]\gt ; [ 1041 XXI . .
th \ldquo ; cr W. N. HARTLEY , F.R.S. , ssor of ( of Scicncc ) , Rccciyed Jnnc 11 , ) 1 .
ical Notes . . . . . . . .
1011 Description of the instrnment the method cf spectra of Bessemcr 3 .
tions nlad at Crcwc at in ) ) . . . . . . . . . . . .
) 4 .
Spectra of the Bcssemel flamc with }nlCj ) } 5 .
Lincs of the alkali of observod lOli7 6 .
Lines and bands of and of cal'bon monoxidc .
7 .
Bands and lines of mangancse the Bes , scnlc ] .
8 .
Lines the spectrum of the ] in ) and with irou lines .
9 .
The con titution of the spcctrum .
10 .
Cause of the of lines tho thu blow\ldquo ; .
The temperatul .
C of the Bcsscmcl metal 12 .
cause of the ) ance of ) Il tho " " period " " .
The technical , of this tion . . .
14 .
and conclusions .
\mdash ; \mdash ; Introduction a is well known used in the cture of steel ) the ) ) character ; in bhe first ) it is ] \ldquo ; yellow hue at one phase of its existence , of amethyst coloured .
In uary , If .
L. DCCCXCIV .
' Pruc .
, ' vol. 3 , p. 57 ) stated that he had } ) during the in an interestiliation of the spectrum ) odnced l this flame , observed a complicated but characteristic series of bright lines and dark absorption sodium , lithium , and potassium lines were most conspgny other lines of undetermined .
In a ]ecture delivered titution he described the Besselner flame spectrum more poin ted out the existence of lines which he believed to be due to the , iron , sodium , lithium , potassiu ] ] , hydrogen , and nitrogen .
These serva ions led to the discovery that the exact point of decarburization of the metal could be detel.nlined by ) of the oscope with much reater exactitude than he lllere arance of the flame , and for ) ' the point at which it was to this instrument was in constant use at heffield in 1863 , itnd introduced in to the steel works of the LondoIl and North-Western Railway .
F. , in a lecture Polytech .
deliyered in recent in the Bessemer process , stated endeayours to lllake spectrum analysis applicable to the Swedish process not led to any useful esult .
, in Poly b , stated up to the close of the previotls year the observations of the flame and sparks issuinlgfi.om the converter and test pieces of the metal , gay better indications spectrum analysis .
In 1867 , made observations on the spectruhl of the Bessemel flame at the works of the Southern at Sitzungsberichte der .
Akademie der Wissenschafteo , ' Vienna , vol. 56 , Part II .
, June ) , hich led to the practical application in Austria of the use of the spectroscope to the control of the Bessemer process .
The ) , as described by LIELEGG , consists of pours of sodium , potassium , and that of the flame of carbonic oxide .
AcconI ) the latter nitrogen , ectrum of this gas appears , nor could any ) of it be ined by compounds which did not contain \ldquo ; ) atnlos)heric air .
It was also shown that though carbon by carbons and , with air or oxygen , yet bonic oxide yields no such spectrtlu .
Carbonic oxide was found to yield only a continuous ) when burnt with air or oxygen , or even with nitrous oxide .
No dark or ) lines were ] .
It represented that the bright lines in the Bessemer flame must result from the much higher telYlpelature which is produced by ustion of the heated at the mouth of the converter , than when laboratory are performed with the same gas .
it occurs in tlJe " " boil up to the end of the " " \ldquo ; period , ibed as follows .
The figures are arbi t scale measurenlents : It noticed that lines of the alkali metals had become visible , an immense number of lines were seen , some fine bright lines , others as intensely dark bands .
Striking evidence was afforded of the cessation of the removal of carbon from the iron ) the disappearance of nearly all the dark lines and most of the bright ones .
Spectrum was to be remarkable from the total absence of lines in the more ible portion ; it extended scarcely beyond the solar line .
It was indispensable that the spectrum should be actually corIlpared with the spectra of the elements sought for .
The of the Bessemer flame was accordingly with the .
spectra:\mdash ; ( 1 .
) Spectrum of electric arges in a carbonic oxide vacuum .
( 2 .
) Spectrum of strong spark b.etween silver poles in air .
( 3 .
) Spectrum of strong spark between iron poles in air .
( 4 .
) The in hydrogelJ .
( 5 .
) Solar spectrum .
( G. ) rbon spectrum , oxy-hydrogen blow-pipe supplied with olefiant gas and The coincidences observed were very few and totally failed to explain the nature of the Bessemer spectrum .
The spectra of neither carbon , nor of carbonic oxide , appeared in the Bessemer flame , either as lines or as absorption bands .
Three lines were traced to iron and a dark absorption band in the red , due to hydrogen , was visible more particularly in wet weather .
J. M. SILLIMAN Phil. pointed out that the progress of the decarbm.izabion in the Bessemer process is determined chiefly by the appearance of the smoke , flame , and sparks emibted from the converter , and that owing rapidity with which the changes takes place it is highly important to catch the exact Inoluent when the blast should be turned off .
The colour and brightness of the stream of gas from the converter when observed by an experienced eye , generally give a sufficient indication , but when pioo iron of a manganiferous character is used , this determination is very difficult ; even those who had had much experien ce made fi.equent mistakes , and found it impossib]e to produce the same quality of metal at every operation .
Mr. ROWAN , of the Atlas Works , Sheffield , made use of coloured glasses , two of ultramarine blue and one of dark yellow , by which the eye was greatly assisted and the termination of the process rendered unmistakeable .
Though LILLLGG first recognised the fact that the spectrum of the Bessemer flame was not the spectrum of carbon , nor entirely that of carbonic oxide , which gives a continuous spectrum , it was BRUNNER in 1868 who first expressed the view that the *It docs not appear tnined the flame of carbon monoxide burnt with air or HTGH spectroscope cannot be considered as a practical indicator of the decarburization in the Bessemer process , because the lines of the Bessemer spectrum do not belong to carbonlc oxide or to carbon , but to manganese and other elements in the ] ) )-iron Oesterreichische Zeitschrift fiir wesen , .
Next , VON LICHTENFELS Ein sir des Spectrums dcr Bessemerflamme Polyted ] Journal , ' vol. , pp. 21.3 , 215 ) remarked that nature of the several green and blue groups of lines in the Bessemer spectrum was not known , : they had been ttributed to arious substances , with certainty as to their identity .
SIMMLER had described the spectrum of nganese as consisting of broad green bands and a yiolet line ] .
to the yiolet potassium line Zeitschrift fiir ytische ( ; LTCHTENFEL , q , examining the spectrum of olved ilcohol , found the green bands to be composed of groups of ] ines , the constituent } of corresponded exactly with the constituent lines in the of the Bessemer spectrum .
Hc concluded the groups of green line , seen in Bessemer spectrum to anese .
J. SPEAR PARKER made a number of servati o the of .
CHARLES CAMMELL and Co. , of Sheffield Chenical News , with coloured glass , and with the spectroscope .
He was unable to confirm statement tlJat the Besscmer spectrum could 1oe seen when .
was merely being heated .
He thought the spectruln could not be attributed bo es as it had been , of opinion that the most characteristic portion of it would be to be to the presence of carbon in some KUPELWIESER , a lecture the Leobenl the application of the to the Besseme ] .
process , quotes observaand assigns reasons to be correct he ibuted the Bessemer spectrum to carbonic oxide .
The lines belonging to carbonic oxide make their appearance the of ) ) convel.ter has become greatly elevated and the carbon of the to burn ; the second and third periods until coml ) let dtion 1 taken are brightest when the temperature is hest , they vanish some w quickly with the combined carbon , while ) .
proporbion of lnolte pig-iron is added to the be visible , though not collyertcr coke .
Also when slag and ' are off ) the of he had observed the monoxide ] ) ectl ( } cc he obtained ith the ) ic hische Z BRUNNER pointed out that ) of ) onic ( is not line ctrum but a continuous band of rays , though Bessemer spectrum is caused by the higher temperature of the W. N. HART L ON convcrters the green group when heated with coke , one with a new lever does .
The lines of nese he believed to be the constituents of oups and of LIELEGG'S so-called carbonic oxide spectrum and also the violet line of anese , the line , attributed to carbonic oxide .
flame from the ) ] of a blast furnace could play the same spectrum as that of the ) just as well if the be composed of lines of if they were due to carbonic oxide .
) request of the fulne which ises from the neck of converter .
the " " and found ib to be a manganese and ferrous silicate .
blnO This an indication of , and iron beinrro concerned in the formation of isch c Zeitsd .
tion , refully and laboriously carlied out by Dr. lTARSHALL WATTS , to the conclusion thatthe lines visibJe in the Bessemer-flame spectrum are mainly due to lnang.anicox id , not to metallic manganese , as had been stated , nor to .
When chloride , carbonate , or oxide , such as the nJineral yrolusite , is heated in the oxyhydrogen flame , a very brilliant banded spectrum is obtained which is for the nlost part coincident with the Bessemer spectrum .
Observations were further made on the spectrum of the flalne obtained on adding -eisen to Bessemel metal , on the temperature of the flame at different in the process , and on different spectra obtained by the employment of different kinds of iron .
Accurate determinationls were made for the first time of the wave-lengths of lines observed in the spectra of the Bessemer of spiegel-eisen , and of manganese dioxide .
The fact that six lines of iron were present in the Bessemer spectrum was establislled , and considered to be a proof that iron may exist as a temperature below its melting-poinb , since certain experiments led to the conclusion that the not hot to melt -iron .
This work of WATTS is the most xhaustive investigation of the subject up to the P. In GREINLR served in the flalne from rons p , -iron the spectrum of anese as ured by During the meeting of the itish Association , Sheffield , in 1879 , I made a short examination of the Bessemer flame with a small direct vision spectroscope , at the works of Messrs. BROWN , ) DIXON ; I also examined the flame of * Revuc Univel.sclle , ' vol. 35 , p. , 1874 .
AT spiegel-eisen .
Tbis examination convinced lllc that the of ) is no visible , and that the ) ands seen in ) ) feature which uishes them in striking nner from ] the band , of ) amely , they are degraded the , the carbon the blue .
This had already been oticed by of the ) the ) potogr of In 1882 I deyised an instrument lneeting the reqnirements of a serics of observations to be made at steel works , particularly studyi n the ) of flames , and the heated .
It was desirable it ilould give a fair amount of dispersion at the less ible end of the s , be rigid .
A train of four at on a table and stand made entirely of wood , to which fitted , with movement to the frame the dark slide , so several could bc photographed on one plate .
rious tvith showed that to the stand too light , the instrument w unsteady .
In the was mounted on a heavy tripod .
The table to the of stand by a winged screwjoint and counter ) oised , so that it could be placed in ahnost any required position .
The of metal , with an ) iece behind the fraIne for the dark slide , so as to makc- it available for observations with eye , for which it is peculiarly well adapted , to the observer the flame behind him , and therefore he is not embarrassed by the .
In the circular box ) end of the dark slide can be fixed it is of a toothed wheel .
The collimator fitted scalc of .
on the draw-tubes , so that both th slide and ) bc out so as to be equi-distant the lellses the pnrpose of correctly .
The camera can be clamped , its exact ) ition d by of a divided rc on the table .
A with ) scalc , flected f the face of that isnl nearest to the leDb in found useful .
The isms move automatically with the , in to secure the minimum angle of for the ) .
A work in front of the slit , and fixed to prism ) , carries a ) three ches focus .
condensel with crossed angles has been in , a device which was described iu letter tu the JIerr vlcTOIt b. It ] advantage of lines , but it was not employed Crewe , .
to one , the adjustment arises for the ) ) moisture , even th of thu " " blo With .
of the upon the slit , solnetimes the was in verticai line the slit , and sometimes placed diagonally .
This depended very lIlch upon bhe position of the converter and the consequent size of the image .
The slit plate was covered with a plate of lart z to exclude dust and metal plate with a piece cut out of one end slides over this , and serves to shorten or lengthen the slit , and secure a greater or ] number of spectra on one .
In some cases , photograph was taken every half-minute , from the commencement to the termination of the " " blow This could be accomplished only by the use of snch a , as the plates measured no more than inches .
focus the ument various photographs of sun-spectra were taken and the ositions of the djustable parts were recorded .
In order to render the instrument portable case was constructed for carrying the prism table , prisms , collimator , camera , without rranging the adjustment of the instrument .
ioged screw when loosened enabled the whole to be detached from the vertical pillal of the tripod , -be counterpoise , of course , having been detached previously .
The prism table was then fastened in its case .
A second case carried the tripod , counterpoise , , and developing dishes .
This case when empty had a squal.e frame of vood which fitted into the top when the lid was open .
The frame was covered cloth , lined with yellow calico .
In the cloth were arm-holes and sleeves fitted with elastic , which came ] way up to the elbows .
The developing dishes itnd measured quantities of solution were placed ready in the box and the dark slide could be ) ened , the plate removed and aced in the developer , while during meut a could be put into the slide .
Deyelopment was always to proceed a given period which experiment had proved to be sufficient .
During the progress of development the dish was covered with an ebonite to exclude an .
possibility of light the photographic plate .
By such means the plates are hxed without the use of a dark room , but it is , of course , essential that the hands are not withdrawll from the box before the developis concluded .
At Crewe my ssistant h the use of a laboratory , but at Dowlais the operating box always used .
ngements vere m de for ying out the first sel.ies of from a point on the floor of the cupola-house Crewe , situated close to the , from which the ladles of spiegel-eisen are tipped into the converters .
The instrument was placed a very solid bench , which could , be moyed about as required .
From observations with a direct vision instrulnent it decided to take of the sun-spectrum at the laboratory , to sensitise some plates with cyanine , develop the aphs , and , having obtained a good focus , to remove .
by hand to the cupola-house .
It was ) sidered best to raph , the ) iued d the wholc period of the " " boil , and , secondly , the blaze , after the addition of spiegel-eisen .
Unfol.tunately the sun was 1lot always LAME HTGH visible , but on one occasion the of of the flamc the four periods in one " " blow were on one plate .
With ) it easy to see that a number of lines were lyiI in the spectrum .
Upon rged prints , some of which ) made with gelatine-silver brolnide 1 , it easy to ( the ition ( lines and of bands yith respect to the line , these a micrometer screw and microscope in the manner already described ( ' ' 1894 ) .
It became necessary , however , ) tain e ts of , accordingly , the Autotype were to ) , in which the ten easily examined and compared with first map of ) ) solar ) nd MACLEAN'S ) of } another than for lines , , with metre , the ) ands in the a ) ily ulicrometer .
bion curves were reduced to cics .
) ) the spectra less If }ently f)( of the ) The fiducial lines selected were 110 lines in ] of spectrum , between P. It was found to be impossible to sure the spectra and obtain measurelnents giving identical .
This will ' on referring to Plate 6 , Crewe , the of the ) third spectra .
Although is son dilficulty in cmcn t of so precise as we are accustomed to in ) cctra , ( : does ) account discrepancy .
It is , in point of , due to the in , or , in , cases , becoming less distinct at the spectra by the uous r being is ) , ib ] remarked , in the broad lines Phil. Trans Part .
Obscrvcttion at ) at I am much indebted to Mr. F. W. , of oconlot i of London and North-Western Dowlais Iron in ' at Crewe alJd at Dowlais , in in il , to 1 I I I the Bessemer , the instrulnel]t ) ) near to the nlollt of the in ' projected the vess vlle tl my made the , the ) tl first seven minutes is .
tlashes out occasionally .
The temperature is evidently low , not above that of iron .
of graphite and of iron are ejected , but these come from the burning of } ) caused by the eruption of air from the molten metal within the converter .
After seven minutes the spectrum of sodium is constant , at ten minutes the spectrum of becomes visible .
Neithel the hydrogen lines nor those of potassium were as a rule visible .
Two violet lines of potassium appear in all the aphs .
During the " " boil\ldquo ; the sparks few and small , because the metal being much hotter and more liquid there is less spray .
The flame is large and it burlls with a steady roar .
At Crewe , it generally possessed the greenish-yellow colour of the vapour of ) , and maintained this colour until about three secollds before the nination of the process .
Sometimes , however , the mouth of the conyerter showed , during the progress of the " " boil the amethyst-coloured flame for a foot or two , extending from its edge , and there was a further tinge of the same tint beyond the yellowish-green , and quite at the outside of the tip of the bright .
The duration of the blow\ldquo ; varied from fifteen to -one minutes .
The used haematite -iron , and the steel made was intended for boiler plates also for small tyres and wheels .
In certain cases particulars are given , but this is ot a matter of consequence in considering the spectra observed .
In all , twelve plates were osed , some of which were so entirely satisfactory as to reIJder any further vork at Crewe unnecessary , there being ample material for some months of study .
Spectra of the Bessemer at Crewe in , 1893 , with detailed cconnt in of the period of the " " the used , the duration of each The first exposure lasted three minutes .
( EDWARDS ' isochromatic instantaneous plate .
) A second exposure made of another part of the plate , and a different " " blow " " \ldquo ; stal.ted at . . . . . . . . . . . .
1.34 of line . . . . .
1.41 red and green .
1.45 xposurc commenced . . . .
1.46 Continued till eIld of " " blow " " . . . . .
, , of . . . . . . . .
, minutes .
A third exposure .
" " commenced at . . . . . . . . . . .
3 .
rance of sodium lline 3.45 red bands 3.48 , . . . . .
3.51 of . . . . .
, , ation of exposu ] . . . . . . . .
minutes .
LAME SPECTRA AT HIGiI The aphs were valuable as showing the extent of the occurrence of a large nuulber of lines in the hich 1 ) observed in the Bessemer fl me before .
The plate , been PLATE EDWARDS ' atic instantaneous ) , stained with cyanincl .
Commencement of " " blow\ldquo ; 1st exposure comlncnced '2nd \ldquo ; 4th of " " ow\ldquo ; . . . . .
ation of These were ot well lithium line ] , to it were just barely visible .
The ) osition of the of the on the of the spectroscope was PLATE S. " " Blow\ldquo ; EDWARDS ' isochromatic ined with position of the image of flame ou the slit Commencement of " " blow\ldquo ; 1st exposure at 2 nd \ldquo ; \ldquo ; tion of .
' Blow\ldquo ; 48 .
Boiler-plate .
isochronlatic i , stained ) the instrument tered f 111 .
tion of of " " \ldquo ; 1 of the sodiun line : 1 1st expohurc . . . . . . . . . . .
, . . .
11h of .
, . . . .
CTLt and green bands inni of the second ) This plate had each spectrum finely ) ( -eisen was .
shown by the in ) violet .
The importance of tbesc violet is uted .
PLATE 8 .
" " Blow\ldquo ; 69 .
] ils .
The position of the of lit isochrolnatic plat , not stained .
A sun phed exposure .
) of of \ldquo ; 1st osure . . . . . . . . . . . .
1 2 nd . .
: of ' ' blow " " . . .
This is a very useful series of ) being conyenieut for ison .
sccn , lines extending into the -violet .
plate ] convenience in identifying the lines .
PLATE " " Blow " " ) The image of the thu ) instantaneous used , without Comnlencoment of " " blow " " 1st sodiu1n liuc hCCl : 2nd .\mdash ; .\mdash ; .\mdash ; .
4th . . .
.\mdash ; .
of bJow \ldquo ; : ; This series was well focussed PROFESSO W. N. ON PLATE 10 .
" " " " 89 .
Vessel No. 2 .
osition of flame the slit , diagonal .
An Ilford stantaneous plate , not stained , nsed .
of expof , Ul .
C. Comm cement of " " . . . . .
sodiun 1 line . . .
12 1st exposul .
C , both potassiunl { thd ithium l visible . . . . . . . . . .
20 { 3 The " " blow " " . . . . .
23 The conver lowered in to in to cool metal , which had become too " " I3 \ldquo ; started at exposure . . . . .
27 .
, 4th commeuced at 11 SO of " " blow\ldquo ; . . . . .
20 secs .
duration of the ) osure was only 20 seconds .
This is an excellent series 0 PLATI 11 .
" " Blow " " 90 .
Vessel No. 3 .
flame fell diagonally on slit .
The plate used was one of ' instantaneous make , ained with cyanirle .
WilS adjusted so to bring the red end of the spectrum in to ) ] ) focus . .
becb .
1)uratiou of exposul'e .
omlneuccment of ' ' " " 30 was a ubual .
at . . . . .
12 40 line . . . . . . . . .
12 1st expof , . . . . .
12 4 minutcs . . . . . .
12 spectrum of other " " " " was phed on plate .
Blow ' ' 91 .
No. 1 .
The position of the flame before .
ation of exposulc .
Commenccment of the blow oxposnrc at . . . . . .
1 till . . . . .
1 ) ) 3 minutes exposure . . . . . . . . . .
1 Conti ) till . . . . . . . . .
1 44 , , " " was ) the at .
, . . . . . . .
1 :30 Aud up . . . . .
, . . .
1 .
, 4 SPLCTRA AT RATU R ] .
" " \ldquo ; .
Vessel No. ails .
Commencement of ' ' \ldquo ; Appearance of sodinm flame .
Appearance of recl greenl ] ) nds 1st exposure at End of " " . . . . . . . .
tion of , 1 minutos .
12 17 Photora ) of \ldquo ; , 189:3 .
The ntity of metal blown at vlais was than the conuverfer were capable of taking twenty tons of at each and the qn blowll was twelve tons .
The ) iron , slnelted from } ore , contained boub per cent. of manganese , and 2 to cent. silicon , with to.3 ' cent. of carbon .
The blowing was generally very to this circumstance , and heavy charges of metal , there was an immense quantity of fnme , which carried by an easterly wind directly towards the instrument .
year completely enveloped in dust from the fume .
The len ses with dnst and with moisture caused by the condensation of steam , ence the flame the slit were obscured .
Observations were ca ried on with incon venience , would not have been the case the wind been blowing another ection .
PLATE 1 .
The plates used specially prepared by rnited , London .
They were stained with nine .
1st exposllrc P LATE 2 .
Blow commcnced Bands out .\mdash ; .\mdash ; .
.\mdash ; .\mdash ; .
\mdash ; \mdash ; 1st to Blowing interrupted ) e-coulnlenced th .
C very brilliant .
( minuleh ) to I nlinn I Till PBOFFSSOR W. N. HARTLEY ON PLATE 3 .
Blow\ldquo ; commenced . . . . . . . . . . .
12 30 Bands flashing out at The mebal too hot and the converter was turn down .
An ingot mould was charged into it at . . .
12 45 Blowing -commenced at . . . . . . . . .
12 47 1st exposure ( miuutes ) . . . . . . . . . . .
12 49 To . . . . . . . . . . . . . . .
12 2nd exposul.e ( 2 minutes ) . . . . . . . . .
12 To . . . . . . . . . . . . . . .
12 Blow\ldquo ; commcnced . . . . . . . . . .
1 20 , Sodium line . . . . . . . .
1 22 line , , , , . . . . . . . .
1 The continuous spcclrnm extremely tlIe lines manganese bands seen upon the continuous spectrum intcnscly brilliant .
The metal vas { oo hot , and the converter was turned at , . . .
1 26 . . . . . . .
1 30 lsb ( minute ) . . . . . . . . .
1 31 2nd , , , , , , . . . . .
1 32 3rd , , , , , 1 4th , , , , , , 1 34 End of " " ' " " . . . . . . . . . .
1 3 PLATE 5 .
" " Blow " " . . . .
2 3 line flashed out at 2 6 anesc bands ffished out . . . . .
2 7 30 lsb exposure ( 30 seconds ) at . . . . .
2 9 30 ' ' Blow\ldquo ; interrupted , converter turned down 2 10 Blowing -commenced . . . . . .
2 17 15 2nd ( 30 scconds ) at . . . . . .
2 18 Till . . . . . . . . . . .
2 18 30 minutcs ) at 2 20 . . . . . . . . .
2 2.3 PLATE 6 .
eceived four exposures minute to minute each , but there was nothing of ' any upon the .
In all probability this was owing to the large body fumes nled wheu the converter was turned down before the plate had been exposed .
was too hot , may be readily cood from fact hcrt bhe nese lashi only minutes after the commencenlent of the blow .
SPECTRA AT HIGH PERAT U The s , am remark applies to this plate .
The The eceding work had been rried on in the opell air , but it considered better to move to the old Bessemer and be under cover , and from the douds of fnme which were driven towards one by the wind , which ] still continued blow from the east .
Hot blast , grey pig iron , containin not per cent. of silicon and from to per cent. of carbon , ) .
the ) roduction of .
Charge of metal 10 tons .
The position of the meant was abont 4 above the of the converter and 6 feet from it .
PLATE 8 .
Blowing commenced at 10 A. , but the blasb stopped because of blowing going on at other } ) .
The blast turned on at 12 o'clock , the mang nese bands shed out at 12 , and np to this the nothing but that of rbonic oxide with the usual alkali lnetals .
1 ] .
111 .
1st . . . . .
1 , , linnfe lseconds ) 12 8 , , 12 4th , , innte ) 12 10 , , 1 1 10 5th , , nlinutes ) 10 , , 6th , , mill u 12 1 , , At the highest temperatures the flame was perfectly transparent .
A great of fume , condensed to coarse dust , was blown bout , much of this fell into the water used in washin the ates , and on films .
could not be avoi , for it was necessary to develop the the and wasb them as soon as developed .
PLATE 9 .
The same metal as before .
Blowing commenced at 12 .
11 ! .
1st exposul'e from , , . . .
1 . .
1 11 , , 4th , , , 1 1 .
, , , 12 6th 1 , 7th sth , , , 1 1 9th 10th ' ' Blow\ldquo ; ended .
This plate was badly fogged .
PLATE 10 .
onlmenced at 12 " " Blow\ldquo ; in terrupted at -commenced at 12 hrfi .
R. exposure } 12 58 30 to 12 59 30 , , , : . . .
12 30 , , 1 , , , , . . .
1 30 , , 1 1 .30 4th , , , , . . .
1 1 .30 , , 1 2 30 , , , . . . .
1 2 .30 , , 1 3 30 6th , , , , . . .
1 3 .30 , .
1 4 .30 , , , , . . .
1 4 :30 .
, 1 .5 SO 8th , , , , . . .
1 5 .30 , , 1 6 .30 , , , , . . .
1 6 .:10 , : 1 7 .30 10th , , , , 1 7 30 , , 1 8 .30 of blow SPECTRA OF THE BESSEMER FLAME .
of the Spectrum of the Bessemer Flrome , as at , ls93 , Steel Works of the London North-Western Railway The lines of the alkali metals.appear , and are , in fact , the only prominent lines during the first period of the " " blow when the silicon is being oxidised and slags are in course of formation .
During the second period , or the " " boi the flame exhibits a continuous spectrum of bright rays , overlying which is a number of bright bands ; some of these appear to be degraded towards the least refrangible rays , others do not to be degraded , but , bounded by lines , or by very narrow bands , possibly by the sharp edges of bands .
There does not appear to be any distinct fading away of bands in the direction of the least refrangible rays more than towards the blue .
At the commencement of the " " boil " " the bands are not so numerous as afterwards , nor the lines so strong and numerous as subsequently , excepting the lines of the alkali metals .
It is noticeable that certain lines appear only at the commencement of the " " boil others do not appear until the end .
There is no very great difference between the spectra taken at the close of the " " boil and at the commencement of the third period , or " " fining stage But both in the second and third periods there are several lines which are short , and seen only at the base of the flame ; others are broad at the base of the flame , like bands , but towards the tip they attenuate into lines .
By far the greater number of the lines in the spectra are iron lines ; especially rich in these lines is the ultra-violet region .
The prominent bands , for the most part occupying the of less refrangibility than 3000 , dne to metallic manganese .
They do quite .
to the bands of the of pure componnds of lllangill]cse , is rtly owing to the COlJtinuous spectrum of carbon monoxide , to rbon bands , as are exhibited in hydrocarbon flames , and also in part to bands of iron .
This will be readily understood from an examina , tion of the photothc -lengths of the edges of carbon bands , for it is evident that as the of carbon overlie those of manganese , the latter are most certainly obscured by the former , and the former by the latter ; because , while former consist of degraded towards the violet , the latter are composed of similar narrow bands towards the red .
One overlying the other can produce the appearance of a broad band , which is not degraded , but appears as a uous spectrum , in this the only part distinguishable is the more refrangible edge of the manganese band , and the less refrangible edge of that of carbon .
Between the two , neitl ) .
bands nor lines can be distinguished if the are strong .
There are some few lines due to carbon monoxide , and certain due to an oxide of manganese , either .
It is not likely that , which is easily decomposed by heat , could exist in the Bessemer flame ; vaporized oxide must be a substance of no small stability .
Refel.ences to the lines , measured by WATTS , have been inserted opposite to the line of neal.est wave-length in the Bessemer spectra .
Lines not identified by WATTS are either lines of iron , of anese , or carbon ] bands .
The lines of metallic manganese were not identified , probably because WATTS used lnanganese dioxide .
or carbonate , for his comparison spectrum , instead of the metal heated in the ydrogen flame .
Lines in the spectrum of the flame , which are more refrangible than the solar line , have not been nined before , and this portion of the spectrum is ecially interesting .
It extends to some small distance beyond the solar line of , or on some plates as far as .
Most of the lines ycry strong very characteristic groups .
They have nearly all been identified with iron ; they are all strong lines , as seen in the arc spectrum of iron , coincident with lines in the sun .
They have been identified with the lines ) hotographed f the spectrum of TURTON'S tool-steel , but steel in the oxyhydrogen blow-pipe yields no greater number of iron lines than occur in the Bessemer .
Ferric oxide , under the same conditions , exhibits a spectrum with a , and still fewer lines .
It certainly appears as if the temperature of the Bessemer metal during the " " boil\ldquo ; is as high as , or approaches that of , the oxyhydrogen blow-pipe flame , when the oxygen contains 10 per cent. of nitrogen , as the commercial oxygen supplied to us was found to do , and the flame is being used spectroscopic purposes in the manner ] ready described ( ' Phil. rans 1894 .
) .
Co the complicated nature of the spectra , the wave-lengths of bands and lines during different periods of the " " blow\ldquo ; have been tabulated , with a of feature of the spectrum to which the measurements belong , and in parallel lInns there are references to WATTS ' measurements , and hose of other FLAME SPECTRA AT URES .
The Constitution of the Bessemer It will be readily understood from the previous investigations of the flame-spectra of iron , manganese , spiegel-eisen , ferro-manganese , silico-spiegel , , pure man- ganic oxide , cal.bon , carbonic oxide , , that the Bessemer-flame spectrum is not characterised especially by the bands of carbon , would be the case to the views of ROSCOE , nor of carbonic oxide according and KUPEL , nor does it belong entirely to nese as indicated by the BRUNNER , V0N NFELS , and WEDDTNG ; furtherlnore , it cannot be ttributed chiefly to manganic oxide , as stated by .
It is , in fact , a complex spectrun ] , in which the bands of metallic manganese , rbon , carbonic oxide , and possibly also of manganic oxide , are superposed ; and the lines of iron and occur with those of other elements , such as hydrogen , lithiuln , potassinm , and sodium .
The hydrogen line ( in the spectrum ) was } ) hotographed only once , and thel ] during a snowstorm , when it appeared completely reversed .
No absorption were at any visible when tions were made upon ilame only .
No nitrogen bands were seen .
No bands belonging to calcium thd magnesiunn oxides were visible , nor lines of these elements .
No cobalt , nickel , copper , nor chromium were detected .
The lines beyond the line , which had hitherto not been examined , are nearly f$ll lines of iron , as mapped by CORNU and ROWLAND in the solar spectrum , and observed in the -spectrum of iron raphed by KAySER and RUNG .
Cause of the ppearan cof Lines the Comrncncement of the " " SILLIMAN detected thirty-th.ec lines in the Bessemer-spectrum ; some of lines were not observed , and others which he did not record were found .
bands were observed , crossed by bright lines ; and it is suggested that the ] lines tend to make a weak continuous spectrum appeal discontinuous , the dark ) eing merely intervals between the bright ones .
The iron spectrum had not been torily identified .
According to tement , " " the contains yet many mysteries to be solved , among which is the cause of the nonappearance of the lines of the spectrum at the beginning and tion of the blow .
' \ldquo ; WEDDING accounted for the absence of the spectrum the and lination of the ' blow\ldquo ; by the absolute quantity of the substance tilized being these times too small to produce a spectrum .
' ' Zeitschrift fur das Berg- Hiitten- und esen , ' ] .
, 1 ) 117 , 1869 ) Ile based his view upon the , recognized by SIMMLER , that a much ' manganese is required to obtain a reaction in the than thal ) can be detected by the well-known blowpipe test with rbona of he Bnnsen of a milligramme of manganese can be detected MMLER ) ; but the quantity of alkalies is much for instance , of a milligramme of potassium , of lithium , and of sodium ( KIRCHHOF ) .
The flame spectrum of anese is almost entirely a banded spectrum , the peculiarities of which had not been investigated at that time .
STLLIMAN had urged against this view of WEDDING:\mdash ; lst .
If the disappearance of the manganese lines in the Bessemer spectrum be owing to the diminution of the quantity of manganese , we should infer that these lines would gradually grow more indistinct , and then fade away ; but the fact is the contrary\mdash ; the manganese spectrum incl e in brilliancy from its first appearance , and is more intense just before being swept away than at any other time .
The analysis of bhe which appears when the flame ceases , proves that a considerable quantity of manganese is still volatilized ; and it is notable that in manganiferous iron this quantity increases towards the close of the " " blow 2nd .
Tt would be more difficulr to account by this theory for the non-appearance of the sodium line at the beginning of the " " blow as sodium then , in all probability , exists in the issuing gas in sufficient ntity to produce its spectrum at a high temper ture , as it is only by special precautions that we can keep it out of any flame .
3rd .
A still greater difficulty would arise in applying this theory to the spectra of sodium and lithium at the close of the " " blow As has been stated , these lines sometimes disappear at the moment of complete decarburization , and sometimes remain .
In the former case to say that the sodium had been exhausted would not be in accordance with what we know of that element .
WEDDING based his explanation of the non-appealance of the manganese lines upon the analysis made by BRUNNER .
It was found that the manganese contamed in the iron fell from per cent. in pig to , and finally to cent. in the decarburied and that the manganous oxide in the slag first increased from to 3790 per cent. , and then sank to per cent. ; and , furthermore , that certain amount of manganese is to be found in the fume .
SILLIMAN states that since the manganese contained in the pig iron decreases continuously , and that contained in the slag after the termination of the boiling period also decreases , a considerable volatilization of body is probable , just at the bime when the spectrum is best developed .
WEDDING found from BRUNNER'S analysis that some of the manganese is volatilized from the , and it was further considered that the manganese spectrum during the entire process cannot be due wholly to the volatilization of nuanganese directly from the iron ; for while the amount eliminated from the iron grows continually less , the manganese spectrum grows .
If there were a sufficiently large quantity of carbonic oxide flame to render the escaping gases owing , it is evident they would not issue from the converter as dark smoke , but as incandescent vapcur , having its characteristic spectrum .
The lack of sufficient flame therefore account for the disappearance of manganese spectrum .
criticism of WEDDING'S arguments it may be urged that , LALlh iCTRA A ' H1Gt H because the proportion of manganese decreases in the pig in the slag fter the tion of the " " boil " " also decreases , it does not rily follow that the actual quantity of es in the slag is diminished .
wanting which would enable us to decide how thuch manoaneseo .
is volatilized , since we do not know the absolute quantity of slag and iron .
BRUNNER'S analyses do not appeal to me to prove that the absolute quantity of manganese in the slag diminishes or increases during the " " boil since manganese is not the sole basic constituent of the slag .
This question was considered by MARSHALL WATTS who , in experiments both at Crewe and at , always observed a difference between the ordinary Bessemer ] ) ectrum as seen at Crewe and that of eisen .
The difference , which consisted in a relative intensity of the lines , was so great that it was not at first perceived that the spectra in ctny way the same .
At Barrow this erence in the spectrum was not seen , the spectrum of spiegel-eisen being identical with that of Bessemer metal , only more intense .
The ordinary Bessemer spectrum Barrow identical with the spectrum Crewe .
The difference between the Barrow and the Crewe spectra was attributed by WATTS simply to a difference in tempel.ature .
It was stated , however , it might have been connected with a difference in composition of the metal upon .
Experiments made on the temperature of the flme howed t , at the commencement it was below 1300o C. , but it gradually rose reaching 2000o C. From a study of my photographs it appears certain that the whole phenollleno is primarily due to rise of temperatul.e , which takes place rapidly and continuously during the " " boil while at the same time an increasing quantity of cal.bonic oxide escapes from the converter .
The bath of metal is first heated by the oxidation of the manganese and silicon .
Such oxidation oduces an enormous amount of heat ; first , because the heat of combustion of these elements is very high ; secondly , because the products of combustion are solid , at high teml ) liquid , and carry none of the heat awav .
to been first ecognlsed by Lieutenant DUTTON , U. A. , in 187 Chern .
News the carbon burns and yields a amount of heat to the metal , the blast which passes through it , and so increases the rapidity of ) of the carbon , which serves againl to raise the temperature of the .
The contents of the conyerter carbonic oxide and nitrogen , this the manganese and iron are porized , but not oxldized .
lt may be easily un derstood that the ttlre continues to until termination of tbc " " boil because the tenll ) erature of the of and consequently the erature of the it the ) , , so that the temperature of the combustion of the also of the oxide is higher .
We are , in , dealing combustion nnder those in a Sienlen furnace the } I ESSOR W. N. ON the spectrum remains at its brightest until the end of the " " fining stage I conclude , from the evidence of my photographs , is certainly not the case .
Even the eyc can detect a waning in the brilliancy of the spectrum .
See Bessemer photographs 8 and 10 , t , aken at Dowlais ; of the six spectra on Plate 8 the fifth is by far the strongest .
The fourth and sixth are much the same as anese , but the sixth is much stronger in iron lines , and also has a stronger continuous spectrum .
On Plate 10 there are nine spectra , each of which received an expoSure of half a nlinute , the first and last of these are the weakest .
In confirnlation of this the detailed statement may be quoted which refers to these ' plates .
April , old Bessemer plant.\mdash ; Dowlais , grey pig iron , containing not less per cent. of silicon and to per cent. carbon , was being blown for tinplate iron in quantities of 10 tons .
The ) osition of the insbrument just about 4 feet above the mouth of the converter and about six feet away from it .
The blast was turned on at 12 o'clock un til 12.5 .
; the flame contained nothing but the continuous carbonic oxide spectrum , with the usual alkali metals , but at 12.5 .
the manganese began to flash out .
The first exposure was at .
until M. 8 .
12 8 to 12 9 40 3rd , , 4th , , , , 12 10 10 , , 12 11 10 5th , , , , 12 10 , , 12 14 10 , , , , 1215 , .
1217 The flame ) erfectly transparent at the highest temperatures , and it possible to look right into the converter and see the fluid slag thrown up against the mouth and drop back into the vessel .
This is to be loted , the fourth spectrum had an exposure of exactly a minute , while the sixth had two minutes ; as regards the manganese ban ds the spectra are much alike , but in carbonic oxide and in iron lines the sixth is much the stronger spectrtlnl .
This shows that the quantity of manganese vapour was decreasing , but the quantity of iron yapour was increasing .
Though the filt spectrum received an exposure half as long again as the sixth , the spectrum is much more than twice as strong .
AT HIGH ERAT U PLATE 10 .
Dowlais .
April , 1893 .
same pig iron blown for before .
" " Blow\ldquo ; commenced 12.45 ' ' Blow\ldquo ; terrupted at 12.49 Blowing 12.55 P. 1st 2nd \ldquo ; \ldquo ; \ldquo ; 7th \ldquo ; 8th .
, , , . .
, , , . . . .
30 to 1 .
5 :JO , 1 :30 last are exceedingly teeble , spectrtll .
lt is true that No. 11 received ) of -lnly minute , as against one minute for all others , but this would not account for the feeble spectrunl with the very one of No. 10 , and the much No. ( J. It is evident that the mangan es ) nd the lines becoming more prominent .
As is burnt the erature n fall very rapidly , because the ) combnstibIe left is the iron , its of combustion is comparatively low ; temperature of the metal would therefore not be long maintained , and the blast old very soon , under these conditiol ) cool the metal , so as to solidify it , is the when " " skulls\ldquo ; are The metallic vapour within the convertel , as to ) composed of carbonic oxide in excess , would be oxides fume .
GREINER has described uses of a from ) containing phosphorus and large quantity of nesc .
The following figures show composition of the ) used obtained from it * Revue Universelle , ' vol. } , 1874 .
vol. 17 , 1 ) ' Journ. Chem. ' vol. 1 , The indication of the close of the process differs from that in ordinary work , whereas the oscope usually gives distinct indications during the " " fining stage with manganiferous pig it is much difficult to make use of it to advantage .
The excessively brilliant flame due to the combustion of manganese vapour necessitates the use of dark blue glasses to protect the eyesight of the operator .
When the metal is decarburized a thick smoke of brown oxides rises out of the bath , and finaJly becomes so denQe that it hides all other indications .
But before this moment has arrived the lines in the blue disappear , the bands in the green grow weaker and then disappear , while those in the yellow become weaker .
When all the bands have vanished the spectrum becomes continuous .
The blast is continued for a minute or two longer to decarburize the metal as completely as ible , nd a sample of slag and metal is remoyed from the converter by thrusting a wrought-iron .
into the bath ; from the appearance of these samples the nature of the metal is ascertained .
This brown smoke of oxides has been noticed by Dr. , of Osnabriick ( ' Le Ciyil , ' vol. , p. 25 , 1880 ) , when the converter is inclined so that some air passes ovel the surface of the bath of tnetal ; it is , therefore , due to oxidation .
I have observed such fume to be produced in enormous volumes when the charge become too hot , and the convertel is turned down during the " " boil so that the blast instead of the 1nolten passes and sweeps out the ) ours of iron and manganese .
Clouds of foxy-red smoke produced in this manner I have rise to a height estimated to be at least 200 , and even 300 feet .
The smoke is composed of little spherular particles containing oxides of iron and manganese .
can be no doubt that was , in the main , correct in considering that the noll-appearance of the lines of manganese at the colnmencement and terlnination of the " " blow\ldquo ; is owing to the fact that the quantity of tnatel.ial volatilised at these periods is insufficient the production of a spectrum ; but it may also be due , at the close of the process , to the oxidising atmosphere within the converter .
The spectroscopic phenomena of the blow are otedly determined by the chemical composition of the gaseous conten ts of the conyerte and of the bath of metal , the temperature of the metal , and of the issuing gases .
The of rise of rature is to incl.ease the volatilisation of manganese and iron .
An explanation of this will be facilitated by reference to three series of curves .
The first is reduced from the analyses of Mr. G. J. SNELUS , F.R.S. , who gave the original composition ] of pig iron blown , and of the meta ] in the converter at different in the process ; the second series is from the analyses made by the same nist , of the gas issuing from the conIverter a , t different periods .
The third eries of curves is drawn fro ] ] the analyses of metal taken from the commencement until tion of the " " blow " " as quoted by KUPELWIESER .
The pig iron in the latter case was highly manganiferous .
It will be seen from Diagram I. , that the pig-iron contains very little mnnganese , hat this is oxidised immediately during the first six minutes .
A small propor- tion of combined carbon is oxidised the same time .
The larger proportion , however , of this element is in the form of graphite , and this is converted into combined .
The combined carbon is then oxidised , and the oxidation proceeds with increasing rapidity .
The silicon at the same time also oxidises , and the oxidation of the two elements proceeds together rapidly during the " " boil until during the " " \ldquo ; there is little of either left .
Comparin these results with the composition of the uing gas , what do we find ?
First in the slag-forming , when the lines of the lkali metals are not yet visible in the spectrum , the carbon oxidised is all converted into carbon dioxide .
But the temperature rises , there is production of carbonic oxide , a gas which , according to Sir I. LOWTHIAN BELL , has a greater stability in presence of iron at elevated temperatures .
At this period the lines of the alkali metals are seen .
The gases of the converter , under such conditions as exist during the " " boil are those of a reducing atmosphere .
Oxidation of combined carbon to carbouic oxide then continues until near the close of the " " .
stage and it will be noticed that oxidation proceeds with increased ctivity .
If we consider that the temperature of the metal at the commencement of the " " blow\ldquo ; is no higher than that of the melting point of cast iron , that free oxygen passes through it carbon dioxide is evolved , there can be no doubt that the temperature is insufficient to volatilise manganese if it be present , and , therefore , no spectrum of this element is visible .
When manganese cannot be vaporised , iron certainly cannot .
Even the alkali metals are not carried oub of the for some five or six minutes .
When the ases are largely composed of carbonic oxide , and a higher prevails , the metals are volatilised , and the principal lines of sodium and potassium are obsel.ved during different periods up till the termination of the " " blow Why the red and violet lines of potassium are not visible is , no doubt , owing to the overpowering brilliancy of the other portions of the spectrum , and not because they are absent .
The best evidence of this is , that on plates specially sensitised , , he red line appears , and in every case the violet lines have been photographed .
The curves reduced from KUPELWIESER'S analyses are very different in detail , though the same in eneral effect .
FLAMLl SPECTRA AT HIGH The leading difference is owing to ) composition of the metal that of a highly manganiferous iron , containing much carbon and silicon .
Here we have manganese oxidised from the first , but at a low temperature ; the silicon is also oxidised , and a slag is formed from the esulting materials .
carbon , the silicon , and the anese , are dised together during the \ldquo ; when the temperature rapidly rises .
The effect of this is to volatilise manganese anal iron , and the of the con yerter constituting reducing atmosphere at a high temperature , carl.ies the metals of the convertel , where they are burnt in air along with the carbonic oxide .
Towards the close of the " " fining stage " " the manganese in the is reduced in quantity and the iron is incl.eased , so that the lines of the one give place to those of the other metal .
Suddenly the iron spectrum also becomes enfeebled , and the " " blow\ldquo ; is stopped .
Without doubt we have , at this point , a very rapid ution of temperature , because the only combustible material left is iron , and its heat of combination is comparatively small .
The Bessemer il me is not the esult of combustible gases and yapou being mixed with oxygen and burning within the converter , and thus , in a state of combustion , issuing into the air ; but on the contrary , it is a flalne of carbonic oxide at an exceedingly high temperature burning outside the conyertel in a cold atmosphere .
Temperature of and of the Various attempts were made by WATTS to determine the temperature of the and he concluded that , though it above the melting-point of gold , it was below that of platinum .
to some recent measurements made by LE CIIATLLIER ( ' Comptes ' vol. 114 , p. 670 ) , the temperature in the Bessemer conycrter the " " boil\ldquo ; 1330o C. , at the finish ; while the steel in the is C. The scale of temperatures adopted was that of .
:\mdash ; Melting-point of gold 104.5o C. .
1500 , , 1775o C. , , We have no measure of temperature at the period of the " " boil unless the metal in the converter is cooled durin the of the my photographs clearly indicate , it is difficult to how its could be raised by the additionl of the cooler eisell t ) of process of rburiz tion , and the still cooler which is at only a red-heat .
The rise of at this could be accounted the short ' after-blow\ldquo ; of secollds , is intended merely to PROFESSOR W. N. ON kinds of metal , but in no case could it exceed the temperature of the " " boil or that lnoment when the spectrum is most intense .
It must be remembered composition of the iron used , and the more or less rapid at which it is blown , ndoubtedly influence the temperature .
The greater the mass of material upon , the greater will be the rise in temperature .
For instance , a greater heat evolution and a higher temperature would ] the combustion of 10 cwt .
of silicon in 12 toIJS of pig ]uring a blow of lninntes , than from 1 cwt .
of silicon in 5 tons of pig iron blown for 20 minutes .
There is one sig.nificant fact to be obseryed in the spectra of the flame photographed during the " " boil\ldquo ; and the finishing stage\ldquo ; which bears upon the temperature of the metal .
When the spectra of manganese , manganic oxide , iron , oxide are photographed , tlJe number of the lines and bands in the spectra not more numerous than with a Bessemer flame spectrum of onJy one half minute 's exposure , though the above spectra may have received any exposure from 30 to 80 minutes .
WATTS obseryed ( ' Phil. Mag. , ' 1870 ) that the sodium lines 5681 and be employed as an index of temperature , since they are present in the spectrum flame containing sodium the temperature of which is hot enough to melt platinum , but do not pear at 1ow temperatures .
The Bessemer flame does not this double lino , but only the 1 ) lines ; neither does it show the lithium line , }hich appears at a somewhat lower temperature .
We cannot conclude from this that the flame is not hot enough to produce these lines , because in such a case we have to deal , not only with the temperature , but the quantity of material present , and the relative brilliancy and con sequent visibility of the two pairs of sodium lines .
The proportion of in the Bessemer flame is evidently very small , the natl.owness and want of intensity of the lines , and the fact that they were not seen reyersed in any spectrum .
Hence , though the temperature may be enough , the quantity of material present is not sufficiently large to yield the lines 5681 and 5687 .
The elations of the different lines have really not been investigated in flame-spectra , except and alone so far as they apply to total extinction of all lines , which in the case of sodium refers to the lines only , and this in flames no hotter than that of a Bunsen burner .
If we apply the same linle of reasoning to the nppearance of the reversed hydrogen lines in the red and t , he blue , it may be stated that the line in the red , corresponding to solar line , neyer appears in any hydrogen .
hydroburnt with air or with oxygen .
It invariably appears in sparks passed through steam , and it also comes out as a reversed line under suitable conditions in the Bessemer spectrum .
conditions of its appearance do not depend upon an alteration in temperature , but on the presence of a sufficient amount of water-vapour in the blast .
This cel.tainly seems to point to a higher temperature than that accorded to the FLAME CTRA flame by .
Finally , we have this striking fact to consider : when the conyerter nited at the ) .
of the , the potasslum .
' the red lithium , and the yellow sodium lines presenlt e are seen the , but neither the orange line the , sodium pair 5681 , 5687 appear ; yet there must be a lnaximunl amount of in the or in other words all the alkali metals in the concentrated in this maximum , however , being an extremely ntity .
It is worth while to take into account the heat uf combustion ) the elements in pig iron which are remoyed during the " " blow and calculate , so ) able , the absolute heating effect of .
oxidation .
Any calculations of the kind must necessarily be incolnplete , owing to the speciric heats ases at high felll tures being undetermined , though these have been ttely itedb I and also by IATLLIEp .
) ecific ] of ) iron and of nister and also } THE Combustible Elements Iron , the Colnpotlnds , and their 1 [ eat of les ' A. ) ) .
The names of the authorities are indicated as follows : ; ] TROOST and .
and ; B. , ; G. , GRASSI .
Kilo .
} 1 .
of ch 0 .
T. 7 .
H. ; , A. , In iive tons of pig iron there 5,080,240 .
Let the osition o ] ) ) be that which was quoted by Mn . . . .
cent. . . .
C. . . .
3 .
, , as Si . . . .
, ' . . . .
, , olnbi n Total carbon Ann. ] ) ) I I .
de .
vol. 4 , KILOGRAMME Heat-units evolved by the Combustion of the above Materials .
Grms .
Mn 175,776 C. . . . . .
199,653 Si . . . . .
99,573 Kilo .
heat-units Total kilo .
heatevolved by units evolved by each lgrm .
element .
302,335 489 , 150 7 .
mlne heat-units evolved by the coInbnstion of the , carbon , and silicon in five tons of iron GRAMMLS of purities eliminated Five Tons of Pig Iron in the of Monoxide , solid Silica , Manganous Oxide .
' CO SPECIFIC Heats for Equal Weights . . . . .
0.245 N. . . . . . . . . .
0.175 Fe . . . .
0.11379 Mn . . . .
SPECIFIC Heat of the Materials in the Converter and of the Products of Combustion .
Kilo .
Grms .
Sp. heat-units .
Weight of iron 4 , ) CO 465 , . . . .
1 , \ldquo ; silica . . .
two substances form manganese oxide 226,911 a The ecific heat of lass , perhaps , approximates more closely to that of the slag than that of any other substance which has been determined .
This is higher in the molten than in the solid state .
Sp. Ht .
Sp. Ht .
At 21 F. : at 57 F. AT RATU R The of the silica and the nganese oxide\mdash ; Sp. Ht .
kilo .
he -nnits .
grms. 83,3 .
The total kilo .
heat-units developed by combustion in the converter amount to 1,558 , 197 , and the specific heat of the iron and the products of combustion , with also the nitrogen in the air , amounts to kilo .
heat-units , nd the pyrometrical effect is \mdash ; or The temperature attained , according to the foregoing calculations , amounts to 1454o C. above that of the molten cast iron .
This , however , is the value ; we must for th1 specific heats of tlJe gases , the metal , and the elevated temperatures than the } th specific heats were determined .
The specific heat of the conve ] .
must be considerable , but it ) ) ememl ) ered that it is already heated to the temperature of the molten But even if we allow that 50 per cent. of the is ) or conyeyed then we should the temperature bove thal ( ] molten pig thus with grey iron at 1220o C. the metal acquire a teml ) erature of more 1947o C. , which is above the melting point of platinum .
Judging by the number of lines and bands belonging to iron nese which have been photographed in the of the flame , the telnperature must nearly approach that of the oxyhydl .
fl me , even if it does not exceed it in certain cases at the highest temperature of the " " boil At 1)owlais , for instance , where the metal , which rich in silicon , carbon , and manganese , is just a hot-blast furnace and conveyed by in ladle to the con verters , it is probably hotter at the commencement of the " " blow\ldquo ; than if cold pig iron had been merely melted in a cupola .
The parallel colunlls below show the numb ] of line , sobsel'Ved in the of respective substances under difleI'ent conditions :\mdash ; Bessemer ) CO spectl'um .
8lines and of bands .
posure to 3 minutes Mn spectrum , 7 : ; lines and of bands .
Exposure to 3 minutes .
Fe spectrum , 92 lines .
minutes .
CO spcctrum , 10 lin cs and of ban drogen , 1 iincs of Fo octrunl , 9 lines .
xposnrc The temperature to be judged by the iron , because there is ) difference in roport of the metal present in two cases than in the case of the carbonic oxide and manganese .
It is a striking fact that the Bessemer flame quired a much shorter exposure than the oxyhydrogen blow-pipe ilame , but the volume of flame is much larger ; on the other , however , it is so close to the ilJstrument .
The following quotatio *is of particular interest in connection with the temperature of the Bessetner " " blow\ldquo ; :\mdash ; " " Les methodes pyrometriques , j'ai entretenn i diverses reprises l'Acadernie , m'ont permis d'effectuer la mesure precise d.es tempe'ratures developpees dans les foyers iels .
Les resultats ainsi obtenus sont , pour certaines industries , en contradiction avec les estimations faites anterieurement , et no seront sans doute sans contestation .
Je serais heureux si leur publication pouvait provoquer des experiences contradictoires sir le nleme sujet .
Les ratures donnees ici sont bien inferieures a celles qui sont le plus ralement admises pour les industries en question : 2000o pour l'acier ; 1800o pour porcelaine ; 1200o pour le gaz d'eclairage .
ration de ces derniers chiffres tient i plusieurs causes .
Entre differentes de'terminations de temperatures non conntes , on choisit de pref6rence les ) elevces , par suite d'un sentiment instinctif qui conduit admettre un quasi-proportionnalit6 entre la temperature d'un corps et son eclat ou quantity de combustible depense pour l'echauffer , tandis qu'en reality ces deux grandeurs Cl.oissent suivant un fonction extremement rapide de la temrature .
En second lieu , le procede le plus quemment employ jusqu'ici dans l'industrie pour ] mesures pyrometriques , a ete la methode calorimetrique , en se servant des morceaux de fer dont on supposait torb la chaleur specifique invariable .
nfin des causes d'erreurs particulie res sont venues fausser des comparaisons dans llesquelles on utilisait le point de fusion du palladium ou du platine .
Ainsi la temperature du Bessemer avait 6te fix par a 2000o , parce que le ) latin paraissait fondre rapidement dans sa flamme .
J'ai reconnu qu'il no fondait pas , mais se dissolvait dans les gouttelettes d'acier fondn entraine par le courant gazeux .
De lneme , le palladium pass pour fondre dans differents fours ou en realite il se transform , sans fusion , en nne mousse spongieuse , par le fait d'hydrogenation ou d'oxydation passagere of the of the Manganese Spectrum in all cases the ' ' Boil\ldquo ; and the close of th " " Fining stage There .
is one fact equires to be explained in connection with the spectrum of the Bessemet flame .
How can the characteristic and bands in the Bessemer LP vclopp dnns lcs ers \ldquo ; Rendns , ' vol. 114 , p. 470 .
FLAME AT IIIGII } be , accounted for in ( cases if they are ) dne to carbon oxide but to manganese ?
On an examination of the curves the of SNELTTS it will be observed that the nanb ) is burnt out of the molten ) within the first few into ( It is ) vious that nganese b in these .
from of metal ) to the end of the " " fininos ' ' if the to arked indication of them is atively large .
sole sonrce of the ) be quantity of metal reduced from the slag , this reduction cannot take ) once , is the result of continously teltlr ) the che mical ctio of reducing materials such ferrous cal.bide , ( possibly ) In order to hetller s. could ) reduced and rise to teristic spectrum of the Bessemer picce of by the tten I both in the inner outer flames , where in the latter case eduction could take placc by dissociation by ] in the former it might be aided chemical action of the cxccss of .
The photographs obtained ikingly like thosc the at it with line for line in the spectrum , exactly can no whatever that both the proportion of nesc in tite 1 ) ) the in Diagram I , that the in metal is converted the hrst five or six minutes , it is evident the continued ] iancy of the spectrum during the " " boil\ldquo ; must } ) entirely due to its tion from the In order to connect the of ) ectrum with the chelnical h involved in the tion of the iron , we consider sily of the line of " " ( stagc the final ) flame , and it will be bc dn to other cause fill of ture , consequenlt upon a reduction in ] tity of ] ) thinness , ency , want of th ) is the comparatively ntity of ) being caused by an escape of oxygen tl which is ] cloud 0 .
) of hc carbon causes the disappearance of the Let us now consider the case of the ' " " \ldquo ; The spectrum of the in \ldquo ; ) , and those of metallic manganese is ized in ths of ] ] their distinctness , and great brilliancy .
of make , plain , for herc we of in the , .
oxidized during the " " \ldquo ; to PROFESSOR W. N. HARTLhY ON whole remaining period of bhe blow , this metal being of similar composition to that blown at Dowlais .
It must , however , be remembered that the higher temperature at Dowlais would , if acting upon nganese slag , produce a spectrum more similar in character to that obtained from ganese or spiegel-eisen than that obtained at Crewe , but the iron lines in these circumstances would also be stronger .
The connection between the termination of the blow and the drop of the is to be explained exactly as in the case of the spectrum at , the difference in the spectra being due to the quantity of vapour of manganese in the flame .
The Technical Aspect of expelience has shown that in England , in Styria , and at Seraing , in Belgium , the use of the spectroscope has rendered substantial service in determining the end of the operation in the Bessemer converter , notwithstanding that the nature of the spectrum observed not accurately ascertained , nor the cause of its production well understood .
The reason of this is not far to seek , when we consider that towards the close of the " " fining stage\ldquo ; the indication is particularly distinct , for it culminates in the disappearance of the bright lines and flutings of manganese , whether these proceed from the presence in the flame of material vaporized directly from the metal itself , or from the slag which is formed from the oxidation of elements contained in the metal during the first period .
The quantity of grey cast iron , of spiegel-eisen , or ferro-manganese , which is finally added , determines the hardness of the steel , and the composition of the added material is ascertained with exactitude by mealr of frequent analyses .
By this means a metal is obtained which is much more constant in composition than when the process is interrupted before the completion of decarburization , and when , after the result of a test of the mebal , spiegel-eisen is added and blowing is resumed minutes .
lt was the aim of of steel works in the early days of process to blowing before complete decarburization , ordel .
that the sary proportion of cavbon might be left in the charge ; but it was found that no cel.tainty in being able to produce the me quality of metal at each ' ' blow It will be seen the results of this investigation that the thermo-chemical operation involved in could not be always carried out exactly under the same conditions , or within the same precise limits of time ; the initial temperature of the metal , and even the temperature of the converter into which the charge is allowed to flow , can affect the rate of combustion of the carbon , and the oxidation of the mannese and silicon ; and , furthermore , it must be borne in mind that , with successive rges from the same blast-furnace , the composition of the metal varies to an extent which can easily upset all previous calculations .
When the charges come from ditferent furnaces , further complications and increased difficulties arise .
As there is AT JIIGII I no indication , except at the close of the \ldquo ; ) unburnt is reduced to a proportion of ] , it quite evident that the most rational able lnode of is it and then to add a sufficiency of a ) metal which ill produce of required hardness and composition ; by the as well as by spectroscopic yses , the tion of the st \ldquo ; is clearly indicated .
The constitution of the tion , and the canse of tYauese b ' iron lines , even after the nese h ) the metal , which explained , affords scientific reasons not only to ) the has been universally adopted , but of not pritct i of complete tion is 1nost ried out , it yieldc and con tinues to yield , enorlnous q , of ) steel or , in of purity and of remarkable constancy i 1 .
The complex nature of the the of bands of nese , carbon , ic oxide , possibly ( of oxid of the lines of iron , nese , potassium , sodiuln , en .
The of manganese are to some extent obsctl.ed , first , by the of the carbonic , secoll the of carbon ; .
the ganese bands are ) degraded in the opposite direction , that is , blue .
2 .
The cause of the tncc of the in .
th the " " blow\ldquo ; is the little of the 1nolten metal , and .
the free that a seous mixture which ) come from ) inter ) the silicates present in very ) do not disclose they contain til : tl bnt } been detected in do show in the 4 until ) temperature nfiici c There can be ) ) I and iron are freely in cnt ) ic heated state , out of the of lnolen lines and of iron itnd of esc , ) ' ] bands in the of various alloys of tCXClV .
PROFESSOR W. N. ON 4 .
The question of porization of manganese and of manganese oxide from slag is eyond all doubt by actual experiment the blowpipe flame .
This ' the observed by , namely , when a converter is being heated with coke after it ha been used , but not re-lined , the spectrum of the Bessemer makes its appearance . .
The luminosity of the flame during the " " boil\ldquo ; is due , not only to the combustion carbonic oxide , but also to ) presence of the vapours of iron and es in the The of the manganese spectrum at the end of the " " fining ' is due to a reduction in the quantity of heated carbonic oxide escaping oYll the converter , which arises from the diminished quantity of carbon in metal .
When the last traces of carbon are gone , so that air may escape through the metal , the blast instantly oxidizes any either in the metal or in the atmosphere of the converter , and furthermore oxidizes some of the iron .
The temperature must then with great rapidity .
7 .
The entire spectroscopic enomena of the " " blow\ldquo ; are undoubtedly determined by the chemical composition of the molten iron and of the gases within the converter , the temperature of the metal , and of the gases .
8 .
The probable temperature of the Bessemel flame at the finish is produced by the combustion in cold air of carbonic oxide heated to 1580o C. ; that is to say , to the temlJe ature which , according to LE , is that of the bath of molten metal from which gas has proceeded .
bath of ) acts at the same time as a ns of blast , producing the gas , and as a furnace on the regenerative principle which the prior to its combustion .
9 .
If by the lines and bands ' to iron and nese which have beeu measured in phed spectr of the Bessemer flame , the temperature must nearly approach of the flame , and may easily attain the -point of 10 .
spectrum enable from Bessemer slag by the oxyhydrogen flame is composed of precisely the most characteristic features of the flame-.spectrum , as seen from the erter ab Crewe .
The continuous spectrum of carbonic oxide , bands ) lines of compound , of tarry carbon , , as a matter of course , absent .
The ff me at Dowlais differs this , and l.esenbles the spectrum of metallic anganese .
These diffel.en ces may not be entirely due to higher ] ) el.ature at , but to the difference in composition of the metal .
The manganese at Crewe be oxidized to within the fir , seven minutes\mdash ; that is say , the less Sectrun m its appearance in the .
That at Dowlais would not be all removed from the of lnetal b oxidation before the end of the " " \ldquo ; forIner yields such a ) be obtained from slag , the latter , one hich is only spiegel , nganese , pure metallic FLAME CTRA A lIIGlt nese .
There be that , in blowing nganese and iron be by the volatilization of elenlents slag .
11 .
The complete termination of the -ing \ldquo ; is indicated , but is no indication by the flame of the composition of the metal within the nverter at any stage .
As the progress of the " " \ldquo ; is overned by the position of the metal temperature in the converter , and as these cannot be controlled with perfect ctitude during each " " biow it follows the practice of is the best course to pursue , the amount of nese b added subsequently .
I desire to record expression of my .
F. W. , of the Works , and Mr. E. P. of Dowlais , for the they have me in making these observations ; to the Goyernment ratl b Committee , the of cal.rying out this investigation ; and to assistant , for the care he has exercised in carrying out my instructions , in the ) and measuring some of the spectra .
I propose to pursue the in another ection , by a series of obser .
to the basic Bessemel " " blow the , and various ) of Siemens steel furnaces .

