SERVIC~A131~ITy OF GLUE JOINTS
SCHOOL OF FORESTR Y
OREGO1 STATE MIME
CORVALLIS, OREGO N
UNITED STATES DEPARTMENT OF AGRICULTURE
L
OREST
L FOREST SERVIC E
PRODUCTS LABORATOR Y
Madison, Wisconsi n
In Cooperation with the University of Wisconsi n
April 1938
OF GLUE JOIaTS l
o sure tests give data_ vegetable,
asein, and blood type s
-Trr3 .'
By
Bas se , Eprgsir e'or
Introductio n
Tests have demonstrated that among the more important causes o f
failure of well-made glue joints in service are the following : (a) chemical
hydrolysis of the glue,,,(b) destruction of glue by microorganisms, and . (c )
mechanical stresses developed on the glue joint as the wood changes cimensiohs under the effect of changes in moisture content . The relativ e
different factors varies largely, depending on th e
type of glued construction, the variety of glare,-an _service conditi-ens ..
'ha
m
The significance of :W 4i. is and. Sts rel.a-tila to ,a
11tas been deve3apedd for casein glues (1)2 ant the ge p'! prine•id
believed to app ,y to of ier<0ro'tre .llrl es . lager conditions f v al :e
epoorganisms on glues of corn osi ri'ar
decay the destructive eeff'ft
that serve as- fbed~ has b e
' .; o be ?apid and _x r ,trti,•o- . • The ea fewt, how ever, may be gre t .y redlaie '_by twIoe-o> atirrg tea~ic ri ztemilals in the glu e
or by adequate trea ent of tlt--e . entire . g1 ad-'a ffber with toxic. preservatives (2) . The-.lfect of micr•lergatiams on'the glue joint and deterioration due to Yy!ol.ysis ray,, off' *o)tV.se, be ,eliminated by the selection of
adhesives that t not subject •tie.. attack by i .olds and ,that do not hydrolyze .
The effect of met nicalaction, ii Wever, ¢ ia®rt be greatly
retarded by minor changes 14 formulation . t-4holag4D .1
iag mlekWw
l ave_ a remarkedt on b1 ate ,., ' hydro .ysis,
.w .ll the mechanical
eft-eats be changeO'to r ' ~it 1degroe ~ ~_ g . toxic terials t o
the glue ory tireatirrg~ ,
I He'd I.Ambers wi .ly used fo r
14 co
the p<axe.sOvati,on of wood. 'The primary re•q• ' ; t, a
:ttga t,.*e joints ar e
to maintain the4-r strength under mereatefl ehah0p. in tmo; d
i c'eo
ns ,
appears to be'a:-O.ue that ' i=w-fad-0s enough, eithi• in
c fir= ndition
to. withstand 'this Ere *s c ®veiled udei,=sere conditions of service .
+fi
1
tide
,
i ,we
1Published' in Mechanical Engineering, April 1938 .
2
-.Number .s inn .parentheses refer to the Bibliogt'a y. at tflaoJINId o the
R1172
WuBad
t
e e s deve
ade. glue joints un,ela s
li'rst
,
be
the
0t
.
'e'€
as
do
5a
changes in moisture co n4e-i;1 r
magnitude of the stre . e 1 ,d 61Vped, a, ,d s Gnd, upoft 't`hlc ree to-TAB
a of th
the glue softers and: ,weakens When it be*orse :s 4ed,st . The
arta4'e
. eil rage s
medhaAtAOt ettageottdr velopeci-depends on the extaaat of the .
and how far the extremes of •moisture oontegt depart fx+e'rm the moiiM 9ontent of the stock at the time the glue sets ; also upon the densi and
shrinkage characteristics of the species used, .a . the thi I S of the
laminations . The resistance that the glue' lin e offers *o ,ghe4swigMt;eeOBs ,
e glue
may be thought of as depending oa. the ame it of water absorbed .
s~ngth
when
.
softensand
]
pses
its
line, and the degree to which the glue
d
resiiant
will
wa
stan
this absorption occurs . Glues that are, water
laeking
in
tilt*
'
proper
er ty.
exposure to higher humidities than glis that. are
Failure of we.1
9
-
c
Tests were conducted at the FO-r•eet Products -Laa , eI r
mine the relative resistance of some of the commionly used we't~
adhesives to service conditions involving succae•ssive exposii . 0 alternating
conditions of moisture that would develop umeohanical s',r ~ss' o- .nexeasing
magnitude on the glue joint . The purpose ,s gfur :here 4Q-1c._i.
7f'
the conditions of service under whic} ~
main permar~s
eit
to
durable and the cone`,
could be expected.
11.14
~_
In these tests the complications of species characteristics an d
thickness combinations were eliminated by the use of a single species an d
a single thickness combination throughout, although it is recognized tha t
these factors must be considered when general applications of the result s
are made . Further, the tests do not include the more recently develope d
synthetic resin adhesives . Other tests, however, have indicated that
joints made with hot-pressed synthetic resins are highly resistant t o
conditions favoring mold attack and hydrolysis and that they retain a
high percentage of their dry strength when saturated with water or expose d
to high humidities .
;4
.
1
T
n p
a
n e ; 4 0 'r i roast M e
a and comp
three plies of selected. 116-in . yellow birch veneer,
each
of seven glues . They were (a) sodium-silicate-lime-casein glue
t1 according to . for-mina. )4-R of
Forest Products Laboratory, (b) and (c) tw o
sodium-{h drextde-liaae-easein goes, (d) high-grade animal glue having a
jelly st3 eatgth of 303 a and a viscosity .of 1,04 r llipoises, (e) low-grad e
animal glue ha ns a je1iy.O etgth of 190 g am* a 'v 'e 'st' " _-.
`
`, poises, (f) c omtert_9.g v eg r it a .or starch glue a*d_
glue mixed, acrimrtli-n ; t 'o
paraterrnaldeh0cle formu7ld ' - •e'
Laboratory . In Table l,'rgich eves the compositioh. •of they
on a
parts-byr.weight basis, the•ind idi 1 glues -a,re desi ■ted by' ,'e,Letter s
used in the preso,edi-rag sentence ,
.,-
.
R1172
-2-
Jj
,4
Table 1 .--Composition of glues used in test s
Constituents
Ammonium hydroxide
Animal glue
Blood albumin
Casein
Lime
Paraformaldehyde
, processed" starch Sodium hydroxide
Sodium silicate
Pater
(a)
----G 1 u e--- (e) : (f)
(d)
(c)
(b)
(g )
5. 5
100
100
2g
: 100
. 15
: 100
100
: 100
: 15
15
100
: 70
330
12
: 2. 0
.
3
4
. 380
. 225
. 195
: 225
. 190
The veneer was conditioned to approximate equilibrium with 30
percent relative humidity before gluing, which brought it to a moistur e
content of about 6-1/2 percent . The gluing conditions were adjusted t o
fall within limits favorable to the production of good joints (3) with eac h
glue . The gluing operations with animal, casein, and vegetable glues wer e
carried out in the conventional way . The panels glued with blood albumi n
were spread and pressed at room temperatures, allowed to remain unde r
pressure overnight, and hot-pressed the following morning for about 1 0
minutes at 260° F . and under a pressure of 200 pounds per square inch .
After gluing, the panels were again conditioned to approximat e
equilibrium with 30 percent relative humidity and then cut into standar d
plywood test specimens (3) . Each of the 17 panels yielded 30 test speci mens, giving a total of 510 for each glue . Five specimens from each
panel were tested dry, and any panels that gave low or erratic test value s
were discarded .
$ped"imens rfom 't` `fie 'remains g panels of each glue were then mixe d
together to insure random sampling and divided into 6 groups of 60 each ,
one group for each of the tests listed in Table 2 . Air temperatures wer e
maintained at approximately 80° F ., while the soaking water temperature s
of test No . 6 varied with those of the laboratory . Five specimens wer e
withdrawn from each group and tested on a plywood-testing machine of con ventional design (3) at intervals of 1 , 8, 12, 16, 20, 40, 6o, 80, 120 ,
and 160 weeks . Specimens were unprotected, and their dimensions wer e
small enough to permit the wood to attain approximate equilibrium wit h
atmospheric conditions at each period of the exposure cycle .
R1172
Table 2 .--Data on exposure test s
Second stag e
First stage
Test :
1
2
5
6
T ype of
exposure
cycle
: Continuousa
: Alternating
Alternating
Alternating
: Alternating
Alternating
:
Time,
days
:
:
•
Relative .
humidity, :
percent
.. ... .
30
60
g0
90
14
l4
14
14b
97 b
-
2-
Time,
days
14
.
:
Relativ e
humidity ,
percen t
30
30
14
30
14
30
30
14
12
30
I7sed as controls .
kSoaked in water at room temperature .
In analyzing the data from the tests, the average test value o f
each group that remained in 30 percent relative humidity was considered a
n control" average and other test values of specimens glued with the sam e
glue were expressed as percentages of this . As the different exposure s
were continued and test values were obtained at different intervals, thes e
were converted into percentages and plotted . After plotting, the point s
were joined by straight lines to illustrate general trends in strengt h
changes . No attempt has been made to draw curves of mathematical exact . -ness because the number of specimens tested at each period was no t
sufficient to justify such treatment . The tests have extended over a
sufficient period of time, however, to permit confidence in the genera l
tendencies illustrated.
Animal and Vegetable or Starch Glue s
Both animal and vegetable or starch glues are lacking in wate r
resistance, and the durability of joints made with them and exposed t o
water or to damp atmospheres is so similar that the results were average d
together and shown on the same curves, Figs . 1 and 2 . When exposed t o
test No . 6, a soaking-drying cycle, all joints made with either anima l
or vegetable glue separated completely either in the first soakin g
period or in the drying that immediately followed (Fig . 2) . This be havior is to be expected of glues that are not water-resistant .
Specimens, glued with animal or vegetable glue, exposed to tes t
No . 5 did not separate immediately, but at the first test period, the en d
of the first cQnrlete cycle, the average joint strength was less than 4 0
percent of the original . All specimens had failed before the end of th e
fourth cycle, or in somewhat less than 16 weeks . This test does no t
R1172
-4-
ais
involve exposure to free r a,tero but a relative Midity Qf 91 , parrs.
a
ur e
d
g;
sufficient "high to p :eimi t -devalo&rent of me L&s "and-to ? r
content of the wood to about 2 g percent during the "wet _l at" . Of tr i
cycle . Specimens exposed to test No . 4 failed less rapidly than Vh se ex posed to test No . 5 . They had decreased in strength, ho .Wover-,, by sl e •
50 percent by the end of the first test cycle, and all specimemshhad .
failed by the end of 36 weeks . Moisture content of the wa64 du-i_i g- tthe
"wet half" of this circle probably reached about 22 pereen . (Fig . 1); .
In test No . 3, more than 20 weeks of exposure were requied -be fore the average test values had fallen to 50 percent of the . origir,T.
By the end of the eighty-first week, however, all •spetimaens . has f 401.6d
indicating that animal- and vegetable-glued joints cannot be dep cOd
upon to remain permanently durable when exposed to conditions where the
wood may at times exceed about 17 percent moisture -erontant, which. i s
approximately the equilibrium e dition for g0 pegoant relative hu it y
(Fig . 1) . These results may appear contrary to exile ience, for •f tat 00
glued with starch and animal glues often serves sati fdeto ly i .n ;ipi -te of
occasional exposures to relative humidities in excess of go p:eree:r t.
In those cases, however, the proteetion afforded b the finiehh . may
prevent the . weed reaching the equilibrium moisture c'Ontent-tlf 17 percent ,
particularly if the exposure , to dampness is not p :aloage4..
In tests Nos . 1 and 2, no evidence of signifieant loss i n
strength could be detected during the 160 weeks . that the tests were i n
progress . Test No . 2 approximates the changes t : moLst:ix o •onisenti tl att
can be expected in interior woodwork in normal uste in heated bujltnvgs- i n
the northern part of the United States (4) . In this type of .ore e,
properly designed and well-made joints of animal .or vegetable glue shoaa]4
prove permanently durable, and experimental evidence, other than thes es .
tests, supports that belief (5) .
Casein and Blood Glue s
When exposed to conditions that cause alterrna ting : aches
the moisture content of woo y, ;,the casein glues used in these e-x .a iten:t s
displayed no great differem et among themselves in Aur ab 1.it r of jo-t,a: ..
Results from the three casein glues have been avetaged toget er, therefore, and are shown in a single set of curves as Fig . 3 . Tiey are be lieved to be typical of the general. tendencies that would dev~aI i. i-wbol
any of the better grades of casein glues are subjected to• similar `I pigs'..
Results of test No . 6 illustrate that joints made with gain
glue cannot be termed waterproof . In this series of s'oakii -dryiA. ests ,
all specimens glued with casein glue had failed by the end of 15:_voata..
In other similar tests (6), similar results have-been obtained . Mil e
the time of total failure has varied from 3 to 15 months, an eemiy weaken ing of the joints was apparent in all similar tests .
0
Exposure to 97 percent relative humidity followed by d ng ,
test No, 5, also causes rapid weakening of unprotee ed = s a g as _ flint .
R1172
-5-
with total failure occurring in these testre r a t abo u t 1 . . oaOq or 12,
complete cycles . Even the cycle of-2 weeks in 90 pers . rtt re] ' lark . talia ty
followed by 2 weeks drying i 3.0 : percent releegvi humidity, test .40,. :1 ,
-l
s~esi;
was severe enough to cause ultimate failure of unprotected.,
.
The
rate
of
failure
is
slower
than
that
displayed
by
'joints
o f
joints
d
.ons,
but
at
theen
vegetable or animal glue exposed to the same coaditi
.
had
failed
all
casein
joints
of some gl weeks, 20 complete cycles-,
In test No . 3, hemmer, me waidence of significant wpakenpng of
joint strength was disclosed during the ] .60 weeks that the tests sage in
progress . The difference in resistance between casein and animal o r
vegetable glues is demonstrated in this test. Well-made ca :seln.glua
joints are apparently able. to . withstand repeated eaposur_es to-80 lerCen :t
relative humidity while joints made with starch or animal glues u1tir` .t;eL,y
fail in this exposure . Tests Nos . 1 and 2 failed to disclose any weakenring of hell-made casein-glue joints while the tests were in progress an d
none would be expected, regardless of the length of time that the test s
might be extended .
Until the development of artificial-resin glues, the "heteptft 1
paraformaldehyde blood glue was one of the most water-resistant of wealY x
working adhesives . In test No . 6, specimens held together for nearl y
2-1/2 years (Fig . 4) before total failure, and, in other similar test 's ,
specimens have withstood the cycle for as long as 3 years, Even tli .s
resistant glue, however, cannot be classified as waterproof, - far' :sire
moisture changes repeated over a long time will cause failure in the .jei .
i
At the end of 160 weeks, blood-glue joints exp osed to test- M-5
had lost approximately 50 percent of their original strength . From the
downward trend in joint strength illustrated in Fig . 4, total failur e
might be expected somewhere between 4 and 5 years . In other similar tests ,
however, blood-glue joints have withstood this exposure for as long as 6
years and, at the end of that time, the joints retained some 4o percen t
of their original strength . By that time, the wood which had been exposed to the cycle of 97 percent relative humidity followed by 30 percen t
was beginning to show some evidence of decay .
From the trend illustrated in Fig . 4, ultimate failure might b e
expected in blood-glue joints exposed to test No . 4 . In view of the long
resistance to the more severe test No . 5, however, it appears rea oiabl y
safe to assume that these joints might, on more extensive t-eets, prov e
permanently durable as long as the moisture content of the wood did no t
exceed approximately 22 percent (Fig . 1) . These tests produced n o
evidence that well-made blood-glue joints will fail when s- j,eet0d t o
tests Nos . 1, 2, or 3.
41,
Destructive humidities appear to be between 60 and g0 percen t
relative humidity, corresponding to a maximum equilibrium moisture con tent in wood of about 17 percent, for animal and vegetable or starc h
glues ; between g0 and 90 percent relative humidity, corresponding to a
R1172
-6-
O
maximum equilibrium moisture content in wood of about 22 percent, fo r
casein glue ; and between 90 and 97 percent relative humidity, corres ponding to a maximum equilibrium moisture content in wood of about 27
percent, for blood glue .
Bibliography
(1) "Casein and Its Industrial Applications," by Edwin Sutermeister ,
Chemical Catalog Co ., New York City, 1927, pp . 169-217 .
(2)
"Increasing the Durability of Plywood," by Don Brouse, Mechanica l
Engineering, Sept . 1931, pp . 664+-66 .
(3)
"The Gluing of Wood," by T . R . Truax, U . S . Dept . Agr . Bulleti n
1500 . (Available from the Supt . of Documents, Washingto n
D . C ., for 250 a copy . )
( L±) "Moisture Content of Wood in Dwellings," by E . C . Peck, U . S, 'Crept .
Agr . Circ . 239 . (Available from the Supt . of Documents, Washing ton, D . C ., for 50 . )
(5)
"Age and Strength of Glue Joints," by Don Brouse, Woodworkin g
Industries, June 1931 , pp . 32-33 .
(6)
"Behavior of Casein and Blood Glue Joints Under Different Condition s
of Exposure," by Don Brouse, Furniture Manufacturer, Sept . 193+ ,
pp . 9-11 .
•
R1172
-7-
o
.,
o
fti
o
,a
t
o
o
o
DO
-rl
•r' I
)
4
Ell F•~
0
0
a
rd)tird
o b
4-I
O
U1
4
O
0
1i
t
o
c) + )
c3
N
r
o
td
o
o
m
o
C5
i M N N
P4
LC\
I
O
re\
I
I.c'1
N
I
o
N
I
LIM r-I
i
o
rI
(4uaoaed) pooee u2 4u9a.uoo exna.szo n
O
I
I .[1
`
0 °
R,
o
El
co
40 hO
0
ai
O
1'W1Ts
Ill
' il
3
a
j
r
0
CV
r-I
0
0
r-1
a]
-I-
r
„rte F . .~+
rV
o
-b
-0
- 1 --est_Ta
o
cg
.
N
;o a .aGoaed )
(senva Zo :aq.uoo
sent .& 96'} Doodd&j d
o
aCak--307
V
LO
0
0
-0
0
a)
,
o
r-- t
0
ri
a
col
I
E
a
Cu
Test N .,
i7
- f
0
i
0
►
0
Soul
Q
0
0
(sant.9a toa4troo Sat)
eiozad SeTLteA q. sa pooeaSN
.©e a
0
r
u
0
-
0
um
-3
-
N
0
r--4
r
s
t%
p
-- Ca
IX)
0
}
°
LD
r
p
d
CtJ
;0 89el.ueo,zad )
senuA q,se . poom.Eld
(saniVn Zoa~uoo' Lip
0
0