STUDIES ON THE SYNTHESIS OF CHUJRAMINE AND HYDRAZINE
DISSERTATION
Presented in Partial Fulfillment of the Requirements
for the Degree Doctor of Philosophy in the
Graduate School of The Ohio State
University
By
Bussell Stephen Drago, B. S.
The Ohio State University 195*+
Approved hy:
Adviser
Department of Chemistry i
ACKNOWLEDGEMENT
The author wishes to eaqpress his sincere appreciation to
Dr. Harry H. Sisler for having accepted me as one of his students and for having proposed this problem. His able suggestions and continuing interest were a great aid to the progress of this work and his enthusiasm and pleasant disposition were an inspiration to the author. Thanks are due to many other members of the staff who have in one way or another contributed to the accomplishment of this research. Special thanks are due to Dr. T. Rubin and
Dr. J. Calvert for their suggestions concerning the calculations of the thermodynamic quantities.
The author acknowledges with gratitude the fact that the
Davison Chemical Company has, through a contract with the Ohio State
University Research Foundation, given generous support to the research reported in this dissertation.
The author is greatly indebted to his wife, Ruth Ann, both for typing this dissertation and for her continuous support and encouragement • ii
TABUS OF CONTESTS
Page
Introduction 1
Chapter I Historical
A. Chloramine Synthesis: the Beactlon of Chlorine with Ammonia 3
B. The Baschig Synthesis for Hydrazine 6
C. The Synthesis of Hydrazine in Liquid Ammonia 1 1
D. Preparation and Properties of Anhydrous 13 Hydrazine
Chapter II The Synthesis of Chloramine 20
A. Introduction 20
B. Experimental 20
1. Materials 20
2. Experimental Procedure 21
C. Besults 2U
1. Initial Buns 2^
2. Investigations with a New Ammonia Inlet 25
3. Corrected Flow Calculations 28
Investigations to Test the Flow Bate Theory 32
D. Conclusions 3^
Chapter III The Bole of Caustic, Gelatin and Ammonium Ion in the Aqueous Chloramine-Ammonia Beactlon 37
A. Introduction 37
B. Experimental 33
1. Materials 38
2. Experimental Procedure 38 iii
C. Be suits l+o
1 . The Bole of Caustic k 0
2. The Effect of Anmanium Chloride 52
3. The Effect of Gelatin 5 ^
D. Conclusions 5L
Chapter XV The Effect of Fixed Baseand Gelatin on the 57 Chloramine-Ammonia Beactlon in Liquid Ammonia and in Ethyl Alcohol
A. Introduction 57
B. Experimental 57
1. Materials 57
2. Experimental Procedure 57
C. Besuits 58
Chapter V The Separation of Anhydrous Hydrazine from Ammonium Chloride-Ammonia-Hydrazine Mixtures 6 l
A. Introduction 6 l
B. Experimental 6 l
1. Materials and Equipment 6 l
2. Experimental Procedure 63
C. Besults 70
1. Preliminary Experimente 70
2. Liquid Composition-Vapor Composition Data 72 Obtained vlth the Sodium Hydroxide Coated Apparatus
3. Liquid Composition-Vapor Composition Data 73 Obtained vith the Teflon Coated Apparatus
D. Conclusions 75 iV
g&sa
Chapter VI Vapor Pressure-Composition Studies on 76 the System Ammonia-Hydrazine at Elevated Temperatures
A. Introduction. 76
B. Experimental 76
1. Materials 76
2. Experimental Procedure 76
C. Besuits and Thermodynamic Calculations 87
1. Results 87
2. Thermodynamic Calculations 97
D. Conclusions 119
Chapter VII Summary 120
Bibliography 123
Autobiography 126 STUDIES ON THE SYNTHESIS OF CHLORAMINE ANT HYDRAZINE
Introduction^-
The application of hydrazine both as a rocket propellant and as
a versatile raw material for the synthesis of other hydronitrogen
compounds has given rise to a considerable amount of research on the
synthesis, application and physical properties of this Bubstance. There
is a tremendous amount of patent literature concerning application of
hydrazine but its actual industrial use has been limited by the present
high cost of the material. A more economical process for the synthesis
of hydrazine would undoubtedly bring about an increase in the utiliza
tion of this compound.
At present, hydrazine is produced commercially by methods which 2 entail slight modifications of the original Raschig process. In all
instances the reaction is carried out in an aqueous solution and the
very stable hydrazine hydrate is obtained as the product. An additional expense is encountered when anhydrous hydrazine is desired because it
is necessary to distill the hydrate with barium oxide or caustic, or apply some similarly expensive process to remove the water. The current high cost of hydrazine indicates that this synthesis is an expensive chemical process. 3 Mattair and Sisler were the first to report the synthesis of hydrazine in liquid ammonia by the reaction of chloramine and anhydrous ammonia. Chloramine, the Raschig synthesis intermediate, is produced in the Sisler-Mattair process by the gas phase reaction of chlorine and anhydrous ammonia. Since ammonia does not form a stable solvate with hydrazine it was hoped that this synthesis might be developed into an economical process for producing anhydrous hydrazine. Further k 5 studies have been carried out in this laboratory by Boatman , Neth , 6 7 8 Hurley , She liman , and Kelmers .
These investigations have been extended in the present study and the results will be presented in this dissertation. Conditions leading to almost quantitative yields of chloramine by the gas phase reaction of ammonia and chlorine have been discovered. The possibility of obtaining anhydrous hydrazine from the chloramine-ammonia reaction mixture by a high pressure distillation has been qualitatively demonstrated. The partial pressures of hydrazine and ammonia in equilibrium with various liquid mixtures of these two substances have been measured at several elevated temperatures. Activities of the components have been calculated for this system. The effects of the Baechig synthesis additives on the hydrazine-forming reaction between chloramine and ammonia have been studied in water, liquid ammonia and alcohol. A discussion of these topics is presented in this dissertation. 3
Chapter I
Historical
A. CSLORAMHig SYNTHESIS: THE REACTION OF CHLORINE WITH AMMONIA
The earliest observations on the reaction of ammonia and chlorine 9 10 were reported by Simon , Donny, and Mareska , Simon noted that explosions occur when chlorine gae 1b bubbled through concentrated eolutione of aqueous ammonia. When ammonia is present in excess of okve the chlorine the products formed/\nitrogen and ammonium chloride.
When an excess of chlorine is reacted with either aqueous ammonia or an aqueous solution of an ammonium salt, nitrogen trichloride is 10 formed. Explosions also occur when gaseous ammonia is bubbled into 11 liquid chlorine. Valentin! reported that ammonia gas burns in chlorine gas when mixed at ordinary temperatures. The reaction is described by the equation:
8NH3+ 3C12---> 61111*0 1 + N2 (1) 12 These results were later confirmed by Schwarz and Striebich . 13 Selivanoff reported that the stoichiometry for the reaction of aqueous ammonia with an excess of chlorine is represented by the equation:
*+WH3 3C12 -— » NCI3+ 3NH^C1 (2) 1^ This reaction was studied in more detail by Noyes and Iyon , who arrived at the following equation for the reaction in aqueous solution from a quantitative study of the products and reactants:
12NH3 + 6 CI2 --- *N2 + NCI3 + 9NHjjCl (3)
These studies vere performed at an ammonia to chlorine mole ratio of two to one. Bray and Dowell postulated a mechanism for the ammonia-chlorine reaction In aqueouB solution Based on the formation of chloramine as an Intermediate. They stated that chloramine behaves as the intermediate
in the formation of nitrogen from the reaction in basic solution, and also as the intermediate in the formation of nitrogen trichloride in acid
solution. Theee reactions were represented as proceeding according to the following equations:
6NH3 + 6 CI2 ) 6nh2ci + 6HC1 ( M
3MH2 C 1 * Ng +N H 4 CI+ 2HC1 (5)
3NH2 C1 + NH^Cl --- > 3NH3 + NC13 + HC1 (6 )
Reaction (k), the very rapid primary reaction, is followed almost quantitatively by reactions (5) or (6) or both, depending upon conditions. Noyee accepted thiB mechanism and pointed out that the summation of reactions (10, (5)> and (6) results in the equation:
3NH3 + 6 CI2 ---» N2 + NC13 + 9HC1 lU which is similar to reaction (3 ) that he had reported.
Dowell and Bray ^ reported the following reaction between nitrogen trichloride and aqueous ammonia:
1«H3+ NC1 3 --- > N2 + 3WH^C1 (7)
Combination of this equation with (3) yields:
8MH-+ 3 CI3 ---> 6 NHkCl+ N2 5 11 which is equation (l) reported by Valentini for the complete oxidation of ammonia in the gas phase. 17 Noyes and Haw studied the reaction between chlorine and ammonia in the gas phase, as well as in carbon tetrachloride, chloroform and pentane solutions. In these experiments mole ratios ranging from 1.5 up to 5 moles of ammonia to 4 moles of chlorine were employed. They represented the reaction in these concentration ranges by the equation:
Unh3 + 3C12 --- > MH^Cl + NCl^ (8)
Chloramine and dlchloraaine were postulated as intermediate products.
These authors also found that chloramine and dichloramine are produced in the reaction of chlorine with ammonium salts. O The synthesis of chloramine was first reported by Raschig .
^.ualitative tests indicated that chloramine is formed from the reaction of equimolar quantities of aqueous ammonia and sodium hypochlorite.
Several reactions which chloramine undergoes are reported in this reference. 1.8 Markwald and Wille report the preparation and isolation of chloramine. The pure material was prepared by drawing the vapors from an aqueous solution of sodium hypochlorite and ammonia through anhydrous potassium hydroxide at water aspirator pressures. A substance was o ° obtained which upon removal of ammonia at -70 C solidified at -60 C, sometimes to a crystalline form and other times to a glass. The
Q substance exploded at -50 C forming nitrogen, chlorine and ammonium chloride. The material was analyzed by dissolving it in ethyl alcohol and determining the nitrogen and chlorine content. The analysis con firmed the presence of chloramine. 3 Sislor and Mattair prepared chloramine by the gas phase reaction of chlorine and ammonia.
2NH3 + Cl2 >MH2 C1 + NH4 CI (9)
The reaction is carried out by a flow process using a large excess of ammonia and diluting the chlorine stream with nitrogen. They reported that the reaction proceeds very rapidly and that the yield of chloramine based upon chlorine is nearly quantitative. Study of this reaction was continued by Neth , who reported improved yields of chloramine at lower ammonia to chlorine mole ratios(83.5$ NHgCl yield at an ammonia-nitrogen- chlorine mole ratio of 5.6/3 .Vi) • The data obtained showed a con siderable amount of "scatter" indicating that the variables affecting chloramine yield were not entirely understood.
B. RASCHIG SYNTHESIS FOR HYDRAZINE 2 Raschig discovered that hydrazine could be prepared by the action of sodium hypochlorite upon an excess of ammonia in strongly basic, aqueous solution. This synthesis, with slight modifications, represents the most important,current, commercial method for making hydrazine. The process occurs in the two steps represented by the following equations:
NH3 + OCl" > NHgCl + OH' (10)
NHgCl + NH3 + OH“ ---» +Cl“ + H20 (ll)
Reaction (10) proceeds very rapidly. Maximum yields of chloramine are produced in cold solutionsbelow 10°G. The second step (ll) which results in the formation of hydrazine, is brought about by the action of an excess of ammonia upon chloramine. Raschig^ reported that this reaction is very slow at room temperature and must be carried out at elevated temperatures to produce good yields of hydrazine. Attempts by Raschig to carry out this reaction at ordinary temperatures resulted in the production of very little or no hydrazine. Other attempts to perform the chloramine-ammonia reaction at ordinary temperatures using extremely large excesses of ammonia have also resulted in yields considerably 19 20 below theoretical . Nagasawa reportedjon the other hand^that the yield of hydrazine shows little dependence upon the temperature.
In addition to reaction (ll), chloramine reacts with hydrazine in aqueous solution to yield nitrogen.
2NH2 C1 + > 2NHUC1 + (12)
This reaction explains the fact that the yield of hydrazine ie lees than quantitative. Moeller2'*' and BodenBtein22 reported that this reaction
(12) Is very sensitive to the action of certain catalysts, especially cupric ion. These authors reported that ordinary tap water contains enough dissolved cupric ion (lppm) to prevent hydrazine formation unless an inhibitor is employed. The addition of glue or gelatin as an inhibitor brings about a considerable improvement in yield. Moeller claims that inhibitor ie not necessary if extreme precautions are taken to purify the reagents. However, hia results indicate that he was unable to purify hie reagents completely because the addition of gelatin doubled his yield of hydrazine.
The tremendous interest in hydrazine has resulted in a very thorough investigation of the factors affecting the yield of hydrazine in the
Raschig synthesis. The effect of the mole ratio of ammonia to hypochlor ite upon the yield of hydrazine was studied by Joyner 2^ and by Thompson and Joy2^. A plot of the mole ratio against per cent yield hydrazine gives a curve that is exponential in character. Little hydrazine is obtained below ammonia to hypochlorite mole ratios of 5 to 1. The curve also indicates that little is to be gained by using mole ratios above 30 to 1.
A very extensive investigation of the effect of various inhibitors 8 23 upon the decomposition reaction was carried out by Joyner . He reported that gelatin, glue and peptone are equally effective at similar concentrations. Starch, dextrin, sucrose, colloidal stannic acid, silicic acid, animal charcoal, wood charcoal and asbestos powder were found to exert a lesser beneficial effect. Urea, saccharin, sodium azlde, sodium oleate, sodium palmitate, and lithium chloride did not inhibit metal ion catalysis of reaction (12). Joyner also studied the effect of tyrosine, uric acid, tryptophane and sodium salts of glutamic acid. In spite of the ability of these substances to complex metallic ions their effect as inhibitors is negligible. Pfeiffer and 25 Simons found Trilon B (sodium salt of ethylene diamine tetracetic acid) to be effective as an inhibitor but Trilon A was ineffective.
A number of investigators have studied the effect of glue and gelatin 21 23,2k concentrations upon hydrazine yields * * Excesses of gelatin over the 3mall amount necessary to complex the impurities do not affect the yield of hydrazine.
The role of fixed base and the effect of fixed base concentration upon hydrazine yield was not quantitatively studied by the early investigators. The difficulty encountered in evaluating the amount of sodium hydroxide present in sodium hypochlorite solutions prevented a quantitative study. Qualitative reports on the effects o' caustic 23 2 were published by Joyner and Baschig . Baschig reported that the reaction of chloramine with ammonia was also in competition with a decomposition reaction of the chloramine by hydroxide. The reaction of chloramine with caustic was represented by the following equation:
31^01 + 30H“--- » HH3 + N2 + 3 ^ 0 + 3C1” (13) McCoy studied this reaction In hot, etrongly alkaline solutions and confirmed the above equation. He reported that the reaction proceeds in the following etepe:
NH Cl + OH' >NH OH + Cl" (lh) 2 2 NHgCl + 2NH2OH + OH" » HH3 + N2 + 31^0 + Cl“ (15) 23 Joyner , in hie studies of the effect of added salts upon the
Raschig synthesis, noticed that the effect of ammonium ion was noticeable when the quantity of added ammonium salt was sufficient to react with all the fixed baee. The addition of sodium chloride and potassium chloride had no effect upon the hydrazine yields. 1 Most of the early work was of a qualitative nature and no clear insight into the mechaniem of inhibitors or the effect of fixed base on this synthesis had been obtained. In the past year two attempts have been made to explain the mechanism of the Raschig 27 synthesis. Cahn and Powell report that they have succeeded in correlating existing yield data for the Raschig synthesis by assuming simple rate laws for the reactions involved. They assume that the chloramine-ammonia reaction (ll) and the chloramine-hydrazine reaction
(12) are bimolecular, and propose the following mechanism.
NH + 0C1~--- > NH Cl + OH" fast (l6) 3 2 OH" NH Cl + NH.-- » (N HcCl)--- * N H. + Cl" + H O (17) 2 3 2 -> 20H” 2 2 - NH^Cl + NgHjj > (N^Cl)Nl^Cl*— * Ng + 2NH3 + 2C1+2H20
(18)
It is stated that the nature of the intermediate in equation (18) is not known. Preliminary experiments with ieotopically labeled hydra zine indicate that triazane (N-H) probably is not formed. i 5 10
Kinetic expressions were derived, which fit the existing
information on the effect of gelatin, by assuming that gelatin
complexes cupric ion or similar catalytic impurities. The authors 23 interpret the very large decrease in yield obtained by Joyner when ammonium ion was added in excess of fixed base, as being due
to the instability of the copper-gelatin complex below a pH of 11. 28 Cahn and Powell point out that there is polarographic evidence that
the copper-gelatin complex is unstable at lower pH's. At pH 9=55
gelatin acts as a normal maximum suppressor for cupric ion; at pH
lU gelatin shifts the half wave potential by about one volt.
Audrieth et a l ^ report the successful quantitative study of the effect of caustic on the hydrazine-forming reaction by using
t-butyl hypochlorite aB the oxidant instead of sodium hypochlorite.
Very pure t-butyl hypochlorite free from the impurities usually
found in sodium hypochlorite was prepared. These authors emphasize
the necessity of fixed base in the hydrazine-forming reaction to realize
substantial yields of hydrazine. They believe that the active
intermediate in the hydrazine synthesis is the chloramide ion, NHC1 , and propose the following mechanism for the Baschig synthesis and
similar reactions:
NH3 + 0Ci“ => NHgCl + OH" (19)
NHgCl + OH"----- > HHCl"+ Hg0 (20)
NHCl" + B ---- ^HKB + Cl" J2l) where B = NH^, RNHg, RgNH, Hg0 or R0H(?)
This proposed mechanism is similar to the well known Hoffman re arrangement . 11 so A recent publication by Riley et al reports that hydrazine
la not the intermediate in the nitrogen-producing reaction that
occurs when ammonia Is oxidized by hypochlorite. These conclusions
were arrived at by oxidizing under various conditions, ammonia 1*5 containing N v in the presence of equimolar amounts of hydrazine
and analyzing the hydrazine remaining at the end of the reaction for 15 1 *5 N . Insignificant amounts of N y were found in the hydrazine. 31 Sisler et al hare reported that chloramine, produced by the
gas phase reaction of ammonia and chlorine, reacts with excess ammonia
in aqueous solution in the absence of fixed alkalies or any other
additives to give high yields of hydrazine (above 80$). The authors emphasize the necessity of using solutions and reagents which are free
of impurities which catalyze the yield-reducing reactions.
C. THB SYNTHESIS OF HYDRAZINE IN LIQUID AMMONIA
The final product obtained from the Raschig synthesis is present
in dilute aqueous solution. Since a very stable hydrate is formed by hydrazine, a costly distillation of hydrazine hydrate with barium oxide or some other dehydrating agent Is necessary to produce the anhyd rous material. This difficulty could be avoided if the hydrazine - producing reaction were carried out in an anhydrous solvent which does not form a stable solvate with hydrazine. Liquid ammonia is such 3 a solvent. Sisler and Mattair reported that appreciable yields of hydrazine could be obtained by the reaction of chlorine with anhydrous ammonia. The reaction was carried out in three different ways:
a. Gaseous chlorine was reacted with a large excess of liquid ammonia. 12
b. A eolation of chlorine In carbon tetrachloride was added to
a large excess of liquid ammonia.
c. Gaseous chlorine, diluted with nitrogen, was reacted with
large excesses of gaseous ammonia to form gaseous chloramine.
Cl2(g) + 2HH3(g)----» NHgClCg) + NH^Cl(s) (22)
This reaction occurs instantly and the chloramine yields are very high. The ammonium chloride produced by the gas phase reaction was filtered out and the gases were condensed into a large excess of liquid ammonia where the chloramine reacts with ammonia to form hydrazine in accordance with the following equation:
NHgCl + 2NH3 ^N2H^+ NH^Cl (23)
It is believed that this reaction is accompanied by the yield- decreasing side reaction:
2NH2C1 + NgH^ * 2NHj^Cl + (2M)
Considerable excesses of ammonia were reported desirable in all three processes. Process c. produced the highest yields of hydrazine. The gas phase reaction was carried out at room temperature and the liquid phase reactions from -75 to -80°C.
Sisler and Mattair reported further that when the ammonia is allowed to evaporate from the ammonia-hydrazine-ammonium chloride reaction product, a solid residue consisting of hydrazine hydrochloride and ammonium chloride is obtained. Ammonium chloride is always present in excess because of/^additional amount formed from the hydrazine- destroying reaction (210. Hydrazine hydrochloride forme even though the hydrazine is a weaker base than ammonia because the following equilibrium is shifted to the right by the boiling away of the ammonia: 13
»2HU (1) + NH^Cl(s) >N2H5Cl(e) + MS^g) (25)
The shift occurs because of the low volatility of hydrazine compared
to ammonia.
Additional studies on the chloramine-ammonia reaction in liquid 32 ammonia were reported by Sisler et al . Increases in yield at constant
ammonia to chloramine mole ratios were reported as the reaction temp
erature was increased. Increases in hydrazine yield with decreases In
chloramine concentrations were also observed at each of the temperatures
studied. The authors also reported that the presence of hydrazine in
the liquid ammonia at the start of the reaction greatly lowers the
percentage of the chloramine converted to hydrazine. This supports
the proposal that the yield reducing side reaction is:
2NH2C1 ' NgHj^— »N2 + 2NH1+C1 (2k)
Ammonium chloride also reduces the yield of hydrazine from the
chloramine-ammonia reaction at higher temperatures.
Contrary to the beneficial effect of caustic in aqueous solution, 7 Sheliman reported that the addition of potassium metal, sodium metal,
sodium amide, or potassium amide to the liquid ammonia prior to the
chloramine addition results in little or no hydrazine being produced.
A study of the chloramine-ammonia reaction in certain other 31 anhydrous solvents was also reported by Sisler et al
D. PREPARATION AND PROPERTIES OF ANHYDROUS HYDRAZINE 33,31*, 35, Lobry Ae Bruyn first prepared anhydrous hydrazine by the reaction of sodium msthylate with hydrazine hydrochloride in absolute methanol. CH OH RaOCH3 + NgHp.Cl »NaCl + CH^OH + NgH^ (26) Ill
The sodium chloride is only slightly soluble in methanol and most of
it can be removed by filtration leaving a solution of hydrazine
in methyl alcohol. Removal of the alcohol by distillation leaves
essentially pure hydrazine.
Since the only commercial methods for producing hydrazine
involve reaction in aqueoue media, most of the published research
has been concerned with means of concentrating the hydrazine from dilute aqueous solutions. The general methods by which anhydrous hydrazine
can be obtained involve either dehydration of the dilute solutions or
precipitation of the hydrazine in the form of an inorganic or organic 2^,36 derivative. The most common commercial method for concentrating
the solution is distillation. Theoretically, water can be removed until a constant boiling solution containing 58*1? mole per cent hydrazine
is obtained. However, in actual practice it is most economical to
concentrate to the 85 per cent hydrazine hydrate stage. When anhydrous hydrazine is desired the 85 per cent material is dehydrated with sodium hydroxide. In this dehydration, the hydrate is refluxed for several hours with an equal weight of caustic. The hydrazine is then removed by dis- o tillation at atmospheric pressure. Temperatures ranging up to 150 C are needed to produce nearly theoretical yields of the final product, which 37 still contains five to ten per cent water . Anhydrous hydrazine can be
obtained by distillation of this product from barium oxide or Bodium 35,38,39,^0 hydroxide under reduced pressures
Another method, used to produce anhydrous hydrazine commercially, 1*1,1*2 involves the ammonolysis of hydrazine sulfate . Hydrazine sulfate
is only slightly soluble in water and can be precipitated from the dilute 15
Raschig liquor by the addition of sulfuric acid. The precipitate is
then filtered out and added to liquid ammonia where ammonolysis occurs
in accordance with the following equation:
NgHif^SOj, + 2NH^ ?N2Hi^ + (HH^JgSO^ (27)
The ammonium sulfate is insoluble in liquid ammonia and may be removed
by filtration, leaving a solution of hydrazine in liquid ammonia.
The ammonia ie distilled off leaving anhydrous hydrazine behind.
Other methods for concentrating hydrazine from dilute aqueous
solutions are listed below: 1+3, (a) Removal of water by freezing
(b) Dehydration with anhydrous sodium sulfate
(c) Liquid-liquid extraction methods Ho. These methods depend
upon the formation of ketazine by reaction of hydrazine with acetone and
the extraction of ketazine with such solvents as ether, benzene, hexane
and carbon tetrachloride. Water-immiscible aldehydes and ketones have 45 been used to extract hydrazine . These substances serve to convert
hydrazine into azines and also act as extractive solvents. 46 (d) Azeotropic distillation 21,47,48,49 (e) Precipitation as an azine 50 (f) Precipitation as an insoluble double salt 51 Furthur dehydration of concentrated hydrazine by using calcium , 52 53 sodium , and metallic amides has been attempted in an effort to
produce the anhydrous material. These methods were successful but metallic calcium and metallic sodium when present in excess of the water react with hydrazine to form extremely explosive metallic 53 hydrazldee. Stolle aleo found that anhydrous hydrazine could be 16
distilled in a vacuum from very concentrated hydrazine provided a
quantity of eodamide insufficient to react with all the water is
added•
NaNHg + NgH^.HgO ---> NaOH + NH^ + (28)
When excess of eodamide is used explosions occurred as soon as the o temperature was raised to 70 C.
As illustrated in equation (25),
N2H^ + NH^Cl -- > N2H5C1 + NH a separation problem is also encountered when hydrazine ie pro duced in liquid ammonia by the Sisler-Mattair process. Upon evaporation of the solution containing the final products, a mixture of ammonium chloride and hydrazine hydrochloride is obtained. Various procedures for producing anhydrous hydrazine from these salts were 5^ investigated and are summarized below:
(a) Separation with Sodium in Liquid Ammonia
Treatment of hydrazine hydrochloride and ammonium chloride mixtures with an equivalent amount of sodium dissolved in liquid ammonia removes ammonium chloride by the following reaction:
Na + NH^ + Cl > NaCl(s) + NH^ + ^ ( g ) (29)
The precipitated sodium chloride can be filtered out and the mixture of hydrazine and ammonia plus some dissolved sodium chloride separated by distillation. The use of excess sodium must be avoided to prevent the formation of explosive sodium hydrazide.
(b) Separation with Sodium Methylate
Yields obtained from this process were greatly improved, 33 over those originally reported by de Bruyn , by treating the methanol 17 solution of hydrazine hydrochloride with ammonia to set free the hydrazine prior to the addition of sodium methylate.
N H + + NH ----» N H + N K (30) 2 3 2 h U VJ NHjj + Cl"+ Na+ + 0CH3^--- > NH^ + CH^H + NaCl(s)
(31)
(c) Separation with Alkalies if Boatman showed that anhydrous hydrazine can he obtained by distillation of a dry mixture of hydrazine hydrochloride and ammonium chloi'ide to which dry eodium hydroxide has been added. Anhy drous hydrazine is obtained when sufficient anhydrous base ie used to both displace the ammonia and hydrazine and to react with the water produced in the reaction. Addition of lesser amounts of base result in the formation of hydrazine hydrate.
(d) Separation by Fractional Crystallization
The ternary phase diagram for the system ammonia-hydrazine- ammonium chloride has been studied at low temperatures and atmospheric 6 pressure . A solid phase containing ammonium chloride can be crystallized from ternary mixtures in a certain portion of the system. This results in a corresponding enrichment of the solution in hydrazine with respect to ammonium chloride. Ifter removal of the ammonia, the hydrazine in excess of the ammonium chloride present can be distilled off ae the anhydrous material. The hydrazine hydrochloride residue which remains can then be redissolved in liquid ammonia and subjected to furthur fractional crystallization.
(e) Separation with Cyclohexylamina. 55 A process has been patented which involves the use of 18 high boiling amines to liberate anhydrous hydrazine from its salts.
The hydrazine being less volatile can be distilled off, leaving the amine hydrochloride as the residue. Cyclohexylamine has been used successfully for this purpose.
(f) Separation by Solvent Extraction
Since ammonia is a stronger base than hydrazine, extraction of hydrazine from solid hydrazine hydrochloride and from mixtures of hydrazine hydrochloride and ammonium chloride by solutions of anmonia in anhydrous solvents is possible . The ammonium chloride must be insoluble (or of low solubility) in the extracting solution and thus capable of being filtered off. Anhydrous hydrazine can then be obtained by distillation of the solvent. Promising results were obtained when diethyl ether and ethyl cellosolve were used as solvents.
Many thermodynamic properties for anhydrous hydrazine have 56 been measured and are summarized in the literature . The vapor pressure of liquid hydrazine has been studied by several investigators 0 O ( O O 58 in the temperature ranges 0 to 70 C , 20.2 to 11^4.1 C , and o o 35 59 lUO to 380 C . Measurement of the vapor density of hydrazine o 0 over the temperature range 90 to 130 C indicates that there is no evidence of association in the vapor state. This finding is sub- 60 stantiated by an investigation of the infrared spectrum of gaseous hydrazine, which shows no indication of any extensive hydrogen bonding.
The value of Trouton's constant for hydrazine is 25.23 cal/mole- degree. This high value (21 for an unassociated liquid) is an indication of association in the liquid state. 19
Hydraxine le an endothermic compound that undergoes decomposition
under appropriate conditions vith the release of a considerable amount
of energy. Information on the explosiwe properties of hydrazine is 6 1 listed in Table I
TABLE I
Minimum Ignition Temperatures of Hydraslne
Surface Atmosphere Ignition o Test Sample Temperature C CO.
Pyrex glass Air 2 7 0 .0 1
Pyrex glass Oxygen .0 2 ro 0 Platinum Air 226 e
Platinum Oxygen 30 .0 2
Ferric oxide Air 23 .15
Ferric oxide Nitrogen* 23 .15
Black iron Air 132 .05
Black iron Nitrogen 131 .06
Stainless steel Air 1 6 0 .1 0
Stainless steel Air 1 5 6 .1 0
Stainless steel Nitrogen None at U15
♦Flame ignition after contact with air. Chapter II
The Synthesis of Chloramine
A. INTRODUCTION 3 Earlier studies on the gas phase reaction of ammonia and chlorine had indicated that the clogging of the chlorine inlet tube with ammonium chloride was one of the major problems to be oTercome in this synthesis. A ramrod technique in which the inlet tube was punched clear at short intervals was first used to prevent 5 plugging. Later , a device which involved electrostatic precipitation of the ammonium chloride was found to be successful. However, it was hoped that the plugging problem might be more conveniently- solved by change in the design of the chlorine Jet.
It was also reported that attempts to correlate chloramine yield with other variables in an effort to determine the critical factors affecting this reaction was not entirely successful. This indicated that there was not a complete understanding of the variables involved in this process.
Addition of nitrogen as a diluent to the chlorine stream had been found to improve the yield of chloramine. It was hoped that the beneficial effects of nitrogen dilution could be obtained by improved reactor design and thus eliminate the problem of handling large quanti ties of a non-condensable gas in a plant operation. The studies reported in this chapter were carried out with these objectives in mind.
B. gXPERIMEBTTAL
1. Materials
Anhydrous synthetic-grade ammonia manufactured by the Verkamp
Corporation was used in these experiments. It was removed directly from 21 the container as needed. Chlorine, obtained from The Ohio Chemical and Surgical Equipment Company vae distilled from its original container into a stainless steel tank. The chlorine vas thep removed from this stainless steel tank as required. Oil- pumped nitrogen manufactured by Linde Air Products vas used in this study.
2. Experimental Procedure
Narrow, high-velocity jets were designed for introducing chlorine into an ammonia atmosphere in such a manner as to favor rapid mixing of chlorine with a large excess of ammonia. Spinner- ettes, constructed of 90$ platinum and 10$ rhodium were obtained from the Baker Company for this purpose. Three different sizes were obtained with hyperbolic openings of .0018 inch, .0031 inch and .001*3 inch diameters, respectively. It was also hoped that use of the splnnerettes vould minimize the ammonium chloride deposits in the vicinity of the chlorine inlet.
The designs of the splnnerettes and the apparatus used in this study are Illustrated in Figure 1. A glass wool plug located at the end of the reactor tube opposite the chlorine inlet (not illustrated) served to filter out of the gas stream the ammonium chloride produced in the gas phase reaction. The emergent gas stream vas condensed in a series of cold traps kept at dry ice and liquid air temperatures. A water bubbler at the end of the trap system collects the very small amount of chloramlne that was not condensed in the cold traps.
Prior to the start of the reaction, the entire system vas flushed with nitrogen introduced through a reducing valve into the Figure l
Ammonia inlet Pyrex glass reactor tube To Nitrogen Tank V [i»S Class tube Valve Spinnerette
Chlorine inlet Hex. Nut, fjrbrass Tube, ~ brass ' bar stock
Rubber stopper
Spinnerette Detail
Small dia hole
Side View Front View chlorine Inlet. The ammonia flow was then started, and about ten minutes later the cooling baths were placed around the traps. The ammonia flow was measured by a calibrated flowmeter of the differential manometer type. The valve leading to the nitrogen tank was closed
(see Fig. l) and the valve on the chlorine tank was immediately opened, initiating the reaction. The initial temperature of the reactor was o usually about 25 C, but it becomes hot in the vicinity of the reaction zone as the reaction proceeds. After the reaction was stopped, the reactor was flushed with ammonia and nitrogen. The chloramine-ammonia condensate was allowed to stand a sufficient period of time for the chloramine to react completely with ammonia to form ammonium chloride.
The excees liquid ammonia was then allowed to evaporate, leaving in the residue chloride equivalent to the original chloramine. The weight of chlorine converted to chloride in the gas phase reaction is obtained by washing the reactor walls and the glass wool plug with water and analyzing the resulting solution by the Volhard procedure.
The percentage yield of chloramine may be calculated by the equation:
Per cent Yield NH Cl= a-b x 100 (l) 2 172. where a = total weight of chlorine entering the reactor
b - weight of chlorine retained in the reactor as ammonium chloride.
This formula arises from the fact that if the reaction produces a chloramine yield of 100# in accordance with the equation:
Cl2 (g) + 2NH3 (g) > NH^Clfg) + NH^Cl(e) (2) fifty per cent of the chlorine remains in the reactor tube as ammonium chloride. If the reaction proceeds completely to yield nitrogen in 2k accordance vith the equation:
8NH3 (g) + 3C12(S)--- > N2 (g) + 6NHuC1(b) (3) the yield of chloramine Is 0 per cent and all the chlorine le retained In the reactor. The total weight of chlorine used Is obtained from the sum of the chloride in the reactor plus that pro duced by decomposition of the chloramine after standing in the traps.
C. RESULTS
1, Initial Experiments
In these experiments the straight glass tube, illustrated in
Figure 1, vas used as the ammonia inlet. A series of experiments of fifteen minutes duration, without nitrogen dilution, employing the
.0031 inch diameter spinnerette were carried out. The results shown in Table II were obtained.
TABLE II
MoJLe Ratio of Ammonia Per Cent Yield to Chlorine Chloramine
7.8 to 1 36.2
8.8 to 1 31.8
20 to 1 2 9 .O
21 to 1 25.2
22 to 1 21.0
1*3 to 1 91.3
28 to 1 86.5 25 The following experimentb were made with the .0018 inch
spinnerette:
TABUS III
Mole Ratio of Ammonia Per Cent Yield to Chlorine Chloramine 2 l A to 1 78.6
1 6 . to l 39.3
An experiment was also attempted with the .00^3 inch spinnerette.
An 11.2 per cent yield was obtained at a U.lf to 1 ammonia-to-chlorine
ratio.
In all of the above reported runs which resulted in low yields,
the reactor tube became very hot in the vicinity of the reaction zone.
Ammonium chloride formed at the chlorine inlet in all cases.
2. Investigations with a Hew Ammonia Inlet
It was thought that the low chloramine yields and the scatter in
the data reported above might be due to improper mixing caused by the
introduction of all the ammonia into the reactor at one point. A new ammonia inlet was designed to provide a more uniform flow of ammonia around the chlorine Jet (see Figure 2). The following series of fifteen minute runs were carried out: 26
TABLE IV
Spinnerette Inlet Mole Ratio of Per Cent Yield Sixe (inches) Ammonia to Chlorine Chloramine
.0018 39.^ to 1 94.1
.0018 30.1 to 1 42.9
.0018 21.9 to 1 jk.2
.0018 18.0 to l 75.0
.0018 9.1 to 1 5.0
.0031 32-8 to 1 94.0
.0031 23.6 to l 90.9
.0031 21.4 to 1 92.5
.0031 13.8 to 1 64.9
.0031 7-8 to 1 10.2
The yields obtained from the experiments with the .0031 inch diameter spinnerette show the improvement made with the new ammonia inlet. In all of these experiments there were formations of ammonium chloride on the chlorine inlet tip. The experiments that produced good yields all had formations which consist of several hollow cones with their apices on the chlorine tip. The flow from the new ammonia inlet kept the reactor tube clean in back of the chlorine inlet so that growth of these formations could be observed. In those experiments in which low yields were obtained, a yellow flame could be noticed at the chlorine tip. The appearance of this flams coincided with an Increase in nitrogen evolution as observed in a water filled absorption tower at the end of the trap system. The time required Figure 2
Ammonia inlet Pyrex glass reactor tube To Nitrogen z..." , , .. ! Tank mm ^ Pyrex glass tube Valve i Z Spinnerette i m u \\\W /<■/.' / 7 7 t ]T ^1OQ z z z/ /, y "v / TZ. Chlorine inlet " I ' l l mill \ Hex Nut, I brass Tube, £ brass ii
^Rubber stopper . 28 for the flame to appear also could be roughly correlated vlth the yield of chloramine obtained. Runs were observed In which a tip Indicative of good yield began to form and then "burned off" ae poor conditions set In.
The poor yields obtained In some cases with the .0018 Inch diameter spinnerette may be caused by the same effect that was observed when the ammonia was introduced on one side of the reactor.
The small spinnerette delivers about one-third the chlorine flow of the medium sized one; therefore, at a given mole ratio lower ammonia flows are also employed with the smaller spinnerette. With these low flows
It is probable that we have lees uniform motion down the tube. The low flows, the volume decrease resulting from reaction, and the heat of the reaction could produce a turbulent effect that causes the chloramine to be carried back Into the reaction zone and thus reintroduced into the chlorine stream. This then results In further oxidation of the chloramine to nitrogen which in turn causes a further volume decrease probably In accordance with the equation:
6NH^ + NBLC1 + 2C1---- > N0 + 5WH. Cl (4) 3 2 2 2 h These circumstances lead to a "physical autocatalytlc effect" and explain the very sharp decreases obtained once turbulent conditions are realized. Figure 3 ia a plot of the per cent yield of chloramine versus the ammonia to chloramine mole ratio for the results obtained
In these studies.
3. Corrected Flow Calculations
If our above hypothesis Is correct, the total flow should be a more important variable than mole ratio in determining yields of % Yield -100 oe t NHjd j H N ato R Mole iue 3 Figure
2 <3 □ J i i i I i I i i _J 26 26 26 26 03"da Sinrte l N3inlet NH3 old dia. Spinnerette, .003l" .0043"dia. Spinnerette, new NH3inlet Spinnerette, .0043"dia. .0018“dia. Spinnerette, newNH3inlet Spinnerette, .0018“dia. dia. new Spinnerette, NH3inlet .0031 .0031“dia. Spinnerette, new Spinnerette, NH3inlet, .0031“dia. yoemc needle Hypodermic LEGEND 0 2 4 6 8 40 38 36 34 32 30
N2addttion j
VO ro 30
chloramine. The total Inlet flows of the gases were corrected for the
volume decrease caused "by the formation of chloramine and nitrogen
according to the reactions:
2NH3(g) + Cl2 (g)--- >NH2Cl(g) + RHjjCl(s) (2)
3Cl2(g) + 8NH3(g)-- ^ ( g ) + 6WHuCl(s) (3)
In reaction (2) we have a decrease of two moles of gas per mole of
chlorine reacted; on the other hand, (3) corresponds to a decrease
of three and one-third moles of gas per mole of chlorine reacted.
The 65 per cent chloramine yield obtained with the .0031 inch diameter
spinnerette at an ammonia to chlorine mole ratio of 13.8 to 1 will
he used to illustrate a sample calculation:
Onflow - .583 molea/hr.
MH^flow = 8.05 moles/hr.
Total flow = 8.63 moles/hr.
A 100 per cent yield of chloramine would cause a .583 x 2 - 1.166
mole/hr. reduction in gas flow.
A 100 per cent decomposition reaction would cause a .583 x
3 l/3 - 1 .9^ mole/hr reduction in gas flow.
A 65 per cent yield would therefore cause a total reduction
in gas flow of: 1.17 x .65 + 1 .9^ x .35 = l.M* moles of gas per hour.
The corrected total flow is therefore 8.63 - 1.W*- = 7.2 mo lee/hr.
The results of these calculations are illustrated in Figure 4.
All these experimente were performed with the same diameter reaction
tube. Calculations have shown that the flow effect causing poor yields with the aaell spinnerette also correspond to yield-reducing effects with the .0031 inch diameter spinnerette. This indicates the possibility % Yield Chloramine 100 20 30 40 60 90 50 80 70 10 (9 to I) to (9 (18toI) (22 toI) (7.8 to(7.8 I) oa Fo Rt (orce) moles/HF (corrected) Rate Flow Total o (30 to I) I) to (30 6
(39 to I (39 8 o. n aetee idct N3:I ml ratio mole :CI2 NH3 indicate parentheses in Nos. iue 4 Figure
□ 0043 in. spinnerette spinnerette in. 0043 □ • .0031 in. spinnerette spinnerette in. .0031 • .08 n spinnerette in. .0018 o (21.4to I) 10
LEGEND □ (12.8 to I) 12 1 • (23.6 to I)(23.6 •
14 16 (32.
18 o o W 32 of obtaining good yields of chloramine at even lower mole ratios than these obtained with the .0031 Inch diameter spinnerette, providing sufficiently large total flew rates are employed. k. Investigations to Test the Flow-Rate Theory
Further indications of the Importance of flow conditions can be obtained either by increasing the flow rate by the addition of an inert gas to the ammonia stream or by designing a reactor to eliminate obstructions which produce turbulence in the gas stream. Higher yields at lower mole ratios would also be expected from experiments employing the large spinnerette since the total flow of gases in the reactor would be greater.
An experiment vas attempted with the .00^3 inch diameter spinnerette. v e The apparatus had toAmodified to include a longer reaction tube to handle the large amounts of ammonium chloride formed. The length of the run was also decreased to ten minutes for the same reason. A yield of 87 per cent chloramine was obtained at an ammonia to chloramine mole ratio of 12.8 to 1. This increased yield of chloramine over that obtained with the .0031 inch diameter spinnerette at a similar mole ratio is definite support of the total flow theory.
To further test this hypothesis, two experiments using the .0031 inch spinnerette at the very low ammonia to chlorine mole ratios of
7.9 to 1 and 3.9 to 1 were carried out, but with the addition in each case of sufficient nitrogen to give a total gas flow of thirteen moleB per hour. It should be emphasized that the nitrogen was not added as a diluent to the chlorine stream but was added to the ammonia. The tvo gases were mixed in a mixing chamber and then introduced into the reactor* The apparatus was modified to include
two gas dispersion tubes immersed in water at the end of the trap
system to collect the chloramine carried through the cold traps by the
non-condensable gas. The yields of chloramine obtained were 83 and
68 per cent respectirely. The data in Table IV show that without the added nitrogen to increase the rate of flow, a chloramine yield of only
10.2 per cent was obtained with the .0031 inch spinnerette at a mole ratio of 7.8 to 1. Thus, though these two experiments would not quite fall in line with the maximum yield(9 h per cent) reported in
Table IV, they certainly support our hypothesis that the rate of flow of gases away from the reaction zone is a factor of primary importance in determining chloramine yields.
The differences between the maximum yield reported in Table IV and the yields obtained in theBe experiments may be attributed, in part, to decomposition of chloramine on the walls of the reactor.
This would Indicate that ammonia tends to Inhibit the heterogeneous decomposition of chloramine more effectively than nitrogen does. Pre
ferential absorption of ammonia rather than chloramine on ammonium chloride and glass could be the explanation for the inhibitory effect
of ammonia. Similar decomposition of chloramine on the walls of the trap is observed at low ammonia to chlorine mole ratios.
A new reactor was designed to eliminate the obstructions which probably led to turbulence in the spinnerette reactor. A twenty-six gauge hypodermic needle, silver-soldered to a half-inch iron pipe was used as the chlorine inlet in the new reactor. A common pin was used to perforate holes in a fourty-four millimeter diameter filter thimble which was employed as the ammonia inlet. The perforations symmetrically covered the entire area of the end of the
thimble. The holes vere made in four concentric circles about the
hypodermic needle as a center. A hole vas made every one-fourth of
an inch in the first tvo inner circles and every one-half inch in
the outer circles. The apparatus is illustrated in Figure 5. With
the exception of the reactor tube, which was reduced to a size of
fourty-four millimeters inside diameter, the rest of the apparatus and the experimental procedure is the same as that previously described.
These runs were performed without nitrogen addition. Trouble was encountered in controlling the chlorine flow and as a result only a limited amount of data was obtained.
Ammonia to Chlorine Per Cent Yield Mole to Ratio Chloramine
UO to 1 93.0
6.1 to 1 59-3
2.9 to 1 21.9
The above data have been plotted in Figure 3 for comparison purposes. Definite increases in chloramine yield at low ammonia to chlorine mole ratios have been obtained with this new reactor design.
D. Conclusions
The data obtained illustrate the necessity of removing chloramine from the reaction zone as soon as it is formed in order to realize high yields of chloramine. High yields of chloramine can be obtained in the reactor illustrated in Figure 2, without too much dependence upon the ammonia to chlorine mole ratio as long as a total rate of gas flow greater than about eight moles per hour is employed. Of course in all these experiments a considerable excess of ammonia over the Figure 5
Ammonia inlet Pyrex glass reactor tube To Nitrogen Tank inim Glass tube *C 4 (Mill Hypodermic needle '// // // // // // // Z V / / / / ~zr Chlorine inlet Preforated Filter Thimble
Tubet £iron pipe
V\ RubberD..1 stopper
LO VJ1 36
stoichiometric requirements was present.
The reported runs employing the hypodermic needle as the chlorine
inlet also illustrate that high chloramine yields can he obtained
at very low ammonia to chlorine mole ratios with proper reactor design.
These results also show that the decomposition of chloramine by
chlorine is probably the major yield reducing reaction in this synthesis.
The equation for the complete oxidation of ammonia to nitrogen, resulting
in the appearance of a flame in some cases, has been reported by Valentinl. 17 These studies substantiate the report by Noyes and Haw that chloramine
is the intermediate in the complete gas phase oxidation of ammonia to
nitrogen by chlorine. In order to produce high yields of chloramine,
it must be removed from the reaction zone before further oxidation to
dichloramine, nitrogen trichloride or nitrogen occurs.
These results should encourage the use of chloramine in organic
or inorganic synthesis. Since large excesses of ammonia are not required and since nitrogen dilution is unnecessary, chloramine production should be relatively inexpensive.
( Chapter III
The Bole of Caustic, Gelatin, and .Ammonium Ion In The Aqueous Chloramine-Ammonia Reaction
A. INTRODUCTION
A controversy has arisen in the recent literature concerning the role of caustic and the mechanism of hydrazine formation in the reaction of chloramine with ammonia in aqueous solution. Cahn and 27 Powell postulate the following mechanism as the path for this reaction: (OH") NH Cl + N H > (NgH^Cl) ------> N + Cl + HgQ (l) 23 29 2 Colton, Jones and Audrieth postulate a mechanism based on the Hoffman rearrangement in which the chloramide ion NHCl is the reacting species:
NHgCl + 0H~ ^ NHCl" + H20 (2)
N-H-C1~+ B > H-N-B + Cl’ (3) where B = NH , RNH0, R NH, or Ho0. These authors claim that it is 3 2 2 <=- only in solutions containing fired base that chloramine reacts with ammonia to yield hydrazine, for only in such highly basic solutions ig the formation of appreciable concentrations of the chloramide ion 29 NHCl" possible (pK = 15 ± 2) NH2CI Sisler et a l ^ have reported that high yields of hydrazine are produced in aqueous solution in the absence of fixed base and gelatin.
It was decided that a detailed, quantitative study of the effect of caustic upon hydrazine yield was desirable to aid in establishing a mechanism for this reaction. B. EXPERIMENTAL
1. Materials
The caustic used in this study was C. P. grade. All solutions were made with distilled water and saturated with anhydrous synthetic grade ammonia at room temperature. Knox pure gelatin powder wae used as the
inhibitor. Ammonium chloride of high purity was made by passing very pure ammonia and hydrogen chloride into doubly distilled water. This solution was used in the study of the effect of ammonium ion.
The chloramine used in this study was produced by the gas phase reaction of anmonia and chlorine. This method of producing chloramine was chosen because it has definite advantages compared to methods employ - ing other oxidants for this synthesis. When sodium hypochlorite is used as an oxidant (Raschig process) it ie difficult to evaluate the effect of caust in, since some sodium hydroxide is always present in the sodium hypochlorite solutions. The gas phase synthesis for chloramine also eliminates probable complications to a mechanism study arising from the introduction of t-butyl alcohol into the reacting solution as occurs 29 when t-butyl hypochlorite 1b used as the oxidant . In addition, ammonia is added with the chloramine by this method so that the ammonia con centration in the reacting solution is kept at the saturation point during the reaction.
2. Experimental Procedure
The aqueous solutions of ammonia, caustic, and gelatin were pre pared by heating and stirring gelatin suspensions of different caustic concentrations until the gelatin dissolved. Distilled water was added in each case to give a total volume of seventy milliliters and the solution was then saturated with ammonia. 39
The chloramine-nitrogen-ammonia gas stream vas passed into
Jointed traps containing the above solutions. The solutions vere
kept in the temperature range of 21 to 2k°C by a cold vater bath vhile the chloramine vas being added. The total amount of chloramine added vas varied in the different experiments by changing the length
of time that the gaseous mixture containing chloramine vas passed into the solution.
After the chloramine addition, the reaction mixture was alloved to stand at room temperature for thirty minutes and vas then diluted to 250 milliliters vith boiled distilled vater. Aliquots vere taken 62 and analyzed for hydrazine by the acid-iodate method and for chloride 63 ion by the modified Volhard procedure . The amount of chloramine used in any experiment i8 equivalent to the amount of chloride found, for all of the chloramine is finally converted to ammonium chloride. The calculation of the yield of hydrazine is based upon the amount of chloramine added to the liquid. The ammonia concentration corresponding to saturated solutions of ammonia in caustic solutions of various con centrations vere determined by saturating standard caustic solutions vith ammonia at 21°C. After saturation the ammonia flov vas stopped and the solution cooled to prevent the loss of ammonia. The cold solution vas added to a volumetric flask vhlch contained standard acid and an analysis for total basicity vas performed. The total ammonia vas obtained by subtracting the known amount of caustic from the total basicity. The following results vere obtained for the solubility of ammonia in caustic solutions under the conditions of our experiment: TABLE V
Cauetic Concentration Solubility of Ammonia mo lee/liter gas/lOO gms of ^ 0
U2.8
.5020 i«-3.0
.5020 1^3.2
2.1*56 38.1
U .912 26.8
5.010 2 7 .O
The solubility of ammonia in the various cauetic solutions studied was obtained from a plot of the above values (Figure 6). The mole ratio of ammonia to chloramine in the various cauetic solutions vas obtained by using these values for the solubility of ammonia.
C. RESULTS
1. The Bole of Caustic
The data obtained from studies of the effect of an excess of caustic are listed in Table VI. In these experiments there vas at least enough cauetic present to neutralize the acid produced from the reactions of chloramine:
NH^Cl + NH3 + OH" » + Cl"+ Hg0 (k)
2HHgCl + lfgH^ + 20H” >Ng + + 2 ^ 0 + 2Cl“ (5) Solubility of Ammonia grams/lOO grams of water 40 30 10 0 0 oaiy f h Cutc Solution Caustic the of Molarity 2 iue 6 Figure 5 3 4 6 ■r 1*2
TABLE VI
Experi Initial Gelatin Mole Ratio Chlora Per Cent ment # Cauetic Concentra Ammonia to mine Yield Molarity tion gme/l Chloramine Molarity Hydra of water zine
1 0.357 8.6 65.1* to 1 0.379 73.3
2 0.357 7.0 225 to 1 0.110 82.3
3 0.357 7.0 136 to 1 0.182 80.7
5 0.357 8.6 126 to 1 0.196 82.3
7 0.357 8.6 113 to 1 0.218 8 0. 1*
9 0.357 8.6 70.1 to 1 0.372 75.8
15* cauetic 8.6 96.0 O .257 79.fc added ae 16* the reaction8.6 31.8 0.777 62.9 proceeds 17* 8.6 31.1* 0.785 61.6
10 0.095 8.6 225 to 1 0.110 8 3 .0
12 0.095 8.6 1*10 to 1 0.060 9 2 .U
1*1 0.095 8.6 21*7 to 1 0.089 81*.6
19 1.79 7.0 65 to 1 0.371 67.7
20 1.79 7.0 69.5 to 1 0 .31*7 67.2
21 1.79 8.6 119 to 1 0.201* 73.9
22 1.79 8.6 18.5 to 1 1.305 1*7.5
23 1.79 8.6 73.5 to 1 0.328 70.0
21* 1.79 5.7 62.8 to 1 0.381* 6 9 .O
25 1.79 1.1* 67.1 to 1 0.362 67.5
26 1.79 .11* 62.9 to 1 0.381* 68.2
27 1.79 8.6 21*9 to 1 0.097 78.3
28 1.79 none 62.6 to 1 0.386 1*1.6 43 Cont. TABUS VI
Ixperi- Initial Galatin Mola Ratio Chlora- Per Cent mant # Caustic Concentra Ammonia to mine Yield Molarity tion gms/l Chloramine Molarity Hydra of water zine
29 3.58 14 188 to 1 0.110 73.7
30 3.58 8.6 60.0 to 1 0.343 60.0
31 3.58 8.6 57.3 to 1 0.365 58.6
32 3.58 8.6 54.2 to 1 0.380 60.6
33 3.58 8.6 19-7 to 1 1.045 44.8
34 3.58 8.6 59.8 to 1 0.344 62.3
35 3.58 8.6 99.0 to 1 0.209 68.2
36 5.37 8.6 43.0 to 1 0.349 57
37 5.37 12.9 51.5 to 1 0.288 59
38 7.16 8.6 29.4 to 1 0.343 56
These data are illustrated in Figures 7 and 8 . The initial sodium hydroxide concentration is recorded and this concentration does not decrease very much during an experiment for all the curves except those In which the Initial caustic concentrations were relatively lew, viz. 0.095 and 0.357 molar. For the latter curve
(0.357 molar) all of the cauetic Is used up at an ammonia to chloramine mole ratio of 63 to 1. This curve was extended hy the points indicated with asterisks hy the Btepwise addition of sodium hydroxide as the reaction proceeded. This technique enabled us to eliminate the existence of a large concentration of fixed base at any one time during the experiment and still provide enough caustic to neutralize all of Percent Yield of Hydrazine 90 80 70 60 40 20 30 50 10 3 4 5 6 7 8 9 . 11 . 1.3 1.2 1.1 1.0 .9 .8 .7 .6 .5 .4 .3 Moles of Chloramine per Lfter ofSolution Lfter per Chloramine of Moles iue 7 Figure o X V □ 9 •. * Initial NaOH conc. .095 moles per liter per moles .095 conc. NaOH Initial Initial NaOH conc. 5.37 moles per liter per moles 5.37 conc. NaOH Initial liter per moles liter per moles 1.79 conc. NaOH .357 Initial conc. NaOH Initial Initial NaOH conc. 3.58 moles per liter per moles conc.3.58 NaOH Initial Initial NaOH conc. 7.16 moles per liter per moles proceeds as reaction added Caustic 7.16 conc. NaOH Initial L_L _L Legend _L _L
-r ■r Percent Yield of Hydrazine 20 10 30 40 50 60 80 70 - 90 5 0 5 07 9 15 2 15 5 15 8 15 1 25 4 255 240 225 210 195 180 165 150 135 120 105 90 75 60 45 30 15 ai Mls H N Moles Ratio iue 8 Figure 3 in Solutiory^Moles CINH Solutiory^Moles in 0 v O X • * a X asi de srato proceeds reaction as added Caustic Initial NaOH conc. 7.16 moles per liter per moles liter per moles 7.16 conc. NaOH Initial liter per moles 5.37 conc. NaOH Initial 3.58 conc. NaOH Initial Initial NaOH conc. .357 moles per liter per moles .357 conc. NaOH Initial Initial NaOH conc. 1.79 moles per liter per moles 1.79 conc. NaOH Initial liter per moles .095 conc. NaOH Initial X X Legend 2 Added I X X
X X VJI U6
the acid formed by the reactions of chloramine.
These results show conclusively that excess caustic causes a
decrease in the hydrazine yield. It is postulated that this decrease 2 Is caused by the decomposition of chloramine by hydroxide ion.Raschlg
reported the following equation for this reaction:
3NH Cl f 30H*---->3Cl” + N + NH + 3BL0 (6) ? S 3 ^ ^ 2,19 Contrary to other reports , hydrazine yields equivalent to
those reported at similar mole ratios in the Raschig process were
obtained by carrying out the reaction at the low temperatures of
21 - 2^°C. In addition,the reaction was found to proceed very rapidly at these temperatures. Maximum yields were obtained in the
time required to add the chloramine and prepare the solutions for the hydrazine analysis.
Figures 7 and 8 also indicate that the moles of chloramine added per liter of reacting solution is a more important factor affecting hydrazine yields than is the ammonia to chloramine mole ratio. In
Figure 7 where the per cent yield of hydrazine is plotted against the mole ratio, the decrease in hydrazine yield with increasing caustic con centration is apparently reversed for points at 5.37 and 7.16 moles per liter initial sodium hydroxide concentration. This reversal is not observed in Figure 8 , where the moles of chloramine added per liter of solution is plotted against per cent yield of hydrazine. In the concentrated caustic solutions the ammonia to chloramine mole ratio is decreased not only by increasing the moles of chloramine added per liter of solution but also because the caustic decreases the solubility of the.ammonia. Thus, in concentrated caustic solutions, the number of moles of chloramine added per liter of solution Is lover
for a given ammonia to chloramine ratio than In pure aqueous solution,
thus accounting for the fact that the hydrazine yield at a given ammonia to chloramine mole ratio Is higher for 5.37 and 7.16 molar sodium hydroxide solutions than for concentrated alkali solutions.
Calculations vere made based upon the assumption that the decrease in hydrazine yield vith increase in sodium hydroxide con centration at a constant number of moles of chloramine added per liter of solution is caused by chloramine decomposition by the sodium hydro xide as described in equation (6 ). Figure 9 is a plot of the per cent of the total chloramine decomposed by sodium hydroxide against the average sodium hydroxide concentration during the reaction. These calculations vere made at a constant value of 0.20 mole of chloramine per liter, at vhlch value a maximum hydrazine yield of 8 l per cent had been obtained. Excess caustic vas present in all runs and the average excess caustic concentration vas calculated by subtracting one-half the total millimoles of acid formed by the chloramine reactions from the total millimoles of caustic initially present and converting to molar concentration. The amount of chloramine decomposed is calculated by multiplying the decrease in hydrazine yield by 1 .00/0.81 since the decrease in hydrazine yield by the excess alkali represents decomposition of only that fraction of chloramine that vent to produce hydrazine in the most favorable case. This calculation assumes that excess hydroxide is not beneficial to the hydrazine-forming reaction, for if caustic does favor this reaction even more chloramine must be decomposed by hydroxide
ro Percent of Total Chloramine Decomposed by Caustic 9* ro oj
Average NaOH Cone . (m oles/liter) ^9
than 1b calculated above.
Several experiments vere performed In vhich chloramine was added
In excess of the caustic present. These results are reported in Figure
10 along vith pertinent points obtained vlth excess caustic and those
obtained by the addition of oaustio as the reaction proceeded. It
can be observed from this curve that the yield of hydrazine does not begin to depart from the yield-molarity curve in excess caustic solution until the number of moles of chloramine added per liter of solution becomes almost equal to the initial concentration of the caustic. This curve illustrates that excess caustic over that required to neutralize the acid formed by the reaction of chloramine does not favor hydrazine yields. Furthermore, it is Interesting to note that the yields of hydrazine obtained in O.O95 molar caustic and in 0.357 molar caustic solutions fall on almost the same curve as long as the number of moles of chloramine added per liter of reacting solution does not exceed the molarity of the sodium hydroxide. This fact oasts doubt upon any mechanism involving caustic or hydroxide ion in the hydrazine-forming step of the reaction, because, unlike the Hoffman rearrangement, the hydrazine-forming reaction does not appear to be effected by the hydroxide ion concentration.
The sharp decrease in yield that occurs vhen all of the caustic is used up is believed tc result from the fact that the gelatin-hydroxide 27 complex breaks down at this point and the impurities introduced vith the caustic catalyze the decomposition reaction. Once the caustic has been neutralized the decomposition reaction becomes very rapid for this reason. As chloramine Is added beyond this point not only does the Percent Yield of Hydrazine 80 k 0 9 60 70 40 50 30 20 10
Moles of Chloramine per Liter Liter per Chloramine of Moles .2 iue 10 Figure .5 0 X X No caustic .357 mole per liter initial NaOH conc. NaOH initial liter per mole .357 caustic No asi de srato proceeds reaction as Causticadded conc. NaOH initial perliter mole .095 .357 mole per liter initial NaOH conc. NaOH initial liter per mole .357 conc. NaOH initial liter per mole .095 caustic No of .7 _L Solution .8 9 1.0 .9 _L Legend
5 1
per cent yield decrease, tut the total amount of hydrazine present
in the solution is diminished very rapidly. In vlev of these findings, 30 the conclusion of Riley, Richter et al , that hydrazine is not the
Intermediate in the hypochlorite oxidation of ammonia to nitrogen
under the conditions of their experiment deserves additional
amplification. Their conclusion is based upon the fact that the
tagged nitrogen atom of the chloramine does not appear in the hydra
zine isolated at the end of the experiment when gelatin is not used
as an inhibitor. These conclusions were based on experiments which
were carried out in the presence of large initial quantities of
hydrazine. The large hydrazine concentrations used by these authors,
in addition to the metal ion impurities present in caustic may favor
the decomposition reaction of hydrazine (5) with chloramine to such
an extent that the hydrazine-producing reaction (b) can't compete with
it. If this were the case tagged hydrazine would not be produced in
their experiments but the hydrazine-producing step may be an inter-
mediate in the hypochlorite oxidation of ammonia when^hydrazine is not
present.
These studies indicate that it is desirable to have only enough
caustic present to remove the hydrogen ion produced in the chloramine- ammonia reaction plus a slight amount to stabilize the caustic gelatin
complex which removes the impurities added with the caustic. The detrimental effect of the ammonium ion to the hydrazine yield can probably be attributed to the acid catalysis of the hydrazine-des troying reaction (5). 52
2. The Effect of Ammonium Chloride
Several experiments were performed in which chloramine vae added to
aqueous solutions of ammonia containing varying amounts of very pure
ammonium chloride. These results are presented in Figure 11 where the
average ammonium chloride concentration during a reaction is plotted
against the per cent yield of hydrazine, with an approximately constant amount (0.116 to 0.120 moles) of chloramine added per liter solution.
The average ammonium chloride concentration is calculated frcm the sum
of the ammonium chloride added initially plus one-half the ammonium chloride produced "by the reactions of chloramine.
TABLE VII
Experi Chloramine Average Ammonium Per Cent Yield ment # Molarity Ton Molarity Hydrazine
39 .116 .1*03 32.1
1*0 .120 .302 3U.6
1*1 .116 .127 52.0
1*2# .120 .060 60.0 ♦This result was calculated from Neth's data for the chloramine- ammonia reaction in aqueous solution. His results were checked at a different chloramine molarity and found to "be consistent with these being reported.
These results substantiate our previous statement that ammonium by ion decreases the hydrazine yield, probably^acid catalysis of the hydrazine-destroying reaction. Studies in caustic solutions indicate that chloride ion concentrations of the order of magnitude involved in our ammonium chloride studies have little or no effect upon the yield. Percent Yield of Hydrazine
— r\5oj-t>cn
3. The Effect of Gelatin
Our experiments on the effect of gelatin are reported in Table VI
and illustrated in Figure 12. The hydrazine yields were obtained at a
constant initial caustic concentration of 1.79 moles/liter and with
approximately 0.37 moles of chloramine added per liter of reacting
solution. These results are in agreement with those previously 21,23 reported
ExperimsntB were also tried in which gelatin was added but sodium
hydroxide was not. The addition of .6 of a gram of Knox's gelatin
powder to the aqueous ammonia solution had no effect on the hydrazine
yield. The addition of .7 grams of Pure Food gelatin to the aqueous ammonia solution resulted in a considerable decrease in hydrazine yield.
No difference in the behavior of the two brands could be observed when caustic was used in the reaction. It is believed that impurities in the Pure Food gelatin caused the low yield in that experiment. This
is further Indication that gelatin does not complex impurities to any appreciable extent in aqueous ammonia solution in the absence of caustic.
D. C0NCUJSI0NS
In conclusion, the results of this investigation seem to indicate that the presence of alkali in the aqueous reaction of chloramine with ammonia to produce hydrazine serves only two purposes:
1. To neutralize the acid formed in the reaction which would otherwise form ammonium ion which has been shown to adversely affect the yield of hydrazine.
2. To stabilize a gelatin-metal ion complex, the formation of which prevents the metal-ion catalysis of the chloramine-hydrazine Percent Yield of Hydrazine 20 0 4 80 0 , £L 50 60 =,70 0 3 10 1 2 3 4 5 6 7 8 9 1.0 .9 .8 .7 .6 .5 .4 .3 .2 .1 0 rm Gel i 70ml H20 l m 0 /7 tin la e G Grams iue 12 Figure L" —cr "LT" — 55 reaction. There is no evidence from this study or from others reported in the literature to indicate that alkali serree in any other manner in promoting hydrazine formation. In fact, this atudy hae shown that an excess of alkali over that required for (1) and (2) is actually deleterious to the process.
Increases in hydrazine yield over those produced in the Raschig process vere obtained by adding caustic to the reaction as the chloramine vas being added thus avoiding a large caustic concentration at any time during the experiment.
Whereas the present results are not completely definitive vith respect to the mechanism of the chloramine-ammonia reaction, it is our belief that they favor the Cahn, Powell mechanism over that of
Audrieth, Colton and Jones; for, according to the latter, an excess of sodium hydroxide should favor the hydrazine-producing reaction.
A further argument against the Audrieth mechanism is that by its very nature it is restricted to aqueous solution and the chloramlne- ammonia reaction has been shown to produce hydrazine effectively in a variety of non-aqueous solvents. Chapter IV
The Effect of Fixed Base On The Chloramine-Anmonia Reaction In Liquid Ammonia and in Ethyl Alcohol
A. INTRODUCTION
The marked effect of trace impurities upon the hydrazine
decomposition reaction and the effectiveness of gelatin in
inhibiting the effects caused by these impurities in aqueous
solution prompted us to carry out experiments employing sodamide,
potassium amide and potassium as additives In hydrazine synthesis
by the chloramine-ammonia reaction in liquid ammonia and ethyl alcohol. 5,7, These studies differed from earlier ones, in that gelatin was added
in all of these runs. Even though only a very small amount of gelatin
is soluble in liquid ammonia, it vas hoped that the amount would be
sufficient to complex the Impurities Introduced vith the addition of
fixed bases.
B. EXPERIMENTAL
1. Materials
The potassium metal and sodamide used in this study were commercial grade while the potassium amide was prepared by the catalytic (iron wire) decomposition of a solution of potassium in liquid ammonia. Potassium alcoholate was prepared by the addition of potassium to ethyl alcohol.
Anhydrous grade synthetic ammonia vas removed directly from the tank and used in this study. Knox powdered gelatin was used as the inhibitor.
2. Experimental Procedure
Chloramine produced by the gas phase reaction of ammonia and chlorine was bubbled into traps containing liquid ammonia solutions of the base and inhibitor additives at a temperature about -3U°C. One experiment was carried out in ab8olute ethyl alcohol solution which was saturated with ammonia at 21°C.
After the chloramine had been added the volume of the solution was marked and the traps were allowed to stand four hours. The alcohol run was then diluted to 250 ail with distilled water and analyzed for hydrazine and chloride as previously described. In the liquid ammonia experiment the solution was evaporated to dryness and then washed into a 250 ml volumetric for analysis.
C. RESUITS
The results obtained are summarized in Table VIII. Buns A,B,C,D, and F were performed in liquid ammonia and Bun E was performed in ethyl alcohol that had been eaturated with ammonia.
TABLE VIII
Experi Base added, Gelatin, Conc. Chloramine Yield of ment No. Moles/liter grams/liter Molarity hydrazine per cent
A .587(NaNH2 ) k 0.306 ---
B .SUotKNHg) 3 0.233 1 8 . k
C .507 (E) + .507(KNH2 ) 9-3 0 .2^0 2.7
D 1.25(XNH2) ^3 .3 (sorbitol) o.iko 13.1
E .500 (K0CgH5) 12.0 0.500 16.9 --- F * k 0.220
* A liquid ammonia solution of potassium amide was added as the reaction proceeds.
In Experiment B the yield of hydrazine is the same as that obtained in liquid ammonia without the addition of potassium amide. A yield was obtained in ethyl alcohol saturated with amaonla Bimilar to that obtained 59 in Experiment E with potassium alcoholate addition. These similarities in yield might Just be coincidental or they nay Indicate that ammonium ion catalysis of the chloramine-hydrazine reaction is not as important in these solvents at the lower temperatures. Studies on the effect of ammonium ion upon hydrazine yield at the lower temperatures in these solvents should be carried out. It should also be pointed out that in these experiments a large excess of potassium amide and potassium alcoholate was employed which, in view of our work in aqueous solution, could have caused a low yield. In Experiment F, a liquid ammonia solution of potassium amide was added stepwise as the chloramine-ammonia reaction proceeded in an attempt to avoid a large excess of base. No hydrazine was obtained in this experiment, probably because of complete reaction of the amide with ammonium ion and chloramine soon after it is added.
If this occurs fixed base is not available to form a stable gelatin complex in the periods between base addition and the impurities introduced with the addition of potassium amide could then catalyze the decomposition reaction.
No hydrazine was obtained in our experiment with sodamide addition possibly because of too high a concentration of impurities in commercial sodamide to be complexed by the small amount of gelatin that is soluble in the solution.
Low yields were also obtained in reactions where potassium was added to the liquid ammonia. The low yields were probably caused by the reduction of chloramine by the ammonia-potassium solutions.
The possibility of using tertiary amines that are more basic then ammonia should also be investigated since these can be obtained in 60 very pure form. The use of piperidine may be satisfactory for this purpose. At the higher temperatures, where the deleterious effect of ammonium chloride has been demonstrated, addition of a strong base should increase the hydrazine yields if large excesses of these bases are avoided. Chapter V
The Separation of Anhydrous Hydrazine from Ammonium Chloride-Ammonia-Hydrazine Mixtures
A. INTRODUCTION
A complication arises when the separation of anhydrous
hydrazine from the reaction product consisting of ammonia, hydrazine
and ammonium chloride is attempted. Though hydrazine is a weaker base
than ammonia, hydrazine hydrochloride forms upon the evaporation of
ammonia because of the low volatility of hydrazine compared to that of
ammonia•
N2HU(1) + NHjCKs) > NgH^Cl(s) + HH (g) (l)
In liquid ammonia solution the above equilibrium is shifted to the left and the liquid ammonia solution contains essentially free hydrazine and ammonium chloride. Therefore, at elevated temperatures and higher pressures the vapor above these solutions should contain appreciable ammounts of hydrazine, especially in the more concentrated
solutions. It should be possible, then, to find conditions under which hydrazine may be removed from the system in the vapor phase.
B. EXPERIMENTAL
1. Materials and Equipment
Anhydrous, synthetic grade ammonia was distilled from the tank and used as the solvent. Varying amounts of hydrazine dihydrochloride, purchased from the Matheson Chemical Company, were added to liquid ammonia to produce an ammonia solution of hydrazine plus twice its molar equivalent of ammonium chloride.
An Amimco high pressure reaction vessel of 600 ml capacity, fabricated of stainless (Type 3V 7) was employed and is illustrated
in Figure 13. The high pressure valves and lines used were also made of r m
\y/y7/?77,
Teflon Lid
Vapor Sample Liquid Dip Tube Sample
Dip Tube Thermocouple Well
Glass Covering
Pyrex Liner 6 3
Type 3^7 stainless steel. The liquid solutions were contained in a
pyrex liner.
Preliminary experiments showed that the solutions being studied
attack stainless steel but do not effect Teflon (a polymer of tetra-
fluoroethylene). Short and long Teflon dip tubes were therefore used
to remove the vapor and liquid samples respectively. The stainless steel
thermocouple well was protected from the liquid by a glass covering.
The Teflon lid was incorporated in the apparatus midway through
these experiments to help prevent very serious decomposition of
hydrazine vapor at these elevated temperatures. This modification
served to isolate the vapor from the stainless steel walls and resulted
in a considerable decrease in the amount of hydrazine decomposition.
The vapor outside the liner apparently decomposes to some extent forming
a blanket of nitrogen which prevents further decomposition.
2 . Experimental Procedure
Ammonia was condensed in the pyrex liner and hydrazine dihydrochloride
slowly added. The autoclave was flushed with ammonia before the pyrex
liner was inserted. Frost was wiped from the outside of the pyrex liner with ethyl alcohol. The liner was then inserted in the autoclave and
the autoclave sealed. After the liquid warmed to room temperature the
autoclave was gently shaken to mix the contents and a considerable amount
of vapor valved off to remove any traces of air or alcohol Introduced
into the autoclave during the loading procedure.
The autoclave was heated with an amlnco heating Jacket. The temp
erature was recorded by a Leeds and Northrup Micromax self-recording
potentiometer. A relay activated by the Micromax controlled the 6k o temperature to within ±3 C. This circuit is illustrated in Figure lU.
The clock tinier was set to start heating the autoclave at night and the
relay circuit roughly controlled the temperature. During the day the
relay circuit was closed and the temperature of the autoclave adjusted
hy manually setting the variac controlling the heater. By this means the
temperature was kept constant.
The vapor and liquid lines were hoth heated. The liquid line
temperature was kept slightly below the autoclave temperature and the
vaper line temperature was kept slightly above the temperature of the autoclave. These temperatures were manually controlled by Variacs.
The apparatus originally contained a pressure gauge, located between the vapor valve and the autoclave but it was removed because it could
not be properly heated, and, as a result, hydrazine from the vapor
sample condensed there, giving inaccurate results.
The trap systems illustrated in Figures 15 and l6 were used to absorb the liquid and vapor samples. The liquid sample was always
taken first and the vapor sample taken twenty to thirty minutes later.
This time interval was allowed in order to compensate for the small
disturbance in equilibrium caused by the removal of the liquid sample.
The vapor sample was absorbed directly into standard hydrochloric acid, bromcresol green being used as the acid-base Indicator. The entire sample was analyzed for total base and then for hydrazine by the iodate method. The liquid sample was condensed in trap I ^ by
liquid nitrogen. The ammonia was then vaporized into standard hydrochloric acid contained in Trap 11^. As the ammonia boils away the hydrazine and the ammonium chloride react displacing ammonia from the ammonium Autoclave - © M - - M C~ —S V— Relay R - R LCRCL CIRCUIT ELECTRICAL Variac eodn Micromax Recording a.c. I lOv lc Timer Clock ob Switch Bomb Figure J
-1
Y M T her mo ^ couple 65 C ... _ © - ^ J 66
Figure 1 5
LIQUID LINE
N* Tank To Liquid Line
Nitrogen Reservoir
■^Mercury Bubbler
y Liquid Standard Nitrogen Acid
TR A P I L
TRAP H L 67
Figure 16
VAPOR LINE (Front half of line )
N? Tank
Nitrogen Reservoir
Two Way Stops *2^Mercury Bubbler
Heating Tape
Standard Acid
TRAP Zy 6 8 chloride. When ammonium chloride Is In excess of the hydrazine the number of moles of hydrazine must be subtracted from the ammonia to give the concentration of the ammonia in the liquid.
When hydrazine is in excesB, the total hydrazine will again hare to be subtracted from the total bases found in the analysis since hydrazine vill liberate ammonia from all the ammonium chloride present and the remainder of the hydrazine vill be titrated as a base. The contents of Trap 1^ vere analyzed for hydrazine, chloride and total ammonia. The contents of Trap 11^ vere analyzed for total ammonia.
Bromcresol green vas used as the acid-base indicator. Analysis of the contents of Trap II for hydrazine in several runs gave negative X* results. A sample calculation vill be presented below:
Trap II contained 50 ml. of 3»00^M HC1 at the start of the Is sampling, and required 30.36ml. of 1.0012 M HC1 to titrate the excess base present at the end of the sampling period. Trap I vas vashed L into a 1000-ml. volumetric flask. A 25-ml. aliquot was taken for the hydrazine analysis, 9*78 ml. of .1000 M KIO^ being required.
Twenty five ml. of .0996 M A NO vas added to a 25-ml. aliquot e 3 previously acidified vith HNO^s ^.08 ml. of SONS .1012 M vas required for a Volhard analysis of chloride. A 50-ml. aliquot required 1.51 ml. of 1.0012 M H01 for analysis of total ammonia. 69
Total hydrazine:
9.78 x .1000 x 1000 = 39-12 millimoles of H I 25 * H Total ammonia:
50 x 3.001* = 150.20 30.36 x 1.0012 = 30.39 180.59 millimoles of NH^
1.51 x 1.0012 : 1.51 X 20 = 30.20 millimoles of NH in trap 1^ ^
210.79 Total NH analyzed (millimoles) 3 - 39.12 millimoles of hydrazine 171.67 Total millimole8 of NH^ in the liquid sample
Total chloride:
25 x .0996 = 2.1*9 millimoles of AgNO^
1*.08 x .1012 = .1*1 millimoles of K3CNS 2.08 millimoles of Cl- in aliquot
2.08 millimoles x 1000 = 83*20 Total millimoles of Ol"
83.20 millimoles of Cl~ 39.10 millimoles of hydrazine 171.67 millimoles of ammonia 293.97 total millimoles
83.20 _ 28.30 mole per cent NH, Cl 293.97
39.10 _ 13.30 mole per cent I H, 293.97
171.67 - 58.1*0 mole per cent NH 293.97 3 70
C. RESULTS
1. Preliminary Observations
Inconsistent results were obtained in the early studies because
of very eerious decomposition and corrosion problems. However, certain
interesting observations were made during the course of these experi ments and are recorded below:
a. Very rapid decomposition of hydrazine in systems such as were involved in this study occurs in the neighborhood of l80°C.
Evidence for this as obtained in an experiment in which the temp erature was slowly raised to this point, included an almost instant aneous rise in pressure from 1200 psi to 2000 psi and a very rapid temperature rise to above 200°C (the limit of our recording instrument).
When the bomb was taken apart, a green substance which turned rust color upon standing in the air was observed on the walls of the autoclave.
b. In one experiment, it was shown that after leaving the contents in the autoclave at 130°C for twelve hours and cooling slowly overnight, more than ninety per cent of the hydrazine was recovered from the vapor and residue, on valving off the vapor phase at room temperature. However, in a similar experiment in which ths vapors o were valved off at 130 C, only twenty-five per cent of the hydrazine was recovered. This indicated that a major amount of hydrazine decomposition at these temperatures probably occurs in the vapor phase.
c. It was found that a sodium hydroxide coating applied to the autoclave and to the vapor line reduced decomposition. The coating was applied according to the following procedure: 600 grams of sodium 71
hydroxide, C. P. Grade, vas dissolved in bOO grams of distilled water.
This caustic solution was poured into the autoclave which was allowed to
stand twenty to thirty minutes before pouring the solution off. The
apparatus vas then assembled, heated to l80°C under an atmosphere of
nitrogen, and sealed. After cooling, the bomb was opened and the
charged pyrex liner immediately inserted. Another experiment was then
carried out in which the autoclave contents were allowed to remain at
130°C for twelve hours and the vapors were then valved off at this
elevated temperature. A recovery of 89.2 per cent of the original
hydrazine vas obtained in this case, indicating a considerable decrease
in the amount of decomposition. A similar experiment was carried out
at 150°C and 65.6 per cent of the hydrazine was recovered.
d. A sodium hydroxide coating was applied and dried by nitrogen o gas without baking. The apparatus was charged and heated to 120 C. o In the process of withdrawing vapor samples the temperature rose to 13^ C.
Sampling was stopped and the heaters were cut off. The temperature \ o o steadily increased to C and then almost instantly rose above 200 C
and the pressure increased 600 psi. Analysis of the contents revealed
that all but 1.8 per cent of the hydrazine had decomposed, indicating the
importance of baking the sodium hydroxide coating.
e. The Teflon lid described in Figure 13 was constructed and
incorporated into the apparatus. The vapor valve and line were coated with sodium hydroxide as described in paragraph C but the autoclave was not. A recovery of 9^ per cent of the hydrazine was obtained o after standing at 115 C. A considerable Increase in the concentration of hydrazine in the vapor was also observed. 72
2. Liquid Composition-Vapor Composition Data Obtained vlth the Sodium Hydroxide Coated Apparatus.
In these experiments the Teflon lid was used and the vapor valve
and line were coated with sodium hydroxide. The results are summarized
In Tahle IX. These results are not quantitative but do give a qualitative
Indication of the feasibility of this separation.
TABLE IX
Experi Temp. Liquid Composition Vapor Composition ment No. °C Mole Per Cent Mole Per Cent nh3 NH^Cl N H, N2\ 2 4 1—1 0 IT\ • 17 113.5 10.5 67.2 22.3
18 113.5 8.6 63.5 27.9 .105 «=* OJ 0 19A 113.5 2 2 . 4 40.1 37.7 «
19® 113.5 20.8 37-9 41.3 .236 .233
19C 113.5 21.3 3 9 A 39.3 .238 .234
20A 113.5 lit.8 57.1 28.1 .223 .212
20® 113.5 14.8 53.6 31.6 .226 .217
The concentration of hydrazine in the vapor In experiment 17
Is low compared with the other experiments. When the autoclave was
opened at the end of this experiment, It was noticed that a piece had broken off the pyrex tube covering the thermocouple well. This had
exposed the vapor beneath the Teflon lid to a small area of metal
surface and could account for the low concentration of hydrazine in
the vapor. 73
In the last four experiments reported, a two-way stopcock the was introduced into^vapor line so that two consecutive vapor samples could be obtained. The stopcock was heated slightly above the bomb temperature to prevent the condensation of hydrazine. These results demonstrate the very small change in vapor composition produced by sampling.
3. Liquid Composition-Vapor Composition Data Obtained with the Teflon Coated Apparatus.
In all of the experiments reported above a considerable amount of iron oxide was observed in the liquid sample indicating a considerable amount of corrosion by the liquid. In an effort to avoid this, the autoclave, valves and lines were coated with "Teflon One Coat Enamel"
(Du Pont Code Number 851-204) whereever the liquid or vapor could come in contact with the metal. The coating was applied by swirling the enamel in the autoclave until a uniform covering was obtained. The o coating was dried and then baked in an oven at 230 C. A hard, glossy finish was obtained. The data recorded in Table I were obtained with the Teflon Coated apparatus.
TABLE X
Experi- Temp Liquid Composition Vapor Composition *ent No. °C Mole Per Cent Mole Per Cent
21A 114.0 11.6 60.4 28.0 .206
2 IB 114.0 11.0 61.2 27.8 .199
21C 114.0 8.4 62.2 29.4 154 These experiments gave indication that the decomposition
of hydrazine in the vapor phase had been diminished by the Teflon
coating. Run 21A vas taken tvo days after the ahtoclave was loaded.
Run 21B vas taken four hours after 21A5and Run 21 C vas taken tventy
hours after 213. The liquid sample obtained in Bun 21A vas colorless,
that from Run 213 vas slightly rust colored, while that of Experiment
21C vas very highly colored and contained pieces of the Teflon coating.
During the run the glass covering for the thermocouple well had broken and in the places where the liquid came in contact with the
Teflon-coated thermocouple veil considerable attack of the metal had occurred. The liquid sampling valve and lines also had been attacked.
However, the coating on the vapor valve and vapor line,as well as the coating on the autoclave which came in contact with only the vapors,, vas not affected.
It is believed that the hot solution of ammonium chloride in liquid ammonia penetrated the Teflon coating and reacts vith the metal underneath. Although the coating vas stripped from the metal, the
Teflon enamel did not appear to have reacted chemically.
Experiments shoved that anmonia-hydrazine solutions did not attack the Teflon coated apparatus, indicating that the ammonium chloride vas the main cause of the corrosion problem. The data reported in Table X indicate that most of the corrosion occurred after withdrawing Sample 21A.
Information on Kel-F (a polymer of chloro-trifluoroethylene) coatings indicated that several coats could be applied on top of one another. Only one coat of the Teflon Enamel could be applied because 7 5
the second coat vould not adhere to the first. By using several coats of
Kel-F, it vas hoped that a non-porous coating might be obtained. The
coating vas applied by a procedure similar to that employed for the
Teflon coating. Seven successive coatings vere applied and a very
smooth, apparently nonporous coating vas obtained. Unfortunately,
hovever, tests showed that ammonia alone at temperatures in the range o o of 100 to 130 C completely destroyed the coating leaving a black
carbonaceous residue. This vas not a matter of porosity; Kel-F
is itself attacked by ammonia at elevated temperatures. Because of
the problem of devising a method of operation in vhich the liquid Bample
can be removed without coming into contact with stainless steel, this
study vas discontinued.
D. COMCmSIONS
These results have shown that hydrazine can be separated from the reaction product of the chloramine-ammonia reaction by a high temperature-high pressure distillation. Very serious corrosion and
hydrazine decomposition problems vere encountered. Although the
corrosion problem could not be overcome in this study, it can be avoided in a distillation on a commercial scale by using a glass
lined stlllpot. The hydrazine decomposition problem vas solved by
coating the apparatus with Teflon enamel. 7 6
Chapter VI
Vapor Pressure-Composition Studies On the System Ammonia-Hydratine at Elevated Temperatures
A. INTRODUCTION
The study of the system ammonia-hydrazine-ammonium chloride
reported in the proceeding chapter was discontinued because of the
corrosive nature of this liquid. However, preliminary work on the
binary system ammonia-hydrazine indicated that this system could be
studied with our apparatus. Vapor pressure-vapor composition-liquid
composition data for the binary system ammonia-hydrazine were obtained
therefore, at four temperatures, viz. 68.5, 100.3, 114.1, and 124.9°C.
B. EXPERIMENTAL
1. Materials
The autoclave, high pressure valves and lines employed in the previous experiments (Chapter X) were coated with "Teflon One Coat
Enamel", as previously described, for use in this study. The pyrex
liner was loaded with anhydrous synthetic-grade ammonia and purified, anhydrous hydrazine. Matheson, Coleman and Bell, 95# anhydrous hydrazine was redistilled in vacuo over sodium hydroxide to completely dehydrate it. The purity of the product was checked by a freezing point deter- minination and was found to be better than 99 per cent.
2. Experimental Procedure
The apparatus used in this study is illustrated in Figure 17•
The solutions of ammonia and hydrazine vere contained in a Pyrex liner which vas fitted with a Teflon lid. Thin Teflon sheet wae placed on the bottom and around the vails of the autoclave to prevent chipping 7 7
Figure 17
Heating Tape
'////// // //
Teflon Lid Liquid Sample Dip Tube Vapor Sample Dip Tube
Pyrex Liner Teflon Sheet 7 8 of the Teflon enamel by the Pyrex liner. One short and one long
Teflon dip tubes extended into the Pyrex liner for the purpose of withdrawing liquid and vapor samples which were forced through the tubes by the pressure in the autoclave. The inside diameter of these tubes was kept very small to diminish the amount of material tied up in the lines.
The autoclave was immersed in a well stirred, thermostated oil bath. Sohicyl No. 300 steam cylinder oil was used for the bath. The oil was heated by a primary, 500-watt heater which operated constantly and a secondary, 300-watt heater which was connected to the control o circuit. The temperature was kept constant to within +.05 C. with 6k a Klystron tube circuit and a "Merc to Merc" thermoregulator. The temperature remained constant during a given experiment but over longer periods of time drifted slightly because of the distil] vtion of mercury from the thermoregulator into the reservoir. It was thiis necessary to reset the regulators from time to time. The oil-bath temperature was measured by a calibrated ASTM thermometer.
The two super pressure needle valves were wrapped with flexible heating tape. The tapes were connected to variacs for temperature control. The temperature of the liquid valve and liquid line was kept slightly below the bomb temperature, whereas the temperature of the vapor valve and lines were kept 5° to 10°C. above the autoclave temperature.
The temperatures of the valves and lines were recorded by a Leeds
Northup Micromax self-recording potentiometer connected through a switch box to properly located copper-constantan thermocouples.
The Pyrex liner and autoclave were finally dried by flushing them 79
with ammonia for forty to fifty minutes prior to the condensation of
ammonia and addition of hydrazine. The vapor valves and lines vere flushed
with dry nitrogen.
The ammonia was condensed in the Pyrex liner, by a dry ice bath. The
rate of flow of ammonia was always large enough to produce a positive
ammonia flow through the liner. Hydrazine was removed with a dry pipette
from the pure stock material under an atmosphere of dry nitrogen end
added to the Pyrex liner. The solution was then slowly warmed to o ° about -35 to -30 C and then shaken until all of the hydrazine dis
solved and the solution was well mixed. If this mixing precaution
was not observed, several weeks vere required to obtain good mixing
by a diffusion process once the autoclave was loaded. Frost was then
removed from the outside surface of the Pyrex liner by wiping the vessel
with a towel moistened with denatured alcohol. The alcohol vas then
wiped off, the liner vas immediately placed in the autoclave which had
been flushed vith ammonia, and the autoclave was sealed. After the contents
of the bomb had been allowed about thirty minutes to warm up, a large
vapor sample was withdrawn to remove any traces of water vapor, alcohol,
or air that may have been introduced into the autoclave during the
loading.
The system vas left in the oil bath for three days to allow
equilibrium to be attained prior to withdrawal of the first sample.
After an experiment was started, the valve heaters, bath heaters, t> fi st irrer and control circuit had toAkept on constantly in order that
equilibrium would not be disturbed. The vapor pressure, liquid
composition, and vapor composition were measured in that order, at 8 0
several temperatures.
The vapor pressures were measured, with Bourdon type pressure
gauges. Two Ashcroft gauges of this type were obtained for this * study and calibrated with an MIT type dead weight gauge. One
gauge, divided into 25 psi subdivisions, measured pressures in the
range of 0 to 3000 psi. The other gauge was divided into 5-psi
subdivisions and covered the range from 0 to 600 psi. The gauges were connected to high-pressure fittings so that the assembly could be attached to the vapor valve on the autoclave. A high-pressure "tee" was incorporated into the assembly as shown in Figure 18 so that nitrogen could be introduced into the gauge. Nitrogen was added until the pressure in the gauge was about 5 psi below the estimated pressure of the autoclave. By this technique, a pressure determination could be made without removing a large gas sample from the bomb contents and consequently disturbing equilibrium. The addition of nitrogen also prevented the condensation of hydrazine in the pressure gauge.
After the gauge had been filled with nitrogen to the desired pressure, a high-pressure needle valve was closed isolating the nitrogen lead line from the gauge system. The vapor pressure of the system was measured by opening the vapor valve on the autoclave for a very short period of time. The pressure of the system was checked again in 5 to
10 minutes and recorded if the two readings agreed.
This technique for measuring vapor pressure was checked by determining the vapor pressure of a pure sample of ammonia at various temperatures. The measured values are compared with the literature values in Table II.
* The author would like to thank Dr. Mai Trezciak for his help in the calibration of the Bourdon gauges. 81
Figure 18 82
TABLE XI
Temp. Pressure Pressure Per Cent °C (measured) (literature) Deviation absolute pel absolute psi
125.6 1*5* 1*57 -.2
1 1 * . 2 1 1 9 1 1 1 8 9 + .2
100.3 918 91* +.*
The measured values agree veil vith the data obtained from the literature and verify the accuracy of this procedure for measuring vapor pressures.
The liquid composition vas determined by analysis of a liquid sample collected In a trap system similar to that Illustrated in
Figure 19. The steps in the procedure employed for obtaining the liquid sample vill be desoribed In the order In vhlch they are performed. The entire trap system and mercury bubbler vas flushed vlth nitrogen,* a known volume of standard acid vas added to Trap II , (see Figure for trap designations). The liquid L nitrogen bath vas placed around TrapI , and the glass stopcock on Xi the line leading to the nitrogen tank vas closed, Isolating the system. A small amount of the liquid/^valved off Into a vaste trap to clear the liquid valve, lines, and dip tube* The liquid sample, used for analysis of the liquid composition, vas taken immediately after removal of the vaste sample by condensing the sample to be analyzed T in Trap IT . Trap III_ vas empty during the sampling and vas used to trap L L «ny standard acid that may blov over as a result of the liquid sample removal. After sampling, Trap III vas filled vith vater and most L 83
Figure 19 LIQUID LINE
To Liquid N2 Tank Line Nitrogen Reservoir
•: Mercury Bubbler
^ Liquid Nitrogen
Standard Acid TRAP II
TRAP DI l 8k of the ammonia in the sample vas distilled into TrapsII and III by L L vanning the condensed liquid sample to room temperature* Trap II , I« containing standard acid, vas kept aoldlc at all times by adding aoid vhen necessary as shown by brcmcrssol green indicator. After boiling ceased in the first trap, the trap, lines and bubbler vere flushed vith nitrogen. Trap 1^ was cooled and the contents vere immediately washed into a volumetric flask containing a known volume of standard acid. The traps were all washed into the same volumetric and the contents were analyzed for total base and hydrazine. Hydra zine was determined by the aoid-iodate method and ammonia by the difference between the total base and the total hydrazine. Con sistent results obtained from analysis of consecutive liquid samples in addition to the constancy of vapor-pressure measurements made before and after liquid sampling indicate that any ohange in the composition of the system oaused by removal of a liquid sample vas so slight that it vas within experimental error. About 10 to 15 ml of liquid solution was removed for analysis and about 5 ml vas dis carded as the vaste sample.
The vapor line, trap and bubbler illustrated in Figure 20 vas flushed vith nitrogen and the sample vas then absorbed directly into standard acid containing bromoresol green as the indicator. The entire sample vas analyzed first for total base and then hydrazine.
The hydrazine analysis vas performed using .OI667M potassium lodate solution so that a small vapor sample could be removed for accurate analysis. Analysis of several consecutive vapor samples indicated that a snail amount of vapor probably consisting of material Isolated 85
Figure 20 VAPOR LINE (Front half of line)
N , Tank
Nitrogen Reservoir
Two-way Stopcock Mercury Bubbler
Standard Acid
TRAP IV in the rapor Taira, must be discarded prior to collecting the rapor sample* The else of the rapor sample removed for analysis rariea vith the liquid composition range being studied* Analysis of eenseeutlve samples taken after the first one had been dleoarded vere reproducible and Indicated that the rapor composition is Independent of the sample size used in our experiments* 8?
C. KBSBPIJS AIED CALCULATIONS OF THERMODYNAMIC QUANTITIES
1. Results
The results obtained from this study of the system anmonla- hydrazlne are summarized In Table XIII to XVI.
Since, In some of the experiments the temperatures of the measurements varied Blightly from the values 8 8 .5°, 100.3°, llU.l°, and 12^.9°, It vas necessary to make slight corrections for these variations. The columns labeled "corrected pressure" and "corrected vapor composition" contain the results of these calculations con verting all the data to the constant temperatures listed above. These conversions vere accomplished by Interpolating the changes in vapor composition with temperature and vapor pressure with temperature at several different constant liquid compositions for slight temperature deviations. These interpolations vere made from the uncorrected curves of vapor composition vs liquid composition and vapor pressure vs liquid composition at the four different temperatures. The corrected data are plotted in Figures 21 and 22.
The dotted line plotted in Figure 21 is the summation of the partial pressures of ammonia and hydrazine calculated from Roault's
Lav. . 1?i °
The partial pressure^ of component "1” In the vapor equals the mole fraction N 1 of component 1, in the liquid times the vapor pressure
P j 0 of component 1 at the same temperature in the pure state. The following values for the vapor pressures of pure ammonia and pure hydrazine vere obtained by interpolation of the data reported in the literature: Absolute Pressure psi 1500 IOOO 1375 250 500 750 1125 125 875 lO 20 I24.9*C. Rooult'ft Low Colculationi Tb« Sywtaai Tb« Aamonla-Hjrdraziito iud opsto Ml % Hydrazine % Mole Composition Liquid 30 40 Figure Figure 50 21 60 90 80 88 90 100 Figure 22 m Vapor Composition Mole % Hydrazine > ^ o 03 p O P P 6g o o ro O £ Oi o o o o o CD O
Liquid Composition Mole % Hydrazine The System Ammonia -Hydra z ine T A B U III 65 35,58 Temperature Yapor Pressure Yapor Pressure °C of Pure Ammonia of Pure Hydrazine pel pel
88.5 720.3 5.99
100.3 913.8 9.22
U 4 . 1 1186.4 14.7
124.9 1438.8 21.47
Yapor pressure data could not be obtained for solutions with a liquid composItloo greater than eighty per cent hydrazine because considerable decomposition of hydrazine occurred In this range. 9£
TABUS XIII
DATA FOR THE SYSTEM AMMONIA-HYDRAZIHE AT 88.5°C
Expt. Temp. Pressure Pressure Liquid Vapor Vapor No. ° e . absolute, corrected composition, composition, composition psi to 8 8 .5°C mole per cent mole per cent corrected to hydrazine hydrazine 88.5°C., mole per cent hydrazine
28B 88.4 495 496 29.57 0.839 0.842
28C 88.3 494 496 29.37 0.817 0.823
29A 88.4 684 685 4.29 0.194 0.195
30A 8 8 . b 564 564 18.93 0.631 0.631
30G 88.5 559 559 19.89 0.673 0.673
31A 88.5 465 465 35.93 0.991 0.991
3 IB* 88.5 464 464 35.8 0.984 0.984
31C 88.5 463 463 3 6 . 1 2 0.987 0.987
32A* 88,6 215 214 75.3 2.37 2.36
32J 88.5 198 198 77.58 2.50 2.49
33A 88.5 385 385 47.82 ------
33B 88.5 381 381 48.29 1.33 1.33
33K 88.5 324 324 58.77 1.67 1.67
34a 88.5 274 274 65.81 2.03 2 . 0 3
3 4 B 88.5 273 273 65.65 2.01 2.01
34b * 88.5 272 272 66.4 2.03 2 . 0 3
* Only the rapor pressure and -vapor composition were measured in these rune* The liquid composition was determined from the rapor pressure- liquid composition curve. This technique enabled us to show that liquid sampling did not affect the rapor composition results. 92
TABLE XIV
DATA FOR THE SYSTEM AitOflA-HYDRAZINE AT 100.3°C.
Expt. Temp. Preeeure Preeeure Liquid Vapor Vapor No. °C. absolute corrected composition, composition, Composition pel to 100.3 C mole per cent mole per cent corrected to psi hydratine hydrazine 100.3°C., mole per cent hydrazine
2 k C 100.0 ------25.if7 1.06 --- 2lfD 100.0 ------26.57 1.11 --- 25A 100.2 76if 766 13.35 ------26B 100.2 79if 796 10.2lf ------26C 100.2 791 793 10.10 ------26d 100.2 792 79*f 9.69 0.if98 0.501 26E 100.2 793 795 9.76 ------27A 100.3 791 791 10.00 0.538 0.538 27B 100.3 788 788 10.5U 0.537 0.537 27J 100.5 761 758 13.93 ------28D 100.5 603 601 31.77 ------28E 100. if 601 600 31.91 ------29B 100,3 85if 85if if.86 0.315 0.315 30B 100.3 705 705 19.65 0.901 0.901 30F 100.3 703 703 2 0 .1f7 0.928 0.928 31D 100.if 566 565 36.79 l.if2 l.if2 32D 100.3 258 258 75.80 3.20 3.20 321 100. if 225 22lf 79.51 3.67 3.66 33C 100.3 if68 lf68 if 8.88 1.76 1.76 33J 100.3 390 390 60.21 2.20 2.20 3^C 100.3 329 329 67.6if 2.7if 2.7if 3l*C* 100. if 327 326 68.20 2.78 2.77
♦Only the vapor pressure and vapor composition vere measured in this run. The liquid composition was determined from the vapor pressure-liquid composition curve. 93
TABUS I V
DATA f o r t h e SYSTEM AMMONIA-HYDRAZINE AT 114.i °c .
Ezpt . Temp. Pressure Pressure Liquid Vapor Vapor No. °C. absolute corrected composition composition composite psi to 114.1 C mole per cent mole per cent corrected psi hydrazine hydrazine 114.1°C mi per cent hydrazine 22C 114 ------10.16 0.913 22D 114 ------10.62 0.940 --- 22E 114 ------10.70 0.939 --- 22F 114 ------II.96 1.01 --- 22G 114 ------12.69 1.05 --- 23B 114 ------6.04 .603 --- 23C 114 ------5.76 .567 --- 23D 114 ------6.95 .676 --- 24a 114 ------22.76 1.48 --- 24B 114 ------23.82 1.46 --- 25C 114,0 955 956 14.75 ------27C 114.0 995 997 10.97 ------27D 114.0 991 993 11.87 ------27E 114.1 988 988 12.08 ------27F 114.1 985 985 12.40 0.993 0.993 271 114.1 946 946 15.71 1.13 1.13 28F 114.0 740 741 34.82 1.82 1.82 2SG 114.1 729 729 35.33 1.79 1.79 29C 114.0 1090 1092 5.18 O.56I 0.564 29D 114.0 IO87 IO89 5.36 0.562 0.565 30C 114.1 883 883 20.88 1.41 1.41 3 IE 114.1 709 709 37.86 1-91 1.91 32S 114.1 311 311 76.52 4.11 4.11 32F 114.1 309 309 77.66 ------32H 114.0 259 259 81.85 4.94 4.95 33D 113-9 573 575 50.28 2.29 2.30 33E 113.9 556 558 51.83 2.41 2.41 331 113-9 469 471 61.43 2.93 2.94 34D 114.1 400 400 69.54 3.50 3.50 343)* 114.1 395 395 70.00 3.55 3.55
♦Only the vapor pressure and vapor composition were measured in this run. The liquid composition was determined from the vapor pressure- liquid composition curve. 94 TABLE! XVI
DATA FOR THE SYSTEM AMMONIA-HYDRAZINE AT 124.9°C.
Expt. Temp. Pressure Pressure Liquid Vapor Vapor No. °C. absolute, corrected composition, composition composition psi to 124.9°C. mole per cent mole per cent corrected to psi hydrazine hydrazine 124.9°C. mole per cent hydrazine
2 J G 125.2 1146 ll4l 14.22 --- 2 J E 125.1 1137 1133 15.17 --- 28H 124.8 828 829 37.65 --- 281 125.O 821 820 39.34 --- 29E 124.9 1265 1265 7.42 1.03 1.03 29F 124.7 1248 1252 7.65 0.959 0.965 30D 124.8 1031 1032 23.39 1.88 1.88 30E 124.9 1033 1033 22.83 1.84 1.84 31F 124.9 798 798 40.01 2.47 2.47 32G 125.1 293 292 83.31 6.03 6.02 33F 125.0 646 645 54.17 3.05 3.04 33G 125.1 537 535 62.18 3.31 3.30 33H* 125.0 534 533 62.50 3.50 3.49 34e 125.3 444 44i 72.40 ------34f 124.9 442 442 72.11 4.63 4.63 34g 124.9 399 399 74.99 4.90 4.90 34g * 124.8 314 315 81.80 5.91 5.92
* Only the vapor preseure and vapor composition were measured in this run. The liquid composition was determined from the vapor pressure-liquid composition curve. 95
The graph of liquid composition versus vapor composition at
constant pressure has been constructed and 1b presented in Figure 23 •
The data plotted in this figure vere obtained by reading the liquid
compositions at a constant pressure corresponding to the four temp
eratures 88.5°, 100.3°, llb.l0, and l2b.9°C. and reading the vapor
composition corresponding to this liquid composition from Figure 20.
These data are presented In Table XVII.
TABLE XVII
Temp. Pressure Liquid composition, Vapor compoi °C. absolute, mole per cent hydrazine mole per cei psi hydrazine
88.5 250 69.8 2.15 100.3 250 76.7 3.35 llb.l 250 82.3 b .86 12b. 9 250 86.0 6.60
88.5 375 b9-9 1.39 100.3 375 62.2 2.33 llb.l 375 71.6 3.70 12b .9 375 77.0 5 .1b
88.5 500 29.2 0.850 100.3 500 bb.8 l.6b llb.l 500 58.1 2.73 12b. 9 500 66.5 3.93
88.5 625 11.2 O.bb 100.3 625 29.2 1.19 llb.l 625 b5.3 2.16 12b. 9 625 55-2 3.10
100.3 750 lb.6 0.72 llb.l 750 33.6 1.79 12b. 9 750 bb.5 2.63
100.3 875 2.9 0.19 llb.l 875 21.8 l . b o 12b.9 875 3b .6 2.29 Vapor Composition Mole % Hydrazine Liquid Composition Mole %Hydrazine Mole Composition Liquid The Sjetem Aaeionia-Hydrazine Sjetem The Figure23 LEGEND 750psi 625psi psi 0 5 2 psi 5 7 8 500psi psi 5 7 3 os Press. Const 0 9 96
100 97
The curves plotted In Figure 23 are at constant pressure with
both the temperature and liquid composition varying along the curve.
As oaleulated from the phase rule, the system has two degrees of
freedom* Therefore, at constant pressure the choice of temperature establishes the liquid composition or vice versa* The curves all
have a point in ocmmon at a liquid composition and a vapor composition
of 0 per Cent hydrazine, the temperature at this point being that necessary to produce the vapor pressure of ammonia for each constant- pressure curve.
2. Thermodynamic Calculations
Sufficient data has been obtained to enable calculations of activities and activity coefficients for ammonia to be made* Ammonia is Chosen as the solvent and the activity of the solvent in any solution is given by the expression:
a „ J:Q (2 ) 1 f 0 1 where a. ^ is the activity of solvent, f^ is the fugacity of the vapor of solvent over the solution and f^0 the fugacity of the vapor of pure solvent at constant temperature.
In order to make these calculations the fugacities of the gases in 66 the mixture must be known. Lewis and Randall assume that gases form perfect solutions and that the fugacity of the i th component in a gas mixture at total pressure P can be represented as
(*) = H ^ 0 (3) whore is the mole fraction of the 1 th component In the gas solution and t ^ ° la the fugacity of the pure gaa at the same temp erature and total pressure. 67 Friedman has measured the compressibility of hydrogen, nitrogen and their mixtures and has shown that this formulation is inaccurate at high pressures. An empirical relationship has been 68 derived from which the virial coefficient B mix and therefore the fugacity of a binary mixture can be calculated from a knowledge of the virial coefficients of the components and Bg.
■ ^ = ' i 2b1 + 2H1',2B12+ «S\ W where
B = 1/2 (B + B2 ) (5) 12 The virial coefficient for ammonia was calculated from the equation
69,70: D/T2 V)= Bo - V” • <6) where B q = .0^087 liter/mole 2 2 A = 3.237 liter atmos/mole
D = 7 . 7 ^ x (T)~2
T = Temperature in degrees absolute 99 TABUS XVIII
Virial Coefficients for Ammonia
Temperature°C Virial Coefficient B
88.5 ".1563
100.3 -.1*31
lll+.l -.1297
12U .9 -.1206
Virial coefficients for hydrazine have not been reported in the literature. Limited P-V-T data for hydrazine were obtained 59 from experiments on the gas density of hydrazine . The virial coefficient was calculated from this data by the equation:
PV = riRT ■+ BP (7)
TABUS I H
Virial Coefficients for Hydrazine o Temperature C Virial Coefficient B
90 .0052
95 .0085
100 .0087
110 .0105
120 .0102 100
Substitution of these virial coefficients into equations (k)
and (5), using the max1nun value for mole per cent hydrazine in the
vapor obtained in these experiments ® result for the virial
coefficient of the mixture which differs from the virial coefficient
of ammonia by only five per cent. Therefore, for this study the
fugacity of the mixture will be calculated from the preeBure-fugacity
data for ammonia.
The pressure-fugacity relationship for ammonia was derived from 69 66 the equation of state for ammonia(9 )and the equation (8 ) : . r P RT In = Jp/VdP at constant temperature (8 )
P = RT + BP (9) T T
P - RT V-B
dP - -RT 0 (V-B) dv
Substituting into (8)
la I f - If V_! V f '(V-B1 )2 dv
Upon integration the following expression is obtained:
In f = -B u. B + In (V'-B) - In (V-B) f . v ’-B V-B
let V ’ ----- ><» so f 1 ---- »P' then 1 ----^ 0 V'-B = RT y» p»
In f ;= B + In RT - In (V-B) f ' V-B f '
In f-In f ' = _B_ + In RT - In f • - In (V-B) V-B In f = B + In RT V-B V-B 101 substituting th® value of V from the equation of state (9 )
V= BT + B P the following expression Is obtained:
In f - BP + In P (10) BT
(ID
where B Is defined by equation (6 )
The data obtained from equation (1 1 ) for the pressure-fugacity relationshipvere converted from atmospheres to pounds per square inch and are presented In Tables XX and XXI. 102
TABLE XX
Preasure-Fugaolty Data for Ammonia 88.5°C 100.3°C Pressure Fugacity Pressure Fugacity psi psi psi psi
14.7 14.6 14.7 14.6
73.5 7 1.6 73.5 71.8
147.0 139.5 147.0 140.3
220.5 203.7 220.5 205.7
294.0 264.6 294.0 267.8
367.5 322.1 367.5 327.1
441.0 376.5 441.0 383.4
51^.5 427.9 514.5 436.9
588.0 476.3 588.0 487.9
661.5 521.9 661.5 536.1
735.0 564.9 735.0 582.0
808.5 652.3
882.0 666.5
911.0 682.1 103
TABLB XXI
Pressure-Fugacity Bata for Aamonla 114.1 124.9 Pressure Fugacity Pressure Fugacity psi psi pel psi
14.7 14.6 14.7 14.6
73.5 72.0 73.5 72.2
147. o 114.1 147.0 141.7
220.5 207.4 220.5 20 8 .9
2914-. 0 27 0 .9 294.0 273.1
367.5 331.9 367.5 335-1
44i.o 390.1 441.0 394.7
514.5 446.0 514.5 452.2
588.0 499.5 588.0 507.3
661.5 550.5 661.5 560.2
735.0 599.3 735.0 6 1 1 .1
808.5 645.9 808.5 659.9
882.0 690.3 882.0 706.8
955.5 732.8 955.5 751.6
1029 773.2 1029 794.7
1103 811.7 1103 835.8
1176 848.2 1176 875.2
1250 913.0
1323 948.9
1397 983-4 1470 1016.2 These data vere plotted on large scale graph paper and the fugacity corresponding to the measured total vapor pressure was obtained. The partial fugacity of ammonia vas obtained by multi plying the total fugacity by the mole fraction of ammonia In the vapor and the partial fugacity of hydrazine is the difference of total fugacity and the partial fugacity of ammonia.
These results were used to calculate activities and activity coefficients for ammonia. The results of these calculation are presented in Tables XXII to XXV. t a b u : x x i i CALCULATIONS FOR 88.5°C f° = 557 NHq Liquid Total Total Vapor Compo Partial Activity Activity Partial composi- Vapor Fugacity sition Fugacity of coeffi Fugacity tion mole Proesure psi mole per cent of Ammonia cient of of fraction Absolute Ammonia ammonia Hydrazine ammonia pel NH3 N2 \
• 950 677 531 99.76 .235 529.7 .951 1.001 1.25
.900 633 501* 99.60 .1*00 502.0 .901 1.001 2.02 00 m 0 • 593 1*79 99.1*6 .51*0 1*76.1* •855 1.006 2.59
.800 558 1*57 99.31* .660 l+5l*.0 .815 1.019 3.02
.750 525 1*35 99.24 .761* 1*31.7 .775 1.033 3.32
.700 495 1*15 99.11* .863 1*11.1* .739 1.056 3.58
.650 1*68 396 99.02 .980 392.1 .701* 1.083 3.88
.600 1*37 371* 98.90 1.10 369.9 .661* 1.107 l*.ll
.550 1*06 351 98.76 1 .21* 31*6.6 .622 1.131 1**35
.500 375 328 98.61 1.39 323.1* .581 1.162 l*.56
.1*50 31*5 305 98.1*1* 1.5b 300.2 .539 1.193 4.76 105
.1*00 317 283 98.25 1.75 278.0 .1*99 1 .21*8 4.95
.350 282 256 98.05 1.95 251.0 .1*51 1.289 4.99
.300 21*7 226 97.85 2.15 221.1 • 397 1.323 4.86 TABLE XXIII CALCULATIONS FOP. 100.3°C >o =- 682 NH3 Liquid Total Tdtal Vapor Compo Partial Activity Activity Partial Composi Yapor Fugacity sition Fugacity of coeffi- Fugacity tion mole Pressure psi mole per cent of Ammonia cient of of fraction Absolute NH3 n2h^ Ammonia ammonia Hydrazine ammonia psi
• 950 851 649 99.69 • 31 647.0 .949 .999 2.01
.900 792 615 99.46 .54 611.7 .896 .996 3.32 O CD e 1.008 4.30 VJ1 746 589 99.27 .73 584.7 .857
.800 701 561 99.10 .90 556.0 .815 1.019 5.04
.750 658 534 98.94 1.06 528.3 .775 1.033 5.66
.700 616 506 98.79 1.21 499.9 .733 1.047 6.12
.6 5 0- 577 481 98.64 1.36 474.5 .696 1.071 6.54
.600 539 454 98.51 1.49 447.2 .656 1.093 6.76
.550 4 99 425 98.35 1.65 418.0 .613 1.115 7.01
.500 462 398 98.18 1.82 390.8 .573 1.146 7.24
.450 428 373 98.00 2.00 365.5 .536 1.191 7.46
.400 390 344 97-79 2 .2 1 336.4 .493 1.233 7.60 & .350 350 313 97.50 2.50 305.2 .41+8 1.280 7.83
.300 309 280 97-17 2.83 272.1 • 399 1.330 7.92 TABLE XXIV CALCULATIONS FOR ll4.1°C
Liquid Total Total Vapor Compo Partial Activity Activity Partial Composi Vapor Fugacity sition Fugacity of coeffi- Fugacity tion mole Pressure psi mole per cent of Aamonia cient of of fraction Absolute NH N H Ammonia ammonia Hydrazine amonla psi 3 2 4 pel
.950 1095 808 99.46 .54 803.6 ,942 .992 4,36
.900 1017 767 99.12 .88 760.3 ,891 .990 6.74
.850 952 731 98.86 1.14 722.7 847 .996 8.33
.800 893 697 98.66 1.34 687.7 806 1.007 9-34
.750 838 664 98.49 1.51 654.0 767 1.023 10.0
.700 786 632 98.33 1.67 621.4 728 1.040 10.6
.650 735 599 98.17 1.83 588.0 689 1.060 11.0
.600 683 565 98.02 1.98 553.8 649 1.082 11.2
.550 631 530 97.86 2.14 518.7 608 1.105 11.3
.500 578 492 97.68 2.32 480.6 563 1.126 11.4
.U50 529 457 97.45 2.55 445.3 522 1.160 11.7 107 .400 483 422 97.15 2.85 4io.o 481 1.203 12.0
.350 439 389 96.81 3.19 376.6 442 1.263 12.1+
.300 393 353 96.42 3.58 340.4 399 1.330 12.6 T A B L E X X V CALCULATIONS FOR 121.9°C f = 1002 NH Liquid Total Total Vapor Compo Partial 3 Activity Activity Partial Composi Vapor Fugacity sition Fugacity of coeffi Fugacity tion mole Pressure psi mole per cent of Ammonia cient of of fraction Absolute NH N H, Ammonia Ammonia Hydrazine ammonia pel 3 2 if .950 1316 9I6 99.21 .76 938.8 .937 .986 7.29
.900 1218 897 98.81 1.16 886.6 .885 -.983 10.1
.850 1137 855 98.51 1.16 812.5 .811 .989 12.5
.800 1068 817 98.28 1.72 802.9 .801 1.001 11.1
.750 1000 778 98.06 l.9l 762.9 .761 1.015 15.1
.700 936 7I0 97.87 2.13 721.2 .723 1.033 15.8
.650 868 698 97.70 2.30 681.9 .681 1.018 16.1
.600 807 659 97.51 2.16 612.8 .612 1.070 16.2
.550 7I6 619 97-35 2.65 602.6 .601 1.092 16.1
.500 687 578 97.ll 2.86 561.5 .560 1.120 16.5
.If 50 628 536 96.91 3.09 519.1 .518 1.151 16.6
.100 567 192 96.60 3.10 175.3 .171 1.185 16.7 o .350 513 I51 96.19 3.81 133.8 .133 1.237 17.2 CO
.300 lf62 if 12 95.69 1.31 391.2 .393 1.310 17.8 109
Activities and activity coefficients were not calculated for hydrazine because sufficient data were not obtained in the region of low hydrazine concentration to permit the determination of the fugacity corresponding to the standard state of hydrazine in dilute ammonia solution.
The activity of one of the components in a binary mixture is related at constant temperature and pressure to the activity of the second component by the Gibbs-Duhem equation:
N^d In a^ = N^d In a^= 0 (1 2 )
-d in a - N 1 d In a N£ 1 (13)
In a^ can be obtained upon integration between limits
'-'A/a ' This equation can be solved by graphical integration of a plot of
® 1 ^ 2 T8 al U8*n 8 reported in Tables XXII to XXY for the activity of ammonia. The results of these calculations are listed in
Tables XXVI to XXVIII. 110
TABUS XXVI
Gibbs - Duhoin Calculations for 88.5 C
Mole Log Value of Antilog j Fraction NV Actirity the integral Ratio of Ammonia, 2 of between the Activity Ammonia limits .950 and ,A
• 950 19 -.0218
.900 9 -.01+53 -.3111+ 2 .01+
.850 5.667 -.0680 -.1+722 2.97
.800 1+.000 -.0888 -.5712 3.73
.750 3.000 -.1110 -.61+72 If. 1+1+
.700 2.333 -.1311+ -.7027 5.01+
.650 1.857 -.1521+ -.71+68 5.58
.600 1.500 -.1778 -.7895 6.16
.550 1.222 -.2062 -.8281 6.73
.500 1.000 -.2358 -.8619 7.28
.450 .8182 -.2681+ -.8916 7.79
Aoo .6667 -.3019 -.9167 8.26
.350 .5385 -.31+58 -.9^30 8.77
.300 .1+286 -.1+012 -.9701 9 .31+ I l l
TABLES XXVII Gibbs - Duhem Calculations for 100.3 C.
Mole N Log Value of Antilog Fraction % Activity the Integral Ratio 0: Ammonia, 2 of between the Activit: Anmonia limits .950 *nd H ^ a
.950 19 -.0227
.900 9 -.01*77 -.3385 2.18
.850 5.667 -.0670 -•1*753 2.99
.800 1*.000 -.0888 -.5795 3.80
.750 3.000 -.1107 -..655!* 1*.52
.700 2.333 131*9 -.7191 5 .21*
.650 1.857 -.1571* -.7661 5.81*
.600 1.500 -.1831 -.8093 6«1:5
.550 1.222 -.2125 -.81*90 7.06
.500 1.000 -.21*18 -.8812 7.61
.1*50 .8182 -.2708 -.9076 8.06
.1*00 .6667 -.3071 -.931*8 8.61
.350 .5385 -.3^87 -.9600 9.12
.300 .1*286 -.3990 -. 981*1* 9.65 112
TABUE ZZVIII Gibbs - Duhem Calculations for Hl*..l8C
Mole K. Log Value of Antilog A~ Traction V Activity the Integral Ratio of Ammonia, H of between the Activities Ammonia limits .950 and N a
950 19 -.0259
900 9 -.0501 -.3260 2.12
850 5.667 -.0721 -.1*821* 3.01*
800 If. 000 -.0937 -.581*8 3.81*
750 3.000 -.1152 -.6592 U .56
700 2.333 -.1379 -.7185 5.23
650 1.857 -.1 6 1 8 -.7667 5.81*
600 1.500 -.1878 -.8111* 6.1*8
550 1.222 -.2 1 6 1 -.81*96 7.07
500 1.000 -.21*95 -.8911* 7.79 1 ro B 1*50 .8182 1 0 -.9212 8.31*
1*00 .6667 -.3 1 7 8 -.91*76 8 .8 6
350 • 538S -.351*6 -.9 6 9 8 9.33
300 .1*286 -.3990 -.9865 9.69 It should he possible to compare the ratios of the activities
obtained from the Glbbs-Duhem equation with the measured ratios of
the fugacltles of hydrazine, because the standard state is canceled
out when a ratio of activities is determined. This comparison may
be made from the data listed in Tables XXIX to XXXI. Listed in the
second column in each table are the ratios of the activities of
hydrazine at the ammonia mole fractions listed in the first column to
the activity at an ammonia mole fraction of .950 calculated by the
Gibbs-Duhem equation.
The third column contains similar ratios determined from the measured values of the partial pressures. The fourth column contains
the value for the calculated ratio divided by the measured ratio at each mole fraction. ill*
TABUS XXIX Activity Ratios at 88.5 C Mole Ratio of Acti- Ratio of Calculated Traction vities calculated Measured ratio______Anaemia from the Gibbs- Activities Measured Duhem Equation ratio
.950
.900 2 .01* 1.62 1.26
.850 2.97 2.07 1.1*3
.800 3.73 2.1*2 1.51*
.750 1*.1*1* 2 . 6 6 1 . 6 7
.700 5 .01* 2.86 1.76
.650 5.58 3.10 1.80
.600 6 .1 6 3.29 1.88
.550 6.73 3.1*8 1.91*
.500 7.28 3.65 1.99
.1*50 7.79 3.81 2.05
.1*00 8.26 3.96 2 .0l*
.350 8.77 3.99 2.20 .300 9.3U 3.89 2.39 115
TABLE XXX c Activity Ratios at 100.3 C
Mole Ratio of Activities Ratio of Calculated Fraotion Calculated from the Measured ratio Ammonia Gibbs-Duhem Equation Activities Measure ratio
.950
.900 2.18 1.65 1.32
.850 2.99 2.15 1.39
.800 3.80 2.51 1.52
.750 *.52 2.82 1.60
.700 5.21* 3.01* 1.72
.650 5.81* 3.25 1.80
.600 6.1*5 3.36 1.92
.550 7.06 3.1*9 2.02
.500 7.61 3.60 2.11
.1*50 8.08 3.71 2.18
.1*00 8.61 3.80 2.27
.350 9.12 3.90 2.31*
.300 9.65 3.9^ 2.1*5 116
TABLE ZZZI Activity Ratios at lll*.l°C
Mole Ratio of Activities Ratio of Calculated Fraction Calculated from the Measured ratio_____ Ammonia Gibbs-Duhem Equation Activities Measured ratio
-950
,900 2 .1 2 1.51* 1*37
850 3.0k 1.91 1.59
,800 3.8k 2.ll* 1.79
750 U .56 2 .2 9 1.99
700 5.23 2.1*3 2.15
650 5.8k 2.5 2 2.31
600 6.U8 2.57 2 .5 2
550 7.07 2.59 2.73
500 7.79 2 .6 1 2 .9 8
1*50 8.3U 2 .6 8 3.11
koo 8 .8 6 2.75 3 .2 2
350 9.33 2.81* 3 .2 8
300 9 .6 9 2 .8 9 3.35 117
Considerable deviation of the ratio of the activities for
hydrazine calculated on the basis of the Gibbs-Duhem equation from
the experimentally measured activity ratios for hydrazine is observed.
Ratios for both the calculated and measured values are taken on the
basis of the activities measured at a mole fraction ammonia of .950
in the liquid. Therefore, an error in the determination of activities
at this concentration would cause a difference between the calculated
and measured ratios. The experimental error in the determination of
the vapor composition is greatest in the liquid composition regions which are dilute with respect to hydrazine, for only very small amounts of hydrazine are present in the vapor. Measurement in this
region requires both the withdrawal of large vapor samples which
disturbs equilibrium and the analysis of email amounts of hydrazine.
However, these experimental errors would become less important as the temperature Increases. An opposite trend is observed indicating that
experimental measurements are not the main cause of the deviation.
Dividing the calculated ratio by the measured ratio should give a factor the constancy of which is independent of any one chosen standard and therefore, should give an indication of the true deviation of the measured from the calculated activities at various concentrations. As can be observed in Tables XXIX to XXXI this factor increases as the concentration of hydrazine in the liquid Increases.
If the measured fugacity of hydrazine is lower than the true value in the regions more concentrated in hydrazine, the measured activity of ammonia would be higher than the correct value. These deviations 11 8 would be magnified by the G-ibbe-Duhem equation because the high value
for the activity of ammonia gives a large calculated ratio for the activity of hydrazine, while the low measured fugacity of hydrazine
gives a low value for the measured ratio of the activities.
The differences between the values calculated from the Gibbe-Duhem equation and the measured values can be explained by discrepancies in the assumptions used to calculated the activities of ammonia and the fugacities of hydrazine. Initially it was assumed that the fugacity of the gaseous mixture could be calculated from the pressure-fugacity relationship for ammonia. A lack of validity of this assumption would cause greater error in the activity calculations as the concertration of hydrazine in the vapor increases. The hydrazine concentration in the vapor is increased both as the liquid concentration of hydrazine Is increased and as the temperature is increased. The greater error in the activity calculation correspondingly occurs at the higher temperatures and in the liquid regions more concentrated with respect to hydrazine. Accurate calculations correcting for these errors require more accurate P-V-T data for hydra zine than are currently available. A different relationship for calcula ting the viral coefficient of a mixture from the viral coefficients of the components may also be required.
An additional assumption which might not have been valid was invoked to calculate the partial fugacity of ammonia and hydrazine from the total fugacity. It was assumed that the partial fugacities were eeual to the 66 mole per cent of the component times the total fugacity t Considerable deviation from this assumption have been reported at pressure above twenty- 7'/ five atmospheres 119
The errors involved In making these assumptions are thus too great to permit a thermodynamic evaluation of the system.
D. CONCHJSIQRS
The vapor pressure and vapor composition in equilibrium vith liquid mixtures of hydrazine and anmonla of measured composition were determined at 88.5, 100.3, llk.l, and 12k.9°C, These studies covered the liquid composition range of 0 to 80 mole per cent hydrazine.
Bo b u Its were not obtained above liquid compositions of 80 mole per cent hydrazine because serious hydrazine decomposition occurrs in this region even at the lowest temperature studied, 8 8 .5°C.
Activities were calculated for the components and were compared to the ratio of activities calculated from the Gibbs-Duhem equation.
Considerable differences in the measured and calulated activity ratios were observed. These deviations were attributed to assumptions made in calculating the activities from the experimental data. The activities calculated were not accurate enough to permit the evaluation of other thermodynamic constants. Chapter VIII
Summary
A. The synthesis of chloramine by the reaction of ammonia and
chlorine in the gas phase was studied to determine the factors which affect the yield of chloramine. The results of these studies led to
the following conclusions.
1. The total flow of gas is a more important factor in determining the yield of chloramine than is the ammonia to chlorine mole ratio. In the presence of a considerable excess of ammonia over the stoichiometric requirement, high yields of chloramine can be obtained independent of the ammonia to chlorine mole ratio as long as the total rate of gas flow is above the critical value for the reactor. A high yield of chloramine was obtained at the low ammonia to chlorine mole ratio of 8 to 1 .
2. The decomposition of chloramine by chlorine is probably the major yield-reducing reaction in this synthesis. The complete oxidation of ammonia to nitrogen results in the appearance of a flame in some cases.
3. Proper design of the ammonia inlet is necessary to eliminate currents which carry chloramine back into the chlorine stream as soon as it is formed in order to realize high chloramine yields. This was accomplished efficiently in the reactor employing the filter thimble as the ammonia inlet.
4. The use of the spinnerette for the chlorine inlet made it possible to carry out the reaction for indefinite lengths of time without plugging of the chlorine inlet by ammonium chloride. 121
B. A quantitative study of the effect of caustic in the reaction of chloramine with aqueous ammonia to produce hydrazine has shown the following:
1. Caustic serves only two purposes in this reaction (a) to neutralize the acid formed in the reactions which would otherwise form ammonium ion which has been shown to adversely affect the yield of hydrazine and (b) to stablize a gelatin-metal ion complex, the forma tion of which prevents the metal-ion catalysis of the chloramine- hydrazine reaction. The gelatin is ineffective as an inhibitor when fixed base is not present to stablize the complex.
2. Excess alkali over that required for (a) and (b) is deleterious to the process. Increases in hydrazine yield over those produced in the Raschig process were obtained by adding caustic to the reaction as the chloramine was being added thus avoiding a large caustic concentration at any time during the experiment.
3. These results have given support to the Cahn-Powell mechanism for the Raschig synthesis, compared to the mechanism proposed by Audrieth et al; for according to the latter an excess of sodium hydroxide should favor the hydrazine producing reaction.
k. The reaction producing hydrazine proceeds very rapidly at temperatures in the range of 21 to 21t°C. Yields similar to those obtained in the Raschig process at elevated temperatures were obtained in this study at the low tempe rat vires of 21 - 2h°c.
C. The addition of fixed base to the chloramine-ammonia reaction in liquid ammonia and in ethyl alcohol has been studied. The addition of gelatin and potassium amide to the liquid ammonia solution, as veil 122
as gelatin and potassium alcoholate addition to ethyl alcohol solutions saturated with amnonia, was shown to have no effect upon the yield of hydrazine.
The addition of potassium and gelatin to liquid ammonia reduced the hydrazine yield produced by the chloramine-ammonia reaction. The low yields were probably caused by the reduction of chloramine by the potassium.
D. Experiments have shown that hydrazine can be separated from the reaction product of the chloramine-ammonia reaction by a high temperature-high pressure distillation.
Very serious decomposition of hydrazine occurred in the stainless steel autoclave at the elevated temperatures. This decomposition problem was overcome by coating the apparatus with Teflon enamel.
E. The composition and pressure of the vapor in equilibrium with liquid mixtures of hydrazine and ammonia were determined at 8 8 .5 ,
100.3, 114.1, and 124.9°C, throughout the liquid compositions range of 0 to 80 mole per cent hydrazine. Activities were calculated for the components and compared with the values calculated from Gibbs-
Duhe®, Equation. The data obtained did not permit the calculation of other thermodynamic constants. 123 BIBLIOGRAPHY
1. Audrleth, L. F. and Ogg, B.A., The Chemistry of Hydrazine. John Wiley and Sons, New York, New York, 1951, pp. 224-232.
2. Raschig, Schwefel-und Stlckatoffetudien. Verlag Chemie G. m. b. H., Leipzig, Berlin, 192^.
3. Mattair and Sisler, J. Am. Chem. Soc., "JA, 1619 (1951).
U. Boatman, Claude E., Dissertation, The Ohio State University, (195I).
5. Neth, Floyd T., Dissertation, The Ohio State University, (1951).
6 . Hurley, Forrest R., Dissertation, The Ohio State University, (195U).
7. Sheliman, Richard W., Thesis, The Ohio State University, (1953).
8 . Kelmers, A. Donald, Thesis, The Ohio State University, (1953).
9 . Simon, Scherer's Journal, 2, 583 (1802).
10. Donny and Mareska, Compt. Rend., 20, 817 (l81*5).
11. Valentini, Gass. Chim. ltd., 1^4, 2lU (188U).
12. Schwarz and Striebich, 2. fur Anorg. Allgem. Chem., 22k, 29-32 (1935).
13. Selivanoff, Ber., 2£, 3617 (1892): 2J_, 1*33, 1010(189^).
1^. Noyes and Lyon, J. Am. Chem. Soc., 22, 1*60 (1901).
15. Bray and Dowell, J. Am. Chem. Soc., 22> 905 (1917).
16. Bray and Dowell, J. Am. Chem. Soc., 39» 896 (1917)*
17. Noyes and Haw, J. Am. Chem. Soc., k2, 2l6^ (1920).
18. Marckwald and Wilie, Ber, 56B , 1319_25 (1923).
19. German Patent 198,307, C.A., 2, 2629 (1908): J Chem. Soc., 9** 11. 1029) (1908)
20. Nagasawa, J. Chem. Soc. Japan, Pure Chem. Sect., 6 9 , 17 (19^8)
21. Moeller, Kgl, DanBke Videnskab. Selskab. Math-fys Medd, 12(No.l6) 1-61 (193*0 •
22. Bodenstein, Z. physik. Chem., 139A. 397-^15 (1928)
23. Joyner, J. Chem. Soc., 122. 111^-1121 (1923).
21*. Thompson and Joy, Report No. AL-312. North American Aviation, Inc. 12U
25. Pfeiffer and Simons, Chem. Ber., 80, 127-128 (1927).
26. McCoy, J. Am. Chem. Soc., 7 6 , 1UU7 (195U).
2 7 . Cahn and Povell, J. Am. Chem. Soc., J6, 2565-67 (195*0.
28. L. Meitea, J. Am. Chem. Soc., 21, 3269 (19U9 ).
2 9 . Audrleth, Colton and Jonee, J. Am. Chem. Soc., j6, 11*28-31 (195*0.
30. Riley, Richter, Rothram, Todd, *fyers and Nuebaum, J. Am. Chem. Soc., 26, 3301 (195*0.
31. Slsler, Boatman, Neth, Smith, Shellman, and Kelmers, J. Am. Chem.
Soc., 7 6 , 3912 (195*+).
32. Slaler, Neth and Hurley, J. Am. Chem. Soc., j 6 , 3909 (I95U).
33. Lobry de Bruyn, Ber., 28, 3085-3086 (1895)*
3l*. Lobry de Bruyn, Rec. tray. Chlm., 13. l*33~*+^0 (189**).
35. Lobry de Bruyn, Rec. tray. Chlm., 1^, 17*+-l8** (1896).
3 6 . Office of the Publication Board, Department of Commerce, Washington, D. C., C.I.O.S. Trip 215, Report 80.
37. Raschig, Ber., 1*3, 1927 (1910).
38. Hale and Shetterly, J. Am. Chem. Soc., 1071-1076 (1911)*
39. Walden and Hllgert, Z. physik. Chem., 165 A . 21*1-271 (1933).
1*0. Welsh, J. Am. Chem. Soc., 1*97-508 (1915).
1*1. Browne and Welsh, J. Am. Chem. Soc., 1728-173*+ (1911).
1*2. Browne and Houlehan, J. Am. Chem. Soc., 33.173^-17^2 (1911).
1*3. Mohr, Thesis, University of Illinois, (19^7).
1*1*. Audrleth and Mohr, J. Phys. Chem., 53. 901-906 (l9*+9)
1*5. Audrleth and Johnstone, unpublished work based on Chemical Engineering Reports by A. Tellez (19*0), G. Goodman (19*0)> R. Scharmer and H. Mueller (19*^8), W. E. Henderson (19**9).
1*6. Hurd and Bennett, J. Am. Chem. Soc., £1, 265-269 0929).
1*7. Curtius and Jiff, J. prakt. Chem., (2), 22, 27-58 (1 8 8 9 ).
1*8. Curtius, German Patent 1*7,600, Chem. Zentr., 60 ii, 719 (1889). 125
1+9. Audrleth and Nickles, Inorganic Synthesis, 1, 92-94 (1939).
50. Curtius and Schrader, J. prakt. Chem. (2), 3 0 , 311-346.(189I+)
51. Moncrleff, Mfg. Chemist, 18, 177-182 (19U7 ).
52. Schlenk and Weichselfelder, Ber., 48, 669-676, (1915).
53. Stolle', J. prakt. Chem., 8^,200 (1911).
54. Sisler, Neth, Boatman, and Shellman, J. Am. Chem. Soc., J6, 3914 (1954).
55. Schvarcz, U . S. Patent No. 2,537,791*
5 6 . Audrleth and Ogg, Loc. Clt. PP 67-8 5 .
57* Scott, Oliver, Gross, Hubbard, and Huffman, J. Am. Chem. Soc., 7 1 , 2293-2297 (19^9)-
5 8 . Hieber and Woerner, Z. Elektrochem, 40, 252-256 (1934).
59. Glguere and Rundle, J. Am. Chem. Soc., 6^, 1135-1137 (1941).
6 c . Fresenius and Karveil, Z. physik Chem.. 44 B 1-13 (1939).
61. Scott, Jones, and Levis, Report of Investigations 4460, May, 1949, U. S. Department of Interior, Bureau of Mines.
62. I. Kolthoff, J. Am. Chem. Soc., 46, 2009 (1924).
6 3 . Caldvell, J. and Moyer, H. V., Ind. Eng. Chem. Anal. Ed.. 7 . 38 (1935).
64. Garrett, A. B., Ind. Eng. Chem., 10 324 (1938).
6 5 . Cragoe, Mpyere, and Taylor, J. Am. Chem. Soc. 42, 206, (1920).
6 6 . Levis, G. N. and Randall, M. Thermodynamics. McGrav-Hill B00K Co. Inc. New York, N. Y. (1923).
6 7 . Friedman, A. S., Dissertation, The Ohio State University (1950).
6 8 . Fowler, R. and Guggenheim, E. A., Statistical Thermodynamics. Cambridge University Press, Cambridge England (1949).
6 9 . Stockmeyer, J. Chem. Fhys.,
70. Keyes, J. Am. Chem. Soc; 6©, 1761 (1938)
71. Glasstone, Textbook of Physical Chemistry, D. Van Nostrand Company, New York, N. Y. pp 306-308 126
AUTOBIOGRAPHY
I, Russell Stephen Drago, was horn In Turners Falls,
Massachusetts, November 5> 1928. I received my secondary school education in the public schools of Turners Falls, Massachusetts.
My undergraduate training was obtained at The University of
Massachusetts, from which I received the degree Bachelor of
Science in 1950. I was called to active duty by the Air Force in 1951, at which time I enrolled in The Ohio State University
Graduate Center at Wright-Bitterson Air Force Base. Upon comp letion of my tour of active duty in 1952, I received an appoint ment as Research Fellow in The Ohio State University, where I specialized in the Department of Chemistry. I held this appoint ment for two years and then accepted an appointment as Graduate
Assistant while completing the requirements for the degree Doctor of Philosophy.