Vol 17, No. 2 ANALYTICAL EDITION February, 1945
INDUSTRIAL
AND
ENGINEERING CHEMISTRY
S C I E N C E IS POW ER
R eprint of Editorial M atter
Advertising O m itted
INDUSTRIAL a n d ENGINEERING CHEMISTRY
A N A L Y T I C A L E D I T I O N
W A L T E R J . M U R P H Y , E D IT O R ISS U ED F E B R U A R Y 22, 1945 ♦ V O L. 17, NO. 2 C O N S E C U T IV E NO. 4
?. 105 %
M anuscript Editor: G . G l a d y s G o r d o n
R . P. C h a p m a n
J. R . C h u r c h i l l
B. L . C l a r k e
Lawrence T. Hallett, Associate Editor M anuscript A ssistant: S t e l l a A n d e r s o n
A dvisory Board T . R . C u n n i n g h a m
G. E . F . L u n d e l l M . G . M e l l o n
M ake-up Assistant: C h a r l o t t e C . S a y r e
R . H . M ü l l e r
B. L . O s e r
H . H . W i l l a r d
SY M PO SIU M ON SPECTROCHEMICAL M ETH ODS OF ANALYSIS
In tr o d u c to r y R em ark» . . . . R. Bowling Barnes 65 A n a ly tica l A p p lica tio n s o f E m issio n S p ectro m etry
J. Raynor C hurchill 6 6
M a ss S p e c t r o m e t r y ... H. W. W ashburn, H. F. Wiley, S. M. Rock, an d C. E. Berry 74
L ig h t A b so rp tio n S p e ctro m etry . . M. G . Mellon 81
B ea rin g C orrosion C h a ra cteristics o f L u b rica tin g O ils. In d ia n a S tir r in g C orrosion T e s t ...
C . M. Loane an d J. W . G aynor 89
A n a ly sis o f W ood S u g a r s ...
Jerome F. Saeman, Elwin E. Harris, an d A lbert A. Kline 95 M odified D rop p in g F u n n e l Milton O rch in 99
T h o r iu m N itr a te T itr a tio n o f F lu orid e w ith S p ecial R eferen ce to D e te r m in in g F lu o rin e a n d S u lfu r in H yd rocarb on s . . M. P. Matuszak an d D. R. Brown 100 D e te c tio n a n d E s tim a tio n o f S te a m -D is tille d W ood
T u r p e n tin e in G u m S p ir its o f T u r p e n tin e . . . . Sidney R. Snider 107
A n a ly sis o f S ilver P la tin g S o l u t i o n s ...
J. N. G regory an d R. R. H ughan 109
D e te r m in in g T races o f B is m u th in C opper by M ean s o f D i t h i z o n e ...Yu-Lin Yao 114
F a cto rs A ffectin g D e te r m in a tio n o f P o ta sh in F e r t i l i z e r s ...H. L. Mitchell an d O . W. Ford 115
A p p aratu s for P rep arin g S a m p le s for A n alysis.
R apid P u lv erizin g a n d M ixin g o f S m a ll S olid S a m p le s ...
Louis Lykken, F. A. Rogers, an d W . L. Everson 116
D evice for F eed in g L iq u id s a t Specified R a te s . . . Harvey Diehl and Clifford H ach 119
A ltern a tin g C u rren t-O p erated T h erm io n ic T itr im - e te r w ith A d ju sta b le R an ge an d S e n s itiv ity . . .
Edmund M. Buras, Jr., an d J. David Reid 120
R apid D e te r m in a tio n o f F a t in M ea t a n d M eat P r o d u c t s ...R. B. O esting an d I. P. Kaufman 125
L ow -L ag T o lu e n e T h e r m o r e g u la t o r ...
Paul E. Snyder and H arry Seitz 126
M ICROCHEM ISTRY
G row th S tim u la n t s for M icrobiological B io tin A ssay . . . Virginia R. Williams an d E. A. Fieger 127
C olor T e st for O ils a n d R esin s, U sin g H irsch so h n R e a g e n t for C h o lestero l . . H oward C. Brinker 130
T h e A m erican C hem ical Society assum es n o resp o n sib ility fo r th e s ta te m e n ts an d opinions ad v an ced b y c o n trib u to rs to its p u b licatio n s.
31,400 copies of th is issue p rin te d . C o p y rig h t 1945 by A m erican C hem ical Society.
P u b lish e d b y th e A m erican C hem ical S ociety a t E a s to n , P a . E d i
to ria l Office: 1155 16th S tree t, N . W ., W ash in g to n 6, D . C .; telep h o n e, R ep u b lic 5301; cable, Jiech em (W ash in g to n ). N ew Y o rk E d ito ria l B ran ch : 60 E a s t 42 n d S tree t, N ew Y o rk 17, N . Y ., telep h o n e, M u rra y H ill 2-4662.
C hicago E d ito ria l B ra n ch : 310 S o u th M ichigan A venue, C hicago 4, 111., te le p h o n e, W ab ash 7376. B usiness Office: A m erican C hem ical Society, 1155 16th S tree t, N . W ., W ash in g to n 6, D . C. A d v ertisin g Office: 332 W est 42 n d S tre e t, N ew Y ork 18, N . Y .; telephone, B ry a n t 9-4430.
E n te re d as second-class m a tte r a t th e P o s t Office a t E a sto n , P a ., u n d e r th e A c t of M arch 3, 1879, as 24 tim es a y e ar— In d u s tria l E d itio n m o n th ly on th e 1st, A n aly tical E d itio n m o n th ly on th e 15th. A cceptance fo r m ailin g a t special r a te of po stag e p ro v id ed for in S ection 1103, A ct of O cto b er 3, 1917, a u th o riz e d Ju ly 13, 1918.
R em itta n ce s a n d ord ers fo r su b scrip tio n s an d fo r single copies, notices of changes of ad d re ss an d new professional connections, a n a claim s for m issing n u m b ers sh ould be se n t to th e A m erican C hem ical Society, 1155 16th S treet, N . W ., W ash in g tp n 6, D . C. C hanges of ad d ress fo r th e In d u s tria l E d itio n
m u s t be received on or before th e 18th of th e preceding m o n th a n d fo r th e A n aly tical E d itio n n o t la te r th a n th e 3 0 th of th e preceding m o n th . Claim s fo r m issing n u m b ers will n o t be allow ed (1) if received m ore th a n 60 d ay s fro m d a te of issue (owing to th e h a za rd s of w artim e delivery, n o claim s can be honored fro m sub scrib ers o u tsid e of N o rth A m erica), (2) if loss was d u e to failu re of n o tice of change of ad d ress to be received before th e d a te s specified in th e preceding sentence, or (3) if th e reason fo r claim is “ m issing from files” .
A n n u al su b scrip tio n — In d u s tria l E d itio n a n d A n aly tical E d itio n sold only as a u n it, m em bers S3.00, n onm em bers $4.00. P o stag e to co u n tries n o t in th e Pan -A m erican U nion $2.25; C an a d ia n p o stag e $0.75. Single copies—
c u rre n t issues, In d u s tria l E d itio n $0.75, A n aly tical E d itio n $0.50; b ack n u m b ers, In d u s tria l E d itio n $0.80, A n a ly tic al E d itio n prices on re q u e st;
special ra te s to m em bers.
T h e A m erican C hem ical Society also p ublishes Chemical and Engineering News, Chemical Abstracts, a n d Jo u rn a l o f the A m erican Chemical Society, R a te s on req u est.
INDUSTRIAL a n d ENGINEERING CHEMISTRY
P U B L I S H E D BY
♦ / » T H E A M E R I C A N C H E M I C A L S O C I E T Y W A L T E R J. M U R P H Y ,
E D I T O R
S Y M P O S I U M
V
Spectrochemical Methods of Analysis
Introductory Remarks
f N U R I N G the past ten to fifteen years important strides have
^ been m ade in the ap p lication of p hysical instruments to analytical problem s of a chem ical nature. The p o p u larity of the various techniques w hich make up the science of instrumental analysis has increased from year to year and particularly during the present em ergency have the m any advantages o f these analytical techniques becom e w ell known. Through their use a lo n g list of analyses important to the accelerated production schedule of the industries of this country is b e in g perform ed d aily.
M a n y of these analyses were formerly considered to be im possible ; others, w hich are n ow com pleted auto
m atically or within a few minutes at most, previously required up to several hours.
In d e v e lo p in g and ad ap tin g these various instruments of physics to the problem s and requirements of the chemist, the physicist and the instrument maker have rendered the subject of analysis an invaluable service. It has indeed been interesting to observe the various steps in the c o n version of these research instruments of physics labora
tories into com m ercially available devices for routine analysis and in certain cases into instruments for industrial control purposes. A lt h o u g h o n ly a b e g in n in g has so far been m ade in the instrumentation of the subject of analysis, the successes already achieved are of such m agnitude as to indicate much further w ork a lo n g these lines in the future. A ft e r the close of the war, much of the information now secret w ill u n d ou b te d ly be directly a p p lica b le to
these problem s, and m any interesting instrumental d e velopm ents m ay be expected.
The im portance of instrumental analysis is attested b y the number of pertinent pu blication s w hich have appeared in recent years and b y the frequency with w hich sym posia on som e phase of the subject have been sponsored b y scientific societies. A s the various m ethods reached higher and higher states of perfection, so also d id the dem ands increase for more detailed information regard
ing the instruments, the techniques, the p o ssib le fields of ap p licaticn , and the potentialities of the methods.
In direct answer to such dem ands, the D iv isio n of A n a ly t ic a l and M ic r o Chem istry and the D iv isio n of Physical and Inorgan ic Chem istry held a joint sym posium during the N e w Y o rk M e e t in g of the A M E R IC A N CH E M IC A L So c ie t y on Septem ber 15, 1944. Sin ce o n ly one d a y cou ld be devoted to this sym posiu m , it was im possible to cover m any of the instrumental d e v e lo p ments w hich w ou ld have been interesting. The discussions were therefore lim ited to the general subject o f Spectro
chem ical M e t h o d s of A n a ly s is . The success of this sym posium is a matter of record, and the officers of the two divisio n s and the speakers are to be congratulated.
Papers b y Churchill (p . 6 6 ), W ashbu rn, W ile y , R o ck , and Berry (p . 7 4 ), and M e llo n (p . 8 1 ) are printed in this issue,- other papers w ill appear soon. It is h o p e d that their pu blication w ill prove interesting, helpful, and stim ulating to m any w ho were unable to attend the sym posium .
R. B O W L I N G B A R N E S Stamford Research Laboratories, A m erican Cyanam id Com pany, Stamford, Conn.
65
Analytical Applications of Emission Spectrometry
J. R A Y N O R C H U R C H I L L
A lu m in u m C om pany of A m erica, N e w Kensington, Pa.
The status of emission spectrometry as an analytical technique in the general analytical laboratory is discussed with particular attention to nonroutine applications. A number of techniques, involving qualitative and semiquantitative tests on a wide variety of materials, are described. Specific examples, drawn from the files of A lum in um Research Laboratories, are given to illustrate the usefulness of the
I
N T H E complete m odern analytical laboratory, emission spectrom etry occupies a place com parable in im portance to gravim etric analysis or titrim etry. Used in com bination w ith oth er analytical tools, the spectrograph often provides the analyst w ith inform ation which m ight otherwise be unobtainable, and protects him both from failing to detect unexpected metallic elem ents and from m aking tedious, time-consuming, and sam pleconsum ing tests for metallic elem ents which are absent or present in insignificant quantities. Occasionally, the laboratory en
counters a problem which can be solved entirely by spectroscopic m ethods. More often the spectrograph is used as a prelim inary to, or in combination with, other analytical techniques. The frequently encountered question, “ Will th e spectrograph even
tually largely replace the chem ist in m etallurgical analysis?” , is sim ply a m anifestation of a popular misconception as to the n atu re and use of the spectrograph in the analytical scheme.
W ith no more absurdity, one m ight inquire w hether the analytical balance will eventually replace the analytical chemist. Both th e spectrograph and th e chemical balance are sim ply tools of th e chem ist’s a rt, and either can be used effectively in simple, repetitive analytical procedures by routine operators who can lay no claim to being analytical chemists.
D uring recent years the m ost highly publicized phase of emis
sion spectrom etry has been its application to high-speed or large- volume routine q u an titativ e analysis. T he im portance of this work is n o t exaggerated by the seemingly extravagant claims of those engaged in this work, or by the enthusiastic aggressive advertising of the instrum ent m anufacturers. I t is literally tr u e th a t hundreds of thousands of dollars have been saved in ap
plications, such as in the alum inum industry, in analytical costs alone, to say nothing of th e benefits resulting from the speed of obtaining results. T o the analytical chemist, however, routine q u a n tita tiv e analysis is of only casual interest, except in so far as he m ay be engaged in the supervision of such work, the develop
m e n t of routine methods, or the evaluation of the stan d ard sam ples upon which the routine m ethod is based.
T he m ost im p o rtan t service of the spectrograph to the an
alytical chem ist is the supplying of qualitative inform ation w ith respect to th e m etallic and sem im etallic elem ents. N o other technique provides so m uch inform ation w ith so little w aste of tim e an d effort. Q ualitative analyses are alm ost completely objective, expected and unexpected elem ents being detected w ith th e sam e ease and certainty. Moreover, sm aller quantities of m ost elem ents are detectable by spectrography th a n by or
d in ary w et chemical tests. T o be of an y real valùe to th e an
alytical chemist, a so-called qualitative procedure m ust actually yield sem iquantitative inform ation. T he degree to which a spectrographic analysis is q u an titativ e depends on the nature of th e m aterial analyzed, the conditions of excitation, the famil
ia rity of the analyst w ith th e p articular elem ent and m atrix in
volved, an d the ex ten t to which stan d ard sam ples are employed.
I n th e roughest sort of q ualitative analysis, th e analyst is gen
erally able to place th e elem ents present in their proper orders of m agnitude. U nder very favorable circumstances, he m ay be
spectrograph in general qualitative analysis in the identification of alloys, in seeking explanations for differences in physical and chemical properties, in the analysis of coatings and platings, in corrosion investigations, and in a variety of other special applica
tions. The limitations, as well as the advantages, of spectrographic methods in qualitative and quantitative analysis are pointed out.
able to m ake q u an titativ e determ inations of an accuracy com
parable to or exceeding other m ethods of analysis. H igh quan
tita tiv e accuracy is to be expected only in cases where th e sam ple is compared directly or indirectly w ith very sim ilar sam ples of known composition and when a m ethod of excitation, peculiarly suited to th e particular ty p e of sam ple a t hand, has been de
veloped. In handling miscellaneous sam ples of a highly varied nature, it is n o t only im practical, b u t v irtu ally impossible, to place a very large proportion of th e spectrographic analyses in a general laboratory on a strictly q u a n tita tiv e basis because of th e problems of standardization an d excitation involved.
I t is true th a t alm ost every analytical laboratory is called upon to perform th e functions of a routine laboratory, a t least oc
casionally. T his circum stance m ay arise whenever the laboratory is required to analyze a large num ber of sim ilar sam ples. In such cases, th e laboratory m ay be able to develop a sufficiently economical q u a n tita tiv e spectrographic m ethod to ju stify the time and expense involved in th e prelim inary experim entation, standardization, an d calibration work required. Such pro
cedures are routine in n atu re and are not w ithin th e scope of this paper. Of greater in te re st to th e analytical chem ist are the q u an titativ e spectrographic analyses which can be used to good advantage in nonroutine work. There are four general cases where precise q u a n tita tiv e spectrographic analysis is feasible in nonroutine w ork: (1) when all other available m ethods of test are so difficult or tim e-consum ing th a t the prelim inary work in
volved in th e spectrographic analysis is justified; (2) w hen no other m ethod will give sufficient sensitivity of detection or suffi
ciently high accuracy; (3) when th e sam ple available is too sm all to be analyzed by other m ethods; an d (4) when a previously de
veloped procedure and a suitable set of stan d ard sam ples happen to be available when th e sam ple is received. All four situ atio n s m ay occur rath er frequently in the general analytical laboratory.
A P P L IC A T IO N S O F E M IS S IO N S P E C T R O G R A P H Y
From th e point of view of th e analytical chem ist w orking w ith a diversity of m aterials in a nonroutine laboratory, th e following qualitative or sem iquantitative applications of emission spectrography are of greater interest and usefulness th a n the strictly q u an titativ e applications.
Pr e l i m i n a r y Te s t i n g o p Un k n o w n s. Probably th e most im p o rtan t function of the spectrograph in the general analytical laboratory is th e prelim inary analysis of miscellaneous sam ples as a guide to q u an titativ e analysis which is to follow. In m ost m odern, general analytical laboratories every unknow n sam ple is first examined spectrographically by a simple, more or less universal procedure which yields a dependable, sensitive test for th e m ajority of th e m etallic elem ents and provides useful sem iquantitative inform ation. From th e spectrographic d a ta , th e analyst n o t only learns w hat elem ents he should analyze for b u t is able to select th e procedure, sam ple weight, an d qu an tities of reagents m ost suitable for the am ounts of the elem ents present.
T he analyst is protected not only from missing rare or un
expected elem ents b u t also from errors in analysis which m ight 66
A N A L Y T I C A L E D I T I O N 67 have resulted from the interfering effects of these unexpected
elem ents in his o th er q u a n tita tiv e determ inations. Moreover, he is spared from having to ru n tim e-consum ing chemical tests for elem ents th a t are n o t present in significant qu an tity .
Id e n t i f i c a t i o no f Ma t e r i a l s. T he general analytical labo
ra to ry frequently receives sam ples on which the only d atu m re
quested is a n answer to th e simple question, “ W hat is th is m a
terial?” T he m aterial in question m ay be anything from a mas
sive casting to a m inute inclusion detected microscopically in a piece of m etal. I t m ay be a th in plating, a sam ple of rock, a p a te n t medicine, or m ysterious precipitate. Such problem s are often solved completely by a spectrographic test. In other cases a spectrographic test m ay be used in com bination w ith chemical tests for acid radicals, organic compounds, or o th er constituents which the spectrograph will n o t detect. Some of th e analytical jobs falling in th is general category m ay require an approxim ate q u an titativ e analysis, as, for example, in th e identification of an alloy, or the determ ination of th e grade of some sta n d a rd product for which a grading system based on p u rity has been established.
Co m p a r i s o n s b e t w e e n Si m i l a r Ma t e r i a l s Sh o w i n g Di f f e r e n c e s i n Ph y s i c a l o r Ch e m i c a l Be h a v i o r. A nother very im p o rtan t category of problem s in which the spectrograph is highly useful is in th e com parison of m aterials of sim ilar ty p e b u t m anifesting different properties. M etallurgical laboratories freq u en tly receive sam ples of alloys having unusually good or unusually poor mechanical properties or resistance to corrosion.
In such cases a simple spectrographic com parison betw een nor
mal and abnorm al m aterials often yields im p o rtan t inform ation, in some cases n o t only solving th e problem a t h an d b u t oc
casionally providing inform ation which m ay lead to im prove
m ents in the product. Sim ilar applications are frequently en
countered in com paring satisfactory and unsatisfactory reagent chemicals, catalysts, p ain t coatings, an d a host of oth er m aterials.
Ch e c k i n g Ch e m i c a l Se p a r a t i o n s. In th e developm ent and im provem ent of w et chemical m ethods an d in checking th e validity of certain existing m ethods, emission spectrography pro
vides an excellent m eans of studying th e completeness of separa
tions, th e p u rity of precipitates, an d the n atu re of any fractions such as insoluble residues, or precipitates of unknow n composition which m ay occur in experim ental analytical work.
Mi s c e l l a n e o u s Qu a l i t a t i v e a n d Se m i q u a n t i t a t i v e St u d i e s. In investigational work, num erous special applications of the spectrograph, n o t included in the foregoing categories, are encountered. F o r example, th e spectrograph has been very valuable in corrosion studies, bo th in diagnostic analyses of the corroded m etal an d corroding m edium an d in tests designed to determ ine th e likelihood of corrosion b y testing b o th th e m etal surface an d its environm ent. A nother valuable ty p e of applica
tion is in m aking q ualitative or sem iquantitative surveys in
volved in studying the occurrence an d distribution of the rarer elem ents, in tracing th e origin of m inor im purities in products whose m anufacture involves th e use of a large v ariety of raw m aterials, in following the progress of purification processes, and in m aking large-scale statistical studies. In m any of these ap
plications, th e spectrograph provides th e only feasible m eans of obtaining th e desired inform ation. T he economy an d speed of th e spectrographic m ethod often enable th e an aly st to o b tain a m uch greater volum e of d ata, thereby m aking th e final results m ore conclusive.
A P P A R A T U S
T he m inim um basic spectrographic equipm ent of a general analytical lab o rato ry includes a spectrograph, a darkroom, a simple excitation source, such as the direct current arc, an d a few sundries an d accessories necessary for the developm ent of films or plates, th e preparation of samples, and th e viewing of spectro
gram s. W ith this m inim um equipm ent, th e analyst can do’much of the q u alitativ e and sem iquantitative spectrographic work dis
cussed. To obtain the m axim um usefulness o u t of the spectro
graph, considerably m ore elaborate equipm ent is required and the additional cost is alm ost always am ply justified. T oday, the general analytical laboratory cannot be considered complete w ith
o u t th e following:
A flexible, relatively high dispersion spectrograph, eith er one of th e larger grating instrum ents or one of th e so-called “ au to m atic” q u artz prism instrum ents.
A direct cu rren t arc source.
A conventional sp ark source or, b etter, one of th e newer, highly flexible excitation u n its which will produce th e equivalent of an y of the conventional sources.
E quipm ent for accurately controlling th e developm ent and processing of films an d plates.
A com parator or projector suitable for th e visual exam ination of spectrogram s, preferably equipped w ith a w ave-length scale an d w ith some provision for com paring tw o or m ore plates or films.
A reliable densitom eter, also preferably equipped w ith wave
length scale a n d a provision for comparison w ith a reference spectrogram . T he com parator an d d en sito m eter m ay be com
bined in one instrum ent.
Chem ical an d m echanical facilities for th e preparation of samples.
A large v ariety of accessories, including arc stands, spark stands, electrode holders, tongs, forceps, electrical and optical testing equipm ent, tim ers, etc.
W ave-length tables, spectrum atlases, an d reference books.
. T he specific m akes of instrum ents and the design of th e various accessories desirable in a p articu lar laboratory depend on the scope and n atu re of th e analytical work anticipated. More specific recom m endations along these lines can best be obtained from laboratories th a t have had spectrographic experience in a p plications of sim ilar character.
G E N E R A L TEC H N IQ U E S
Us e o ft h e Di r e c t Cu r r e n t Ar c. D espite th e em phasis on various types of controlled sparks an d on elaborate new sources recently placed on th e m ark et, th e direct cu rren t arc rem ains th e m ost widely used an d m ost generally applicable excitation source in nonroutine applications. W hile th e controlled spark is m uch m ore widely used in routine q u a n tita tiv e work an d per
forms functions for which the direct cu rren t arc is inadequate, th e direct current arc is m ore nearly indispensable th a n th e spark in general q u alitativ e an d sem iquantitative analysis. T h e arc owes its general superiority in such applications to th e high sensitivity of detection it provides an d to th e fact th a t it can be used conveniently on alm ost a n y ty p e of sam ple.
All direct current arc techniques are very sim ilar an d th e dif
ferences th a t exist are generally superficial modifications neces
sary to increase th e stab ility of th e arc in a p articular application or to favor th e detection of elem ents of special in terest in a p ar
ticular test.
In v irtu ally all cases a sm all portion of th e sam ple (usually n o t over 10 m g.) is volatilized in an arc operating a t from 4 to 15 am peres. T he light from th e entire arc or from th e central portion of th e arc is ordinarily used for th e analysis. (Special techniques, such as th e cathode layer, will not be considered here.) I f th e subm itted sam ple is of suitable dim ensions and properties, one o r b o th of th e electrodes m ay be specimens form ed from th e sam ple. I n th is case, th e arc is o p erated for a tim ed interval determ ined by previous experience. M ore often, a sm all portion of th e sam ple is placed in a crater form ed in the end of a graphite electrode. T his electrode is used as th e lower electrode in th e arc an d is usually positive. A nother graphite rod is used for th e negative upper electrode. An arc of th is ty p e is m ore generally applicable to th e v ariety of sam ples received in a general laboratory and is usually more satisfactory th a n an arc using self-electrodes because it enables th e operator to vola
tilize com pletely a known a m o u n t of sam ple an d because it affords greater sensitivity in th e detection of sm all am ounts of th e elem ents less easily excited in th e arc.
T he arc is essentially a therm al phenomenon. T he elem ents present in th e sam ple m u st be vaporized before th ey are excited an d th e more volatile elem ents will be volatilized an d excited first. T he tem p eratu re a tta in e d by th e electrodes is controlled
68 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol. 17, No. 2 by th e am perage applied, the relative am ounts and boiling points
of the elem ents present, and th e conditions surrounding the arc which affect the transfer of heat. W hen th e sam ple placed in the electrode crater contains constituents of widely different boiling points, th e elem ents of highest boiling p o in t volatilize relatively slowly, if a t all, in th e early stages of the exposure and do not a tta in m axim um excitation until the more volatile substances have been vaporized. W hen self-electrodes are used, m inor elem ents of lower volatility th a n th e m atrix are often very diffi
cu lt to excite, and sensitivity of detection is correspondingly low.
F o r example, molybdenum is n o t detected in alum inum alloys containing up to a few hund red th s per cent when self-electrodes are used, b u t can be detected in am ounts less th a n 0.0 0 1% when a sm all sam ple of th e alloy is com pletely volatilized in a graphite arc.
T he successive excitation of th e elem ents in th e inverse order of volatilities m akes possible an increase in sensitivity of detec
tion for any given elem ent by photographing only th a t portion
■of th e arcing cycle during which th e intensity produced by th a t elem ent is a t a m aximum . T o do this, by m eans of a single ex
posure requires th a t previous d a ta be available which will enable the an aly st to select the proper portion of th e arcing period for the test. A more practical scheme, when extrem e sensitivities are required, has been used occasionally in some laboratories.
This system is generally referred to as th e moving plate technique an d consists of simply m oving the photographic plate vertically, either a t frequent regular intervals or in a slow continuous mo
tion during the arcing cycle. While th is technique is useful in p articular cases, it is n o t generally necessary or particularly desirable in m ost spectrographic analysis because of complica
tions and uncertainties introduced in attem p tin g to integrate intensity over a period of tim e in m aking q u an titativ e estim ates and because simpler, less elaborate techniques are generally adequate.
G raphite is alm ost universally used for arc electrodes in general qualitative and sem iquantitative work. U nlike m etallic elec
trodes, it will n o t m elt or form a crust of oxide. M oreover, it is easily formed to the desired shape and, m ost im portant of all, it is obtainable in th e extrem ely high p u rity necessary for m any purposes. G raphite electrodes could not, of course, be used in testing for carbon. However, carbon tests are rarely made spectrographically and are considered im practical in m any spectrographic laboratories.
F o r certain purposes, th e so-called “ regular” grade spectro
scopic graphite is useful an d electrodes of th is grade are obtain
able from a t least four m anufacturers in th e U nited States.
These electrodes are very inexpensive (approxim ately 6 cents per foot for 0.25-inch diam eter rods) and m ay be purified by chemical tre a tm e n t or by high-tem perature exposure to various atm ospheres to increase th eir applicability to th e analytical problem s of th e general laboratory. However, laboratory m eth
ods of purification are not com pletely satisfactory and even-the purified electrodes are unsuitable for general q ualitative work when m inor im purities are im p o rtan t. A very high-grade graphite, sold under th e designation “ special spectroscopic graph
ite ” , is th e only graphite available in sufficient p u rity to be generally applicable in q u alitativ e analyses for m inor con
stitu en ts. As far as is know n to th e author, graphite of «this grade is obtainable from only one m anufacturer an d th e cost is understandably m uch higher th a n th e “ regular” or interm ediate grades available elsewhere.
In A lum inum Research L aboratories and in m any other laboratories having a sim ilar diversity of problems, it has been found expedient to standardize on th e special spectroscopic graph
ite rods, n o t only because a large proportion of th e analyses m ade actually require graphite of th e highest p u rity b u t also because of th e risk of mixing the grades in a busy laboratory in w hich a large num ber of analysts are doing sim ilar work. T here are, of course, occasional instances w here oth er supporting electrodes are used.
Since th e special spectroscopic graphite has been available, the use of supporting electrodes o th er th a n graphite has been largely restricted to tests for carbon, tests in which th e cyanogen bands produced in th e graphite arc are objectionable, and specific tests in which less expensive electrodes, such as copper, can be used.
Some spectrographers prefer ungraphitized carbon rods to graph-
Figure 1. Typical Graphite Electrodes U sed with the Direct Current A r c
I to 10, lower electrodes I I to 13, upper electrodes
ite rods. The extensive use of ungraphitized carbons has been prevented by th e difficulty in shaping electrodes from this m a
terial.
There are alm ost as m any different shapes of supporting elec
trodes as there are laboratories doing arc work. M ost of these differences are sim ply m inor variations in dimensions an d are not particularly im portant. Figure 1 shows a typical assortm ent of electrodes representing th e m ost generally used types.
No. 1 is a simple shallow cup drilled in th e end of a 0.25-inch electrode. This is probably th e m ost widely used type for m is
cellaneous qualitative and sem iquantitative analysis. N o. 2 has a deeper cavity and is widely used when th e sam ple is introduced into the cavity as a solution or suspension. In such cases the electrode is dried before arcing. Nos. 3 an d 4 are modifications of Nos. 1 and 2, respectively. T he addition of th e center post reduces the w andering of th e arc. Nos. 5, 6, 7, an d 8 are the so- called necked electrodes. N ecking th e electrodes aids in a tta in ing higher electrode tem peratures by reducing th e w ithdraw al of h eat by conduction along th e electrode. Nos. 9 and 10 are special purpose electrodes, developed by an d sold in fabricated form by Applied Research L aboratories. N o. 9, th e boiler electrode, is potentially useful in specific tests for th e m ore volatile elem ents and in th e moving plate technique. T his electrode is specifically designed to enhance preferential distillation effects in th e arc. No. 10, know n as th e platform electrode, has been used in th e q u an titativ e analysis of refractories, ore m aterials, and steel samples.
Nos. 11 an d 12 are th e m ost generally used upper electrodes, No. 11 being simply a sh o rt length broken or sawed from a graphite rode, and N o. 12 having been sharpened in a pencil sharpener or o th er cutting device. N o. 13 is an upper electrode developed for use w ith the platform electrode, and necked to aid in attain in g high arc tem peratures. W hen used w ith samples containing large am ounts of alkali, th is form of upper electrode is effective in preventing th e crawling of th e arc up th e sides of the upper electrode.
T he m ajority of laboratories use 0.25-inch rod in the preparation of electrodes, b u t some use 0.125-, 3/i«-, and 6/ i 6-inch rods. There is also a rath er wide range of cavity depths, center post heights, and wall thicknesses in use in various laboratories. While the thirteen electrodes show n illustrate th e m ost im p o rtan t distinct types, special shapes are often devised for handling individual samples or for certain specific tests. In testing for sm all am ounts
February, 1945
of m ercury, arsenic, or other volatile constituents, a very deep cavity (sometimes 0.5 inch or more) of sm all diam eter has been found effective. Such a n electrode perm its th e distillation of the volatile elem ents from a relatively large sample.
T he center post an d th e pointed upper electrode are b o th ef
fective in reducing th e am ount of w andering of the arc during ex
posure. I t is particularly desirable to keep the arc well centered w ith respect to th e electrodes when th e arc im age is focused directly on th e slit of th e spectrograph. I t is m uch less im portant w hen a n image is focused on th e grating or collim ator of th e spectrograph. I n th is case a magnification can be selected a t which th e arc image will rem ain w ithin the collim ator or grating as th e arc w anders around th e periphery of th e electrode tip and the use of center posts an d pointed upper electrodes becomes unnecessary. A center post m akes an electrode som ew hat less convenient in loading and slightly enhances the band spectra produced by th e carbon compounds formed in th e arc.
Sim ply placing a portion of th e sam ple in a n electrode crater an d applying current will n o t usually produce an arc suitable for general analytical purposes. M ost m etal sam ples will sp it and sp u tter and, in some cases, accum ulations of oxide will form and interfere w ith the stab ility of th e arc. H y d rated salts or oxides will evolve w ater vapor so rapidly th a t th ey will be largely ex
pelled from the arc before a representative spectrum has been obtained. A sim ilar perform ance is to be expected for any sam ple containing highly volatile salts or com pounds th a t de
compose w ith th e evolution of gases or react w ith violence under th e influence of the sudden rise in tem perature when th e arc is struck. Samples containing com pounds of low volatility or con
taining elem ents which form such com pounds in the arc often form sm all m olten beads which frequently roll o u t of th e cavity during arcing.
M ost of th e above effects can be largely elim inated by the proper preparation of the sam ples for arcing an d by th e adm ixture of certain substances w ith th e samples. T he first and m ost ob
vious step in th is direction is the prelim inary heating of the sam ple to elim inate w ater an d to decompose an y unstable com
pounds tending to cause sp u tterin g or loss of sam ple in th e arc.
Sometimes a chemical tre a tm e n t is necessary to convert th e orig
inal compounds present into compounds more am enable to arcing.
In m any cases it is fu rth er desirable th a t the sam ple be diluted w ith some noninterfering m aterial w hich will prevent the form a
tion of mobile beads of m olten salt, oxides, or m etals, and will assist in the volatilization of elem ents of high boiling points or elem ents existing in extrem ely nonvolatile compounds.
Chem ical analysts in th e field are sharply divided as to th e type of m aterial m ost suitable in accom plishing these ends. One group prefers th e addition of a volatile or highly reactive m a
terial which will m echanically spray th e sam ple up into th e arc.
Ammonia salts are th e m ost common additions of th is ty p e and m any an aly sts add am m onia salts in all general q ualitative and sem iquantitative work. W hile th is m ethod represents an im
provem ent over th e simple arc in w hich no diluent is used, it does n o t provide th e best sensitivity an d dependability in dealing w ith such m aterials as alum ina, zirconia, or Columbia. T he other group of analysts, of which th e a u th o r is a m em ber, a d vocates doing all possible to p rev en t loss of sam ple by sp ittin g or spraying and accom plishing th e excitation of elem ents forming refractory oxides by chemical reduction.
B oth ends are accomplished by th e addition of pow dered graph
ite. F o r best results th e sam ple is finely ground an d very in ti
m ately mixed w ith a t least tw ice its w eight of graphite. Such a m ixture yields a steady arc w ith a m inim um of spraying. T he m ost refractory oxides are decomposed an d th e sam ple is vola
tilized in a relatively sh o rt exposure. T he m olten salts or ox
ides are largely prevented from form ing beads. R a th e r high reproducibility of excitation is achieved and th e sensitivity in detecting such elem ents as zirconium , m olybdenum , an d colum- bium is m uch im proved over th e mechanical spraying m ethod.
In m ost cases th e spectrum of carbon and carbon com pounds is actually w eaker th a n when no graphite is added, since th e ex
posure tim e required for complete volatilization is m uch reduced.
M etallic sam ples m ay be handled in a sim ilar way, th e finely divided particles required being generally obtained by filing.
In m any cases it is desirable or necessary to introduce the sam ple into th e electrode as a solution. T his situ atio n arises frequently when the m aterial to be analyzed is too sm all to be conveniently handled as a solid and in com parative analyses where the am ount of th e sam ple introduced into th e electrodes can be controlled m ost conveniently by m easurem ent of volume.
T he use of solutions often provides a convenient m eans of adding an internal stan d ard when q u an titativ e or sem iquantitative d a ta are sought and facilitates th e preparation of synthetic standards.
In some cases the solution is sim ply introduced into th e elec
tro d e cavity and dried by heating; in this case a large p a rt of th e sam ple will be absorbed by th e electrode. Some spectrog- raphers arc th e electrodes prior to introducing th e sam ple, for th e double purpose of purification and prom oting th e absorption of th e sam ple. O ther spectrographers prefer to tr e a t th e elec
tro d e cavity w ith some noninterfering m aterial which will prevent absorption. E ith er an acetone solution of cellulose acetate or kerosene is satisfactory for m ost cases. Sealed electrodes, in general, will yield higher sensitivities of detection, b e tte r re
producibility, and spectra w hich are weaker in band stru ctu re from carbon compounds. W hen su b stan tial qu an tities of alu
m inum , zirconium, titan iu m , or other elem ents form ing oxides of relatively high boiling points are present, th e perform ance of the are is im proved by introducing pow dered graphite or carbon into th e cavity before adding th e sam ple. T he graphite powder prom otes th e volatilization of such elem ents an d reduces the tendency tow ards form ing m olten beads of oxides.
In m ost laboratories, sam ples are introduced into the electrodes as solutions only in special cases. W hen a sufficiency of sam ple is available, b e tte r results are generally obtained by evaporating the solutions to dryness, igniting to rem ove w ater and to de
compose th e less stable salts, intim ately mixing th e residue w ith graphite or carbon, and introducing the m ixture into the electrode as a solid.
Al t e r n a t i n g Cu r r e n t Sp a r k. Spark excitation m ay be of use to th e general analytical laboratory in the following cases;
W hen th e results obtained on th e direct current arc are not sufficiently q u an titativ e.
W hen a m etal specimen is to be analyzed w ith a m inim um of danger to th e specimen.
W hen the te s t is to be restricted to a m etal surface or to a p ar
ticu lar sm all region of a m etal sam ple.
M ost analyses can be m ade som ew hat more precisely from a q u an titativ e standpoint b y using a condensed spark discharge.
Figure 2. Petrey Spark Stand
70 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol. 17, No. 2 Sem iquantitative analysis of a som ew hat higher grade th an th a t
provided by th e arc is o ften required in th e general laboratory in the identification of alloys and in seeking sm all differences in composition between various types of samples. I n the case of m etallic samples, bo th spark electrodes m ay be cut from the sam ple and sparked directly. Cases are frequently encountered where the size or shape of th e subm itted sam ple is such th a t only one electrode can be prepared conveniently from th e sam ple itself. In such cases an alter electrode of graphite is used. In Aluminum R esearch Laboratories and in m any o th er general laboratories, it is general practice to use th e sam ple as one electrode (usually the upper) and a graphite rod as th e other electrode. T he P etrey stan d (Figure 2), w hich is now widely used in q u an titativ e spectrographic analysis, was originally de
veloped for general spark work and is so used in a num ber of laboratories. Any sam ple having a surface large enough to cover the apertu re in th e top plate can be sparked by sim ply laying it on th e spark stand. Various clamps an d ad ap te rs are used to hold very sm all sam ples in place.
While spark spectra have been used less frequently for in
creasing q u an titativ e .accuracy on nonmetallic samples, th e spark can be applied very effectively to salts, oxides, rocks, and mis
cellaneous powders. T he sam ple is mixed w ith de-ashed natural graphite ( 1 p a rt sample, 2 p arts graphite) and pressed into a pellet which is used as one electrode in th e spark. This m ethod yields results alm ost equaling those obtained in sparking m etal sam ples in some cases. Q u an titativ e analyses of m aterials for which no chemically analyzed stan d ard s are available can be carried o u t w ith synthetics by the pressed pellet technique, pro
viding synthetics having sim ilar composition an d states of com
bination of com ponents can be prepared. F o r example, a m etal sam ple can be converted to oxides b y chemical treatm en t and com pared w ith sam ples of oxides prepared from volum etric solu
tions of m etal salts.
M any laboratories are frequently asked to analyze m etal speci
mens whose intrinsic or historical value or whose practical u tility m akes it necessary to analyze th e specimen w ith as little dam age as possible. O ften spark excitation is th e best m eans of obtaining th e inform ation desired w ith a m inim um of dam age to th e specimen. In cases of this kind, a low-powered spark is played on a selected area of th e sam ple from a graphite electrode.
T he spectrum so obtained is com pared w ith spectra prepared in the sam e w ay from sam ples of know n composition. W hen th e tests are complete, th e last evidence of the sparking can usually be rem oved by polishing.
Localized tests on m etal specimens often provide valuable inform ation not easily obtainable by o th er tests. T he identifica
tion of th in coatings and platings is a frequent problem in spectro
graphic analysis.
Such identifications are easily m ade by playing a low-powered spark discharge on the surface. O ften it is possible to identify each of several layers of a coating, as in th e case where one or m ore different m etals have been successively plated onto th e surface. I n such cases, a technique analogous to th e moving p late technique is som etimes useful. T he sp ark is allowed to impinge on th e given spot on th e specimen for a sufficient period to p en etrate all of th e coatings, and th e sp ectro g rap h 'p late is racked a t regular intervals during th e sparking period, giving a num ber of spectrogram s each representing a different d ep th of p enetration. Such tests are som etimes valuable in studying the diffusion of th e base m etal into th e coating. B y using a sim ilar sparking technique, it is often possible to obtain valuable inform a
tion on m etal surfaces which m ay aid in studying th e effective
ness of cleaning m ethods, th e presence of surface im purities introduced in fabrication, th e presence of surface contam inants which m ay have a bearing on corrosion problem s, an d th e n atu re of stains o r deposits of unknow n origin. In some such problem s, A luminum R esearch L aboratories have found a m oving electrode very helpful. By m oving th e sam ple slowly back an d forth during sparking, th e exposure m ay be madef to represent any desired area of th e surface. T his provides a useful technique in identifying extrem ely th in platings, coatings, or deposits which are quickly p enetrated by th e spark, and in obtaining inform a
tion as to average surface composition over a selected area.
Ot h e r Ex c i t a t i o n So u r c e s. M any analytical laboratories m ake extensive use of excitation sources other th an th e simple d irect current arc and the condensed spark. M ost im p o rtan t of these from a stan d p o in t of current usage is th e altern atin g current arc. While th is source is n o t a n essential piece of equip
m ent in m ost laboratories, it has been very useful in specific ap
plications. I n general, th e sensitivity of detection, stability, an d field of application are interm ediate between those of the direct current arc an d th e condensed spark. W ith o u t an y in
tention of derogation, the alternating current arc is om itted from this discussion because th e techniques used, th e results obtained, and th e scope of application are largely included in th e general rem arks pertaining to th e other tw o sources. In m ost applica
tions, th e alternating current arc is a good compromise between th e alternating current spark and the direct current arc, an d as such it will serve a few purposes b e tte r th a n either of th e other two sources.
Looking to th e future, th e m ost im p o rtan t excitation sources, other th a n the direct current arc or th e alternating cu rren t spark, are the highly flexible, all-purpose units such as the M ultisource.
While “ M ultisource” is th e trad e nam e of a particu lar excitation unit, it represents a ty p e which m ay become popular in th e future.
T he M ultisource is highly flexible and is equipped w ith all th e m ost effective devices for regulation and control of th e discharge.
I t is capable of producing discharges equivalent in m ost respects to the direct current arc, th e condensed spark, and the alternating current arc, as well as a num ber of interm ediate discharges not available on other units. This u n it can be used in place of other excitation sources in m any applications, an d th e M ultisource or equivalent ap p aratu s m ay eventually replace th e source units now more or less stan d ard equipm ent in m any spectrographic laboratories.
L a b o ra to ry A lu m in u m C o m p an y of A m e ric a
R e p o rt
A lu m in u m R e se a rc h L a b o ra to rie s No. 0 0 0 0 0 A n aly tical D ivision
N ew K e n sin g to n , P a. June 1 , 19 4 4
Sample O f Miscellaneous M aterials
Received From J o h n D o e
Date Received J u n e 1 , 1 9 4 4 Marked
Q u a lita tiv e S o e c tro g ra p h ic A n a ly sis
Clay
Ash of Decorator* s
Taoe
Ash o f Plum Pudding
( Cnnnp.d ) Major C onstituents: Silioon
Aluminum Zinc
Calcium Sodium
Potassium
1 to 100: Iron
Calcium Titanium
S ilico n Aluminum Phosphorus Sodium Boron
Calcium Aluminum Phosphorus Magnesium S ilio o n
0.1 to 1.00: Sodium
Potassium Magnesium
IronMagnesium Copper Barium Manganese
IronBoron Strontium
.01 to 0.10: Strontium
Barium Vanadium Manganese Zirconium
Fluorine TinAntimony Titanium LeadChromium Strontium Nickel
Manganese Copper TinLithium
Less than .010: Lithium Chromium Nickel Gallium Molybdenum Germanium Copper Zinc
Lithium Molybdenum S ilv e r Gallium Cadmium
LeadBarium Gallium ZincChromium Nickel S ilv e r Lead
Original To
John Doe S ilv e r
Copies To
A ™ h ™ l By A.B.C. D . t , June 1 , 1 9 4 4 Approved By Chief Chemist
Figure 3. Typical Report of Qualitative Spectrographic A n a ly se s
A N A L Y T I C A L E D I T I O N 71
SPECIFIC A P P L IC A T IO N S
T he rem ainder of this discussion is devoted to practical examples illus
trating the usefulness of emission spectroscopy to the analytical chemist.
All this m aterial is based
on work perform ed a t A lum inum R esearch Laboratories.
Ge n e r a l Qu a l i t a t i v e An a l y s i s. All miscellaneous sam ples subm itted for com plete chemical analysis are first tested quali
tatively as follows:
A portion of th e sam ple is prepared for arcing. O n non- metallic sam ples this usually involves ignition to rem ove w ater, grinding to a fineness approxim ating 20 0-m esh, and mixing w ith graphite (2 p arts graphite to 1 p a rt sam ple). In special cases where sm all am ounts of m ercury, arsenic, or oth er volatile ele
m ents are likely to be present, a separate sam ple is prepared w ithout ignition, an d bo th ignited and unignited sam ples are analyzed. In th e case of m etal sam ples, th e preparation usually consists of preparing a sm all am o u n t of fine filings from the sam ple and mixing w ith graphite. In some cases graphite m ay be om itted if the m etal is one which is easily volatilized w ithout spitting. In th is case a sm all fragm ent of th e m etal m ay be used instead of filings. In o th er cases th e m etal m ay be con
verted to salts or oxides by chemical treatm en t, an d th e salts or oxides prepared for arcing as described for nonm etallic samples.
A portion of th e sam ple is placed in th e crater of a n electrode of th e ty p e designated as N o. 1 in Figure 1. T his electrode is used as the lower electrode in th e arc an d is positive. An arc current of 12 to 15 am peres is used on all sam ples which are mixed w ith graphite. On m etal particles analyzed w ithout graphite, th e arc current is usually reduced to 4 or 5 am peres, a t least for th e initial stages of arcing. In all cases, except in specific tests for the more volatile elem ents, th e arcing is continued u n til the sam ple is completely volatilized.
T he spectrum is recorded on a n E astm an T y p e 103-L o r T ype I-L plate. T he te s t is repeated a t tw o or more different settings of th e spectrograph to cover a range of w ave lengths including sensitive lines of all of the elem ents sought. On a G aertner two-lens spectrograph, tw o exposures are required, one extending from 2275 to 4000 A., an d th e o th er from 2800 to 8500 A. If th e sam ple available is sufficient for only one exposure, th e region 2490 to 6000 is used. On oth er spectrographs th e w ave-length ranges will be different and three exposures m ay be necessary to include, a sufficient w ave-length range. An iron arc spectrum and occasionally one or more additional reference spectra are photographed in juxtaposition w ith each spectrum or group of spectra in each region. T he plate is developed an d processed by conventional m ethods.
For interp retatio n , the spectrogram is viewed on a viewing.box or, more often, projected on a screen a t a magnification of a b o u t 10 to 1. T he an aly st identifies th e lines of the various con
stitu e n ts present by the positions of th e lines relative to th e iron spectrum . Q uantities are estim ated by relative blacknesses of line images and th e estim ation is largely based on th e previous experience of the an aly st and com parisons w ith spectra of sam ples of known composition.
Three typical q u alitativ e analyses reported on the form used by Aluminum Research L aboratories are shown in Figure 3. T he percentage groupings are based on the visual estim ates of the analyst and are intended to give the approxim ate percentage of any one elem ent w ith respect to the to tal am ount of all elem ents detected. While th is sam e ty p e of report is used on all general qualitative analyses, th e report m ust frequently contain ad
ditional d ata, such as loss on ignition, to ta l solids, or o th er in
form ation which m ay be required for interpreting the report.
T he interpretation of q ualitative spectrogram s is the m ost difficult job in the spectrographic laboratory. T he quality of the results is directly proportional to the experience and judgm ent of the analyst. H e m ust be thoroughly acquainted w ith the prin
cipal lines of all of th e elem ents and m ust know which lines are interfered w ith by lines of other elem ents, an d to w hat degree.
H e m ust be thoroughly fam iliar w ith th e iron spectrum and m ust have a trem endous accum ulated m em orabilia of d a ta which will enable him to exercise dependable judgm ent in the estim ation
Spark Spectra of Five Typical A lu m in u m A l l o y s
of quantities. H e should have a broad knowledge of the typical compositions of the m yriad different m aterials he m ay be called upon to analyze and m ust be acquainted w ith the effects of dif
ferent m atrices and different techniques of excitation on the spectral response obtained from the elements.
Al l o y Id e n t i f i c a t i o n. O ften a fam iliar alloy can be identi
fied by a simple q ualitative analysis m ade by th e technique in th e foregoing paragraphs. Of course, in m ost cases a complete qualitative exam ination is not required, since m inor im purities are usually of little im portance in determ ining an alloy. How
ever, it is often necessary to m ake an analysis of greater quan
tita tiv e accuracy th a n th e usual q ualitative exam ination in order to distinguish betw een alloys of sim ilar composition. G eneral practice in this laboratory is to m ake a simple q ualitative exam
ination by m eans of arc spectra to determ ine the general ty p e of alloy, and to follow this w ith a spark te st in comparison w ith appropriate sam ples of known composition. Figure 4 shows th e spark spectra of a num ber of alum inum alloys. These spectra were prepared by using a m achined or filed surface of the sam ple as th e upper electrode on the P etrey stan d and a hem ispherically tipped graphite rod as th e lower electrode.
I t has been claimed in several m etallurgical laboratories th a t th e determ ination of alloys by the spectrograph would be am ple justification for the cost of the equipm ent, if there were no other use for the apparatus.
A case in point occurred in a p la n t where a large num ber of alum inum tubes had been c u t very accurately to specific lengths an d given a special polishing treatm en t. In a batch of several thousand tubes it was suspected th a t a few of the tubes were of th e wrong alloy. Since th e tubes had already been cut and care
fully finished, th e rem oval of analytical sam ples would have m ean t the rejection of all tubes sam pled. T he problem was solved by taking all th e tubes to th e laboratory and identifying th e alloys by spectra prepared by playing a weak spark dis
charge on th e surfaces. T he tubes of th e incorrect alloy were easily sorted out and th e slight blem ish left by th e spark was re
m oved by polishing th e rem aining tubes.
Very frequently an alloy identification is required on a sample which is too sm all for positive identification by other means.
This frequently occurs in th e case of alum inum alloys, where the alloys are often very complex. T he chemical identification of an alum inum alloy m ay require tests for eleven different con
stitu en ts and even after testing for the eleven common additions, there is the risk of missing some unusual constituent, especially on very sm all samples.
Co m p a r i s o n s b e t w e e n Ma t e r i a l s Sh o w i n g Di f f e r e n t Pr o p e r t i e s. A typical example of a siinple analytical com pari
son, which had far-reaching consequences, occurred in Aluminum R esearch L aboratories several years ago.
One of th e w orkers in th e m etallurgical lab o rato ry was in
stru cted to prepare several batches of an alum inum alloy and was given the compositions to be used in synthesis. T he series was to contain various am ounts of copper and manganese, and the usual im purities present in th e range of 25S alloy. In subse
q u en t tests, one of the alloys produced was found to have ex
ceptionally high properties for this alloy and was inconsistent w ith th e other mem bers of th e series. A chemicSl analysis had been m ade for th e elem ents added and all th e im purities to be expected from th e raw m aterials used. No explanation for the unusual properties was obtained and th e m a tte r was largely for
gotten until a year or so later when a spectrograph was installed in the laboratories. One of th e m etallurgists still had a speci
m en of th e m etal showing th e unusual properties an d subm itted Figure 4.
