Fibers from Regenerated Collagen

Warren C. Schimpf and Ferdinand Rodriguez.

School of Chemical Engineering, Cornell University, lthaca, New York 14853

Fibers are formed from enzyme-treated, acid-soluble collagen by coagulation of extrudates in ammonium hy-
droxide followed by cross-linking with glutaraldehyde. A tensile strength of 60 MPa is attained when a 5 % colla-
gen solution (pH 3.1) is extruded at an average velocity of 11.8 cm/s through a die which is 0.89 mm in diameter.
Coagulation for I min in 0.2 N ammonium hydroxide is followed by cross-linking for 3 min in a mixture of 0.5%
glutaraldehyde, 9.5% water, and 90% ethanol. Sodium chloride was inferior to the ammonium hydroxide as a
coagulant as far as development of tensile strength was concerned. The effects of die diameter, coagulation,
and cross-linking times and concentrations are explored.

Introduction Testing Machine. This machine provides a constant rate of

Collagen is a biomaterial which constitutes 20 to 30% of
the total body protein in vertebrates. I t is found in high con-
centrations in bone, tendon, and skin (Rubin and Stenzel,
1968; Chvapil et al., 1973). Enzyme treatment of calfskin with
simultaneous acid solubilization yields a soluble collagen free
of the telopeptide appendages which are believed to cause
most of the immunological response when collagen from one
species is introduced into another. Acid-soluble collagen can
be extruded to form a fiber which is coagulated by an increase
in p H or by a high salt concentration. A major use of such fi-
bers could be as absorbable sutures in surgery. The ultimate
tensile strength (UTS) is only one important parameter. Knot
strength and rate of absorption in the body are equally critical.
However, the present work focuses on the optimization of
strength as a function of coagulation and cross-linking con-
ditions.
Because of equipment limitations, the fibers were not under
tension after leaving the extrusion die so that no drawing could
be accomplished. Even with this disadvantage, the effects of
coagulation and cross-linking conditions were marked enough
to act as guides to the production of relatively strong fibers.
Collagen can be cross-linked either chemically or by UV
light in order to increase strength. UV-induced cross-links can
be impractical on an industrial scale, however, since the
cross-links must be formed under a nitrogen atmosphere so
that the free radicals do not react with oxygen before partic-
ipating in the cross-linking reactions (Miyata et al., 1971). The
cross-links which are formed are also less stable than those
formed chemically (White et al., 1973).
Cross-links can be introduced chemically into the collagen
molecule by such reagents as glutaraldehyde, formaldehyde,
acrolein, and chromium salts. The aldehydes are better re-
agents for blood compatibility than the chromium salts,
however, and it has been found that glutaraldehyde introduces
more cross-links than the other reagents mentioned (White
et al., 1973; Stenzel e t al., 1974). Collagen cross-linked with
glutaraldehyde has shown increased resistance to bacterial
collagenase and the biological lifespan is likewise increased
as susceptibility to enzyme digestion is decreased. I t would
appear that glutaraldehyde is a suitable cross-linking agent
to use if the collagen fibers are to be used as sutures.
Experimental Procedures
The collagen used in this work was prepared using en-
zyme-acid solubilization of calf-hide (Hamed and Rodriguez,
1975; Takahashi, 1974). Initial concentration was about 5%
solids by weight and the p H was adjusted to 3.1 with hydro-
chloric acid.
The basic apparatus used in this project is the Instron
extrusion during formation of each fiber and is also used to
test each fiber for ultimate tensile strength. The extrusion and
testing is done in a room a t a constant temperature of 2 1 f 1
"C, and constant relative humidity of 55 f 5%.
A 10-ml plastic syringe, fitted with a short hypodermic
needle, is used to hold the collagen solution during extrusion.
The loaded syringe is driven by the cross-arm of the Instron
machine. The end of the needle is below the surface of the
coagulating solution in a 50-ml glass cylinder. After extrusion
and coagulation, the soft fiber is transferred to a glass tray and
cross-linked by solutions of glutaraldehyde in 90% ethanol.
The fibers are dried on plates of glass which are sprayed with
silicone oil to prevent sticking. When dry, the fibers are
measured and weighed so that the cross-sectional area can be
determined. They are stored a t room temperature in glass jars
containing distilled water for 24 h before being tested in a
wetted condition on the Instron machine to determine ulti-
mate tensile strength. The wet rather than the dry condition
is thought to be a reasonable approximation to conditions of
actual use.
With a gauge length of 75 mm, the fibers are wrapped
around the spindles (each 9.5 mm diameter) four times and
then fastened to the spindles with cellophane tape. The total
length of the fibers tested is approximately 38 cm. A jaw
separation rate of 51 mm/min was used for all tests.
Tensile strengths are reported in megapascals (MPa) based
on the dry cross section of the fibers. A density of 1.34 g/cmi
for collagen is used (Tanioka et al., 1973). In other common
units the strength is given by
g/denier
1 MPa = 145 psi = 0.846 X
Standard conditions are shown in Table I.
Results and Discussion
1. Extrusion Rate. Tensile strength increases with de-
creasing die diameter (Figure l). This is expected since a
smaller diameter a t constant extrusion rate increases the
orientation of the collagen fibrils along the axis of extrusion
and ultimate testing. There is, of course, a practical limit to
the smallest needle diameter which can be used since even-
tually the pressure drop through the needle will be greatly
intensified and the equipment will not bear the stress exerted
on it. The same is true of the allowable magnitude of extrusion
rate for a specific diameter needle. As extrusion rate increases,
pressure drop may increase beyond a practical limit or flow
instability may occur. The tensile strengths a t increasing ex-
trusion rates passed through a maximum and leveled off,
suggesting increased orientation a t first and then perhaps a
competing mechanism taking over as extrusion rate was in-

90 Ind. Eng. Chem., Prod. Res. Dev., Vol. 16, No. 1, 1977
Table I. Standard Conditions
Extrusion rate
Cross-arm speed 1 in./min
Volumetric rate 4.39 cm,"/rnin
Average velocity in 11.8 cm/s (23.2 ft/min)
0.89 mm D die
Die
Diameter 0.089 cm (0.035 in.)
Length 2.0 cm (0.80 in.)
Coagulation
Concentration 1 N NH40H
Time 1min 1 0'2 0'5 i i 5 IO 20
Cross-linking Coagulation Tlme, mlnutes
Concentration 0.5%glutaraldehyde
Time 3 min Figure 2. Increasing coagulation time decreases strength under
standard conditions.

--o/.

20.-
D =I6mrn
.o----- o~-'----n --__o-
-. .v/-v--- V
-
. v-
0.2 Ornm
I - i I i

i i
O
t%z ' 0; ' 1 2
\

4
5
Normality of Coagulant

Figure 3. Optimum ammonium hydroxide concentration is about 0.2

creased. It is beneficial to work at a lower extrusion rate where
tensile strength is still high since the fibers produced are more
uniform in cross-sectional area and are easier to handle than
those produced a t higher rates. I t should be noted, however,
that microscopic examination of sample fibers formed at all
experimental extrusion rates revealed no cases of pronounced
flow instability.
The extrusion rates investigated extend upward into the
range of those used in many industrial spinning processes.
Polyglycolic acid sutures, for example, are extruded at a rate
of 150 ft/min, while other wet-spun fibers such as polyacry-
lonitrile are extruded in the range of 50-300 ft/min (Schmitt
and Polistina, 1967; Alexander, 1966).
The diameters of the needles through which the collagen
was extruded were also in the range of the diameters of the
spinneret holes used in industry. These holes usually vary in
diameter from about 0.002 in. to about 0.030 in., depending
on the fiber material and desired characteristics (Atlas et al.,
1967). The needle diameter of 0.035 in. used for most of the
experimental tests is close to this upper limit. As might be
expected, a longer die length gives a somewhat higher strength.
Increasing the length from 0.80 in. to 1.60 inch increased the
strength by about 10%.
2. Time of Coagulation. A minimum time of 15 s was used
and fibers coagulated under this condition exhibited maxi-
mum tensile strength (Figure 2). Fibers which were not
coagulated at all fell apart when stored in water and could not
be tested. The results indicate that at longer coagulating
times, hydrolysis and possibly chain scission become more
important than the coagulating effect of the ammonium hy-
droxide and serve to weaken the collagen structure. The long
chain collagen structure can then be broken down in a way
similar to that caused by prolonged UV irradiation (Miyata
et al., 1970).
Experimental limitations made coagulating times of less
than 15 s impractical. Therefore, it was impossible to deter-
N for a coagulation time of 1 min.
mine the time of coagulation a t which UTS exhibited a max-
imum. There is reason to believe much shorter coagulating
times are possible before a marked decrease in UTS occurs.
Viscose used in the production of rayon and staple fiber, for
example, is extruded into coagulating baths which range in
length from ten to several hundred inches (Ott and Spurlin,
1963). At spinning speeds of about 350 ft/min, this corre-
sponds to coagulating times of approximately 0.15 to 4 s, which
are much lower than those attainable in the present work.
3. Coagulating Agents. Ammonium hydroxide is superior
to sodium chloride in terms of yielding fibers of high tensile
strength a t almost all concentrations (Figure 3). This is
especially true at low concentrations as an optimum concen-
tration of 0.2 N was found for ammonium hydroxide. Below
0.5 N, sodium chloride did not coagulate the collagen in the
time span measured (1 min), and above 2 N, the fibers coag-
ulated in NaCl were extremely crusty and brittle. In both
cases, tensile strength passes through a maximum, supporting
the argument made earlier of competing mechanisms since
a higher coagulant concentration corresponds to a longer
coagulating time. It would therefore appear from these results
that ammonium hydroxide should be chosen as a coagulating
agent over sodium chloride and that, for this geometry at least,
a concentration of 0.2 N is the optimum.
The main purpose of the coagulating bath is to neutralize
the solvent which contains the collagen, causing it to precip-
itate out of the solvent system. Cross-links may also be formed
by the coagulant's action on the fiber material, thereby
strengthening it further. Rather than using an acid coagulant
to neutralize an alkaline solvent (or vice versa), however, salt
solutions can also be used, as was done experimentally, to
dehydrate the fiber and thus cause precipitation in this
manner. This is common practice in the formation of regen-
erated protein fibers as well as cellulose and other staples (Ott

Ind. Eng. Chem., Prod. Res. Dev., Vol. 16, No. 1, 1977 91
 jfo//;
7ooopsl

4 4
UTS
,/\
/ O
\
0
' 0

\ 0\
0.

1
I 2 5 1 0
Glutaraldehyde concentration, W t 96
O V ' 10
I 1 102 103
Crosslinking time, minutes Figure 5. At a cross-linkingtime of 3 min, the optimum concentration
Figure 4. Strength decreases at cross-linking times greater than 1 h. appears to be about 0.2% by weight. Cross-arm speed was 2 in./
Cross-armspeed was 2 in./min. min.

and Spurlin, 1963; Rebenfeld, 1967). I t should be noted that Table 11. Commercial Suture Materials. Strength
concentrations of these coagulating solutions usually vary Measured in Wet Condition
from about 1N to 5 N, much higher than the optimum of 0.2
N NH40H but nearer the optimum concentration of 1N NaCl Tenacity,
determined experimentally for the collagen system. Material gldenier"
4. Cross-Linking Time. Standard conditions were used Polyglycolic acid 4.75
(Table I) except the cross-arm speed was 2 in./min. Fibers Silk 2.45
which were not cross-linked fell apart after storing in water Catgut 1.79
and could not be tested. A minimum cross-link time of 1min Chromated regenerated collagen 1.70
was used and this resulted in the highest tensile strength Glutaraldehyde--cross-linked collagen 0.49
(Figure 4). Strength is not decreased greatly by times up to from present work (41 MPa = 6000 psi)
1 h. The decrease of UTS with long cross-linking times again ' Denier based on dry fiber.
suggests a competing mechanism whereby the cross-linking
reaction is eventually replaced by the chain scission of the sutures produced industrially often undergo a rigorous tan-
collagen molecules. Prolonged ultraviolet irradiation also leads ning process, which may limit the use of collagen as a bioma-
to chain scission and it appears that the cross-links produced terial. Perhaps the present work can be a step toward truly
in this way are less stable and have a lower IJTS than those absorbable, biocompatible sutures.
formed by the glutaraldehyde (Stenzel et al., 1974).
5. Glutaraldehyde Concentration. The results obtained Acknowledgment
for the effect of glutaraldehyde concentration with variable The authors are grateful to F. Takahashi, E. W. Chapman,
concentration a t a cross-linking time of 3 min support the C. W. Zvanut, and T. Miyata for their assistance and for
previous results of the effect of cross-linking time since a low helpful advice and criticism.
glutaraldehyde concentration corresponds to a short cross-
linking time (Figure 5 ) . A cross-arm speed of 2 in./min was L i t e r a t u r e Cited
used. Tensile strength passs through a maximum a t 0.2% Alexander, A,, "Man Made Fiber Processing, 1966", Noyes Dev. Corp., Park
glutaraldehyde concentration and then decreases, exhibiting Ridge, N.J., 1966.
Atlas, S. M., Mark, H. F., Cernia, E., Ed., "Man-Made Fibers", Voi. 1, p 130,
the familiar competing cross-linking-chain scission mecha- Wiley-lnterscience,New York, N.Y.. 1967.
nism. Chvapil, M., Kronenthal, R. L., van Winkle, W., Jr., ht. Rev. Connect. Tissue Res.,
6, l(1973).
6. Comparative Results. Some common suture materials Hamed, G., Rodriguez, F., J. Appl. Po/ym. Sci., 19,3299 (1975).
were obtained from the Cornel1 Medical Center in New York Miyata, T., Sohde, T., Rubin, A. L., Stenzel,K. H., Biochim. Biophys. Acta, 229,
672 119711.
City and tested in wet condition. Results are shown in Table Ott, E.,' Spuriin, H., "Cellulose and Cellulose Derivatives". pp 1010, 1014,
11. The fibers produced in the laboratory did not compare well Wilev-lnterscience.New York. N.Y.. 1963.
with the suture materials but this is not surprising since they Rebenfkld, L.. in "Encyclopedia of Polymer Science and Technology", Vol. 6,
p 548, Wiley, New York. N.Y.. 1967.
have not undergone the processes used to make the sutures, Rubin, A. L., Stenzel, K. H., M.I.T. Techno/.Rev., 71 (2). 3 (Dec 1968).
and, particularly, they have not been drawn after extrusion. Schmitt, E., Polistina, R., (to American Cyanamid Company), U.S. Patent
For example, the commercial collagen suture materials tested 3 297 033 (Jan IO, 1967).
Stenzel, K. H., Miyata. T., Rubin. A. L., Ann. Rev. Biophys. Bioeng., 3, 231
here are actually braided fibrils which have been treated with (1974).
a chromic tanning process. The collagen sutures themselves Takahashi, F., private communication, 1974. (Collagenwas prepared at Cornell
by Mr. Takahashi while on leave from Japan Leather Company.)
rank last in tensile strength out of the four materials tested Tanioka, A., Jojima, E., Miyasha, K., ishikawa, K., J. Po/ym, Sci. Phys., 11, 1489
but the other properties mentioned in the Introduction still (1973).
make collagen a promising suture material. I t is hoped that White, M. J., Kohno. I., Rubin, A. L., Stenzel, K. H., J. Biomaterials, Med. Devs.,
Artif. Urgs., 1, 703 (1973).
collagen fibers can exceed the typical strengths of many re-
generated protein fibers which are only in the range of 0.3-0.7 Receiiled for review May 24, 1976
g/denier when tested wet (Rebenfeld, 1967). The collagen Accepted November 29,1976

92 Ind. Eng. Chem., Prod. Res. Dev., Vol. 16, No. 1, 1977