M

I
1
‘

}
I

M

133
820

 

TH '

SPECTRAL LlNE INTENSITIES OF
TITANIUM IN STEEL

Thesis fur the Degree of M. 3.
MICHIGAN STATE COLLEGE

George Parker Koch
1942

«.1 9‘,

‘ .1 .9!

153141.: ‘

- . 4

 

II..‘

S
E
m
m
m
L
E
T
A
.l
s
m
Pu
H
m
M

31293 017718317

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 cy-‘ClRCrDateDue p65-p 14

SPECTRAL LINE INTENSITIES OF TITANIUM IN STEEL

BY

GEORGE PARKER KOCH

A THESIS

Submitted to the Grafluate School of Michigan
State College of Agriculture and Applied
Science in partial fulfilment of the
requirements for the degree of

MASTER OF SCIENCE

Denn“tflent 0P Chenithv

19d?

ACKNOWLEDGEMENT

The author wishes to express his

appreciation to Dr. D. T. Ewing

under whose direction this work
was carried out.

#3
p.
39
CL”:
C?
CL

The purpose of this investisation was to
find, if possible, one or more titanium lines
in the spectra of titanium-steel alloys which
mipht be used as the basis of a procedure for
the spectrogranhic analvsis of titanium in
steel, and to investigate the suitabilitv of

these lines for that purpose.

CHOICE OF LINE

There are many factors which enter into
the choice of a line or lines to be used for
a spectrographic analvsis. Perhaps the most
important of these is the position of the line
in the spectrum, that is, its wave length.
This will determine the type of photographic
emulsion which must be used in the analvsis,
and it is of course obviously simpler to work
with a line which falls into an easilv accessible
photographic region.

The character of the line itself is also
of prime importance. It must possess sufficient
inherent intensity to avoid a period of ex-
posure which is impractical or which.will pro-
duce a background of sufficient intensitv to
alter the density of the line. It must also
be separated from adiacent lines by an amount

great enough to insure that it will be resolved

from these other lines when the slit of the
spectrograph is Opened to that width to be
used in the analysis.

If the concentration of the element souaht
is relatively large, and if the intensity of
the line is of itself great, the photoaraphic
image may suffer reversal unless an extremely
short eXposure is taken or unless the amount of
light entering the instrument is decreased by
some means, as for example, by a rotating
sector. If on the other hand the concentration
of the element sOuaht is relatively small,
the use of an intense line is necessary, and
in the case of verv small concentrations, the
use of a line which is both strong and per-
sistent may be in order.

The effect of fluctuations in the excitation
conditions on the intensities of the lines to
be used for the working line and the internal
standard line should also be investiaeted.
This may be done by determining their intensities
under various accurately controlled conditions.
It is well to choose for the working line and
the internal standard line lines which do not
varv in intensity for considerable fluctuations

in the excitation conditions.

In this connection the use of an homo-
logous pair for the working line and the in~
ternal standard line may be advantageous.

An homologous pair has been defined as two
lines having approximately the same excitation
conditions so that for a considerable variation
in excitation conditions any change which takes
place in one will take place to the same ex-
tent in the other. This phase of the problem
of choosing a line was not investigated in

this thesis.

In choosing a spectrum line to be used in
an analysis some attention should also be given
to the amount of energy required to produce
that line. under the usual conditions of ex-
citation only those lines coming from neutral
or singly ionized atoms are obtained. Sources
of unusually high energy are required to ob-
tain lines from atoms in a higher state of
ionization.

The first step undertaken in this in-
vestigation was to obtain a number of titanium
lines which, according to the above reouirements,
might be suitable as working lines. For this
purpose spectra of the titanium-steel containing

the largest concentration of titanium were compared

with spectra of a sample of Armco iron of
very high purity. Two types of excitation
were used, a direct current arc and a con-
densed spark. The spectrum range covered
in each case was roughly from 2500 to 5000
angstroms. The search for titanium lines
was greatly facilitated by using a Judd-
Leris comparator, an instrument rhich.makes
possible the simultaneous comparison of
spectra on tro different plates.

The first part of the search for suitable
titanium lines was carried out by comparing
the spark spectra of the pure Armco iron with
the spark spectra of the titanium-steel sample
containing the largest per contage of titanium.
All the lines appearing in the titanium-steel
spectra and not in the pure iron spectra were
carefully noted and marked if they appeared to
be sufficiently separated from adjacent iron
or titanium lines and possessed an appreciable
intensity. The wave lengths of the titanium
lines which satisfied these reouirements were
determined by comparing the titanium-steel
spark spectra with.what may be called an iron
arc-spark reference plate. This plate con-
sisted of an iron are spectrum and an iron

spark spectrum photographed in juxtaposition

through a Hartmann diaphragm. The range of
wave lengths covered on this plate was the
same as that covered by the titanium-steel
spark spectra. In determining the wave length
of a given titanium line the spark spectra on
the two plates were lined up and the wave
length of the titanium line determined by
interpolation between known lines in the iron
are spectrum. This procedure was necessary
because the spectrum charts available listed
the wave lengths of the iron spectrum in
terms of an arc source and it was found im-
possible to correlate the iron spark and the
iron are spectra because they were so widely
different. Because the maximum resolution
was desirable in the determination of the
wave lengths, a slit opening of five microns
was used.

The slit width was then increased to
forty microns and spectra of the titanium-
steel and of the reference iron standard
taken at this larger slit Opening were com-
pared. The purpose of this step was to
actually determine if the lines chosen were
still resolved from adiacent iron or titanium

lines.

The procedure outlined above for the
spark was carried out over the same wave
length range using a direct current arc, ex-
cept of course that the use of an iron are-
spark reference plate was not necessary.

Two lines of titanium were found which
fitted all the requirements. Their wave
lengths according to Harrison are 3361.213
and 3372.800A and their intensities 600 and
400 respectively.(l). The 3361.213 line is
a persistent line of titanium. The values
of the intensities given are those obtained
using a spark source. Because the intensities
are appreciably larger in the spark than in
the d-c are it was decided to concentrate on»_
this type of excitation. Both of these lines
arise from the singly ionized titanium atom.

Within a range of three angstroms on
either side of the titanium line at 5361.215A
there are no titanium or iron lines having an
intensity of greater than 5 with the exception
of the titanium line at 3361.963A. This line
would not of course be resolved, but its small
intensity(50 as compared to 600) made it im-
probable that it would affect the intensity of

the stronger line to a measureable extent.

Within the same range on either side of
the titanium line at 3372.800A there are no
titanium or iron lines with an intensity of
greater than 30 with the exception of the
iron line at 3370.786A having an intensity
of 200. This line is resolved from the
3372.900A titanium line at a slit Width of

forty microns.

CHOICE OF PHOTOGRAPHIC EMULSION

The choice of a photographic emulsion
depends upon the wave length of the line
which is to be used for the analysis, the
range of concentrations over which the
analysis is to be made, and the accuracy
required in the analysis. The characteristics
of three different Eastman photographic
emulsions, the Spectrum.Ana1ysis No. 1, the
33, end the 40. are expressed in terms of
their H k D curves in Fig. I. These curves
were prepared by exposing the various plates
to visible light through a step density tablet,
the steps of which transmit light in known
amounts, the ratio between successive steps
being the square root of two. The densities

of the several steps were determined on an

Eastman densitometer which reads directly in

densities.

It can be seen that the Spectrum Analysis
No. l emulsion would give the greatest accuracy
since for a relatively small increase in in-
tensity there would be a relatively large in-
crease in density. However, the range of con-
centrations over which this emulsion may be
used is smaller than that for which either
a 53 or 40 emulsion would be suitable.

The H & D curves give no information on
the spectral sensitivities of the different
emulsions. The Eastman 40 was found to be
somewhat more sensitive than the Eastman 33
throughout the region from ?300 to 5000A.

The Eastman Spectrum Analysis No. 1 emulsion

is less sensitive than either the 33 or 40

in the region from 9300 to 5000A, and much less
sensitive in the region from 3500 to 5000A.

In this investigation it was desired to
cover the maximum concentration range possible
with accuracy a secondary factor. This was

the reason for choosing the 40 emulsion.

CALIBRATION OF PLATE
In order to insure that one is working on
the straight line portion of the H & D curve of

the photOgraphic emulsion, the range of densities

over which there exists a linear ratiO' be-
tween the density and the logarithm of the
intensity(or exposure) must be known. The
emulsion may be calibrated conveniently and
accurately by the use of a logarithmic stepped
sector placed in front of the slit. A
necessary prereouisite of the use of such a
sector is that the slit be evenly illuminated,
a condition which was not exactly met in this
work. Nevertheless, from the calibration
curves prepared the density level at which a
linear ratio between the density end the log—
arithm of the intensity(exposure) was no
longer available was evident. It then re-
mained to eXpose the samples and develop the
plate in such a way that the range of densities
obtained for the analysis lines did not ex-
ceed this value, which.was about 1.8. If
this was done, however, a considerable amount
of background was obtained. It has been
shown(2) that a background of sufficient in-
tensity will affect the density of a spectrum
line. For this reason the exposure was reduced
considerably. An exposure of either 9 or A
seconds was found to give some continuous back—
ground, but presumably not enough to alter the

density of the lines. These exposures were

10

made using a slit width of 60 microns. It
may be noted here that a slit width of 40
microns was used in the preliminary work, but
this was increased, as stated above, to 60
microns for the analysis. This change was
instituted in order that the slit of the
density comparator would be more completely
covered by the spectrum lines. In this way
the line could be scanned effectively. The
same conclusions regarding the separation of
the working and internal standard lines from
adiacent lines which held at a slit of 40

microns are strictly valid at 60 microns.

STANDARD SAMPLES

A series of standard samples of titanium
in steel was obtained from the Bureau of
Standards. They were numbered and had the

following concentrations:

Number Titanium
8 1.19
7 .61
6 .33
5 .004
4 .06
1 .04
2 .00

11

The work inthis particular investigation was
concerned only with these concentrations above
0.04% titanium.

The samples as obtained were cylindrical
rods 5/6 inches in diameter over 3% inches of
their length and S/B inches in diameter over
the remainder of their length-lé-inches. The
length of the smaller diameter was the part
exposed to the action of the Spark. The
surface was not renewed for successive ex-
posures during the preliminary investigations,

but a fresh surface is desirable for greater

accuracy, as well as for greater reproducibility.

SPECTROGRAPH AND DENSITOMETER

The spectrograph used in this investigation
was a Banach & Lomb 160 cm. Littrow. It was
found that a definite improvement in the sharp-
ness of the spectrum lines on the photographic
plate could be obtained by using values of the ,
tilt and focus slightly different from those
given by the manufacturers. The improvement
is to be noticed chiefly when a series of lines
very close together, such as the configuration
of iron lines at 3100A, is observed. A more
distinct image of the spectrum lines has the
effect of increasing the resolution of the

instrument.

12

The densities of the epertrum lines and of
the various steps in the step sector patterns
were determined on a Bausch & tomb density com-
parator. A Leeds & Northrup type R d'Arsonval
galvanometer having the following characteristics
was used in coniunction with the comparator:

Sensitivity: 0.00047 muA/mm.
Coil Resistance: 500 ohms
C. D. R. X.: 10,000 ohms
Period: 6 seconds
In order to obtain a critical damping of the
galvanometer, about 9000 ohms was added in
series to the external resistance of the
circuit. A deflection of about 70 mm. was
then obtained at a distance of two meters as
the difference between the clear plate and the
black reading. It was possible to read the
scale to 0.5 mm.
The densities of the various lines were

calculated from the formula

2 ‘Igo
73-102
g'-go
where the g's represent deflections on the scale
of the instrument, 3 the clear plate reading,

go the black reading, and g' the line reading.

15

PROCESSING OF PLATES

A11 plates were developed in EKC D-19
developer and fixed in EKC acid hypo fixing
solution. The 40 plates were developed for
five minutes with agitation at fifteen second
intervals. They were left in the fixing
solution for twice the length of timereouired
for complete clearing. When the time re-
quired to completely clear the plate reached
ten minutes the solution was discarded. A
convenient test for the exhaustion of the
fixing solution may be made by mixing 50 cc.
of the solution with 5 (so. of 4% KI. The
formation of a yellow cloudiness indicates
that the solution is exhausted.(3).

The effect of the exhaustion of the
developing solution on the gamma of the photo-
graphicemulsion may be seen in Curves A,

Fig. II. Curve (2) is the H & D curve of a
plate exposed and developed under the some con-
ditions as the nlate whose H k D curve is (l),
but which was developed after 98 more plates

had been put through the solution. It is
evident that the gamma is increased as the
so‘ution is used. With the small tank of dev-
eloper used and the extensive use ofthe solution

by'many others, the decrease in the volume of the

14

solution because of drag-out and evaporation
presented quite a serious problem, so much so
that it was necessary from time to tine to
bring the level of the solution back to its
original height in order that the photographic
plate would be completely immersed. In this
connection it was desirable to maintain the
potency of the solution as nearly as possible.
Curves B and C in Fig. II indicate that the
and water
addition of eonal parts of fresh developer,br
of voter alone accomplishes this purpose.
Curve (1) in each case represents the H & D
curve of a plate develOped before the level of
the solution was restored, curve(°) that of
a plate deve10ped immediatelv after renewal of
the solution.

In all probability the reason why the
addition of water onlv has the same effect as
the addition of equal parts of water and fresh
developer is that the water alone was added
When a greater degree of exhaustion hadbeen
reached. This would indicate, therefore, that
the prOper proportion of water and developer which
should be added to the solution depends upon the
degree of exhaustion of the solution, and that
the proportion of developer to be used decreases

as the exhaustion becomes more complete.

15

All plates were processed at 18° C, the
tanks containing the develOper and fixer
being maintained at this temperature by a
water bath. Between five and ten minutes
is required for the solutions in the tanks to
reach temperature equilibrium with the
water bath after the external temperature is
brought either up or down to the standard
temperature. A small amount of stirring

hastens the attainment of eouilibrium.

16

DISCUSSION

The results obtained in this study are
shown in Figs. III and IV. The ratios of
the densities of the two titanium lines
used as working lines tothe density of the
iron line 3570.10A. the internal standard
line, are plotted against the logarithms of
the per cent titanium in the standard samples.
The possibility of using the titanium lines
3361.913 and 3372.800 in a spectrographic
analysis is clearly indicated.

In the curves shown it can be seen that
the densities obtained for the same lines and
the same exposures are not eoual. This
difference would be corrected for in any
analytical method by referring the densities
of the lines to a standard calibration curve
or by using the differences in the densities
of the working and internal standard lines as
the basis for the working curve.

The regular decrease in the intensities of
the 3361.213 and 3379.800A titanium lines as the
concentration of titanium decreases is apparent
in Fig. V.

Since the concentrations obtained from

the working curves constructed were only within

17

30 t of the concentrations given for the
standard samples, the results could not
properly be called euantitative. An error

of this magnitude could, however, be tolerated
in a survey analysis.

The same two titanium lines used in
this study have been used for the estimation
of titanium in steel by the internal com-
parison method.(4)

The relation betreen the intensity of the
titanium line at 53R0.30A and the concentration
of titanium in titanium-steels has been in-
vestigated by Brody.(5). The iron line at

3369.549A was used as the internal standard line.

SUMMARY

Curves showing the ratio of the densities
of the titanium lines ss61.als and ssva.a00A
to the density of the iron line 3570.09'7A ”as
a function of the concentration of titanium in
steel have been presented. The results indicate
that the titanium lines given could be used in
a method for the spectrographic analysis of

titanium in steel.

(1)

(2)

(4)

(5)

REFERENCES

"M. I. T. Wavelength Tables," George R.

Harrison, Ed. pp. 270, 272. John Wiley
& Sons, Inc., New Ybrk, 1959

L. W. Strock, "Spectrum Analysis with
the Carbon Arc Cathode Layer, " p. 39.
Adam Hilger, Ltd., London

C. B. Neblette, "Photography, Its
Principles and Practice," Rd. 3, p. 342.
D. Van Nestrand Company, Inc., New York,
1939.

P. G. Barker, J. Iron Steel Inst., 139,
211 (1939) “'-

J. K. Brody, "Spectral Line Ratios of
Iron-Nickel and Titanium-Iron." Thesis
for the degree of M. S., Michigan State
College, 1941.

 

£4

T2. 0

£6

 

[MI/77’

70.8

 

Flax: COMPAR/erfl 0!
£4 JTMA N PZAI'IJ‘

.SA.#/
33

40

d‘ I'EP NUMBER
3 12 I6

 

 

 

 

xvv 3. NM . .. V j

m
\Q 3 S m o
\9\ mx 3 m. . o
kmms§< amt. v.9
m u.
3 , 3
\Q \\\
S \kSSwQ 3
o
Q Q w .1
«Q

tfllpquwQ kg
«.33sz ¢ 22.25 355% $ N
.HGQ

 

 

 

 

a Q *9 Ly “no- 3. a: .3. 3d

 

 

§< 3&33. Qtwqiwxs ‘ as C Q \ 6
. . .R R. 69%
No
so
‘6
Ba
b6
.. at e .
59.693. w - s 32.2. {Q
bxx wk YNQ .
N\b~a\..ka
: m I V
.333... w - e me
fix “£31
Ck g
C
H. 6Q 3

fix

 

 

 

 

 

 

 

MC 9. . Q? E- In...) EN: U.
S< 31$}. omwosxfie- C .1 CR. one .. _ S
N G
we
0
no
3
.. Due
33.0w.» I be
e: Exec.
\.§kL.H UKQ
.. m- .V 3.2.5.:
2333‘- e . Q
.3 EVE .
xwk e.
o.
\.
NH «.Q 3

m4

 

 

it“ * I

i

M

V
v

H

W

W

III. !‘II lllll “Ill Illnl

l1 ll Illlllllll II I!” Mllllllllll

.llll III lllll lllll'lll Illl

.
,
1
I
I

II I Illllllll Ill Illl tllllll Illl

IIIIIIIHIIIIII II II "III

.m .I It “I lllllll m I H IIIllll lllllllllllllfl u I
I nu II H mm llll Mlllllll llllllllllllllll

‘1 I M

t
\

l lllllll IIII I ll Illlll lllzllllllllllll II II'

lIllllll IIHI ll lllIII H]

}
F
a

W H

U

1

1'“ WM

1 'IHIIII rl “I“.IIHI

FIG. V.

lfllllll illllllilHllH il I. ?III HI

I

I
J

—=

um»

HM
‘1‘

H i '11:

NH.

'7
ll
1

. I

'i

. . 4””!!!

[li

NH! ~ HI

1

Illllllll II II II

lllllllll II II I;

 

all,!v-

gun... . . . . _ .m.fiw%m$»..uo {Flh‘tlt-gj .
. .. : ”in... ..‘ i .. ..,.

z 1.... 1......«6

 

MICHIGRN STRTE UNIV. LIBRQRIES

 

II I IIII IIII III [III III ,

312930177 83 7