Dec., 1916
T H E J O U R N A L OF 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
relative efficiency of different types of samplers. For this purpose two bags of fertilizer were carefully made at one of our factories containing 7 . 2 0 per cent ammonia or a guaranteed percentage of 7 per cent. P a r t of t h e nitrogen was derived from nitrate of soda and t h e remainder from organic ammoniates. This fertilizer was sampled b y three different persons a t different times with three different types of samplers and t h e samples analyzed. The results of t h e analyses were as follows: SAMPLED BY Assistant Manager. Chemist ...................... Superintendent
Per cedt Ammonia No. 1 No. 2 No. 3
. . . . . . . . . . . . 6.48 6.90 ................ 6.43
Average
......................
6.67
6.90 7.21 7.17 7.09
.... ....
6.27
6.27
The results under No. I are from samples taken with old style half round tube; No. 2 new style round sampler, which opens and closes; No. 3 what is known as a rice sampler. The first result under No. z is lower t h a n t h e others for reason t h a t t h e sample was not properly manipulated; t h e tube was inserted into t h e fertilizer without being closed. The proper way t o handle t h e sampler is t o push t h e closed tube into t h e fertilizer t h e full length of t h e bag, open, t u r n u n t i l filled, close a n d withdraw. T h e results of these different analyses show t h a t t h e sampler which closes gives results on ammonia nearest t o theory, while with t h e other types t h e results are very much too low. As many of t h e agricultural departments still use t h e old style sampler, it is more t h a n probable t h a t many of t h e deficient analyses on samples taken b y official inspectors are a result of using improper sampling devices. This is a very important matter from t h e manufacturers’ standpoint and, while there has been no intentipnal neglect on t h e part of t h e fertilizer controls, i t is t o be hoped t h a t they will make a thorough investigation of t h e whole system of sampling a n d make such changes in existing methods as may be found necessary. VIRGINIA-CAROLINA CHEMICAL COMPANY RICHMOND, VIRGINIA
BETTER SAMPLES FROM THE FERTILIZER FACTORY1 B y A. J. LAWRENCE Received October 20, 1916
A great many useless analyses are made through lack of cooperation between t h e laboratory and the factory and even where there is a strong desire on the part of t h e superintendent t o have representative samples sent t o t h e chemist, trouble often arises through incorrect methods of sampling. I n t h e fertilizer industry t h e methods of collecting samples vary t o some extent in nearly every plant and we believe t h a t by standardizing these methods, much trouble could be eliminated and more accurate a n d reliable information would result. I n receiving raw materials, some contracts specify as t o t h e number of bags to be sampled. The large majority of sellers, however, leave t h e actual method 1 Presented at 53rd Meeting of the American Chemical Society, New York City, September 25-30, 1916.
1 I45
of sampling t o t h e sworn weigher. We have found considerable variance in t h e methods used by different weighers, many of whom do not realize t h e importance of their work a n d t h e trouble t h a t may arise between buyer and seller owing t o disagreement in analysis. An example of this is shown in a case in which there was considerable difference between t h e buyer’s and t h e seller’s chemists on a cargo of garbage tankage. The source of trouble lay in t h e fact t h a t t h e sworn weigher, without consulting t h e chemist or sampleboy, had quartered t h e fluffy, dusty material during a very high wind, in which a large amount of t h e fine particles were blown away, and in leaving, told t h e sample-boy t h a t he had gotten his sample and t h e remainder was in the barrel. The boy got his sample, not mentioning t h e incident, t h e results of which he did not realize. Also, a t times, where t h e samples are carefully mixed and quartered and t h e duplicates and triplicates agree excellently, still, lack of judgment in quartering on moist or dirty floors, or using brooms or shovels which are contaminated with foreign materials, will result in t h e analysis of t h e goods differing t o some extent from t h a t actually received. A considerable loss of money can very easily result through using t h e “bought” analysis on a lot of goods, which analysis may be slightly different from t h a t in the pile. I n one instance, a low-grade tankage was sampled, t h e sworn weigher requiring t h e unloaders t o take a handful from each cart and drop same into a burlap bag tacked t o t h e side of t h e scales. A large tin bucket was put on t h e other side of t h e scale house, a n d t h e men were required carefully t o take a cupful of sample from each cart and place same in t h e bucket. This experiment was tried because i t was believed t h a t in dropping t h e tankage into t h e bag, some of t h e fine ammoniated particles were lost b y being blown away from t h e mouth of t h e same; also it was noticed t h a t a considerable portion sifted through as each handful was thrown in. On several occasions t h e sample in t h e tin bucket showed a higher analysis, running from 2 j t o go cents per ton in valuation above t h a t obtained on t h e sworn weigher’s sample. The observance of these points would save considerable trouble later on should analysis be made in the piles and found t o disagree with the “bought” test. I n batching stock brands which are to be shipped after being cured, we do not believe t h a t enough stress is laid on t h e method of sampling. In t h e majority of plants, t h e usual plan is for t h e superintendent t o run up several tons of a brand and either have a sample drawn from t h e mill or t h e carts as they dump into t h e pile, or have t h e sample-boy dig down a few inches into t h e pile at several places. I n most instances, with careful sampling, fairly reliable results can be obtained in this way. However, from several experiments we find t h a t quite often very bad analyses result from this method of sampling and cause numerous retests in t h e laboratory as well as new samples from t h e pile. We believe t h a t this trouble can be eliminated t o a large extent by using a large prospecting auger
I 146
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
such as is used in sampling soils and clays. This consists mainly of two blades arranged t o cut a &in. hole a n d so bent a s t o pick u p t h e material each time. Sufficient amount of handle can be attached t o enable a pile of nearly a n y depth t o be thoroughly sampled. I n this way uniform sections can be bored into a pile in several places a n d duplicate samples invariably check, showing t h a t if t h e materials are p u t into t h e mixture, a correct sample of same can be obtained. T h e following tests were made on four brands of about 1000tons each. Four samples were bored from each bin with t h e sampler as well a s a sample being drawn from t h e mill during t h e run of t h e entire lot. The brands being figured t o t h e exact analysis, t h e slight overrun is evidently due t o shrinkage. "IM" BRAND P,Os KzO
..,21,r5.' BPAND P301 G O 12.55 4.96 12.42 5 . 1 0 12.48 5.12 12.635.03
10.05 7.91 9.92 8 . 0 8 9.96 8.10 9.938.04
12.52 5.04
Mill sample ( ~~ ~~ t~11.964.66 d ) )
Av. (Bored)
"1-8-4"
"2-8-lo" BSAtZD
BR*ND
NHt P.0. K10 1.15 8.05 4.00 0.99 8.18 4 . 1 6 1.03 8.15 4 . 0 6 1.138.024.06
NHI P.06 G O 2.18 8.15 10.20 2.01 8 . 1 0 10.20 2 . 1 1 8.17 10.01 1.968.1810.14
9.97 8.03
I .07 8.09 4.07
2.06 8.15 IO. 13
10.67 7.83
1 . 1 0 8 . 5 0 3.60
2.16 7.80 10.30
_ _ -_
_ _ _ __-
On these "mill" samples it would be necessary t o change t h e formula which would result in incorrect analyses, whereas t h e bored samples show t h a t t h e piles actually r u n about a s figured. Bad analyses on bored samples can invariably be traced t o f a u l t y mixing i n which either wrong amounts were weighed in or t h e actual materials fail t o analyze as figured. I n sampling bagged shipments, we believe t h a t much more reliable information can be obtained as t o their analyses b y t h e use of t h e double tube sampling trier which is inserted into a bag, closed, a n d withdrawn, giving a n exact section of t h e material therein. If methods of sampling could be a s rigidly enforced in t h e factory as in t h e laboratory, we believe t h a t more dependable information would result. The average factory chemist does not bestir himself sufficiently t o see when samples are being drawn correctly and t h e average superintendent does not realize t h e importance of t h e small details i n sampling, which have a vital bearing on t h e plant and its reputation for uniform analyses. CXBMICAL L~soamoau.F. S. ROYST*.GUANOCOUPANY N o s ~ o ~Vmarlrrr n,
Vol. 8, No.
12
position on t h e scale of t h e instrument, representing its width a n d intensity by t h e width and height, respectively, of t h e drawina. T h e spectrum map paper here described differs from t h a t ordinarily employed in t h a t there is provided above each scale a cross-sectioned area one centimeter in height on which t h e maps are made, using t h e heavy line a s t h e base line. The vertical lines of t h e crosssectioned area are useful as guides in mapping spectrum lines, while t h e horizontal lines simplify t h e comparison of t h e intensity of lines and hands in different parts of t h e spectrum. The spectra t h a t are shown were mapped a s a n illustration of Bunsen's method a s applied t o this new map paper. The compounds employed were respectively: lithium carbonate, sodium carbonate, potassium carbonate, rubidium carbonate, caesium alum, calcium chloride, strontium chloride, barium chloride, boric acid, a n d manganese chloride. The spectra were produced in t h e usual fashion by introducing t h e sub-
stances into a Bunsen flame on a platinum wire, I n t h e case of t h e chlorides of t h e alkaline earths, where rapidly shifting spectra are first obtained, t h e m a p s represent t h e fairly stable spectra t h a t are produced after a preliminary heating of t h e chlorides on t h e wire.
Received July 17. 1916
This map paper can also be conveniently employed for t h e mapping of absorption spectra a t different concentrations for t h e determination of t h e point of "minimum brightness" of t h e substance in solution.
The method of mapping emission spectra which was suggested by Bunsen' has proven of value in t h e course in Opticochemical Methods a t Cornel1 University as a n aid in familiarizing t h e student with comparatively simple emission spectra. Bunsen's method consists in plotting t h e lines a n d bands of a spectrum on a millimeter scale, numbered t o correspond t o t h e arbitrary scale of t h e spectroscope. Each line or band is placed at t h e position on t h e scale t h a t corresponds t o its
The illustration shows a sheet of t h e spectrum m a p paper about one-third natural size. This paper is furnished by t h e Cornell Cobperative Store, Morrill Hall, Ithaca, N. Y . , in packages of fifteen sheets, together with three sheets of millimeter cross section paper, a t a cost dependent upon t h e price of paper. Until recently a package was sold for fifteen cents, postage prepaid, with lower prices on ten or more packages.
A NEW SPECTRUM MAP PAPER BY R. P. A N D B ~ S O N
1
Ann. d n P h w . und Cham.. 106 (1863). IO.
CORN%LL U N I Y B R ( V I I .
Im*cA. NHw Yoax