Vol. 57, No.

4 / 1 February 2018 / Applied Optics Engineering and Laboratory Note

Practical splicing of poly-methyl-methacrylate

plastic optical fibers
MICHELE C. ZANON,1,2 VINICIUS N. H. SILVA,1 ANDRÉS P. L. BARBERO,1 AND RICARDO M. RIBEIRO1,*
1
Laboratório de Comunicações Ópticas, Departamento de Engenharia de Telecomunicações, Universidade Federal Fluminense,
24.210-240, Niterói, Rio de Janeiro, Brazil
2
e-mail: michelezanon@gmail.com
*Corresponding author: rmr@pq.cnpq.br

Received 25 September 2017; revised 16 December 2017; accepted 27 December 2017; posted 2 January 2018 (Doc. ID 302855);
published 31 January 2018

This Engineering and Laboratory Note describes a simple, fast, cost-effective, and practical splicing technique for
poly-methyl-methacrylate (PMMA) plastic optical fibers. We believe this technique can be useful for home net-
works, laboratory research, and educational uses. It employs a widely available, low-cost transparent adhesive com-
bined with a suitable alignment sleeve. The fiber tips did not need to be cleaned or polished, and the polyethylene
jacket did not need to be removed. The alignment between the cleaved fibers was performed by means of the jacket.
Low insertion losses were measured for the main PMMA-based fiber types. © 2018 Optical Society of America
OCIS codes: (060.2300) Fiber measurements; (060.2310) Fiber optics; (160.2290) Fiber materials; (160.5470) Polymers; (220.4880)
Optomechanics.

https://doi.org/10.1364/AO.57.000812

1. INTRODUCTION The splicing technology for silica fibers is well developed.

Fiber splicing techniques have been developed since the 1970s
and may be divided into two broad categories: fusion splicing
and mechanical splicing. Both techniques require that the fibers
have smooth end faces obtained by proper termination—i.e.,
cleaving or polishing. Fusion splicing is performed by aligning
two terminated fiber ends and applying localized heating
(e.g., by means of an electrical arc). Mechanical splicing
requires that the two terminated fiber ends be held in align-
ment by some mechanical device, such as capillary sleeves or
V-grooves. A transparent adhesive or an index-matching gel
is then applied to fill the gap between the fiber ends [1].
Poly-methyl-methacrylate (PMMA)-based plastic optical
fibers (POFs) with the fiber and polyethylene cable diameters
of 1 mm and 2.2 mm, respectively, are the most commonly
manufactured and used plastic fibers around the world, espe-
cially for data communications [2,3]. Such POFs are practical
in the visible spectrum only [2,3].
The high optical attenuation of PMMA-POFs is balanced by
their sole use for short links up to 500 m, most commonly to
100 m without repetition [2,3]. As a result, the technical benefits
of POFs are generally not at all related to optical attenuation
because their attenuation is so very much higher than glass fibers.
Instead, their benefits are related to their lighter weight and
greater flexibility—which allows for a tighter bend radius and
therefore more compact deployment—and the fact that less skill
is required to perform terminations and interconnections.
POFs usually use much simpler tools and handling procedures,
but the available connection and splice techniques do not result
in the same quality. In addition, POFs are suitable for home
networks and a sophisticated splicing technique is not reason-
able [2,3]. Mechanical splicing has resulted in insertion losses
(ILs) down to 0.5 dB that could be reached after careful
cleaving and polishing of POF ends [2–5].
This paper reports a simple, fast, cost-effective, and practical
mechanical splicing technique for the main PMMA-based POF
types, by using a widely available transparent adhesive. It de-
scribes an unusual approach in which the manufactured sleeves
provide the passive alignment of the POFs by means of their
jacket and not their cladding or core as is usually adopted for
silica fibers. A new user, who learned the techniques in a
few minutes, purposefully performed the splices and measured
their IL for the first time. Low IL values were measured, and the
not-yet-reported result for the graded-index (GI) POFs is also
presented.
2. EXPERIMENTAL
A. Overview
The technique described in this section provides a good balance
between simplicity and efficiency. It is simple because it uses an
easily controlled cure adhesive that is widely available on the
market and an easily manufactured sleeve that allows passive
alignment of the POFs by means of their jacket, without

1559-128X/18/040812-05 Journal © 2018 Optical Society of America

 Engineering and Laboratory Note Vol. 57, No. 4 / 1 February 2018 / Applied Optics 813

polishing. The technique is efficient because it results in robust

mechanical splices exhibiting low ILs even when performed by a
new user. Section 2.B describes the adhesives used. Section 2.C
describes the sleeves and the passive alignment. Section 2.D ex-
(a) (b)
plains the joining of the adhesives and the POFs with the sleeves,
thus describing the procedures to build a permanent mechanical
splice. Finally, Section 2.E describes the experimental apparatus
used to measure each of the mechanical splices, mounted accord-
ing to the explanations given in Section 2.D.
B. Adhesives
Loctite 454 gel adhesive (−40∕!120°C) made from cyanoacry- (c) (d)
late and mixed with an unspecified fraction of rubber was used
for splicing [6]. It is widely available at a low price (∼US$2/2 g
blister), and the cure is almost fully complete in ∼2 min. This
allows for an easy manual fit between the POF ends inside
the sleeves while the curing proceeds. As a result, a low IL is
achievable. The refractive index is assumed as n " 1.49 [7],
which perfectly matches with the widely accepted 1.492 for (e) (f)
the step-index (SI) and double step-index (DSI) PMMA-
POF cores [2,3]. All the tested cabled-POFs and POFs present Fig. 1. Mounting procedure of the POF splices.
diameters of 2.2 mm and 1 mm, respectively.
The well-known cyanoacrylate Loctite 407 is a fast-cure
adhesive [8] that was tested to splice standard PMMA-POFs,
but with poor results. The cure time of this adhesive is known
to be very fast, and it was difficult to keep the longitudinal align-
ment between the POF ends. As a result, the expansion due the
polymerization caused a significant longitudinal misalignment
between the fibers. The measured insertion losses of a few deci-
bels (dBs) were very high, unacceptable for practical purposes.
The Epotek-301 (Epoxi Technology Inc.) adhesive was also Fig. 2. Experimental setup to measure the IL of each mounted POF
tested, again with poor results. The cure of such adhesive is fin- splice as illustrated in Fig. 1.
ished after 24 h and longitudinal misalignment between the
POFs was observed after the completion of the polymerization.
C. Sleeves and the Passive Alignment The PUT was then cleaved [Fig. 1(b)]. Two segments of
Passive alignment without the need for polyethylene-jacket re- jacketed POF to be spliced were produced: POF1 and
moval was made possible using aluminum cylindrical sleeves of POF2 . The two cleaved ends were not polished.
15 mm length and 6 mm diameter, each concentrically drilled The POF1 was pulled and fitted, ensuring the end remained
by a 2.2 mm diameter driller. The tolerance is estimated to be near the middle of the sleeve [Fig. 1(c)]. A drop of fast-cure
∼0.1 mm because the polyethylene jacket is not too hard. cyanoacrylate Loctite 407 [8] adhesive was applied around
Therefore, because the aluminum cylindrical tube can be the jacketed POF [Fig. 1(c)]. A small drop of Loctite 454 [6]
drilled with a 2.2 mm diameter driller to produce the sleeve, gel adhesive was applied on the cleaved end of the POF2 . No
and it is easy to introduce the 2.2 mm cabled-POF inside the crimping was required, and consequently, no additional loss
cavity sleeve. was introduced.
The jacketed POF2 was carefully introduced into the sleeve
D. Procedures to Build the Permanent Splices until it touched the jacketed POF1 [Fig. 1(d)]. The POF2 was
Figure 1 illustrates the procedures used to build a permanent manually pushed and pressed against the POF1 for ∼2 min,
mechanical splice in a “POF under test” (PUT). The latter was forming a thick glue film between them.
kept inserted in the apparatus for IL measurement (see Fig. 2). A drop of fast-cure cyanoacrylate Loctite 407 [8] adhesive
After each step illustrated in Fig. 1 was completed, the output was applied to the jacketed POF2 so that it could be glued to
optical power was measured by using an optical power meter the internal wall of the sleeve [Fig. 1(e)]. In order to obtain an
(OPM) according to the apparatus outlined in Fig. 2. In other even stronger adhesion, additional high-strength glue, such
words, the experimental setup of Fig. 2 was built and the as the two-component epoxy Araldite Standard [9], may be
procedures to perform the mechanical splices as depicted in applied, as shown in Fig. 1(f ).
Fig. 1 were implemented on the PUT.
The PUT was introduced into the sleeve [Fig. 1(a)], passing E. Measurement of the ILs
through a cleaver (Fiber Optic Cutting Block—Edmund Figure 2 shows the experimental setup to measure the IL of each
Optics) with fresh blades [5]. The sleeve fit tightly around mounted POF splice. The light source was a LED emitting
the 2.2 mm jacketed POFs. at 520 nm. The IL depends upon the mode distribution for
 814 Vol. 57, No. 4 / 1 February 2018 / Applied Optics
multimode fiber and thus on the coupling conditions [2,3].
Therefore, the light traverses a standard JIS6863 mode
scrambler before launching the PUT. Finally, the POF pigtail
is connected to an OPM.
The P ref 1 reference power level was then measured. The
PUT was cleaved, and both ends were placed into the sleeve
without the adhesive. The P d 1 power level of transmitted light
was measured (the “d ” stands for “dry connection”). The splic-
ing procedure was performed (as explained in Fig. 1), and the
P g1 power level of transmitted light was measured (the “g”
stands for “glued splicing”). During the Loctite 454 [6] adhe-
sive curing process, the POF2 was pushed and pressed against
the POF1 while the transmitted optical power was monitored.
Hereafter, the POF segment containing the mounted splice was
cut off and discarded. The remaining PUT was reconnected to
the OPM (see Fig. 2). A new P ref 2 reference power level was
measured. The remaining procedure to obtain the P d 2 and P g2
was exactly the same as the procedure described previously. The
quantities P d 2 and P g2 refer to a second splice; that is, they refer
to the repeated IL measurement but for a second built splice.
This procedure was repeated up to the tenth splice for each
of the PMMA POF types. Because 10 splice samples were
prepared for each POF type, there were a total of 120
measurements.
3. RESULTS AND DISCUSSION
The hILd i and hILg i, the average (dB) of the “dry connection”
and the “glued splicing,” respectively, may be calculated from
Eq. (1):
!" #$
P d ;g
ILd ;g " 10 log : (1)
P ref
Table 1 shows the average and standard deviation calcula-
tions of the ILd and ILg measurements for each POF type. NA
stands for numerical aperture.
The hILg iSI " 0.4 dB is comparable or even better than the
previously reported results—e.g., the 0.6–1.6 dB IL range
specified by several manufacturers of polished connectors
[2–5,10–14]. Moreover, the technique was proven to be
“do-it-yourself ” by a new user in the field who learned the tech-
nique in a few minutes. The hILg iSI " 0.4 dB can be accepted
as a good result. The 0.5 standard deviation is high when com-
pared with the hILd iSI " 1.4 dB. This is because each cleavage
Table 1. Average and Standard Deviation Calculations of
ILd and ILg for Each of the PMMA-POF Typea
PMMA-Based POF ILd (dB)
Standard high-NA-SI (Toray 1.4 $ 0.5
PFU-CD1001-22-E1 1 mm
fiber, 2.2 mm cable)
Low-NA-DSI (Toray PMU- 1.9 $ 0.5
CD1002-22-E1 1 mm fiber,
2.2 mm cable)
GI (Optimedia OMJ-Giga/ 2.2 $ 0.6
FF-GI-SE100 1 mm fiber,
2.2 mm cable)
a
Each measurement was calculated after averaging over 10 splice samples.
Engineering and Laboratory Note
does not allow for good control of the POF surface, and the
resultant scattering varies remarkably. When the Loctite 454
adhesive [6] was applied, the standard deviation was signifi-
cantly reduced to 0.1 dB. Such adhesive [6] eliminated the
air gap, thus matching the refractive index between the
POFs. It also filled the micro-voids of the surfaces, which
are not perfectly flat.
As expected, both the hILd iDSI " 1.9 dB and hILg iDSI "
0.6 dB values are higher than what was achieved for the
high-NA-SI-POF. The standard deviations are comparable
(see Table 1). The slight increase from hILg iSI " 0.4 dB to
hILg iDSI " 0.6 dB may be due to the more complex index pro-
file of low-NA-DSI-POFs. This hinders the alignment between
the POFs during the adhesive [6] curing. The ILg;DSI " #0.6 $
0.1% dB is also a good result, as it can be compared with the
#0.61 $ 0.18% dB [15] for low-NA-DSI-POFs connections
that are performed using polishing.
The GI-PMMA-POFs exhibit higher attenuation and ∼15×
more bandwidth than the standard SI-POFs, but their other
characteristics are kept [2]. The 1 mm core size of PMMA-
GI-POFs makes them easier to handle for home networks com-
pared to perfluorinated-GI-POFs [16]. As shown in Table 1,
hILd iGI " 2.2 dB, hILg iGI " 0.8 dB, and standard deviations
are all greater than what was achieved for the low-NA-DSI-
POF. Because the GI-PMMA-POFs presents graded-index
profile and NA ∼ 0.2, there is an even greater sensitivity to
lateral alignment during the adhesive [6] curing, and greater
precision is required for splicing.
The 980 μm core diameter of the PMMA-POFs is signifi-
cantly higher than the 62.5 μm that is one of the main stan-
dards for multimode silica fibers. As a result, a deviation from
the POF-cable concentricity affects the 1 mm standard POFs
much less. Undoubtedly, if the jacket would be stripped and the
POFs themselves would be passively aligned by means of their
cladding when using a properly designed sleeve, a splice could
be achieved with an IL in the 0.2–0.3 dB range—but hardly
less than 0.2 dB [10] obtained for a single sample using the
adhesive formulation described therein [10]. However, this
would require using a tube with smaller inner and outer diam-
eters, which would be more challenging. Our technique, using
sleeves and the proposed adhesive, provides without difficulty
an IL " 0.4 dB for the splicing of standard PMMA-POFs.
PMMA-POF links are limited to ∼500 m, but in practice
lengths around 100 m are more common [2]. Most such links
contain no splices or connectorization. When needed, such
links may contain one or two splices but rarely more than
two, i.e., 0.8 dB maximum splicing insertion loss.
From the optical standpoint, the optical performance im-
ILg (dB)
provement of the “glued” splice when compared with the “dry”
0.4 $ 0.1 splice may be due to two reasons:
A—The refractive index matching between the PMMA-
0.6 $ 0.1
POF ends. Furthermore, the cured adhesive ensures more
robust axial and lateral alignment between the fibers.
B—PMMA-POFs are rheologically and mechanically dis-
0.8 $ 0.4 tinct from the silica fibers. This issue is addressed in [17], which
shows the modification of the rheological and mechanical prop-
erties of PMMA by the addition of inorganic fillers (SiO2 ). The
cleavage techniques for each type of fiber are also distinct [1,2].
 Engineering and Laboratory Note
The attainment of a somewhat rough surface after cleavage of a
PMMA-POF has been observed to scatter light [2,4,5,12–14].
On the opposite side, the cleavage of silica fibers produces
higher-quality surfaces [1] when compared with the PMMA-
POFs, thus reducing light scattering. As a result, when a prop-
erly transparent adhesive is applied between the butt-to-butt
contacted cleaved POF ends, the voids that arise due to the
surface roughness become filled. This drastically reduces, but
does not completely eliminate, the light scattering. The latter
is an important insertion loss mechanism of splicing between
PMMA-POFs [2,4,5,12–14].
Thus, as pointed out in “B” above, the light scattering con-
tributes to causing an IL > 0.2 dB irrespective of the type of
PMMA-POF. This is why the measured optical losses were
0.4 dB or more for standard PMMA-POF, depending on the
type of POF. Alignment is more difficult to obtain when joining
POFs with complex refractive index profiles (i.e., DSI or GI).
Furthermore, “A” and “B” can also explain the significant IL im-
provement (decrease) from the “dry-splice” to the “glued-splice.”
The measured results shown in Table 1 support these
conclusions.
It should be observed that the “tips” presented here may
simplify the accomplishment of PMMA-POF splices concern-
ing research and development and education activities in labs
and home networks for common end-users who are consumers
of broadband services (i.e., non-critical environments). The
POF-cable jackets are not rigid and might not be manufactured
with highly accurate dimensions and circularity. Despite such
imperfections, we have obtained good results that are accept-
able for the aforementioned non-critical situations. The non-
rigid POF-cable jacket is not a serious drawback as the sleeve
uses circular fixtures that uniformly tighten around the jacketed
PMMA-POF cable.
4. CONCLUSIONS
We have presented a practical, simple, secure, cost-effective,
fast and efficient, permanent, low-IL mechanical splicing tech-
nique for the main PMMA-based POFs, with “proof-of-
operator’s-skill” using Loctite 454 gel adhesive [6]. The POFs
to be spliced are aligned by their jacket.
The hILg iSI " 0.4 dB and hILg iDSI " 0.6 dB are compa-
rable or even smaller than those already reported using more
laborious techniques, such as careful polishing. The critical lat-
eral alignment may explain the greater but still reasonable
hILg iGI " 0.8 dB, which can be compared with those obtained
using carefully polished POF connectors. No published results
for measured IL of permanent splicing in GI-PMMA POFs
were found. The procedure presented here has shown good re-
sults despite omitting the use of polishing. Of course, a trained
operator may decide to use careful cleaning procedures, a pol-
ishing stage, and even more sophisticated techniques and tools
(e.g., microtome cut [2] or POF-press-cut techniques [18]). In
this way, IL < 0.3 dB is likely to be achieved.
The technique here presented was only applied to 1 mm/
2.2 mm PMMA-POFs, which are the most commonly manu-
factured and used around the world [2,3]. A suggestion for
future research is to test PMMA-POFs with smaller fiber/cable
diameters using the same described procedure.
Vol. 57, No. 4 / 1 February 2018 / Applied Optics 815
Our technique produces resilient splices for the intended
applications. It can be accomplished after the application of
a high-strength adhesive [9] inside the cavity in order to provide
greater adhesion between the polyethylene surface of the POF
cable and the inner surface of the sleeve. From a qualitative
perspective, we observed the mechanical strength of the splices
we performed in comparison with the connectorized POFs. As
an example, plastic connectors commonly used for POFs
deployed in home networks [2] are not very mechanically
robust; that is, they are poorly resistant to “a pull.” Concerning
the application of the presented technique for a POF link in
critical environments—e.g., industrial facilities—future re-
search would be required to quantitatively evaluate the splices
under longitudinal strain.
We believe the splicing technique reported here could be
useful at least in non-critical environments using POFs, such
as home networks, developments in laboratory and educational
settings. Furthermore, we have identified no drawbacks to ex-
tending our technique for PMMA-POFs with dimensions
larger than 1 mm/2.2 mm, commonly employed for imaging
and illumination, and sometimes also for data communications
and sensors.
Funding. Brazilian R&D Agencies (Faperj, CNPq/MCTI).
REFERENCES
1. J. M. Senior, Optical Fiber Communications: Principles and Practice,
1st ed. (Prentice-Hall, 1985), Chap. 4, pp. 156–163.
2. O. Ziemann, J. Krauser, P. E. Zamzow, and W. Daum, POF
Handbook: Optical Short Range Transmission System, 2nd ed.
(Springer-Verlag, 2008).
3. S. Abrate, R. Gaudino, and G. Perrone, “Step-index PMMA fibers
and their applications,” in Current Development in Optical Fiber
Technology, S. W. Harum and H. Arof, eds. (Intech,
2013).
4. T. Stäber, H. Poisel, and O. Ziemann, “POF connector prepara-
tion by polishing,” in 13th International Conference on Plastic
Optical Fibres, Nuremberg, Germany (27–30 September 2004),
pp. 552–558.
5. M. A. Losada, F. A. Domínguez-Chapman, J. Mateo, A. López, and J.
Zubia, “Influence of termination on connector loss for plastic optical
fibres,” in 16th IEEE International Conference on Transparent
Optical Networks, Graz, Austria (06–10 July 2014), paper Mo.C7.4.
6. Loctite, “Technical data sheet,” Loctite 454 (2012).
7. Dynatex, “Dynatex technical data sheet,” Dynatex Super Glue Gel
Part Number 49433, paper 011649433.
8. Loctite, “Technical data sheet,” Loctite 407 (2009).
9. http://www.aralditeadhesives.com.
10. S. D. Carson and R. A. Salazar, “Splicing plastic optical fibers,” Proc.
SPIE 1592, 134–138 (1991).
11. Y. Mizuno, S. Ohara, N. Hayashi, and K. Nakamura, “Ultrasonic
splicing of polymer optical fibres,” Electron. Lett. 50, 1384–1386
(2014).
12. A. Ankele and W. Langhoff, “Process for the fiber end facet termina-
tion of polymer optical fibers using halogen light,” in Eurocable
Conference, Stuttgart, Germany (5–7 June 2000), pp. 286–291.
13. M. V. P. Ghirghi, V. P. Minkovich, and A. G. Villegas, “Polymer optical
fiber termination with use of liquid nitrogen,” IEEE Photon. Technol.
Lett. 26, 516–519 (2014).
14. R. Oliveira, L. Bilro, and R. Nogueira, “Smooth end face termination of
microstructured, graded-index, and step-index polymer,” Appl. Opt.
54, 5629–5633 (2015).
15. R. Raman and J. Cirillo, “Termination studies on low NA plastic optical
fibres,” in 5th International Conference on Plastic Optical Fibres,
Paris, France (22–24 October 1996), pp. 70–77.
 816 Vol. 57, No. 4 / 1 February 2018 / Applied Optics
16. F. Forni, Y. Shi, H. P. A. van der Boom, E. Tangdiongga, and A. M. J.
Koonen, “Multiband 4G and gigabit/s baseband transmission over
large-core GI and SI POFs for in-home networks,” in 25th
International Conference on Plastic Optical Fibres, Birmingham, UK
(13–15 September 2016), paper OP15.
17. E. S. Silva, L. A. Ribeiro, M. Carolina, B. C. Nascimento, and E. N.
Ito, “Rheological and mechanical characterization of poly(methyl
Engineering and Laboratory Note
methacrylate)/silica (PMMA/SiO2) composites,” Mater. Res. 17,
926–932 (2014).
18. D. Moll and H. Poisel, “Polymer optical fiber termination—a never
ending story?” in 9th International Conference on Plastic Optical
Fibres, Post-Deadline Paper, Cambridge, MA (5–8 September
2000).