44 CUCUO W-CALIBRA TION OF REFERENCE STANDARD HYDROMETERS J.S.C.I.

,68, February, 1949

equally applicable to the generation of BHC smokes. T h e The present method, in contrast, pennits the calibration of
efficiency of the composition hydrometers of any range in a single liquid whose density is
Thiourea
Urc! .... .... X6.9
I ' 4 parts
BHC ..
pans
Potassium chlorate 24.5 parts
.. [Link]
not deliberately altered throughout the whole process, SO
effecting considerable economies in time and materials. In
was found to be 89.50,;. Since, however, BHC is somewhat addition the density of this liquid becomes susceptible of being
more stable to heat than is DDT, it is probable that this figure determined to high accuracy by means of a special hydrometer
could be exceeded by determining the efficiencies of 3 series of unusually open scale which need be calibrated in the basic
of compositions in a manner similar to that described for DDT. manner at intervals only.
In addition to determining the density of this liquid, how-
Acknowledgements ever, two other experimental determinations have to be made,
Much of the practical work involved in this investigation and a considerable amount of computation has to be under-
was performed by Mr. N. L. Bailey and Miss B. I. Grisley, taken, so that the method is relatively complex. Nevertheless
while the analytical work was mainly performed by Miss M. V. it is considered that the disadvantages of complexity are out-
Liddell. weighed by the economies effected.
The author wishcs to thank Mr. E. W. Bateman for his Outline of the present method
interest in this work, and the Chief Scientist, Ministry of
Supply, for permission to publish this paper. In calibrating a hydrometer we have to determine the density
of a liquid in which the hydrometer would float with a particular
Chemical Defence Esperimental Establishment, graduation mark in the plane of the level surface of the liquid.
Porton, This is equivalent to saying that we must determine the density
Wilts. of a liquid in which the hydrometer experiences an upthrust
Received Oitobcr J, r9jB equal to its weight when immersed to the degree stated. NOW
* Bateman, E.W. and Heath, G. D., J.S.C.I.19.)7,66,325 the upthrust on a body immersed to a given extent in a liquid
? Scholcfield, P.G.,Bowden, S. T., and Jones, W. J., J.S.C.I.1946, is proportional to the density of the liquid. It follows that if
65,354 we weigh the body immersed to that extent in two media of
known density, we may find the density of a medium in which
the upthrust on the immersed portion would equal the weight
A NEW METHOD OF HIGH ACCURACY of the body by linear extrapolation.
FOR THE [Link] OF REFERENCE In the present method, hydrometers are weighed in air and
STANDARD HYDROMETERS xylene of determined densities, under prescribed conditions of
immersion in the latter case, and the liquid density required
By F. \V. CUCKOW for free flotation is obtained from these weighings and the
In this method, standard hydrometers are weighed first in air associated densities by linear extrapolation.
and then when immersed to the level of the graduation mark in Since an accuracy of one part in a hundred thousand is aimed
question in xylene of determined density. Thc density of a liquid at, many secondary factors have to be considered as indicated
which would cause thc hydrometer to float with this mark in the
plane of the lcrel surface of the liquid is caiculatcd, essentially by in the next section.
linear cxtrapolation, from these weights and the associated densities
of air and xylenc. Furthcr weighings with m c r e n t degrees of Theory of the method
immersion enable calibrations to be made at a number of points on (I) Equilibriirni equations
rhc hydrometer scalc in quick succession, the working liquid (a) In the case of a hydrometer floating freely in a liquid at
remainingnominally unchanged in density and tenIperaNre through-
out. By comparison, the basic method of working, employing a zoo
ZDy, -
freely floating hydrometer, calls for a liquid of different density for
each point to bc calibrated. Calibrations are made exactly at PN +-g
- VpL +v pG1.. ........ ,.. . ( I )
graduation marks.
The rate of working of the present method is at least four times where PPI = mass of hydrometer
higher than that of the basic mrthod and an intercomparison of D = stem diameter at meniscus level
calibrations effected by the two methods shows that there is no
detcctable loss of accuracy. An open scale hydrometer, which y1 = surface tension of liquid
can be read to [Link] g./ml., is described for observing the g =: acceleration due to gravity
density of the xylenc. A nomogrjm is described for computing V = volume at zooof hydrometer below the plane of the
the surface tension factors of hydrometers in terms of their weight. level surface of the liquid
In the basic method of calibrating reference standard hydro- v = volume at 20' of hydrometer above the plane of the
meters their scale readings are compared with the densities of level surface of the liquid
the liquids in which they float freely, these densities being pL = density of the liquid '

determined by weighing in the liquids sinkers whose volumes p~ = density of the air surrounding the stem of t h e ,
are known from previous weighings in air and water. Such hydrometer.
'
calibrations CM bc effected with very high accuracy when (b) If this same hydrometer is weighed in air at t o we have
suitable precautions are taken, including the use of specially
designed hydrostatic balances and thermostatically controlled
baths,' and the adjustment of the density of the liquid so that
the graduation mark under calibration is at the reference level, . . * . .(z) 9 . . ..
i.e. in the plane of the level surface of the liquid (see Fig. I). where 0: = coefficient of cubical expansion of the material of
This basic method of calibration has the merit of simplicity the hydrometer
with a corresponding degree of freedom from error. It is, and W, = apparent weight of the [Link] air of density
however, time-consuming in that the density of the liquid p~~ against weights of density A.
minure in which the hydrometer is floated has to be specially (c) Consider now the conditions of equilibrium when this
adjusted for each point of the scale calibrated. Also consider- hydrometer is weighed partly immersed in a liquid of density
able quantities of the liquids comprising the mixtures are d, temperature 20' and surface tension ye. The degree of
consumed. immersion is to be the same as in paragraph (a) above, that is
J.S.C.I., 68, Ftbruuy, 191) CUCKO If'-CALIBRATION OF REFERENCE STANDARD HYDROMETERS 45
a volume V of the hydrometer is to be below the plane of the are very small compared with the associated quantities W?
level surface of the liquid and a volume v of the stem is to be
above that plane. and (W, - wd)y they may be taken as equal to and r*p
Under these conditions E R
-
respectively. If, further, the precaution is taken of arranging
ttf-!-?*= VdfVpG,fW,f(,I --?) . . . . ( 3 ) that FG, = PG3 (this is achieved by weighing the hydrometer
in air immediately before weighing it in the liquid) the formula
where p~~ = density of air at the time of this weighing for pL reduces to
and Wd = apparent weight of the hydrometer under these
conditions.
W , + g a&, -
From the three equations (I), (2) and (3) we have to determine
pL by elimination of the unknown quantities ,try V, and v.
Some simplification is required before this can be done.
(2)Simplification of equi1ibrium equations where PG is the air density at the times of weighing the hydro-
(a)The equation (2) may with sufficient accuracy be written meter.
+
m = ~ p cvpGZ +
~ W,(X - '2) . . ... .(4) (4) The ef/ect of making wekhings in a liquid at temperatures
the neglected terms being equal to drgeritrgfrom the standard 20'
+
(V v) (a TO) pG2 If the temperature of the liquid at the time of weighing the
hydrometer in it is t', equation (3) should be written
whose value taking extreme conditions (V 4- 01 = 100 ml.,
Q = 0~000026,(t - 20) = f 10' and pc2 = 0.0012 [Link].)
is 0 ~ 0 0 0 0 3g. Even this extreme quantity may be neglected
as the weighings need be accurate to the order of a few tenths where V,is the immersed volume of the hydrometer at 1' and
of a milligramme only. d is the density of the liquid at to.
(6) The terms w ~ ) G VPG-
~, and U P G ~in the three equations Suppose that to indicate the density [Link] liquid we have
(I), (2), and (3) represent the air buoyancies acting on the an instrument whose coefficient of expansion is the same as
emergent stems of the hydrometer at various times. Variations that of the hydrometer being calibrated, namely z. Let it
in the density of the air surrounding the stems of hydrometers indicate correctly at 20'. Then it is well known-that at t o
are neglected in normal hydrometry, the resultant changes in its indication dA will equal the true .density of the liquid d
buoyant upthrust being too small to affect the hydrometer multiplied by (I + t o ) so that
rcading significantly. This procedure is equivalent to regarding
terms of the form c p as ~ equal to a constant, say 'k.' d= L .--A-
The equations (I), (4) and (3) may then be written I+at-20
~.
Substituting this in equation (9) together with V,= Va
m + - F = VpL+k ...................... -+
( 5 ) (I zrt - 20) we obtain
m = VPG,+~+W,(I -?) ......(6) )If f sDy, = Vs d~ fvPG3
g.
w d (I - PGs/A) * (10) +

which is identical in form with equation (3). That equation

is therefore valid for weighings made at temperatures other
(3) Solution of the equilibrium equations than 20° provided that the apparent density of the liquid (as
indicated by an instrument correct at 20' and of the same
Solving equations (9, (6) and (7) for pL by elimination of coefficient of expansion as the hydrometer being calibrated)
myV and k which are all difficult to mzasure, we obtain without is used instead of the real density.

"JI-
further approximation,

0-0 9 ImL
0.86095 +I.
This can be written
0 4 6 100
w2+xDy1 x --I 0.06105 OlmL
0.06110 Olml
0.06115 Olml.

0.86120pIml

FIG.I . Fic.8 F10.3

FIG.I. Graduarion mark at rsfermce leuel '

FIG.2. Sccriott rhrough lbdrornefer used in xyhw
FIG.3. Hydrontercr scalc ( x y h e ) attd corresponding rraditgr
BS
46 CUCKO W-CALIBRATION OF REFERENCE STANDARD HYDROMETERS J.S.C.I., 68, Eebrwry, Im9
It follows that the final formula for pL given at the end method on a hydrostatic balance, the hydrometer being at that
of paragraph (3) is valid for weighings in liquids at temperatures time partly immersed in the calibration liquid contained in the
other than 20' if d is taken to be the apparent density of the thermostatically controlled bath underneath the balance.
liquid instead of the real density. . y2 : The surface tension of the liquid used in the calibration,
should be determined experimentally. In- the author's ex-
( 5 ) IQdrometers with scales corresponding fo liquids less dense perience the surface tension of the liquid here used remains
than the calibrafiori liquid unchanged over long periods and should therefore need to be
The density ranges of hydrometers normally made are such determined only occasionally.
as to cover all common liquids and of necessity, the first and
last hydrometers in the series will have points indicating Working details of the method
densities rather below and rather above the lowest and highest (I) Hydrostafic balance and thertnostaficallycontrolled bath.'-
density liquids available. It follows that no liquid will be Xylene was chosen as the most suitable calibration liquid in
available in which all hydrometers will sink, and the apparent view of its constant surface tension and reasonably constant
weight of certain hydrometers of very low density range is density. It is contained in a glas jar some 20 cm. by 8 cm. in
therefore necessarily negative. The negative value can be cross-section and 40 cm. deep, which forms part of the equip-
substituted in the formula for pL as the appropriate measure ment to which reference has already been made. The jar is
of the weight of the hydrometer in the liquid used. surrounded by water which can be constantly circulated
(6) Stirfnce tension factor of hydrometer through an outer vessel fitted with plate-glass windows so that
The surface tension factor of a hydrometer is defined as the interior of the vessel containing xylene can be seen freely.
the change in its indication corrcsponding to unit change in This outer vessel contains a toluene-in-glass thermostat, with
the surface tension of the liquid in which it is floating, other mercury as the contact liquid, controlling the current passed
factors remaining constant2 and should be determined so that
appropriate corrections can be applied when the hydrometer
is used in liquids whose surface tensions diRer from yl,the
value for which the hydrometer was standardized.
The surface tension factor can be related to the mass of the
hydrometer by difkentiating equation (I). This gives (with
slight approximation)
dPL - "DPL
-
dY 'W
This quantity approximetes closely to the surface tension
factor as defined above, so that
S.T.P. = -ZDPI.
nag I \
Since in = W, within about I in 1030,each of the quantities F S
D, pL and N I is determined to adequate accuracy in the course
of standardizations by the present method, and the surface FIG.4 FIG.5
tension factor can therefore be calculated without further
FIG.4. Staitrless steel clattip for Irydrotnorer stem
measurements.
FIG.5. Sitspetibs device
Practical requirements
In calibrating a hydrometer by this method we have there- to electric heaters immersed in the circulating water. A small
fore to determine the quantities d, PG, W2, D, yl, Wd and yz steady stream of cold water is allowed to flow into the circulating
required in formula (8) for the computation of pL. ,
water and the thermostat maintains a balance between the heat
d : The density of the liquid in which the hydrometer is lost to the cold water and the heat gained from the heaters.
weighed, is most easily determined by means of a very sensitive The sensitivity is such that the temperature of the xylene in
hydrometer. Since the density of this liquid remains nominally the inner jar remains constant to well within 0 . 0 1 ~over a
unchanged throughout the whole process, the hydrometer can period of hours. Uniformity of temperature throughout the
have a very open scale of no great length. xylene is ensured by means of a pump which removes xylene
pa : The air density at the time of weighing the hydrometer from the surface where its temperature is most easily affected
in air and in the liquid, can be obtained from tables in terms by variable conditions and injects it into the body of the
of the prevailing temperature and barometric pressure, it being liquid.
sufficient to assume an average value for the relative humidity. This system is surmounted by a hydrostatic balance fitted
W,:- The weight of the calibration hydrometer in air of with special stirrups in addition to the normal scale pans so
density p~ against weights of density A, is determined by simple that wcighings of bodies immersed in the bath underneath can
counterpoise weighing on a balancx of known sensitivity and be effected by means of rods passing down from the stirrups
capable of giving the weight with an accuracy of the order of through the floor of the balance case. The balance beam is
0.0002 g. fitted with a mirror so that its oscillations may be observed
D : The diameter of the stem of the hydrometer at meniscus with precision by means of an optical lever, Weighings can
level, is obtained by means of a simple dial gauge. be effected with an accuracy of the order of 0.0002 g. for total
y, : The surface tension of the liquid in which the hydro- loads of a few hundred grams.
meter is normally intended for use, will be a known quantity. ( 2 ) Open scale hydrometerfor the determination old, the density
W,I : The rppomnt wight of the hydrometer in the calibra- of xy1ene.-Figure a shows a cross section through the axis of
tion liquid 4ain8t weights of density A in air of density PG, the hydrometer constructed for the determination of the density
will be obtained by weighing the hydrometer by a counterpoise of xylene. The overall length of the hydrometer is 40 cm.,
J.S.C.1, 68, February, i ~ g CUCKO W-CALIBRATION OF REFERENCE STANDARD. HYDROMETERS 47
while the stem is 5 cm. in length and 2.5 mm. in diameter,'
and the bulb is 4.5 cm. in diameter.
Figure 3 shows the lay-out of the scale contained in the
stem of the hydrometer and the figures alongside indicate the
densities corresponding to the seven main graduation marks.
The procedure in determining the
quantity W,, the apparent weight of the
hydrometer when partly immersed in
xylene, is as follows.
The xylene is brought to a steady tem-
fa
Each of the six main divisions is subdivided into five parts perature and the xylene circulating pump
each corresponding to a density interval of o.00001 [Link]. is shut off,so allowing currents in the
These thirty sub-divisions are each about I . I mm. long and liquid to die out. Sufficient weights are
since the scale reading corresponding to the plane of the level placed on the side of the balance from
xylene surface can be estimated to a tenth of one of these which the hydrometer is suspended to
divisions, the density of the xylene can be read with a precision ensure that this side falls when the beam
of about [Link] g./ml., that is, about one part in a million. release mechanism is operated. The
The hydrometer was calibrated by observing its reading in release mechanism is then operated grad-
xylene whose density was determined by weighing in it sinkers ually until the optical lever system indi- K
of known volume. The determination of the sinker volumes cates a particular deflexion of the balance
was carried out by weighings in air and in distilled water. beam. The hydrometer suspension de-
The accuracy attainable in the determination of the density vice is then adjusted so that the grad-
of xylene by means of this open scale hydrometer is indicated uation mark on the hydrometer stem at
by the results of twenty-six independent calibrations of the which a calibration is required is just
hydrometer spread over a period of five weeks. The determined below the plane of the level surface of
corrections to the scale readings required to give the density the xylene. A slight alteration in the
of xylene at to0 ranged in this series of observations from
- 0~0oooo4to - 0 ~ 0 ~ 0 0 1g./ml.
2 with a root mean square
level of the surface of the xylene then
serves to bring it into exact coincidence .
-
deviation from the mean of o 000002 g./ml., that is, a deviation
of about two parts in a million on the density of xylene.
with the graduation mark. This fmal ,
adjustment is made by slowly with-
( 3 ) Determinaiion of the apparent weight of the hydrometer drawing a solid glass rod from the xylene
.first in air and secondly when partly immersed in xy1ene.-The so making the surface level fall. This
top of the stem of the hydrometer undergoing calibration is setting of the graduation mark to coin-
gripped in the stainless steel clamp depicted in Fig. 4. The cidence with the plane of the level
bearing faces of the clamp carry small pads of velour cloth surface of the liquid can be made with
(attached with a nitrocellulose cement) to avoid excessive great precision. The weights on the
pressure on the hydrometer stems. The hydrometer is then hydrometer side of the balance are ad-
justed until the balance beam swings Fu.6
hung on the device shown in Figs. 5 and 6 and this in turn is symmetrically about the particular datum Suspnirimi system
suspended from the special stirrup on one side of the hydro- position used in the earlier adjustment of
static balance. A k e d weight is placed on the opposite side the degree of immersion of the hydrometer and threc successive
of the balance and this remains unchanged throughout the turning points are noted SO that the weight which would be
whole work. required to bring the balance beam exactly to the datum
The weights on the hydrometer side of the balance are position may be calculated using the balance sensitivity
adjusted until the beam swings symmetrically about a certain
datum position which is the same as that used when weighing corresponding to these conditions. The apparent weight of
the hydrometer in xylene and is more closely defined later. the partly immersed hydrometer is then obtained by difference
Three successive turning points are noted. The weight which from the weights required for balance in the absence of
would be required to bring the balance beam exactly to the the hydrometer, as determined earlier.
datum position is calculated using the sensitivity of the balance The density of the xylene, as indicated by the special open
for such weighings in air. With the hydrometer only removed scale hydrometer, is then recorded so that simultaneous values
from the suspension device, further adjustmcnts to the weights are obtained of the xylene density and the apparent weight of
and observations of balance beam movements are made so that the hydrometer undergoing calibration when immersed therein
the weight of the hydrometer in air may be obtained by dif- to .the prescribed .degree. Calibration at other graduation
ference. The barometric pressure and air temperature are marks on the hydrometer stem can be effected immediately
noted at the time of these weighings in air. by additional weighings with the hydrometer further immersed
The suspension device shown in Figs. 5 and 6 consists of to the appropriate degree.
two stainless steel rods R and S each equipped at one end with Hydrometers which would not normally sink in xylene (see
a small hook. The distance between the upper and lower p. 46, 5 5 ) are made to do so by the addition of Iplummet
hooks can be varied at will by turning the control knob K on which is sufficiently heavy to overcome the upthrust exerted
the assembly A. This assembly comprises two plates P and by the xylene. The difference between the ap arent weight
Q linked by distance pieces X and Y and providing bearings
of the plummet alone and the apparent weight ofthe plummet
for two pulley wheels one of which can be turned by the control and hydrometer gives the apparent weight of the hydronjeter,
knob K. The pulley wheels have steep sided V-grooves in which in this case is a negauve quantity,
which rest the rods R and S. The rods are retained in position Accuracy of the method
by two spiral springs pulling on two flat sided bars E and F
which in turn bear on the flat outer faces of the rods R and S. The accuracy of the method can be illustrated by comparing
The friction between the pressure bars E and F and the rods calibrations effected by its use with those effected by the basic
R and S is sufficient to prevent one rod slipping with respect method permitting free flotation of the hydrometer.
to the other when a load of about 200 grams is applied to the The lirst illustration concerns hydrometers having scales
lower hook. This device serves to suspend the hydrometer in approximately 140 mm. long, each covering a density range of
the xylene and to permit the degree of immersion to be adjusted 0.05 g./ml. The scales are subdivided into one hundred
to within a fraction of a millimetre. parts so that each subdivision corresponds to an interval of
B6
 CUCKO W-CALIBRATION OF REFERENCE STANDARD HYDROMETERS [Link], 68, February, Z M ~

0.0005 g./ml. and occupies a scale length of approximately Table I1

x 4mm. Table I shows for these hydrometers :(I) the nominal Determined density
scale reading; (2) densities corresponding to the nominal scale Hydrometer Scale reading by basic method by present method
reading as determined by the basic calibration method using a No. at 30 dyncs!cm. g.[ml. at 20' g./mI. at zoo
freely floating hydrometer (each value quoted represents an 18110 I .025 1.024 (50 1.02461
independent determination) ; and (3) densities corresponding 1.024 60 1.02461
to the nominal scale readings as determined by the present
,
, I .030 1.029 64 1.029 66
1.029 64 1.029 65
method. (Again each value quoted represents an independent 12815 1.018 4 1.015 04 1.or5 04
determination). 1.015 04 1.015 02
I2822 1.023 2 1.019 62 1.019 60
Table I 1.019 61 1.01g 60
Dctcnnincd density The calibration was then repeated on two separate occasions
Hydrometer by basic method by present method
but the weighings were carried out in tetrachlorethane (density
No. Scale reading
12713 0.650at 17 dynesicm.
g./ml. a t 20'
0.649 25
g./ml. at 20'
0.649 26 -
approximately x 6 g./ml.) in place of the normal xylene (density
0.86 g.]ml.). The density of the tetrachlorethane was deter-
0.649 24 0.649 26
12714 0.650 at 19 dyncs/cm. 0.649 80 0.649 78 mined directly by sinker weighings.
0.649 76 0.649 78 The results of these three sets of observations are set out
12717 0.810 at 25 dyncs!cm. 0.809 21 0.809 22
0.809 ?,I 0.809 22 in Table 111.
,, 0.820 at 25 dynes,'cm. 0.81902 0.81903 Table III
0.819 01 0.81904
,, 0.830 at 25 dyncs!cm. 0.829 09 0.829 09 Detcrznined density
0.829 08 0.829 09 by present method
by present method using
Table I1 shows a further series of results on three hydro- Hydrometer Scale reading using xylenetetrachlorethane
meters covering ranges of 0 . 0 1 0 ~0.008 and 0.008 g./ml. No. at 35 dyncs!cm. g.!ml. sit 20"g./ml. at zoo
respectively with scales 60 mm. long. These scales are sub- 1.799 32
divided into twenty, forty and forty parts respectively so that 1.799 27
1.809 52
the subdivisions correspond to intervals of 0.0005~ 0-0002and I . % 49
0.0002 g./ml. and occupy scale lengths of approximately 3, 1.829 48
1 - 5 and 1.5 nun. respectively. 1.819 46
Further c o n b a t i o n of the accuracy of the method is 1.829 61
1.82958
provided by a series of observations on a hydrometer covering 1.839 82
the range 1.80 to 1-85g.\ml. The scale was approximately
160 mm. long and was subdivided into one hundred parts. -
1.839 80
I 849 86
1.849 84
The hydrometer was hlibrated at six points on its scale by
the present method, [Link] made as usual in xylede Concluslons
and in air. The various groups of figures shown in the last column of
Tables I and I1 make it clear that the precision attained in the
f- present method of calibrating hydrometers is high and ap-
proaches one part in a hundred thousand. It is noteworthy
that in the case of the hydrometers listed in Table I this
corresponds to only one fiftieth part of a scale division and to
a scale length of 0.03 mm. This [Link] less than the thickaess
of a graduation mark, The cloiie agreement between the
results from the two methods of working as listed in the third
and fourth columns of Tables I and I1 shows that the accuracy
also is high and again approaches one part in a hundred
thous and.
Further evidence of the reliability of the method is given
by the results shown in Table 111. Although these calibrations
have been made in liquids of such widely different densities as
xylene and tetrachlorethane the mutual consistency is good.
The rate of working of the present method is at least four
times that of the basic method using a freely floating hydrometer
and it has been demonstrated above that this increased rate of
working has been obtained without detectable loss of accuracy.
Moreover the present method has other advantages over the
slower method, in particular the ease with which calibrations
can be effected exactly at specified graduation marks, economy
in liquids, and incidental determination of all the data necessary
for the derivation of the surface tension factor of the hydro-
meter.
Nomogram for surface tension factors of hydrometers
It was shown on p. 46, sub-section 6, ;hat the surfacc
tension factor may be obtained from the expression

FIG. 7. The noiiiogroirr

 J.S.C.I., 68, February, 1949 HARRISON-COMBINED CARBON IN CAST IRON 49
A nomogram has been constmcted which facilitates d c u l a - THE DETERMINATION OF COMBINED
tion of this factor from the determined values of 0, pu and m ; CARBON IN CAST IRON BY MEANS
the nomogrcm has four circular logarithmic scales.
The scales for pL and D are carried on an inner plate, as OF THE SPEKKER PHOTO-ELECTRIC
shown in Fig. 7,* which can be rotated about the common centre ABSORPTIOMETER*
of the two semi-circles forming its pedmeter. The pL scale, By T. S. HARRISON
ranging from 0.6 to 2.0 g./ml., is carried around an arc of The determination of combined carbon in cast iron m a y be
approximately 140” of the larger semicircle and the D scale, completed absorptiometricallysince Beer’s Law is obeyed. As the
ranging from 2 - 5 to 8 mm., on an arc of the smaller semi- intensity of colour produced by a given sample varies with con- .
ditions, e.g., acid concentration and time of heating the solution,
circle extending over about 120‘. This inner plate is attached a standard must be included in each batch of determinations. Thc
by a pivot passing through the centre to the rectangular base optimum conditions for colour dcvelopment are noted. Colour
plate. stability is good.
The m scale is inscribed on the base plate on a circle of the The total carbon content of cast iron usually exists in two
same radius as the smaller semicircle on the inner plate so that : (a) graphitic carbon, and (6) combined carbon. I n a
the D and m scales are juxtaposed. This m scale ranges from forms grey iron the major portion is prerent as graphite, whereas in a
10 to 200 grams and occupies almost the full 360”. The S.T.F.
scale is inscribed on the base plate on a circle of the same radius [Link] iron all the carbon can exist in the combined form as in
Total carbon is determined by combustion, whilst the
as the larger semicircle on the inner plate. This results in the graphitic carbon figure is obtained by dissolving the sample in
pL and S.T.F. scales lying immediately adjacent to each other. nitric acid, sp. gr. 1-20, filtering through previously igaired
A straight-line cursor is also attached to the centre pivot so
that the line lies along a diameter of the circles on which the asbestos to collect the graphite, drying the precipitate and pad,
and igniting in the cornbustiontube in the usual manner. Cam-
scales are inscribed. bined carbon is then calculated by a simple subtraction. The
In using the device, the cursor line is set over the appropriate result so obtained is accurate, and this procedure is used for
’
value of m and the centre plate is rotated until the corresponding preparing combined carbon standards. In routine p r a a i a
value of D on the adjacent scale also appears under the line. combined carbon is determined by a simple colorimetric test
The cursor is then moved until the ljne passes through the is more rapid and covenient but less accurate than the
value of pL and the surface tension factor corresponding to the which above. But if all due precautions arc taken, the colour test can
three values used for myDand pL is read off the adjacent S.T.F. be of the greatest practical value. When cast iron is dissolved
scale at the point of intersection with the cursor line. in dilute nitric add, sp. gr. I -20, the combined carbon yields a
In constructing the nomogram, the relative positions of the brownish-yellow colour due to the formauon of an organic
four scales have been so arranged that the constants x and g compound consisting of carbon, hydrogen, nitrogen and oxygen.
appearing in the formula are automatically included in the Furthermore, if certah conditions are fulfilled, the depth of
calculation. colour is proportional to the combined carbon content.
Previous work Theoretically, it should also increase in direct proponion with
The present method of calibrating hydrometers was devised the amount of sample used, but this is not so in practia.
in 1937 but, for a variety of reasons, has remained unpublished Further, the tint and depth of colour vary wi+ a great number of
apart from brief notes? conditions such as the temperataure and duraaon of heating prior
Subsequent search of the literature has shown that a basically to the matching of the COIOUS, the concentration of acid em-
similar method had been used earliefi but in a very restricted ployed, the presence of other elements normally found in cast
manner only. Whereas the present method permits the Cali- imn, e.g., phosphorus, and the manner in which the iron has
bration of hydrometers of any range in one liquid, this earlier been made and its subsequent physical treatment. Hence it is
work called for a different liquid for each hydrometer but then evident that for reasonably accurate results it is essential to
permitted the calibration in that liquid of any part of the scale employ standards of closely similar composition and history to
of the hydrometer. Further, no attempt appears to have been the samples under test, and to treat the sample and standard
made there to obtain higher accuracy by setting graduation identically during the determination. The colour method is
marks to exact coincidence with the plane of the level surface less reliable for cast iron than it is for steel, owing to the more
of the liquid. complex nature of the former. It is as yet inapplicable in the
Ackaowledgementrr presence of any element such as nickel or chromium the solution
of which is coloured.
This work was carried out as part of the General Research The object of this research was to develop an absorptiometric
Programmeof the National Physical Laboratory andis published finish to the determination, to ascertah the degree of colour
by permission of the Director. stability, and to investigate the influence of certain variables.
The author desires to acknowledge the comments and sugges- The colorimetric method was first devised by Eggertzl and
tions he received from his colleagues in the Metrology Division appenrs in most textbooks on iron and steel analysis?. 3,dl 6 It
of the National Physical Laboratory. is described by Allport too, in his excellent book entitled
National Physical Laboratory, “ Colorimetric Analysis.”o The accounts given by Arnold and
Teddington, Ibbotsonz and by Blairs are very full, and were used as a basis
Middlesex. Received April 12, 2948 for this work.
In the Fig., “c,” “D,” and “ M ” correspond napcnively to Pnllmkury experiments
pL, D, and tn as used III the text
Standards containing 0.60% and 0.73% respectively of com-
1 National Physical Laboratory, Report for the pu 1923, H.M.S.O. bhed carbon were selected. These figures were obtained by
f Science of Petroleum, Oxford University Press (“ Hydnrmcters,” by combustion, duplicate determinations of the total and grephitic
Verney Stott)
J National Physical Laboratory, Report for the year 1938, H.M.S.O.; carbon Wing made in each case. The combined carbon con-
ibid. 1939 tents of several works samples were similarly obtained. It was
bci der Untenuchung von
4 “ Bcdeutw dcr Kapillaritritsd8~reazen
Aramcters. Kaiserlichen-Nonnal-~Chu~s-Komrmssion : Wise. Abh. Abaunaed from a thesis prcscnted to the Univcnity of London
1902~3,2 in Octot,.r, 1946, and approved for the nward of the Ph.D. De-.
87