576 ILVDliSTRIALA,VD ELVGISEERISG CHEMISTRY T'ol. 21, S o .

6
quick-frozen packaged products. This precludes containers placing four pads or liners within an ordinary slotted carton
from such materials as cork board, balsa wood, and the of the proper dimensions (Figure 4). The container should
various proprietary insulating materials. be as nearly cubical as possible, so as to expose the minimum
Corrugated fiber board has been found to be a very satis- surface t o heat penetration: and, when possible, the corruga-
factory material from which to manufacture insulated ship- tions in the liners should run horizontally so as to lessen
ping containers, and considerable quantities are now being convection currents. Obviously, the shipping container
used for that purpose. It has approximately the same should be packed solidly full, because the heat leakage into
heat-insulating value as cork board or balsa wood, is light, the container is approximately in proportion t o its surface
strong, and inexpensive. Fiber board may be made with area, whereas the amount of heat which csn be absorbed
either straw or chip corrugations, and with kraft, iute, by the product with a given temperature rise is in proportion
chip, or straw liners. It has approximately the same heat- t o the volume of product (Figure 5 ) .
insulating value no matter from which materials it is Figure 6 indicates the effectivenes9 of corrugated fiber-
made. board cases in protecting their contents from outside heat.
Although the construction and thickness of the corrugated It will be noted that the contents of the 50-pound container
containers used by different producers of quick-frozen prod- required 7 days to reach a temperature of 50' F., a t which
ucts vary greatly, an excellent container may be made by time the fish was still perfectly fresh.

The Condensation of Steam'

D. F. OthmerZ
UNIVERSITY OF MICHIGAN,
ANN ARBOR,h.lICH

An apparatus has been built for studying the effect of position expressed as per cent air by volume, T is tem-
temperature, concentration of small amounts of air, and perature in degrees Fahrenheit, and the other letters rep-
'
temperature drop on the rate of condensation of steam resent empirical constants. These individual equations
on an isothermal condensing surface. have been combined with certain simplifying assumptions
Within the experimental range and errors, the following into a single equation:
empirical equations define the coefficient as determined
by these individual independent variables : log f = log AT [1.213 - 0.00242Tl + [m
log A T
- 11 X
f = u(AT)* +
[lOg(C 0.505) - 1.551 - 0.009TI
I Several theoretical reasons for the variations indicated
f = d C + e by the equations have been pointed out in the light of
the resistance concept.
f = FgT The process of condensation of impure steam and that
wheref is coefficient, AT is temperature drop, C is com- of gas absorption have been compared,

T HAS been realized for almost a century that the rate of Later work2 to l 8 has not changed the conclusions of Reynolds

I heat transfer between fluids and solids is dependent

almost entirely on the physical properties of the fluids
which cause them to form relatively stagnant films a t contact
and other early experimenters, and many reports are to be
regarded more as efficiency tests on specific commercial
apparatus than as sources of data of general interest. Xu-
surfaces. However, experimental work has not always been inerous studies have been made, but they have not satis-
planned so that the mechanism of heat transfer through one factorily determined the effect of non-condensable gas, tem-
film could be investigated independently, and in studies of the perature, and temperature drop from steam to tube (or
condensation of vapors there has usually been included amount of heat flowing) on the rate of condensation of steam.
another physical operation, the heating of water. I n these However, the literature shows that the coefficient varies be-
reports references have been made to the amount of air pres- tJyeen 1500 and 3500 (B. t. u. per square foot per hour per 'F.),
ent, but nowhere has its effect been closely scrutinized. and it is the purpose of the present work to show the quantita-
Reynoldsipx in referring to air in condensing steam, wrote: tive effect of these conditions on this variation.
A priori it seemed probable that i t might retard condensation Apparatus
very much; for when pure steam comes up to a cold surface and
is condensed, i t leaves an empty space which is immediately BoILER-The ends of a 38-inch section of standard 12-inch
filled with fresh steam; so that the passage of steam up to the iron pipe were closed by welding on steel disks fitted with
cold surface is unobstructed.* * * If, however, the steam is
mixed with air, then the steam will he condensed and the air valves, packing boxes, etc., as shown in Figure 1. The inner
be left between the fresh steam and the cold surface; so that surface was nickel-plated after the ends were machined for
after condensation has commenced that surface will be protected the fittings.
by a stratum of air, and fresh steam will have either to displace \-.4PORIZATION SPACE AND CONDENSER-If water is passed
this or pass through i t before i t in turn can be condensed.
through a tube surrounded by condensing steam, the water
1 Abstracted from a thesis submitted in partial fulfilment of the re-
is heated and hence the temperature of the tube is not con-
quirements for the degree of doctor of philosophy a t the University of stant. I n order to avoid lengthwise variation in tube tem-
Michigan. Presented before the Division of Industrial and Engineering perature, q-ater was boiled inside the tube a t a lower pressure
Chemistry a t the 76th Meeting of the American Chemical Society, than that of the steam outside, and the vapor formed was
Swampscott, Mass., September 10 t o 14, 1928.
2 Present address, Eastman Kodak Company, Rochester, N. Y.
suitably condensed and the condensate returned. (To pre-
*Numbers in text refer to bibliography a t end of article. vent ambiguity the outer 12-inch shell will be hereinafter
June, 1929 ISDCSTRIAL d S D ESGISEERIAVG CHE-IlISTRY 577

identified as the “boiler,” and the tube system as the “rapori- zation tube would boil toward one end, the boiler rested in its
zation space.”) steel and asbestos setting a t an angle of about 5 degrees with
The construction of this evaporator tube, its auxiliary the horizontal (not shown in Figure 1). Two gas heaters
condenser, and fittings is indicated in Figure 1. The tubular were made by drilling small holes in two sections of pipe
members were brazed together and the 3-inch 0 . d by 47-inch supplied with city gas and compressed air (D in Figure 1).
evaporator tube was soldered to thin copper collars which The hot gases passed through a hood to the stack.
absorbed expansion strains. These collars were soldered to The two 2000-watt General Electric heaters (B, Figure 1)
the boiler ends. Thi2 tube was fitted with thermocouples were installed as auxiliary and more readily adjustable sources
previously attached by a nickel-plating process already of heat. Their small heat capacity and high rate of heat
described. l9 conduction to the water allowed a n almost instantaneous
COOLIKG-WATERSYSTEM-A constant-level tank ( V , change in the steam temperature t o be produced by a small
Figure 1) was installed three floors above. After passing the change in the current. The boiler temperature could be kept
valve ( J l ) and before entering the two 7/s-inch tubes, the constant within 0.01” C. for days a t a time by means of a
water flowed by a thermocouple junction and a thermometer. thermostatic regulator described in another article.20
After the tTTo 7/s-inch tubes, the water flowed back through AIR ;~IANO~IETER-A device for the analysis of the gas
the jacket, then by the otheS junction of the thermocouple and phase in situ mas necessary. The partial pressure of air
to the orifice chamber. might be found by subtracting the vapor pressure of water
An orifice chamber of 2.5 gallons capacity ( Y ,Figure 1, and from the total pressure, but the multiplicity of corrections
Figure 2) was constructed as shown. Water rose in the open and calibrations necessary to obtain this small difference
manometer attached to the can and indicated the hydrostatic accurately were undesirable. Other possible niethods21~22
pressure above the orifice plate. The orifice chamber was were also discarded and a n apparatus was devised to indicate
usually operated half filled with air, so that the compressi- the partial pressure of the air directly on a manometer.23
bility of the gas would smooth any slight fluctuations of THERMOCOUPLE SYsTmi-The accuracy of the results
pressure. The calibration curve of the rate of flow against depended largely on three measurements: (1) the tempera-
manometer readings was a straight line on logarithmic paper. ture of the steam, (2) the temperature of the tube, and (3)
The three systems of water used in the apparatus may be the rise in temperature of the cooling water.

A-Boiler body of 12-inch iron pipe A’-Thermometer for entering cooling water
B-Electric heaters, 2 General Electric 2000-watt 0-Water-difference thermocouple wires
C-Boiler drain valve P-Tube thermocouple wires
D-Gas heaters Q-Boiler manometer
E-Gas valve R-Evaporative space manometer
F-.Air valve S-Water gage in evaporative space
&Water t r a p on manometer T-Evaporative space drain valve
H-Blow-off valve U-Exhaust tube for evaporative space
I-Peep glass and hand hole packing plate V-Constant-level t a n k
J-Iron stand f o r boiler W-Float valve on city water line
K-Hood a n d flue for hot gases X-Water orifice manometer tube
L-Asbestos insulation Y-Orifice chamber
M-Valve regulating cooling water Z-Standard steam thermocouple wires

summarized: (1) the water boiling in the boiler, the steam The steam temperature and the rise in temperature of the
from which condensed on the lower tube of the vaporization cooling water were measured with thermocouples of copper-
space in which (2) water was boiling at a lower temperature constantan. The steam thermocouple was a n ordinary two-
and pressure, the steam from which was condensed both on junction type, and the “water-difference” thermocouple was
the inner wall of the upper 3-inch tube and the outer walls of composed of two of these in series. These thermocouples
the two 7/~-inchtubes through which (3) cooling water was were calibrated with a n A. H. Thomas thermometer graduated
circulated. The lower 3-inch tube was the one on which the t o 0.1’ C., accurate to 0.02” C. and standardized by the
heat transfer experiments were made, the upper tubes serving Bureau of Standards. Their hot junctions were tied to the
merely to discharge this latent heat to the cooling water, thermometer bulb and immersed in a liquid, the temperature
where it could be measured as sensible heat. of which could be varied. The cold junctions were immersed
HEATIXG SYSTEM-h order that the water in the vapori- in melting ice in a Dewar flask. (The melting point of the
578 1NDUXTRIAL AND ENGINEERING CHEMISTRY Vol. 21, No. 6
n
ice used in this and later work PROTECTION AGAINST HEAT LossEs-The boiler above the
and that of ice frozen from con- setting was covered with a layer of magnesia insulation 1*/*
ductivity water was compared inches thick. The vaporization space and its water jacket were
with an accurate resistance ther- similarly insulated, and both surfaces were given several
mometer and found to differ by coats of a varnish suspension of aluminum powder. Heat
less than 0.01" C.) The free losses from the vaporization space were small, not only be-
ends were connected by a selec- cause the temperature difference between it and the room was
tive switch to a special Type K small, but also because over half of the copper tubular rec-
potentiometer. The sensitivity tangle outside of the boiler was water-jacketed, and the heat
of the galvanometer was such as which otherwise would have been lost through the insulation
to allow accurate readings to a was absorbed by the cooling water. I n most cases the
microvolt. The standard cell average temperature of the water was lower than room tem-
was checked a t intervals against perature, and the small amount of heat flowing into the cool-
an Eppley standard cell recently ing water from the outside would tend t o balance the small
c h e c k e d a t t h e B u r e a u of amount flowing out from the rest of the surface.
Standards.
About fifty readings were Operation
taken in the desired range. The CLSSSIFICATION OF EXPERIhZENTS-In a Systematic study
method adopted for using these two of the three variables defining the coefficient must be
and similar data was to plot kept constant while the effect of the third is studied. The
them on sheets of millimeter temperature drop was altered more easily than either the
graduated cross-section paper temperature or composition of the steam, and it was changed
50 cm. in width, as a series of during progressive experiments. When it also was fixed and
nearly parallel lines, each of a steady state was reached, the several readings were taken.
which represented 10" or 20' C. Several of these runs taken a t constant steam temperature and
The thermometer used t o meas- composition, but with different temperature drops, were
ure the temperature of the enter- called a "series." To find the effect of the temperature or
ing cooling water as a zero-point composition it was necessary t o hold one of these constant,
compensator was graduated to with a different value of the other for each of several series. By
half denees. and after checking-
Y I suitable plotting and cross plotting, the coefficient could then
F i g u r e %-Orifice C h a m b e r and against the standard thermom- be studied in its relation to any single one of the three variables.
Manometer
eter was installed next the ther- REMOVAL OF AIR-The boiler was filled with distilled water
mocouple junction in the water line. The steam thermo- which was boiled to expel dissolved gases. When about half
couple was installed about 3 inches from the tube and was had been vaporized, the exhaust valve was closed and the
so shielded as to prevent loss of heat by radiation. temperature maintained above 100" C. After a few hours
TUBETHERMocouPLEs-The tube thermocouple system the air manometer indicated the presence of a few hundredths
has been described.19 The three thermocouples not damaged per cent of a non-condensable gas, and it was impossible to
in the installation checked closely and were connected in prevent a slow evolution of this gas even after weeks of boiling.
parallel. The boiler and tube were heated and maintained Nofe-The absorption of air by water*'t'j and the slowness of its evolu-
a t a constant temperature by the electric heaters. This tion from a boiling liquid26 have been mentioned by various writers. A
temperature was indicated by the standard steam thermo- liberation of absorbed gases from the surface film of water or, more prob-
coude.* , and the e. m. f . of the three-tube
thermocouples. About fifty sets of
readings were taken between 70" and
120' C. On the resulting plots the
farthest of the points was within 0.2 " C.
and the majority were within 0.1" C. of
the curve which was drawn through
them.
The temperature measurement de-
sired in this work was that of the tube
surface. I n practice, the temperature
as indicated by this calibration was the
temperature of the nickel-copper inter-
face and not that a t the surface. The
thickness of this nickel plate was known
and the temperature gradient through it
was found by Newton's law
AT
p=kA-
AL
where g is the rate of heat flow, k is ther- Figure 3
mal conductivity of nickel, A is the sur-
face area, AT is temperature difference, and AL is thickness. ably, the action of steam on iron incompletely covered by nickel t o form
hydrogen would account for t h a t found in these experiments.
Nofe-This formula may be used instead of the logarithmic equation
for the cylindrical flow of heat because the thickness of the plate is very When gas-free steam was desired, a valve on the end (not
small compared with the radius. shown in Figure 1) allowed a slight exhaust. This prevented
This correction was always evaluated and used in finding the the "air" from increasing to a measurable concentration
temperature of the surface of the tube. during the time necessary to make a series. Air was ad-
June, 1929 INDUSTRIAL AiVD EAjTGINEERINGCHEMISTRY 579

mitted when desired by connecting this valve to a source of minutes later established the value of the coefficient under the
compressed air. The slight evolution of non-condensable gas given conditions. The tables in this report include only such
was not consequential when a relatively large amount of air data which by duplication indicate the attainment of a steady
had been purposely introduced. state.
VAPORIZATION SPAcE-The boiler conditions being fixed, CHANGIKGTHE TEMPERATURE DROP-Every series was
the vaporization space was evacuated and the lower tube filled made with the temperature drop as the single independent
with water. Heat W:LS supplied to this tube by the condensa- variable. The largest temperature drop of the series was
that of the first run after which it was
necessary to lessen it. Two methods
were available-the first to reduce the
flow of cooling water, and the second
to diminish the coefficient of heat
transfer from the low-pressure steam
to the cooling water by introducing
air into the vaporization space. This
air would reduce the rate of heat
transfer from the vaporization space
to the condenser almost instantly, the
temperature of the tube would im-
mediately rise, and the heat supplied
to the boiler would have to be dimin-
ished. A new temperature drop would
be established and the second point on
the curve of coefficient versus tempera-
ture drop could be obtained. This new
temperature drop could not, however,
tion of steam on its outer surface and a part of the water was be arbitrarily set and exactly obtained, because it was not
boiled off to remove the air present. possible to know how much air to admit, but the amount
COOLING WATERSusTmf-The water in the tank and lines (as shown by the manometer, R in Figure 1) could be roughly
was drained to waste until it came to a constant temperature. judged and the new temperature drop would be approxi-
The flow was then diverted through the condenser tubes and mately that desired. A finer adjustment was sometimes
adjusted to its proper value. made by changing the water velocity.
READIKGSTaxEw-Heat was allowed to flow from the This work continued gave values of the coefficient as a
boiler to the vaporization space and thence into the cooling function of the temperature drop. At the end of a series the
water until it was thought that a steady state had been condenser and vaporization space were drained of water, the
reached. After a reasonable period, readings were made. air manometer was read, and the composition of the vapor
These were always taken in the same order: (1) steam ther- calculated a t thermal equilibrium.
mocouple, ( 2 ) tube thermocouples, (3) cooling-water differ- D A T Tax~x-Probably
~ twice as many runs as are here
ential thermocouple, and (4) water-orifice manometer. reported were made before the experimental difficulties were
Throughout the series readings were also made on the entering thought to be under control. These preliminary data have
cooling water thermometer and the air manometer. When been disregarded, but all subsequent work is tabulated except
there was no air in the boiler the air
manometer was unchanged, but when
air was present the readings during a
run were so erratic as to be useless.
The important readings of the air
manometer were those taken before
and after a series when there was no
heat flowing through the tube wall.
These were perfectly definite and their
average was taken as indicative of the
air present during the runs.
Note-Smiths reported a similar variation
of the amount of air and attributed the differ-
ence between t h a t calculated a t equilibrium and
t h a t calculated a t a steady state of heat flow
t o large deviations from Dalton's law of partial
pressures Instead, there is probably a gradient
of composition throughout the steam space of
the boiler. Thus the total amount of air pres-
ent could be found during a time when heat
was flowing only by the integration of the com-
position a t a large number of points. It must Figure 5
be noted in subsequent discussion t h a t the
average amount of air before the run is not the same as the a >erage amount those individual readings taken before the attainment of a
in the free space during a run since some will be held in a g.is film around
the tube. This work makes no distinction and its results are based on the
steady state. The three sets of series reported include (1)
total amount of air present. coefficient versus temperature drop a t a very low or "zero"
concentration of air and constant temperatures of 100O , 105",
After the first readings the apparatus was allowed to run 10 110", and l l 5 O C.; ( 2 ) coefficient versus temperature drop a t
or 15 minutes before a second set. If these checked, a steady a constant temperature of 100" C. and various constant
state had been reacheid and the coincidence of a third set 5 percentages of air; and (3) the same for 110" C.
580 INDUSTRIAL A N D ENGILVEERING CHEXISTRY Vol. 21, KO. 6
Sample Calculations If the equation for heat flow is rearranged and the appropriate
values substituted, the correction for the drop through the
The results presented were calculated from the values read nickel is 0.43" C. and the corrected value of 4T is 20.11 - 0.43
on the calibration curves with a 20-inch slide rule. The data = 19.68"C.
taken were not all in one set of units, as is unfortunately the The steam film coefficient may not be calculated.
case rather often in engineering practice. The thermocouples
It is
equal to
were calibrated in Centigrade degrees and the water manom-
eter in pounds of water per second. The potentiometer f = 4
= 2053 3600 = 883 B. t. u. per square foot
AAT 2.36 X 19.68 X 1.8
was read to the closest microvolt, and the manometer to the per hour per ' F.
closest millimeter, although it was possible to estimate to one-
tenth these values. The set of data s h o m in the tables for The relation
approximately 110"C. steam, 1.07 per cent air, and tempera- B. t. u. per hour per square foot per O F. X 4.882 = kg. cal. per
ture drop of about 20" C. will have the calculations indi- hour per square meter per O C. and the reciprocal factor
cated. 0.2048 are often useful.
The readings of the air manometer before and after the
series and their arithmetic averages were respectively: 28.05,
27.65, and 27.85 cm. The zero point-i. e., the reading when
there was no air in the boiler-was 29.10 cm. a t 110" C. The
difference between this and the average is 1.25 cm., which
when corrected for the water above the mercury is 1.25 -
1.25/13.6 = 1.157cm. of mercury corresponding to the partial
pressure of air in the boiler. Peabody's tables give the
pressure of saturated steam a t 110" C. as 107.45 cm. The
total pressure in the boiler is the sum of the partial pressures,
and the air present is thus

1.157 + 107.45
loo = 1.066 per cent

Since the temperature of the steam was held within 0.1" C. Figure 7
by the thermostat, the average of the e. m. f.'s for a series mas
converted to degrees instead of using the temperature for each An Empirical Equation
run individually. The values from the respective charts
follow: Several empirical relationships between the coefficients
obtained and the values of the independent variables were
c. noticed: (1) With constant temperature and concentration
Steam temperature 110 04 the logarithm of the coefficient is a linear function of the
Interface temperature 89 93
AT 20 11 logarithm of the temperature drop; (2) with constant tem-
Rise in water temperature 22 50
Weight of cooling water, 0 507 lb per second
perature drop the reciprocal of the coefficient is a linear
function of the composition expressed as per cent air by vol-
ume; (3) with constant temperature drop and concentration
the logarithm of the coefficient is a linear function of the
temperature.
The lines representing these data were drawn (Figures 3 to
7) without recourse to the method of least squares, but in
relation to all of the data and not merely that of the individual
series. These generalizations may, respectively, be expressed
by the equations:
Logf = log a f b log AT (1)
-1 = s c + p (2)
fog f = log h + (log F)T (3)
where f is coefficient, AT is temperature drop, C is compo-
sition (per cent air by volume), T is temperature of steam in
degrees Fahrenheit, and the other letters are constants to be
evaluated from the geometric properties of the graphs. The
logarithms in these and later equations are to the base 10.
Figure 6 Note-The points of Figure 6, used in ( 2 ) are obtained from values of
the coefficient a t 1' F. on the lines of Figures 3, 4, and 5 . The points of
The term AT represents the difference between the tem- Figure 7 used in (3) are obtained by reading values of coefficient a t a tem-
perature of the steam and that of the nickel-copper interface perature drop of 1' F. from Figure 5 where there is no air present in the
steam. The third conclusion and Equation 3 i s thus based only on experi-
of the tube, and is corrected by subtracting the temperature mental data for zero air. Values of the coefficient, when a measurable
drop through the nickel plate which is found, as mentioned amount of air was present, were taken a t only two temperatures of t h e
above, from the relation g = k A AT/4L. The heating sur- steam and the third generalization may be definitely made only for the case
face is 2.36 square feet (cylinder 35.75 X 3.025 inches) and the where there is no air present in the steam. However, the final equatlon
is applicable a t these two temperatures t o the data taken with air present.
thermal conductivityof nickel is 0.1133 B. t. u. per square foot Equation 2 and Figure 6 state t h a t if the reciprocals of values read
per second per " F. per inch. g is the product of the weight of from the lines of Figure 3 or 4 are plotted against the per cent air, the re-
water per second flowing through the condenser by the in- sulting lines will be straight. If t h a t is accepted, i t may be shown t h a t t h e
crease in its temperature, or in English units lines of Figure 3 or 4 cannot converge t o a common point although t h e
intersections of any pair of lines may be infinitesimally close. Within t h e
0.507 X 22.50 X 1.8 = 20.53 B.t. u. per second admittedly large experimental errors, this is of no consequence and since
June, 1929 I&YDCSTRIAL A.VD E;VGINEERIXG CHEJfIISTRY 581

t h e d a t a indicate t h e possibility of such convergence and t h e mathematical the temperature difference across the interface is found to be
handling is greatly simplified b y its assumption, these lines are drawn a s
very small and the corresponding resistance negligible, which
sheafs. This common intersection is, however, too far t o thti right t o have
any physical significance. T h e intersections have the same abscissas for agrees with what Reynolds1 said: “In fact there is no limit t o
both steam temperatures studied, b u t t h e ordinate of t h e higher steam the rate a t which pure steam will condense but the power of
temperature is the smaller. If t h e lines of Figure 5 representing different the surface to carry off the heat.” The total resistance may
temperatures are extended through their pole, t h e intersecl.ions with t h e
common ordinate of t h e poles of Figures 3 and 4 m a y be considered as poles
therefore be Considered as due to the gas and condensate films,
of other sheaves. X plot of the ordinates of such intersections, constant and the formula corresponding to Ohm’s IaxT inighi be written
.A in Figure 8 , is a straig’llt line.

The equations inay be combined if the geometric properties

pR AT

of the lines are considered, giving:

where g / A is the rate of heat flow per unit of surface, AT is
logf =
log AT_[log m 4- (log i O T ] Ilog (C t r) 2 log(1’v) - log ii - T log F ] the temperature drop from steam to tube, and R is the sum of
__ ii
log u resistances of gas and liquid films. Few published experi-
x [log AT - log n1 (4) ments have been expressed in units of resistance.
The constants in Equation -1 have been tabulated in English The mathematical analysis of the mechanism of heat trans-
units. fer through the liquid and vapor films is hindered by the large
number of physical properties to be evaluated. Such things
GEOMETRIIIAL as the rate of mutual diffu3ion of steam and air; the viscosity,
COXSTANT SIGKIFICAXE GRAPH vALUE
log 2, Intercept 1 Constant .A vs. tem- 4 173i apecific heat, thermal conductivity, and density of the air,
log ii Slope j perature (Figure 8 ) kiis
i- Ilistance betw8:en inter- Reciprocal of coefficient
section of l / j Y S . C vs. composition (Fig- 0.505
lines and 1/,’ axis ure 6 ) , + - ’ -
250
log L1,’r) 0.29iO
1
i
log h Intercept Coefficient vs. tempera. 0 1848
log F Slope 1 ture (Figure i ) 0,00000
log B Logarithm of abscissa Coefficient vs. tempera.
of intersection of lines ture drop a t constan’: 3.439
of different compo- T (Figures 3 and 4,
sitions

If these values are substituted in Equation 4 there follows:

log f = log J T [ 1 213 - 0.00242 1’1
LIOZ (C
+ [
Is - 13 X
+ 0.505) - 1.531 - 0.009 TI (51 Figure 8
This relation expresses in B. t. 11. per square foot per hour
per O F. t’lie logarithm of the coefficient as a function of tem- condensate, and steam; and the latent heat of the latter
perature, temperature drop, and cornposition of the steam. determine the amount of heat passed. These values all
The constants are individually evaluated in the metric system change with varying temperatures through the films, and the
and in kg. cal. per square meter per hour per O C. mean values can only be approximated. A knowledge of the
thicknws of the condensate film and of the air concentration
logf log A T l l . 4 4 2 - 0 0047Tl + r03g-
1:
- 11 and temperature a t its surface would also be valuable. Be-
[log (C f 0.505) - 2.6568 - 0.0139Tl (6) cause most of these data are not accuratrly known, a basic
theory was not formulated for comparison with the experi-
Physical Discussion mental observations. However, the effect on the coefficient
of a change of any one of the independent variables may be
Most problems dealing with potential gradients may be predicted and checked against the data.
attacked by a method understood by workers in heat transfer The first relation which n-as considered was the linear
and the analogous ficld of gas absorption t o be based on the function of the logarithms of the coefficient and the tempera-
“resistance concept.” Resistance is the reciprocal of con- ture drop. The graphs (Figures 3 and 4) start with a tem-
ductivity, and the total resistance of several substances in perature drop of 1 F., but some data were taken with lower
O

series is the sum oi tlte separate resistances, or in the transfer temperature drops. Although the errors of measuring these
of heat from a condensing vapor to a solid: small temperature differences made the coefficients so un-
1/s = l/,fa 1/JC l/jL + + reliable that they are not plotted on the graphs, they indicate
no change of the relationship.
wherej = coefficient from steam to tube
j G = coefficient across gas film Steam, in condensing, covers the tube with a film of water
fc = coefficient a t the condensation interface which is drained by gravity, but which increases in thickness
f~ = coefficient across film of condensate as the rate of condensation increases. The resistance to heat
The mechanism of condensation may be studied by Lang- transfer is increased and therefore the coefficient is decreased
niuir‘s equation2’ for the Condensation of metallic vapors. with increasing temperature drop. The condensate film is in
This may be coinbirled with the Clausius-Clapeyron equation viscous flow, and if the coefficients a t zero air could be divided
and the gas lails to give: by the film thickness, the results would be approximately
constant-i. e., the coefficient of thermal conductivity of the
‘=v.%&x dp dT
X (Tr3/2 - TL3/2) condensate. Since the coefficient is more easily determined
than the film thickness, the latter may be calculated from the
where q is the rate of heat transfer, T V and T L are absolute thermal conductivity of water which according to JakobZu
temperatures of the vapor and liquid sides of the interface, +
varies as k = 0.001325 (1 0.002984t), where k is the con-
-11is molecular weight, R is the gas constant, and dP/dT is ductivity in calories per sq. em. per second per ’ C. per em.,
the slope of the vapor-pressure curve which is assumed to be and t is Centigrade degrees up to 80” C. Preliminary work
a straight line in this very narrow range. When the experi- in this laboratory has verified this equation up to 120’ C.
mental values of g are substituted with the other quantities, The plot of the coefficients obtained with practically pure
582 I N D U S T R I A L A N D ENGINEERING CHEMISTRY Vol. 21, No. 8
steam (Figure 5) show comparatively large irregularities. but those which were taken several hours later a t lower tem-
The large amount of heat passed in these runs caused an un- perature drops are considerably lower than expected. This
even boiling in the vaporization space and a larger variation shows that a slow evolution of non-condensable gas during a
in tube temperature. Another reason was that a very slight few hours had reduced the coefficient, and the line as drawn
amount of air makes a large difference in the coefficient when is based only on the points which were taken first, when the
the steam is practically pure. The series a t 212O, 221", and steam was substantially pure.
230' F. were made with a small amount of steam exhausting, Some experimenters have indicated that a change of tem-
to prevent a building up in the boiler of an appreciable con- perature of pure steam did not appreciably affect the coeffi-
centration of non-condensable gas. The series for steam a t cient. I n the present work the coefficient is shown to increase
239" F. was obtained just after the steam had been freed from with increasing temperature by the line in Figure 7 derived
air by "boiling out," but with the exhaust valve closed during from the lines in Figure 5. This might be expected since the
the series itself. The first points obtained on this last series properties of water vary with temperature. At a given
are those with the highest temperature drops, and they are temperature drop the condensate drains faster and forms a
seen to be consistent with the lines of lower temperatures, thinner film when the viscosity is lowered by an increase of
Data
WATER WATER
TUBE DIFF. A9 AT AT TUBE DIFF. Ae AT AT
TEMP. TEMP. WATER AT NICKEL(COR.), (COR.) TEMP. TEMP. WATER AT NICKEL(COR.) (COR.)
c. a C. Lb./sec. e. c. c. 'F. OC. OC. Lb./sec. OC. 'C. 'C. O F . f
AVER A G E T E M P E R A T U R E OF STEA M 11U.04° C.-LO66 P E R CENT AIR
AVERAGE TEMPERATURE O F STEAM 99.97O C.-1.42 P E R CENT AIR
89.93 22.50 0.507 20.11 0.43 19.68 35.42 883
89.93 21.86 0.521 20.11 0.43 19.68 35.42 882 80.78 19.02 0.421 19.19 0.30 18.89 34.00 647
92 46 33.70 0.295 17.58 0.37 17.21 30.98 881 82.12 18.70 0.422 17.83 0.30 17.55 31.57 686
92.46 33.43 0.298 17.58 0.37 17.21 30.98 873 82,34 18.51 0,421 17.63 0.29 17.34 31.20 686
92.95 17.98 0.542 17.09 0.36 16.73 30.12 888 83.22 18.12 0.421 16,75 0.29 16.46 29.61 707
92.85 18.00 0.540 17.19 0.36 16.83 30.30 88! 83.89 17.93 0.421 16,08 0.28 15.80 28.44 729
98.01 14.82 0.542 12.03 0.30 11.73 21.12 1040 93 69 7.90 0.420 6.28 0.12 6.16 11.09 822
102.10 15.23 0.400 7.94 0.23 7.71 13.88 1205 93.69 7.99 0.420 6.28 0.13 6.15 11.07 832
102.20 15.69 0.400 7.84 0.23 7.61 13.70 1258 93.55 7.87 0,420 6.42 0.12 6.30 11.34 so0
102.28 15.83 0.400 7.76 0.24 7.52 13.54 1284 98.02 3.26 0.419 1.95 0.05 1.90 3.42 1097
102.17 16.58 0.402 7.87 0.25 7.62 13.72 1334 98.02 3.39 0.419 1.95 0.05 1.90 3.42 1142
106.30 9.38 0.399 3.74 0.14 3.60 6.45 1585 97.92 3.33 0.419 2.05 0.05 2.00 3'60 1065
A V E R A G E T E M P E R A T U R E O F STEAM 110.03' C.--1.96 P E R C E N T AIR AVERAGE T E M P E R A T U R E OF STEAM 99.97' C.--3.47 P E R C E N T AIR
84.35 17.34 0.538 25.69 0.35 25.34 45.28 561 90.14 10.04 0.317 9.83 0.12 3.47 17.48
84.35 17.61 0.540 25.69 0.36 25.33 45.28 572 91.31 9.24 0.317 8.66 0.11 8.55 15.39
84.57 17.29 0.541 25.47 0.35 25.12 45.20 568 91.48 9.16 0.317 8.49 0.11 8.38 15.08
91.32 25.68 0.297 18.72 0.29 18.43 33.17 631 91.58 8.98 0.317 8.39 0.11 8.28 14.90
91.55 25.56 0.297 18.49 0.28 18.21 32.78 636 95.77 4.92 0.317 4.20 0.06 4.14 7.44
91.31 25.73 0.297 18.73 0.29 18.44 33.22 632 95.90 4.95 0.317 4.07 0.06 4.01 7.21
96.00 21.34 0.296 14.04 0.24 13.80 24.85 698
96.15 21.22 0.296 13.89 0.24 13.65 24.59 702 AVERAGE T E M P E R A T U R E O F STEAM 99.99' C.-6.21 P E R C E N T AIR
96.05 21.19 0.297 13.99 0.24 13.75 24.75 697 92.18 5.71 0,302 7.81 0.06 7.75 13.95 340
100.65 16.07 0.295 9.39 0.18 9.21 16.58 784 93.18 4.64 0,302 6.81 0.05 6.75 12.15 317
100.45 16.84 0.295 9.59 0.19 9.40 16.91 806 93.41 4.58 0,302 6.58 0.05 6.53 11.75 323
100.13 16.97 0.293 9.91 0.19 9.72 .17,51 780 97.50 1.98 0,301 2.49 0.02 2.47 4.44 368
99.98 16.97 0.295 10.06 0.19 9.87 17.78 774 97.55 1.93 0,301 2.44 0.02 2.42 4.35 366
103.85 11.68 0.294 6.19 0.13 6.06 10.91 864 97.45 1.90 0,301 2.54 0.02 2.52 4.53 346
104.00 11.75 0.294 6.04 0.13 5.91 10.64 892
107.40 6.29 0.293 2.64 0.07 2.57 4.62 1094 AVERAGE T E M P E R A T U R E O F STEAM 100.U6° C.--11.44 P E R C E N T AIR
107.40 6.18 0.293 2.64 0.07 2.57 4.62 1075 74.32 10.85 0,303 25.74 0.12 25.62 46.09 196
A V E R A G E T E M P E R A T U R E OF STEAM 110.04° C.-2.89 P E R C E N T AIR 74.32 10.85 0,303 25.74 0.12 25.62 46.09 196
80.10 19.31 0.470 29.94 0.34 29.60 53.30 468 74.08 11.04 0,304 25.98 0.13 25.85 46.49 198
80.10 19.28 0.470 29.94 0.34 29.60 53.30 467 84.57 8.23 0.303 15.49 0.09 15.40 27.70 247
80.45 19.24 0.470 29.59 0.34 29.25 52.65 472 84.12 8.57 0,303 15.94 0.10 15.84 28.50 250
92.48 13.67 0.470 17.56 0.24 17.32 31.18 566 84.12 8.50 0,301 15.94 0.10 15.84 28.50 247
92.62 13.72 0.470 17.42 0.24 17.18 30.94 572 84.29 8.33 0,303 15.77 0.09 15.68 28.20 246
94,82 11.92 0.471 15.22 0.21 15.01 27.06 57 1 AVERAGE T E M P E R A T U R E O F STEAM 99.96' C.-ZERO AIR
94.72 11.92 0.471 15.32 0.21 15.11 27.20 567
94.72 11.99 0.471 15.32 0.21 15.11 27.20 570 89.50 23.38 0.508 10.46 0.45 10.01 18.01 1810
99.98 13.73 0.298 10.06 0.15 9.91 17.83 630 89.10 24.11 0.508 10.86 0.46 10.40 18,72 1797
99.28 14.55 0.295 10.76 0.16 10.60 19.08 617 89.10 23.97 0.508 10.86 0.46 10.40 18.72 1787
107.92 4.63 0.294 2.12 0.05 2.07 3.73 1004 89.32 23.80 0.508 10.64 0.45 10.19 18.34 1810
922 92,l.i 18.so 0,508 7.81 0.36 7.45 13.41 1957
107.72
107.76
4.67
4.63
0.294
0,294
2.32
2.28
0.05
0.05
2.27
2.23
4.09
4.02 932 . 91.63
91.86
19.24
19.05
0.507
0.507
8.33
8.10
0.37
0.36
7.96
7.74
14.33
13.93
1871
1904
AVERAGE T E M P E R A T U R E O F STEAM 110.04' C.-4.53 P E R C E N T AIR 93 88' 16.36 0.507 6.08 0.31 5.77 10.39 2196
78.55 20.77 0.345 31.49 0.27 31.22 56.20 350 93.9s 16.5s 0,507 5.98 0.32 5.66 10.20 2266
78.55 20.65 0.345 31.49 0.26 31.23 56.20 348 93.98 16.71 0.506 5.98 0.32 5.66 10.19 2280
78.63 20.65 0.345 31.41 0.27 31.14 56.00 349 96,70 12.32 0.506 3.26 0.23 3.03 5.45 3138
87.94 16.00 0.344 22.10 0.21 21.90 39.40 383 96.70 12.06 0.506 3.26 0.23 3.03 5.45 3071
87.94 15.98 0.343 22.10 0.21 21.89 39,40 382 96.70 12.68 0.505 3.26 0.24 3.02 5.43 3238
87.94 16.05 0.344 22.10 0.21 21.89 39.40 384
11.39 0.342 14.51 0.14 14.37 25.85 413 AVERAGE T E M P E R A T C R E O F STEAM 1U5.02° C.-ZERO AIR
95.53
95.53 11.35 0.342 14.51 0.14 14.37 25.85 412 98.44 19.11 0.502 6.58 0.36 6.22 11.20 2363
99.28 9.98 0.301 10.76 0.11 10.65 19.17 430 98.81 18.38 0,493 6.21 0.34 5.87 10.57 2355
99.05 10.18 0.302 10.99 0.11 10.88 19.58 431 99.04 18.25 0.493 5.98 0.34 5.64 10.15 2435
9 s . 95 10.31 0.302 11.09 0.12 10.91 19.75 432 102.96 10.04 0,491 2.06 0.18 1.88 3.38 4000
98.80 10.31 0.302 11.24 0.12 11.12 20.00 427 103.64 8.25 0,489 1.38 0.15 1.23 2.21 5005
102.95 7.03 0.302 7.09 0.08 7.01 12.62 462
102.82 6.96 0.302 7.22 0.08 7.14 12.85 449 AVERAGE T E M P E R A T U R E O F STEAM 1U9.97' C.-ZERO AIR
102.87 7.03 0.301 7.17 0.08 7.09 12.76 455 101.91 23.21 0.506 8.06 0.44 7.62 13.71 2352
106.50 3.92 0.304 3.54 0.04 3.50 6.30 519 101.96 22.89 0.507 8.01 0.44 7.57 13.62 2339
106.50 3.93 0.304 3.54 0.04 3.50 6.30 520 101.82 23.28 0,506 8.15 0.44 7.71 13.88 2332
A V E R A G E T E M P E R A T U R E O F STEAM 110.U4° C.-6.69 P E R C E N T AIR 103.36 20.51 0,508 6.61 0.39 6.22 11.20 2546
103.27 20.27 0.506 6.70 0.38 6.32 11.38 2477
71.42 12.90 0.410 38.62 0.19 38.43 69.20 210 0.39 6.26 11.27 2553
38.36 0.20 38.16 68.64 213 103.32 20,70 0.506 6.65
71.68 13.02 0.410 103.36 20.60 0.506 6.61 0.39 6.22 11.20 2557
71.85 12.95 0.410 38.19 0.19 38.00 68.40 213 3.89 0.31 3.58 6.44 3463
9.35 0.407 23.01 0.14 22.87 41.18 254 106.08 16.07 0.506
87.03 105.88 16.58 0.505 4.09 0.31 3.78 6.80 3380
87.08 9.32 0,407 22.96 0.14 22.82 41.10 233 0.32 3.98 7.16 3260
23.09 0.14 22.95 41.32 259 105.67 16.84 0.505 4.30
86.95 9.51 0,409
96.00 6.06 0.412 14.04 0.09 13.93 25.12 273 AVERAGE T E M P E R A T U R E O F STEAM 115.04' C.-ZERO AIR
95.77 6.01 0.413 14.27 0.09 14.18 25.53 267 11.91 2600
0.413 14.27 0.09 14.18 25.52 267 108.00 37.35 0,302 7.04 0.42 6.62
95.77 6.00 36.61 0.303 7.04 0.42 6.62 11.91 2559
101.68 7.11 0.300 8.36 0.08 8.28 14.90 393 108.00
15.82 372 107.93 36.52 0.306 7.11 0.42 6.69 12.05 2550
101.17 7.11 0,302 8.87 0.08 8.79 0.48 7.94 14.29 2454
101.63 7.11 0.302 8.41 0.08 8.33 14,99 393 106.62 23.86 0.535 8.42
7.91 14.24 418 106.91 23.68 0,535 8.13 0.47 7.66 13.79 2523
102.05 7.20 0.301 7.99 0.08 8.01 0.47 7.54 13.57 2513
101.78 5.92 0,303 8.39 0.07 8.32 14.98 329 107.03 23.21 0.535
6.34 445 112.82 11.98 0.473 2.22 0.21 2.01 3.62 4300
106.58 3.39 0.303 3.56 0.04 3.52 2.00 3.60 4443
108. 54 3.52 0.303 3.58 0.04 3.54 6.61 459 112.82 12.32 0.473 2.22 0.22
6.59 446 112.73 12.32 0.474 2.31 0.22 2.09 3.76 4260
106,54 3.53 0.303 3.70 0.04 3.66
 June, 1929 INDUSTRIAI, ,4ND ENGINEERING CHEMISTRY 583

temperature, and from Jakob's equation the thermal con- crease of 0.5 per cent air decreases it to a third, a third in-
ductivity increases. Both tend to increase the coefficient a t crease of 0.5 per cent decreases it to a fourth, etc. I n prac-
higher temperatures. tice, air in a concentration of only a few per cent in the steam
The enormous decrease of the rate a t which nearly pure seriously reduces the capacity of steam-condensing equipment.
steam condenses when a slight amount of air is added shows The points with t'he largest amounts of air in both the 212"
that the non-condensable gas must be localized around the and 230" F. series are seen to be off the line as determined by
tube. The decrease of the coefficient by 50 per (cent caused the other points in the graph in Figure 6. This is not believed
by adding 0.5 per cent air cannot be explained if the air is to indicate the failure of the relation between the coefficient
evenly distributed, since there would then be almost no and the composition for steam with 6 or more per cent air,
change in the partial pressure of steam a t the interface and but was due instead to experimental difficulties. Until the
condensation would take place unimpeded. Thc>reis, how- runs with large concentrations of air were made, the reading
ever, a backward diffusion of the air against the flow of steam made with the greatest ease and controlled with the greatest
and a steady state is reached when the rate of diffusion of the facility was that of the steam t'emperature. When, however,
air away from the tube is balanced by the rate at which it is the air exceeded 4 per cent, this temperature varied consider-
carried up to it. Obliously, the steady state is reached when ably during readings, even though the pressure was kept
the rate of diffusion of the steam through this air film is just constant by the relay circuit. Since it could not be accu-
equal to the rate a t which condensation niay take place as rat'ely determined, the temperature drop and coefficient varied
determined by the resistance of the condensate film. This over a n even greater range. The plotted values at these
rate of heat transfer through the condensate film probably higher concentrations of air were therefore given less weight
depends only on the partial pressure of the steam :it the inter- in drawing the representative lines.
face and is independent of the presence or absence of air if the Throughout the whole of this work no mention has been
partial pressure of steam is unchanged. made of the mass velocity of the steam with regard to the tube.
The proress of heat transfer from a vapor containing a non- With flow of vapor past the heating surface, vapor velocity
condensable gas is seen from these considerations to be ex- will affect the coefficient by changing the thicknesses of the
actly analogous to the problem of gas absorption developed gas and liquid films. It will probably change the gas film
by Lewis and Whitman29 and others. The air film is com- much more than the liquid film and thus in the case where
parable to the stagnant gas film in the absorption of a gas pure vapors are being condensed, its effects will be compara-
from a mixture with one which is inert, and both are caused tively small. From the hydromechanics of condensate drain-
by molecular diffusion. I n the case of a pure vapor passing age it is evident that velocity of steam will change the co-
over a liquid absorbent, Lewis and Whitman show that the efficient differently for every type, angle with horizontal, and
resistance of the vapor film is negligible compared with the length of tube surface.
resistance of the liquid film. The absence of a film between
Acknowledgment
pure vapor and condensate has been shown in another experi-
ment in this laboratory, which will not be described here. The writer desires to express his indebtedness to Prof.
The heat flowing from steam to tube passes first through W. L. Badger, under whose supervision this research was
the gas film if air is present, then through the condensation conducted. An acknowledgment is due to R. V. Miller,
interface, and finally through the film of water. The recip- Edward Ronda, and especially H. R. Wilson, seniors in the
rocal of the coefficient represents the resistance and when Department of Chemical Engineering, for their help in the
plotted (Figure 6) it is seen to increase as a linear function of calibration of apparat'us and in the calculation of experi-
the per cent of air present. These lines d o not intersect a t mental data. Thanks are also extended t o the Swenson
the origin, and the resistance is not zero a t zero air because Evaporator Company, of Harvey, Ill., for assistance in the
of the resistance of the condensate film. The intersection of purchase of the equipment.
the lines on the graph does occur, however, a t a point where
Bibliography
the resistance t o the flow of heat would be zero or the coeffi-
cient is infinity (since 1/03 = 0) and at this point the thick- 1-Reynolds, Proc. Roy. Soc., S o . 144 (1873).
ness of the water film approaches a n infinitesimal. 2-Joule, Phil. Trans. R o y . Soc., 151 (1861).
3-Dalby, Bibliography, Inst. Mech. Eng. (1909).
From the fact that a t zero air the only resistanre is that of 4--Badger, "Heat Transfer," Chemical Catalog Co., 1925.
the condensate film, it is apparent that as the amount of air 5-Smith, Engineering, 81, 395 (1936).
is increased the proportional effect of the condensate resist- 6--Kerr, Mech. Eng., 35, 731 (1913).
ance t o the total resistance becomes less and a given change 7-Robinson, J. IND.ENC.CHEM.,12, 644, 818 (1920).
8--Josse, Engineering, 86, SO2 (1908).
of the thickness of the condensate film affects the coefficient to 9-McBride, J . A m . Soc. diech. Eng., 30, 797 (1938).
a smaller degree. A t a given temperature of steam the 10-Weighton, Trans. Inst. A'QVQ~ Arch., 48, 123 (1906).
thickness of the water film is largely dependent on the amount 11-Royds, "Heat Transmission in Boilers, Condensers, and Evaporators,"
of heat flowing, which is in turn dependent on the temperature D. Van Xostrand Co., 1921.
12-Orrok, J . Am. Soc. Mech. Eng., 34, 734 (1912).
drop. The generalization which may be drawn from this is: 13-Callendar and Ticolson, Engineering, 64, 481 (1897).
Under constant temperature conditions, the rate of change 14-Clement and Garland, Univ. Ill. Eng. Expt. Sta., Bull. 40 (1909).
of coefficientwith respect to temperature drop varies inversely 15-McAdams and Frost, J. IND.ENG. CHElr., 14, 13 (1922).
with concentration of air. Nathematically this may be 16--'i?'ebster, Trans. Eng. Shipbldrs. Scotland, 67, 58 (1913).
17-Porter, Engineering, 107, 125, 140 (1919).
formulated 18--Anon, I b i d . , 107, 145 (1919).
1 19-Othmer and Coats, IND.ENC.CHEY.,20, 124 (1928).
20-Othmer, I b i d . , Anal. Ed., 1, 97 (1929).
al-Geberth, Ibid., 16, 1277 (1923).
The graphs in Figures 3 and 4 show the rate of change of the 22-Palmer and Weaver, Bur. Standards, Tech. Paper 249 (1924).
logarithm of the coefficient with the logarithm of the tem- 23-Othmer, IND. END.CHEX.,Anal. Ed., 1, 46 (1929).
24-Tamamushi, Bull. Chem. Soc. J a p a n , 1, 173 (1926).
perature drop (the slopes of the indicated lines) to decrease 25-Campbell, Trans. F a r a d a y Soc., 10, 197 (1914).
with increasing concentrations of air. 26--Porter, J . ROY. Tech. Coll. (.G ~ Q S P,,
I . ,
J WS)o. . 2. 19 11925)
The curves of the reciprocal of the coefficient against com- 27--Langmuir, Phys. Rev., 2, 331 (1913); 8, 149 (1916); J . A m . Chem. Soc.,
38, 2221 (1916); Trans. F a r a d a y Soc., 17, 607, 621 (1921).
position (Figure 6) show that adding- 0.5 -per cent air decreases 28--Jakob, A n n . Phrsik, 63, 536 (1920).
the coefficient to one-half its value a t zero air, a second in- 29-1,ewis and Whitman, IND. ENC. ~ H E M 16, . , 1215 (1924).