523210
U. S. Department of Agriculture, Forest Servic e
FOREST PRODUCTS LABORATOR Y
In cooperation with the University of Wisconsi n
MADISON, WISCONSIN
SORPTION OF WATER VAPOR B Y
PAPER-MAKING MATERIAL S
1 - Effect of Beating
By C . O . SEBORG
Junior Chemist
an d
ALFRED J. STAMM
Chemis t
Published i n
INDUSTRIAL AND ENGINEERING CHEMISTR Y
November, 1931
rprf
V-
_
L
▪
I
r
I{
I
.
a
III)il
J I
11
II
eh
-
1
111
%J
T
. 1
ti l I ~
JI
1 1II1
1 I~~I
II
41 '1'
SORPTION OF Ws.T .Lt VAPOR BYL PAP. R-Mi KING MaTr IAL
T-
u`r ; ,
I
1
1 . Effect of Beatingl
4I
f
.m
:
-
n
1
- -.
By
r . ` 1I I1 _ `
C . O . Seborg ,
Junior Chemis t
q~
ti
-11 .
and
A .
J . Stamm,
Chemist
1
71'
1
-
Synopsis
•
I
11
n1
I,c
►
1Y
I
.'r[ ' _ 1 1
Measurements were made of the moisture content-relative ° -i r
humidity and moisture content-electrical conductivity relationships o f
pulps and stuffs .
An apparatus was designed for measuring the relative humidit y
relationship of materials at atmospheric pressure, using a method i n
which air, humidified by passing through various saturated salt solution s
to give the desired relative humidities, was circulated about th e
adsorbent material suspended from a sensitive quartz-helix balance . The
extensions of this balance were measured with a cathetometer . The
hydration (hygroscopicity) of the pulp, which the regularity of the curve s
showed to be physical rather than chemical, was not affected by beating .
Electrical conductivity-moisture content relations likewis e
showed no differences between beaten and unbeaten pulp .
1
.'resented before the Division of Cellulose Chemistry at the 82n d
Meeting of the American Chemical Society, Buffalo, N . Y ., August 3 1
r,
to September 4, 1931 .
J
-r' . .
'
M4
I
rl
R967
-1-
I I
1
_ 1
itI•
. ya
m
1
• r
■
J
The so-called hydration of the paper industry is not a true
hydration . It is probably a phenomenon of fiber-fiber bonding rathe r
than of fiber-water bonding .
Introductio n
410
The objectives in this series of studies are to determine the
hygroscopicity of pulps, stuffs, and the various components of wood unde r
normal atmospheric conditions ; to determine the relations of the sorptio n
hydration (hygroscopicity) to the so-called hydration of the pape r
industry, that is, the change in properties of fibers produced by beatin g
which in the past has been assumed to result from a change in thei r
ability to retain water ; to determine the extent to which this sorptio n
hydration affects the fiber-bonding properties ; and to critically study
the desorption-adsorption hysteresis .
The various materials to be studied are sulfite, Icraft, soda ,
and semi-chemical pulps ; bleached pulps ; beaten pulps ; isolated woo d
components, such as lignin, Cross and Bevan cellulose, and alpha cellulos e
from both pulp and Cross and Bevan cellulose ; and furnishing materials ,
such as coatings, sizings, and fillers .
This paper deals only with the effect of beating on th e
sorptive hydration of a sulfite pulp . The word "sorption" is the same as
used by McBain (8) to include both adsorption and absorption phenomena .
Previous Sorption Studie s
A number of investigators have published papers on subject s
pertaining to the hysteresis of the adsorption and desorption of wate r
vapor by various materials, which ranged from silica gel to textiles an d
paper (1, 2, 6, 1, 11, 12, 3 , 15, 16, U, 1S, 19.) a
All of these investigators, with the exception of Patrick (11 )
and his ho-workers, have obtained definite hysteresis effects . Th e
hysteresis effect in non-elastic gels, such as silica gel, has bee n
attributed by Zsigmondy (19) to a difference under adsorption and de sorption conditions in the curvatures of the menisci of the liquid fillin g
the capillary spaces . These differences in curvature are supposedly du e
to differences in the wetting of the capillary walls, which are les s
readily wetted after drying . ix somewhat different explanation, by Urquhar t
and Williams (17), is more applicable to elastic gels, such as cellulose ,
in which the swelling follows adsorption from the start . Thes e
investigators attribute the effect to the free secondary-valence bonds o f
the hydroxyl groups of the cellulose molecules . In the original water-soake ,
R967
-2-
1
0
condition, the free hydroxyl groups are practically all satisfied b y
water . When the cellulose is dried, these hydroxyl groups are freed and ,
as a result of the shrinking, pairs are drawn together so that the , ,
individual groups mutually sat i y each other . Upon adsorption, part of
the hydroxyl groups that hav e
ually satisfied each other are not free d
for water adsorption, thus g i
decre
ption . ~r
. . . .
^ +...i .• .
_ick (11 .
t the hys t
effect obtained by all o
due to th e
.e
of
gases
eni
o .
"
them
have worke d
1'
evacuated systems . Pat r
ments
s a
F ter•
n h h t e greatest precaut .ri~:• .
air from their silica gel . .
'q6T
y sh e
e
hysteresis effects but the curves
6* _ r.1 _ ssure a content -- are parabolic rather th .
~n shae-, .
Elder ( )
I
have studied the adsorption o_
.
epare,o
feren t
{ i L,g, .methods that gave dif ex t .
y
the
Ell
obtained a parabolic curve ; with mo rn
id curves ; and wi t
the most porous gels, parabolic curv
erse curvature .
zt thus appears quite probable tha n
of adsorption, as obtained by Patrick, may h .
only
gels that show parabolic sorption curves . E
~. f
stating that the hysteresis effect in non-ela n1
b
the fact that the last traces of non-condensabl .- gases have no. _
removed, this does not necessarily imply that the hysteresis s
by ' ~
elastic 'gels, such as cellulose, is also due to the presence o ~ ~
■~
condensable gases . Furthermore, since cellulose cannot be rig s
degassed without decomposition, the hysteresis effect may just
accepted as a property of the system as to be attributed to experimenta l
error .
41, .
Ther seems to be little object, therefore, in working in an
1
xcept when diffusion of the vapors is depended upon fo r
tblvment of equilibrium .
iw
ii thug, sees
only justifiable but also advantageous t o
r e four
,+
it e .- - -ntal work an atmospheric-pressure system i n
'which Alit
ied by passing through different saturated salt solution s
to give the desired relative humidities (3), is circulated about th e
adsorbent material suspended from a sensitive quartz helix balance (8) the
intension of which is measured with a cathetometer . With such a syste m
tbm
-nts are normally= garde under conditions corresponding to thos e
data will be used . The method has a further advantage i n
h equilibrium point is obtained under constant vapor-pressur e
ns . In all of the researches cited, in which a vacuum system wa s
recent work of Pidgeon and Maas (12) is excepted .}, a definit e
volume of water vapor was admitted, and, as adsorption progressed, the
vapor pressure diminisle:d . Under such a system the most readil y
accessible part of the i atelHHal may initially adsorb more water vapo r
than the final equilibrt : amount, thus permitting a subsequen t
combination of both adsorption and desorption to take place at the sam e
time . Such a condition might readily affect the normal hysteresis .
For the same reason a number of simultaneous determinations cannot b e
made in a single apparatus under changing vapor pressures, whereas the y
can be made when the humidity and temperature are held constant .
Th~a gimp' 'recent measurements of Pidgeon and Maas (12) hav e
bean made at . ei munstant vapor pressure in an evacuated system, by holding
the temperate of the vapor-supply bulb- (which is left open to the re's't
of the system after evacuation) below that of the rest of the sorptio n
apparatus by -Pm amount--that gives the desired relative humidity in th e
sorption system . Under these conditions the accuracy with which th e
relative humidity is maintained depends directly upon the accuracy wit h
which the two temperatures are held . At high relative humidities, 9 5
per cent for example, a deviation of 0 .2° C . from the required temperatur e
difference between the vapor-supply bulb and the sorption system at 25' G .
causes a deviaViAla in relative humidity of 1 .2 per cent . When saturate d
salt solutions a,rm used to- maintain the relative humidity, , the effect o f
temperature deviation is much less ., For example, a saturated potassium
sulfate solution that give's approximately the relative humidity jus t
mentioned (95 per cent) would require a temperature deviation of 30° C . t o
cause the same deviation (1 .2 per cant) in relative humidity (5) .
Apparatus and _. ]xperirental Methods .
Sorptive Metho d
The apparatus used in this part of the . esearch is shown i n
Figure 1 . -Four simultaneous measurements were made in the sorption tube s
f, E, h, and i . These tubes, connected in parallel at the distributin g
head d, were--in series with d, the spray trap c, and the humidif ll-ng '
train., b . The humidified air, after - circulating abo1t the test s4elples
suspended from quartz heljces .;•. was discharged through the glass tubin g
of the hooka-1, k, 1-, and m,- that were -sealed i .to the top of t
sorptio n
tube's . The conditioning tubes were immersed in a thermostatic, water bat h
having glass windows o two opposite faces . The temperature of the ball
was maintained at 250 t 0 .02° C-: and . that of the r- .osm at 23° t 1.0° C a
In making ;measurements at high relative humidities, t .e s],ht drop in .
temperature' of the exhaust air was sufficient t .o cause condensation i n
the upper part of the exit tubes, which were spbsitantially at :room
temperature ; so in -order te .prtaeent this condensation ha .t was applie d
x967
Figure 1 .--Diagram of sorption apparatu s
' -e. ; inle t
b; humidifying chamber s
c; spray trap
At manifol d
stop cocks
f, _g, h, and ; sorption chamber s
Z k, ] and
hooks of glass tubin g
o, p,. and 21 local heating elements
near the ends
p, and g . The
into a beaker
The pressures
to give equal
of the tubes by means of small electric light bulbs, n, o ,
discharge tubes were bent over with the ends barely dippin g
of water to permit observation of the velocity of discharge .
in the tubes were made equal by adjustment of the stopcock s
velocity .
The helices, which were made of quartz fibers, were of suc h
size that a helix with a diameter of about 1 .0 cm . and 30 to 35 turns
ia ; length of 2 to 3 cm . was able to carry a load of 300 to 1000 mg .
to weigh to an accuracy of 0 .1 per cent . Cathetometer measurement s
were made at the fixed ends of the helices as well as at the deflectin g
ands to avoid any possible errors that might otherwise have been cause d
by creeping of the sealed-in rubber stoppers ; all cathetometer reading s
were made to 0 .005 cm .
The humidifying train b consisted of three glass tubes, 35 cm .
in length and 2 .5 cm . internal diameter, connected in series . The tube s
contained short lengths of glass tubing to increase the solid-liqui d
contact surface of the ascending air bubbles . .k.n excess of the salt wa s
placed in the bottom of the tubes before sealing to insure the maintenanc e
of saturation . A pre-humidifying train with large-bore tubes was set up
outside of the thermostat . This arrangement s
e air t o
ap sz4
#4 ti dessai'qs-iiiadpWif
fying
train b,
vent
ppreciable quantities o f
water frua ILl malt s
d have resulted i n
e y.ta%'1U itala and c
system .
_ ailher
idity in equilibriums with each of th e
NkturaOmil'
was determined by a dynamic method as nearl y
comparAle with the c
under which the humidifying tubas wer e
used as possible . Mr w at a constant speed through one of th e
humidifying systems just described and then through an absorption train ,
consisting of a U-tube filled with calcium chloride, followed by anothe r
filled with phosphorus pentoxide suspended on glass wool . The dry ai r
discharged from the phosphorus pentoxide tube was then passed throug h
another humidifying system which was similar to b except that it wa s
filled with distilled water . The saturated air coming from the distilled water humidifier was passed through another absorption train . Both set s
of humidifying tubas and absorption trains were kept in the thermostati c
bath . The ratio of the amount of moisture absorbed in the firs t
absorption train to the amount of moisture in the second, after makin g
volume corrections for the difference in the total air pressure obtaine d
after pssin through each of the sets of humidifying tubes, gave th e
relative 'ginicity of the air in equilibrium with the particular saturate d
salt solution under consideration . In preliminary tests air was passe d
through the system in increasing speeds to determine the speed at whic h
the humidity of the air was no longer in equilibrium with the humidifiers .
One-third of the maximum permissible speed of 45 liters per hour was use d
in this
if
so as to insure the attainment of equilibrium .
x967
eJ 4
ir
1.
Table 1 .--equilibrium relative humidities of various saturated salt
solutions
Relative humidity
20° C .
. ' 0° C ,
. 100° C . .
25° C .
: Calculated :experimental :0bermiller (10) ;Calculate d
Per cent :
Potassium acid
8748
.
Per cent
.
Per esat
Per cent
953
..........
Sodium chloride, NaCl
•
73 .8
Zinc nitrate ,
Zn(3d0 )2 . 6 H2 O
10 .9
7 4+ .2
'749
53 .4
52 .0
4o . o
42 . o
11 .1
15 .0
75 . 8
11 . 5
In Table 1 are listed the relative humidities obtained at 2 50
0.02° C . in equilibrium with each of the aaturated salt solutions used i n
this research . The experimental values represent the averages of at leas t
six determinations . Variations among individual measurements rarel y
exceed 0 .5 per cent . The results obtained by Obermiller (10), using a
static method, as well as values calculated from vapor pressure an d
solubility data at 0° and 1000 C . obtained from the literature, are give n
for comparison . Unfortunately, data are not available to complete th e
table, In general, the relative humidity decreases slightly with a n
increase in temperature . Obermiller r s value for lithium chloride i s
apparently high, whereas the results of this investigation for potassiu m
acid phosphate and sodium dichromate seem a little high in comparison wit h
the other values . Since good checks were obtained for both of thes e
measurements, the deviations were probably due to differences in th e
purity of the salts .
In these studies the experimental zero humidity was obtaine d
by passing the air through the towers filled with phosphorus pentoxid e
suspended on glass wool .
410
Measurements were made upon an unbleached spruce sulfite pul p
and the same pulp after it had been physically processed for g hours in
a beater . The beaten and unbeaten pulps were thoroughly washed for 2 4
hours with running distilled water and were then made up Into thin tes t
sheets . The wet sheets (containing about 60 per cent water) were cu t
to a size that would weigh approximately 0 .5 gram when ai dry, wer e
rolled into the shape of a cylinder, and prior to the final measurement s
were suspended for about two weeks in an atmosphere of abouc96 per cen t
relative humidity . This was done to remove the large excess of fre e
moisture, thus making it possible to use more sensitive helices as wel l
as to cut down the time required for the actual measurements .
The desorption measurements were started at the moisture conten t
induced by 95 .3 per cent relative humidity and were continued stepwise a t
decreasing relative humidities to zero humidity . The adsorption was the n
followed by using the same sets of humidifying tubes in the reverse order .
Equilibrium was obtained in the desorption measurements at high relativ e
humidities in approximately 2 days and at lower relative humidities in on e
day . In the adsorption measurements equilibrium was obtained somewhat mor e
rapidly, requiring approximately 1 day at high relative humidities an d
considerably less than 1 day at lower relative humidities . In each case a n
additional day was allowed as a factor of safety .
s
Electrical-Conductivity Metho d
The results of the sorptive method were confirmed by electrica l
conductivity-moisture content measurements, using an apparatus previousl y
developed for determining the fiber-saturation point of wood (14) . Thi s
apparatus consisted essentially of a brass screw clamp with circular contac t
faces, 1 .0 sq . cm . in cross section, for making contact with the pulp disk ,
The two contact faces were insulated from each other by 4 cm . of the hardrubber base . The reading apparatus consisted of a portable suspensio n
galvanometer, an ayrton shunt, a key switch, and a single 1 .5-volt dry
cell . The shunt, which allowed 1 .0, 0 .1, 0 .01, or 0 .001 of the curren t
to pass through the galvanometer, was connected in parallel with th e
galvanometer and in series with the dry cell, contact clamp, and key
9 switch .
Readings of specific electrical conductance could be made from 10
to 10 3
mho with this simple apparatus .
R967
-7-
The measurements were made upon circular disks approximatel y
3 .0 rpm . thick and 1 .0 sq . cm . in cross section when in water-soaked and
swollen condition . The disks were dried in air for a few minutes, an d
then held in small weighing bottles for about an hour . in order to approximat e
uniform moisture distribution before weighing and making measurement of th e
electrical conductivity and of the dimensions . The disks were then drie d
for a few minutes, after which the procedure was repeated .
.experimental Result s
•
•
The sorption measurements of four samples of pulp an d
corresponding stuff, conditioned simultaneously, are recorded in Table 2
and plotted in Figure 2 . The results show that beating has a negligibl e
effect upon the moisture content-relative humidity relationship of pulp .
This same result was obtained by Campbell and Pidgeon (2) in an evacuate d
system . The hysteresis loop is not closed at high relative humiditie s
nor does the hysteresis effect disappear at low finite relative humidities .
The fiber-saturation point obtained from extrapolation of the adsorptio n
curve to 100 per cent relative humidity is approximately 29 per cent . Thi s
too is practically the same as the value obtained from the extrapolatio n
of Campbell and Pidgeon t s curve (2) .
The electrical measurements plotted in Figure 3 show tha t
beating has no effect upon the electrical conductivity-moisture conten t
relationship of pulp . Drying and resoaking the pulp also snows no effec t
upon the relationship . This is in harmony with Murphy and Walker's finding s
(3) that there is no difference between the electrical conductivity moisture content relationship of cotton fibers under adsorption and
desorption conditions, thus showing that the electrical conductivity i s
a function of the absolute moisture content and is independent of the mean s
of attaining that moisture content . A linear relationship exists betwee n
the logarithm of the conductivity and the moisture content from the dr y
condition up to fiber saturation . above this moisture content, th e
conductivity changes but slightly . Over the whole range the results fo r
the beaten and the unbeaten pulp are identical within experimental error .
The fiber-saturation point -- the point at which deviation from the linear
relationship starts (28 per cent) -- corresponds quite well with the valu e
obtained by extrapolating the relative humidity-moisture content curve s
to 100 per cent relative humidity .
R967
-8-
Figure 2 .--Moisture content-relative humidit y
relationship of beaten and unbeate n
pulps
20
30
40
50
50
70
Relative Humidity (Per cent )
Figure 2
S()
•
Figure 3 .-- .L ctrica4 conductivity-moisture conten t
relationship of beaten and unbeate n
pulp s
•
S
I'5o
tta.tAi q.onpTTOO
ICY.
TDoz,zq.oata oTIT o2dS
0
'tor
~ri
Ir
I
Table 2 .--2ffectof physical processing on equilibrium moisture conten t
of
-:Relative :
Moisture content (O . D . basis)
:humidity :
Sample :
at
Unbeaten pulp
Stuff (highly beaten pulp )
25o
.
.
'CAM= : Desorption : Adsorption :
Desorption :
Adsorptio n
:Per cent :
1
2
:
:
::
3
4
R967
.
95.3
87 .0
74 .2
53 .4
40 .0
11 .1
0
Per cent :
Per cent :
Per cent
Per cent
32 .7
19,2
13 .2
8,8
5,6
2 .0
0
20 .4
114 .7
10. 14
6 .6
5.9
4 .1
0
143 .0
22 .1
14.9
9.2
6 .3
2 .7
0
21 . 7
14. 6
34 .7
'20 .7
15 .1
9 .5
-6 .-g
2 .7
0
22 .1
114.7
10 .0
6 .7
54 5
2 .4
0
36 .3
21 .9
15 .3
9 .9
6 .6
2 .9
0
23 . 4
15 . 3
11 .9
7 .9
5. 6
2. 3
0
95 .3
87 .0 A
74 . 2
53 . 4 :
40.0
11.1
0
32 .5
2 .0 .5
14.8
9 .4
6 .5
2 .9
0
82 .1
14 .4
10 .0
6 .9
5 .6
2 .7
0
32 .7
21 .7
15 .2
10.2
6 .5
2 .6
0
22 . 2
15. 7
12 . 2
8. 2
6.2
2, 8
0
95 .3
87 .0
74.2
3)4
.0
11 .1
0
32 .7
20 .3
114.5
9 :4
6 .6
2 .7
0
22 .6
14.7
10 .0
7 .3
5.6
2 .3
0
29 .7
19.9
14 .5
9 .5
6 .3
2 .6
0
22 . 3
95 .3
87 .0
714 .2
53 .4
40 .0
11 .1
0
:. .
:
:'
:
t'
.
.
11 . 2
7.4
5. 3
1.8
0
13 . 3
11 . 6
7. 7
5, 5
2 .5
0
Conclusion s
The continuous unbroken form of the moisture content-relativ e
humidity relationship indicates that the hydration of pulp is a physica l
sorptive hydration rather than a chemical hydration in which the bondin g
is in definite molecular proportions .
The desorption-adsorption hysteresis is similar to tha t
obtained by other investigators who used an evacuated system . The
presence of air thus seems to have no effect upon the extent of hysteresis .
The results show that beating has no effect upon the tru e
sorptive hydration of the pulp examined . The total internal adsorptiv e
surface remains unchaed, though the outer surface of the fibers an d
fibrile may be Increased considerably . .It is presumably the increas e
in the, available bonding surfaces .of the f-ibers. : and fibrils for each othe r
that affects the physics,, properties of the resultant sheet of paper (18) .
The so-called byd:rat om of :the paper industry is not a phenomenon of fiber ;o_n
water bonding as the, riche Implies, but appears rather to be a pl e
of fiber-fiber bonding .
•_
The fact that beating has no effect upon the electsica l
conductivity-moisture content relationship indicates that this .••
is a function of the aaount of sorbed water and is not d e.,pen404 upon the
fiber-fiber bonding . Such a relationship is to bw expec. .ed 'AA it i s
realized that the dry fiber is a vary poor conductor (specifier .. onductance :
1o-12 to 10-13 mh.o) .
'11 `
Literature Cite d
(1) Anderson and Kelvi n
1914 . Z. Physik. Chem. 88 : 191 .
(2) Campbell and Pidgeon
1930 . Pulp and Paper Magi Can . 29 :185 .
(3) Hatfiel d
1931 .
J . Abr . Research 42, 301 .
(4) Holmes and Elder
1931 . J . Phys . Chem . 35, 82 .
S
(5) International Critical Table s
1926 . 1, 67 . McGraw-Hill Book Co ., New York City .
(6) Kuyirai, Koboyashi and Toriyama
1923 . Bul . Inst . Phys . Chem . Research 2 :105 ; J . Soc . Chem .
Ind . 42A :8-94 .
(7) McBain and Baker
1926 . J . Am. Chem . Soc . 4g , 696 .
(8) Mason and Richard s
1906 . Proc . Roy . Soc . 78A :412 .
(9) Murphy and Walke r
1928 . J . Phys . Chem, 2, 176 .
(10) Obermille r
1924 .
Ibid ., 109, 145 .
(11) Patrick
1930 . Colloid Symposium Annual 1, 129 ; John Wiley & Sons ,
New York .
(12) Pidgeon and Maa s
1930 . J . Am. Chem . Soc . 52 :1053 .
(13) Sheppar d
1930 . J . Phys . Chem . 34 :l041 .
(14) Stamm
1929 .
Ind . Eng . Chem . Anal . Ed . 1, 94 .
(16) Trouton and Poo l
1906 . Ibid ., 77.x :309 .
(17) Urquhart and William s
.
1924 . Shirley Inst . Memoir's 3 ;197 ; J . Textile Inst . 15T :43 .
1929 . Shirley Inst . Memoirs 8 ; J . Textile Inst . 20T :125 .
(18) Van Bemmelen
1897 .
Z. Anorg . Chem . 13 :3 1