Working PaperPDF Available

Translation of Koch, J., 1901.

Authors:
  • Independent Researcher

Figures

Content may be subject to copyright.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 475
Translated by Alastair B. McDonald 20 June 2017
Öfversigt af Kongl. Vetenskaps-Akademiens Förhandlingar 1901. No 6.
Stockholm.
Communication from Upsala University Physics Department.
Contribution to the Knowledge of Heat Absorption by
Carbon Dioxide.
By John Koch.
(Communicated on 12th June 1901 by K. Ǻngström.)
1. Introduction.
The absorption of radiant heat by carbon dioxide has been the subject of a large number of
investigations. One wants to determine the amount of the total radiation absorbed from different
heat sources and for different concentrations of carbon dioxide, and also the amounts of the
absorption spectrum affected
1
.
Studies of the first type have been carried out by Franz
2
, Magnus
3
, Tyndall
4
, Lecher & Pernter
5
, Lecher
6
, Rontgen
7
, Heine
8
, Keeler
9
, Angstrom
10
, Kurlbaum
11
, and
Arrhenius
12
. Against several of these
1
Because studies of this type are a little off the subject of this small work question, we need only recall
the researches of Angstrom, Paschen and Rubens & Aschkinass through which the absorption spectrum
of carbon dioxide up to 20 µ is known.
2
R. Franz, Pogg. Ann. 94, p. 337, 1855.
3
G. Magnus, Pogg. Ann. 112, p. 514, 1861.
4
J. Tyndall, Contributions to Molecular Physics in the Domain of Radiant Heat,1872.
5
E. Lecher & J. Pernter, Wied. Ann. 12, p. 180. 1881.
6
E. Lecher, Wied. Ann. 12, p. 467, 1881.
7
H. Heine, Weid. Ann. 16, p. 441, 1882.
8
W. C. Rontgen, Wied. Ann. 23, p. 259, 1884.
9
J. E. Keeler, Americ. Journ. of Sc. 28, p. 190, 1884.
10
K. Ǻngström, Wied. Ann. 39, p. 267, 1890.
11
F. Kurlbaum, Wied. Ann. 61, p. 417, 1897.
12
Sv. Arrhenius, Ann. der Physik, 4, p. 690, 1901.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 476
Translated by Alastair B. McDonald 20 June 2017
studies, however, objections can be made. Franz had the carbon dioxide enclosed in long tubes
whose ends were sealed with glass plates, which apparently absorbed most of the infra-red
spectrum. Magnus, Lech & Pernter, and Lech had both the heat source and the meter
(thermopile) placed in the same enclosed space in which the carbon dioxide was included,
without taking into consideration that the behaviour of the thermopile in air and in carbon
dioxide are not directly comparable with each other.
Röntgen and Heine specified absorption by measuring the pressure increase that occurs in an
enclosed volume of gas from heat absorption; how large a fraction of the total radiation was
absorbed, one can hardly calculate from their data.
Moreover, we have worked to determine the absorption in general of carbon dioxide layers at
different pressures. However, it is of great interest to know how the absorption of carbon
dioxide at constant pressure varies with the layer thickness. From Prof. K. Angstrom's studies
previously published here, it is established that the absorption is not, as one would previously
have assumed, independent of the gas density (that it is unchanged when the product of gas
density and layer thickness remains the same)
1
. A reinvestigation of the question of the heat
absorption of the carbon dioxide thus appears desirable for several reasons, which is why Prof.
K. Angstrom asked me to investigate the absorption of carbon dioxide in layers of different
lengths using low temperature heat sources.
2. Method
In my investigations, I have used a method which is the same as that specified by Professor
Angstrom
2
, so a
1
K. Ǻngström, Öfversigt af K. Sv. Vet.-Akad. Förhandl. 58, p. 371, 1901.
2
K. Ǻngström, Wied. Ann. 39, p 267, 1890.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 477
Translated by Alastair B. McDonald 20 June 2017
detailed description would be superfluous - I will only briefly mention the main features.
From a carbon dioxide cylinder, the carbon dioxide was passed through drying tubes containing
phosphoric anhydride into the test tube, provided with a suitable number of junctions, and the
ends were closed with plates of rock salt. The test tube was also connected to an air pump and a
mercury manometer. A blackened platinum coil was used as a heat source, which was set in a
highly reflective sleeve, and by means of an electric current was heated to ~100°C and to
~300°C consecutively, with the current intensity being read on a precision ammeter. In a couple
of experiments, I used a hollow body as a heat source, which was heated by means of steam to
about 100°C. Between the heat source and the test tube was a portable water shield. The
radiation was measured by means of bolometers. The galvanometer employed by me was of an
Angstrom type and had, when measured, an internal resistance of 10 ohms.
Moreover, the whole arrangement was well protected against air currents and rapid temperature
fluctuations by paper screenings and cotton.
In order to determine the absorption of carbon dioxide, the radiation was observed through a
tube filled alternatively with dry air and dry carbon dioxide. It was assumed here that the air
exerts no appreciable absorption, although the air was not freed from its carbon dioxide, but the
error thus generated is entirely within the range of observational errors.
3. Results.
With the above, I have determined the carbon dioxide absorption in each layer of 12, 20, 31, 98
and 389 cm in length
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 478
Translated by Alastair B. McDonald 20 June 2017
and at three different pressures: ½, 1 and 2 atm. In Tables 1 and 2, the results are presented, and
the values here denote averages of several (5 - 10, often more) individual determinations.
Table 1 Source temperature: approximately 100 ° C.
Layer
thickness
in cm.
Absorption in per cent at pressure in
atmospheres of
0.5
1 (=760mm)
2
12
5.9 (6.1)
8.1
-
20
7.9 (8.1)
9.5
11.3 (10.6)
31
8.9 (9.2)
10.2
-
98
9.9 (10.2)
11.7
13.7 (12.9)
389
-
-
-
Table 2 Source temperature: approximately 300 ° C.
Absorption in per cent at pressure in atmospheres
1/2
1 (=760mm)
2
4
-
-
-
-
6.0 (6.2)
7.6
9.6 (9.0)
-
-
-
-
-
8.6 (8.9)
10.5
12.6 (11.8)
16.2
11.9 (12.3)
13.1
16.0 (15.0)
-
From these tables we see that the absorption of carbon dioxide in 98 cm at 2 atm pressure [2 * 98
= 196 cm atm], for example, is greater than the absorption in 389 cm at ½ atm pressure [0.5 *
389 = 194.5 cm atm] of carbon dioxide, as Prof. Angstrom's investigations require, the
absorption increases relatively with the pressure. An approximate measure of this increase, we
obtain the following observation. The absorption of carbon dioxide in 98 cm with 4 atm
pressure [392 cm atm] gave 16.2% (for a heat source of 300°C),
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 479
Translated by Alastair B. McDonald 20 June 2017
while (according to Table 2) 389 cm of carbon dioxide at 1 atm pressure absorbed 13.1% - so
the increase of the absorption due to the pressure is 23.7% or by 5.9% per atm. With the help
of these approximate determinations I have calculated the values at ½ and 2 atm pressure by
converting the 1 atm value to half and double the layer thickness respectively, and the
numbers in () indicate the values thus obtained.
The information in Tables 1 and 2 provides the values for the absorption shown graphically in
Fig. 1. Curve I relates to Table 1 (with the part of the curve designated ─ ─ ─ ─ obtained by
extrapolation), curve II to Table 2. The ordinates are the absorbed heat in percentage of the
total radiation, and the abscissas are the corresponding layer lengths plotted in cm per 1 atm
pressure; here I have reduced the observations made at ½ and 2 atm pressure at 1 atm by using
for layer thickness, the product of length in cm and pressure in atm, and the numbers in () were
taken as absorbance values.
Thus, we see with the length of the layers that in the beginning the absorption increases very
quickly, but then fairly slowly. Thus, 50 cm carbon dioxide at 1 atm pressure, for example,
already absorbs 9.2% of the radiant heat source of 300°C, while at 200 cm only 12.0%, at 400
cm 13.2%, and at 800 cm about 15% is absorbed - the last number is somewhat uncertain.
The agreement between my values and Keeler's determination
1
of the absorption of carbon
dioxide in 340 cm at 1 atm pressure (= 11%) is fairly good. He used as a heat source
1
J. E. Keeler, loc. cit., p. 197.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 480
Translated by Alastair B. McDonald 20 June 2017
a sooted copper sheet at approx. 270°C. Measurements were made without the use of
rock salt plates; it is calculated (see below), that he would have found using these 11.5
-12%, while the corresponding figure with me is 12.8%.
In contrast, my curve differs very significantly from that which Professor Arrhenius
1
found. He determined the absorption in a layer of carbon dioxide 50 cm in length and 1
- 8 atm pressure. In Table 3, I give the absorbance values found by him for a heat
source of 100°C, and for comparison also that obtained by me, the latter curve I, taken
from Figure 1. In column 4, the differences in percentage between the corresponding
numbers are in column 2 and 3; finally, in column 5 are the differences in percentage
per atm.
Table 3
Layer
thickness
in cm.
Absorption in per cent by
Difference
in percent
Difference in percent
per atmosphere
Arrhenius
Koch
50
10.4
10.6
-
-
140
14.3
12.3
16.3
5.8
155
15.0
12.5
20.0
6.4
170
15.8
12.7
24.4
7.2
200
15.9
13.0
22.3
5.6
275
18.1
13.4
35/1
6.4
300
18.3
13.6
34.6
5.8
400
20.0
14.2
40.8
5.1
Average: 6.0
We see that the agreement is also complete for a layer length of 50 cm, where
the pressure of Prof. Arrhenius' investigation was at 1 atm and the
observations were available under the same experimental conditions as for
me, but that his values grow significantly faster with the length of the layer
than mine. We also see from the
1
Sv. Ahhrenius, loc. cit., p. 692.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 481
Translated by Alastair B. McDonald 20 June 2017
figures in column 5 of Table 3 above, when we compare it with the determination of the change
in pressure by atm on page 478, that this deviation can be explained entirely from the influence
of pressure.
Such a disparity prevails between the results of Tyndall's and my estimates, as if they were
made under different experimental conditions. In Table 4, I give Tyndall's estimations
1
for the
absorption of the radiation from a heat source of 100°C in carbon dioxide, converted into
percentages as with those of Prof Very
2
and Prof Arrhenius
3
.
Tab. 4.
(360 units according to Arrhenius' assumption (p. 19 Contrib.)).
Total radiation (
(334 " Very's " (pp. 18-19 " ))
We see that Tyndall's absorption curve has a different look to the mean. However, this is
nothing more than what we could expect in advance. We need to
1
J. Tyndall, loc. cit., p. 37.
2
F. W. Very, U. S. Department of Agriculture, Bull. G., p. 87, 1900.
3
Sv. Arrhenius, ibid, p. 692.
Pressure in
inches (Layer
thickness = 120
cm)
Absorption from
Approx. layer
thickness in cm
(Pressure = 1
atm from
Arrhenius)
Tyndall
in galv. degrees
Very
in percent
Arrhenius
in percent
0.5
5.0
1.50
1.4
2.03
1.0
7.5
2.25
2.1
4.06
1.5
10.5
3.14
3.0
6.09
2.0
14.0
4.19
4.0
8.12
2.5
17.8
5.33
5.1
10.15
3.0
21.8
6.53
6.1
12.18
3.5
24.5
7.34
6.7
14.2
5.0
25.0
7.45
6.9
20.3
10.0
36.0
10.78
10.0
40.6
15.0
42.5 (48)
(14.37)
11.8
60.9
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 482
Translated by Alastair B. McDonald 20 June 2017
remember that these results of Tyndall, have been made at fairly low pressures, and partly
also that Tyndall worked in these experiments with a brass tubes highly polished inside
without diaphragms, in which therefore a disturbing forced reflection was present.
1
I have with Table 4 also wanted to show the uncertainty, which adheres to Tyndall’s
converted, partly due to the fact that Tyndall has not specified the absorption in
percentages, but in galvanometer degrees and without indicating the strength of the total
radiation unambiguously.
In contrast, there is relatively good agreement between Tyndall's and my estimates for the
absorption of the radiation from a heat source 270-300°C. It is important to note that these
were all made at atmospheric pressure.
In table 5, some of Tyndall's values
2
are shown.
Tab. 5.
Thickness
in inches
Absorption
in per cent
Thickness
in cm
0.01
0.86
0.03
0.05
2.1
0.13
0.1
3.3
0.25
0.5
5.7
1.27
1.0
6.3
2.54
1.5
6.7
3.81
2.0
7.6
5.08
12.0
8.4
30.8
33.0
8.0
83.8
The last two items in the above table were not found directly in percentage but are calculated.
According to Table I and II on page 407 of Contributions, in fact the absorption is in 12 inches
carbon dioxide = 37°, and the
1
J. Tyndall, loc. cit., p. 35 - 38.
2
J. Tyndall, loc. cit., pp. 80, 170, 407.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 483
Translated by Alastair B. McDonald 20 June 2017
total radiation = 79.8° (= 440 units to the specifications on page 260 loc. cit.). The absorption is
therefore 8.4%.
We also find in Table 1 p. 80 of Contributions as relative values of the absorption for carbon
dioxide 90, and for ammonia 1195 (layer thickness = 33 Engl. Inches. Pressure = 1 atm.).
Furthermore, Tyndall observed that, when the tube was filled with ammonia and a metal screen
is interposed between the heat source and the thermopile, it increased the deflection very
slightly. Since the number 1195 therefore represents almost 100% absorption, then we find that
the value for the absorption of carbon dioxide in 83.8 cm is about 8%.
As mentioned, here the agreement where the observations were carried out under similar
conditions, is even better.
In addition to the observations stated in Table 1 on page 478 relating to carbon dioxide I
measured absorption in layers of 31 and 12 cm respectively at different pressures, which I
recount only in passing here (Table 6). The results for the 12-inch tube values are obtained by
interpolation between 50 to 60 direct measurements.
Table 6.
Temperature of Radiation: at 100°C
Pressure
In mm Hg
Absorption
31 cm tube
12 cm tube
760
10.2
8.1
647
10.1
7.7
548
9.8
7.2
445
9.4
6.5
345
8.6
5.6
242
7.3
4.4
123
4.7
2.9
99
4.4
2.1
52
3.4
1.1
23
1.3
0.5
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 484
Translated by Alastair B. McDonald 20 June 2017
4. The influence of the rock salt plates on the above results.
The above provisions apply to the absorption of carbon dioxide between 5-7 mm thick rock salt
plates. To determine the influence of these plates on the results found, I carried out some
experiments which I will report briefly here.
Adjoining Figure 2 gives us a schematic representation
of the arrangement.
M is a vertically standing, inside blackened brass tube,
consisting of two parts (I and II), which can be pushed
into each other with friction; into the lower end of
which the bolometer tube B is inserted. S is the heat
source (a hollow body of about 100°C), Gf and Gr are
respectively a fixed and a movable water shield. F is a
fixed diaphragm, two of those are also mounted in the
tube M II. Carbon dioxide can be inserted using tap T
either through the tube or through the tube Rm M and B
in RB.
Fig. 2.
In all experiments using this method in which the tube was thus open upwards, the carbon
dioxide flows through the tube slowly RM to compensate for the loss by diffusion to the upper
end of the tube M.
From Prof. Kurlbaum's investigations
1
, we know that the details of the bolometer are
considerably different for the same irradiation in air and carbon dioxide. To investigate this
change in my instrument, I first went
1
Kurlbaum, loc. cit.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 485
Translated by Alastair B. McDonald 20 June 2017
the following way. The mouth, C, of the bolometer tube was covered with a rock salt (or
fluorite) glass. M was filled with dry carbon dioxide, while air was in B, and the radiation (ul)
determined. Thereafter, the air was replaced by dry carbon dioxide and B simultaneously
pushed down the tube MI to the mark d, so that the carbon dioxide layer on the exposed grid
bolometers also now had the same length as before. The strength (uk) of the radiation was
determined anew. The same quantity of heat was seen in both cases on the bolometer, thus
giving uk /ul as the relationship between the data of the bolometer for carbon dioxide and air.
Below in Table 7, are recorded the values that I have received in this way. Above each column is
given the substance that made up the cover plate.
Table 7
Fluorspar
Thickness: 1 mm.
Rock salt
Thickness: 5 mm.
1.100
1.103
1.104
1.103
1.102
-
1.105
-
1.101
-
This results as the mean value uk/ul = 1.103. (It hardly needs to be said that this value only
applies to the instrument investigated here.)
I have now determined the absorption by carbon dioxide without the use of rock salt plates -
from which with the help of Table 1, we would obviously be able to calculate the correction for
the absorption by the rock salt plates - I had tried to come to this correction in a more direct way.
In the following, with the help of the corrections determined above, I have converted the results
obtained without rock salt plates
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 486
Translated by Alastair B. McDonald 20 June 2017
into the values which they would have yielded with such results, in order finally to
compare these values so calculated with those previously found (Table 1 and Figure 1).
For the determination of carbon dioxide absorption, the cover plate was removed before the
stream C and the observed intensity of the radiation, B and M when the tubes were filled with air
and carbon dioxide. Since we know the change in the particulars of the bolometer definitely, we
can calculate from the absorption by carbon dioxide under these conditions, namely that the
absorption of the air can be neglected. The error, which we commit here, is in any case entirely
within the observational errors. In Table 8, I give the result of the absorption in these conditions.
Tab. 8.
Layer thickness
in cm of 1
Atm. Pressure
Absorption
in percent
54.8
9.7
53.4
9.5
49.2
9.5
48.4
9.4
43.4
9.3
22.8
9.0
To obtain the correction for the absorbance of the rock salt plates more directly, the following
procedure was used.
A system of two rock salt plates - I carried out the same investigation for fluorspar - was with
some space between them, in a frame on the screens F (Fig. 2) placed so that they were always
placed in the same position. The ratio (g) between the radiation with and without plates was
observed, while various carbon dioxide layers were between the heat source and the bolometer.
The values found are reported in Table 9.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 487
Translated by Alastair B. McDonald 20 June 2017
Table 9.
Carbon dioxide
in cm
g
Rock
salt
Fluorspar
0.0
0.756
0.510
22.8
0.749
0.545
42.7
0.749
0.548
49.0
0.748
0.548
From those applied in this experiment, the rock salt plates they were exactly the same as those
applied in the 31 cm tube one was 3 mm and the other 5 mm thick; the two fluorite discs each
had a thickness of 1 mm.
By means of this table, we can now calculate the correction for the absorbance of the rock salt
plates.
For air we had gl = 0.756
and for carbon dioxide gk = 0.748.
Taking a as the uncorrected carbon dioxide absorption and a' as the corrected, then obviously
(1 - a ') / (1 - a) = gl / gk
(so the effects of the reflection at the surfaces is eliminated by this method) from which, since a
is known by the preceding experiments, we obtain a'/a = 0.91.
This result is in exceptionally good agreement with that calculated by Prof. Arrhenius
1
from
the observations of Rubens and Trowbridge
2
(= 0.93). In contrast, Prof. Very
3
has found this
correction significantly greater (= 0.75); However, the difference is explained by the fact that
the latter has not eliminated the effects of the reflection at the surfaces.
1
Sv. Arrhenius, loc. cit., p. 697.
2
H. Rubens & J. Trowbridge, Wied. Anu. 60, p. 724. 1897.
3
F. W. Very, loc. cit., p. 108; Astrophys. Journ., Vol. 8, p. 211, Nov. 1898.
ÖFVERSIGT AF K. VETENSK.-AKAD. FÖRHANDLINGAR 1901, No 6. 488
Translated by Alastair B. McDonald 20 June 2017
Finally, in Table 10, I give for comparison the values converted from Tab 8, which thus here
apply to carbon dioxide between rock salt plates, along with those taken from the corresponding
curve I in Fig 1.
Table 10
Carbon dioxide
in cm at
1 atm pressure
Absorption in percent from
Table 8
(calc.)
Curve I Fig. 1
54.8
10.7
10.7
53.4
10.4
10.6
49.2
10.4
10.5
48.4
10.3
10.4
43.4
10.2
10.4
22.8
9.9
9.6
One sees that the agreement between the corresponding figures in Col. 2 and 3 of the above
table is quite satisfactory.
Thus, in this small work the absorption capacity of the carbon dioxide is determined for
different thicknesses - between rock salt plates - at constant pressure (½, 1 and 2 atm), and also
the influence of the absorption of the rock salt plates on the carbon dioxide absorption.
-------------------------
ResearchGate has not been able to resolve any citations for this publication.
Contributions to Molecular Physics in the Domain of Radiant Heat
  • Tyndall
Tyndall, Contributions to Molecular Physics in the Domain of Radiant Heat,1872.
  • Sv
  • Arrhenius
Sv. Arrhenius, Ann. der Physik, 4, p. 690, 1901.