Determination of Trace Quantities of Nitrogen in Petroleum Fractions 0. 1. MILNER, R. J. ZAHNER, L. S. HEPNER, and W. H. COWELL Research and Development laboratory, Socony Mobil Oil Co., Inc., Paulsboro, N. 1.
b During a study of catalytic reforming of petroleum hydrocarbons, a method was needed to determine nitrogen concentrations of less than 5 p.p.m. A method has been devised which comprises the isolation of nitrogen compounds by extraction with concentrated sulfuric acid. A conventional Kjeldahl digestion of the acid extract is carried out, followed by neutralization of the excess acid and distillation of the ammonia from a basic medium. The ammonia is a b sorbed in an acidic solution and determined colorimetrically by reaction with phenol and hypochlorite to form a blue indophenol. The standard deviation of the method for samples containing from 0 to 3 p.p.m. of nitrogen is 0.1 4 p.p.m.
the method is sensitive compared with acidimetric titration, it is limited in precision to about 3 or 4 y . This is equivalent to no better than 1 or 2 p.p.m. with samples of the size that can be conveniently handled by the conventional Kjeldahl digestion, usually a maximum of 2 or 3 grams of organic matter. .4n obvious approach is to separate the nitrogen compounds from the bulk of the hydrocarbons and then proceed 11ith the Kjeldahl digestion and determination of ammonia. Bond and Harriz (1) successfully sequestered the nitrogenous matter on silica gel. The silica was then digested, and the ammonia was determined by distillation and acidimetric titration. However, large samples are required and complete separation of the nitrogenous matter is attained only a t slow percolation rates. Bumping during the digestion proves troublesome. Another possible technique was extraction; 98% sulfuric acid had been reported to extract nitrogen compounds completely (3). This was particularly convenient, because the acid extract could be digested directly in the normal Kjeldahl procedure. The resulting ammonia could be determined with phenol and hypochlorite to give increased sensitivity over acidimetric titration. A study of this approach has resulted in a method suitable for
0
nitrogen compounds, present in petroleum hydrocarbon fractions, poison most catalysts used in commercial reforming processe.. T o control the nitrogen content of such stocks, it is necessary to be able to determine nitrogen in the range from 0 to 5 p.p.m. with a precision of the order of 0.1 p.p.m. Noble (4) used the reaction of animonia with phenol and hydrochlorite to form a blue indophenol a s the basis of a colorimetric method for determining nitrogen in petroleum. Although RGASIC
Table 1.
Accuracy of Extraction Method
Sample Pyridine, reagent grade Aniline, Eastman redistilled Benzonitrile, Eastman Quinoline, Eastman Naphtha Straightrrun Thermal Coker Pyridine Vanadyl tetraphenyl porphine,
Solvent or Diluent Iso-octanes Iso-octane' Iso-octane& Iso-octanes
.. Kaphthab Kaphthab Naphthab
Total Nitrogen, P.P.M. Found Direct Calcudilated gestion Extraction 2.56 2.19 3.89 3.32
.. 0: 7
i:i 1.4
..
..
.. ..
..
20 26 .. 17
.. ..
8
2.8, 2.4, 2 . 6 2.3, 2.2, 2.1 3.5,3.8, 3.6 3.2,3.4 18 25 0.6 16 1.3 1.4 7.9
Benzenea 1.0 .. 1.3 Xitrogen-free, by extraction analysis. Catalytically hydrogenated mid-continent naphtha; 0.25 p.p.m. nitrogen by extraction analysis. 7.5y0nitrogen
a b
1528
ANALYTICAL CHEMISTRY
the routine determination of trace amounts of nitrogen. DEVELOPMENT OF METHOD
The technique as first tried consisted of the following steps: extraction of the sample with four 10-ml. portions of 98% sulfuric acid; addition of 10 grams of potassium sulfate, 2 drops of mercury, and several Carborundum boiling chips to the acid extract; digestion of t h e extract, followed b y a 3-hour refluxing period to ensure conversion of the nitrogen compounds to ammonia; distillation of the ammonia; and colorimetric assay of the ammonia in the distillate by the Koble procedure. To check the completeness of extraction, tests viere run on pure nitrogen compounds dissolved in hydrocarbon and on certain blends of petroleum stocks. The blends Ivere prepared from naphthas that contained relatively large amounts of nitrogen as determined by thc conventional Kjeldah1 procedure r ith a colorimetric finish (4). These high-nitrogen stocks were diluted with a mid-continent naphtha, that had been catalytically hydrogenated to remove nitrogen, to give blends cont;iining low concentrations of nitrogen in the range of interest. The results in Table I indicate that nitrogen compounds of widely varying structure and basic strengths are extracted almost completely. Additional data from duplicate analyses of typical low-nitrogen gasolines and naphthas (Table 11) indicate a standard deviation of 0.14 p.p.m. for nitrogen concentrations of less than 3 p.p.m. Purification of Reagents. Table I11 shows some souices of the nitrogen blank. To reduce t h e blank, t h e reagents were purified b y recrystallizing potassium sulfate, redistilling water a n d sulfuric acid, and pretreating sodium hydroxide solution with zinc. Recrystallized potassium sulfate retained much of t h e nitrogen present in t h e untreated salt. Consequently, the quantity of potassium sulfate added to the digestion mixture was restricted to 10 grams, one-half the quantity needed to ensure rapid conversion of refractory nitrogen compounds to ammonia, according to
Lake and his associates ( 2 ) . T o compensate for the use of a minimal quantity of potassium sulfate, samples had been refluxed for 3 hours after complete destruction of the organic matter. Because ammonium sulfate might be the impurity in the potassium sulfate, simple heating was tried for further purifying the salt. Table IV s h o w t h a t heating a t temperatures over 550" C. is sufficient to remove nitrogen completely. With the removal of the nitrogenous impurity from potassium sulfate, the salt level could be increased to 20 grams and the final refluxing period could be reduced to the normal 1 hour. Tests with macro amounts of pyridine, considered to be the most refractory type of nitrogen compound, confirmed that nitrogen was completely converted to ammonia under these conditions (Table V). Purified 50% sodium hydroxide solution was prepared b y dissolving 2 kg. of reagent grade sodium hydroxide in 2 liters of redistilled water. T o the hot solution, approximately 5 grams of pondered zinc was added and the solution was blown with a n inert gas (helium or nitrogen) for 2 hours. The solids were allowed to settle overnight, and the solution was filtered into a polyethylene bottle having a drain cock a t the bottom. With the use of redistilled acid and water, heated potassium sulfate, and zinc-treated caustic, total blank values were reduced to 15 to 20 y of nitrogen. Use of 92% Acid. Sulfuric acid (987,) extracted excessive organic matter from olefinic stocks, t h u s slowing digestion. More dilute acid was tried t o determine if this effect could be lessened without affecting t h e completeness of t h e extraction. Also, two 20-ml. volumes were used instead of four 10-ml. volumes. Included in t h e samples tested were iso-octane solutions of extremely weak nitrogen bases. Table VI s h o m t h a t 92% acid effectively extracts trace quantities of nitrogen compounds from a wide variety of petroleum stocks. APPARATUS A N D REAGENTS
Beckman Uodel B spectrophotonieter, equipped with a 2-cm. cell and set a t a wave length of 625 mp. Kjeldahl digestion and distillation apparatus. All connections are groundglass spherical joints which are held together tightly by clamps. Separatory funnels, 250 i d . , with Teflon stopcock. Teflon boiling strips, X 6*12inches, cut from a Teflon sheet lIIEinch thick. Sulfuric acid, 98%, redistilled in allglass apparatus. Water, redistilled from acid permanganate in all-glass apparatus, is used in preparing all reagents and in all operations. Sulfuric acid, 92%. Add to 100 ml. of n ater wfficient 98% redistilled sulfuric
II. Repeatability of Method Sitrogen Found, P.P.M. Test 1 Test 2 Test 1 Test 2 0.9 0.8 0.6 0.3 0.9 0.7 1.0 1.0
Table 111.
Table
0.2 1.2 0.4 0.1 0.4 0.2 1.3
0.1
1.4 0.7 3.0 1.0 0.5 1.5
1.1
0.6 0.1 0.4 0.4 1.4
S.D.
=
dg
=
Nitrogen in Distillation Reagents Charged to distillation flasks: 250 ml. of HzO as indicated below 80 ml. of 50% NaOH soln. made nith
reagent grade pellets
1.1 0.5 3.1 1.3 0.6 1.8
6.5 grams of zinc granules 3 Hengar granules S
Found, pio. 1. Distilled water 40, 51 No. 2. Distilled water, cationexchanged 41, 3T S o . 3. Distilled Kater, cation-
0.14 p.p.m.
exchanged. NaOH solution filtered and b l o m 5 hours 35, 36 No. 4. H20redistilled from acid 15, 17 permanganate S o . 5. HzO, redistilled from NaOH, acid permanganate. zinc added t o solution while hot, blown 2 hours 11, 8, 8
acid so that the final volume a t room temperature is 1 liter. Potassium sulfate, purified. Reagent grade potassium sulfate is heated a t 550' to 700' C. for 16 hours. Sodium hypochlorite solution, reagent grade, 5y0minimum available chlorine. For each bottle of reagent, follow the procedure used by Koble (4) to determine the volume of reagent required to develop maximum color intensity on a known amount of ammonia. This volume must be redetermined each time a fresh bottle is opened or at weekly intervals, whichever is more frequent. Acid-washed iso-octane. Extract -4STRI knock-grade iso-octane with one tenth its volume of %?yo sulfuric acid. and wash with water until neutral. Allow to settle, filter, and store.
Table IV.
Treatment Xone Recrystallized Heatkd 16 hours 550" 700O
Fused
D
0.0 0.0 0.0
486, 657 876, Si0 a Total digestion time was 1.25 hours, including a refluxing time of 45 minutes 10 20
40 40
878 878
after clearing.
Theoretical
Sitrogen Content, P.P.M. Found Direct 98% digestion acid
...
... ...
,.. ...
...
...
Visbreaker gasoline Pretreated mid-cont. naphtha Thermal-mid-cont. naphtha blend Benzonitrile in iso-octane
A B C
c. c.
Extraction of Nitrogen with 92% Sulfuric Acid
Sample Thermal naphtha
Carbazole in is-octane Indole in iso-octane Pyridine in iso-octane Coker naphtha
N per Gram K?SOI 1.3 0.9
PoNitrogen, tassium Sulfuric Added Sulfate, Acid, as Grams M1. pyridine Found
Extraction and Digestion. Place 70 grams of sample in a 250-ml. separatory funnel and extract twice with 20-ml. portions of %yosulfuric acid. If t h e sample contains high concentrations of olefins or aromatics, excessive amounts of organic matter will enter t h e acid phase and cause
A B
y
Potassium
Table V. Effect of Salt Concentration on Nitrogen Recovery"
PROCEDURE
Table VI.
Purification of Sulfate
, ,
... ...
,
...
19.3 1. 9 17
...
...
48 1 84 25 0.5 11,6,11.7
...
1.8
...
12
i i ;i 3
... ...
...
53 47 9 16
... ...
25
... ...
...
...
...
92%
acid
494 83 25 0.5 11.8 19.9 1.8,2.2 15 24 12,12 53 45 8.4 17
ter Meulen Methoda Heavy naphtha A B
86,88 7,7 139, 161
