EXCHANGE 
 
The Adsorption of Ammonia by 
 Silica Gel 
 
 DISSERTATION 
 
 SUBMITTED TO THE BOARD OF UNIVERSITY STUDIES 
 
 OF THE JOHNS HOPKINS UNIVERSITY IN CONFORMITY 
 
 WITH THE REQUIREMENTS FOR THE DEGREE 
 
 OF DOCTOR OF PHILOSOPHY 
 
 BY 
 
 LEVI YORGEY DAVIDHEISER 
 
 BALTIMORE 
 
 June, 1922 
 
 EASTON, PA. : 
 
 ESCHENBACH PRINTING COMPANY 
 1922 
 
The Adsorption of Ammonia by 
 Silica Gel 
 
 DISSERTATION 
 
 SUBMITTED TO THE BOARD OF UNIVERSITY STUDIES 
 
 OF THE JOHNS HOPKINS UNIVERSITY IN CONFORMITY 
 
 WITH THE REQUIREMENTS FOR THE DEGREE 
 
 OF DOCTOR OF PHILOSOPHY 
 
 BY 
 
 LEVI YORGKY DAVIDHEISER 
 
 BALTIMORE 
 
 June, 1922 
 
 EASTON, PA.: 
 ESCHENBACH PRINTING COMPANY 
 
 1922 
 

 CONTENTS 
 
 Acknowledgment 4 
 
 Introduction 5 
 
 Apparatus 6 
 
 Material 6 
 
 Procedure 6 
 
 Table I 7 
 
 Figure I 8 
 
 Table II 9 
 
 Figure II 9 
 
 Figure III 10 
 
 Figure IV 11 
 
 Biography 13 
 
 043931 
 
ACKNOWLEDGMENT 
 
 The author wishes to express his thanks and appreciation to Dr. W. A. 
 Patrick under whose instruction this work was carried on for the aid and 
 suggestions received. 
 
 I also wish to take this opportunity to extend my sincere thanks to 
 Drs. Frazer, Reid, Lovelace and Thornton for their kind assistance in the 
 laboratory and lecture room. 
 
THE ADSORPTION OF AMMONIA BY SILICA GEL. 
 
 This investigation is a continuation of the studies of Patrick of adsorp- 
 tion of gases and vapors by silica gel. In Gottingen he studied the 
 adsorption of ammonia, sulfur dioxide and carbon dioxide. Since then 
 many improvements in experimental technique ha\e been introduced, as 
 well as radical modifications of the theoretical treatment of the experi- 
 mental results. The experimental method and an outline of the theoretical 
 views ha\e been given in a paper dealing with the adsorption of sulfur 
 dioxide. 1 
 
 The adsorption of ammonia was studied for a number of reasons. In 
 the first place, this gas, in the measurements made in Gottingen, was 
 found to be more strongly adsorbed than sulfur dioxide although the 
 latter exhibits a higher critical temperature. In the second place, our 
 previous measurements indicated that the behavior of ammonia was 
 anomalous from the standpoint of the rate at which equilibrium was 
 reached. Again, ammonia is extremely soluble in water and it was there- 
 fore hoped that by a careful study of its adsorption by silica gel, light 
 would be thrown on the question as to the nature of the small amount 
 of water that is always associated with the gel. Furthermore, in am- 
 1 MacGavack and Patrick, /. Am. Chem. Soc., 42, 946 (1920). 
 
monia we have a gas that has been the subject of numerous careful in- 
 vestigations, and as a result the physical constants, such as vapor pres- 
 sure, density and surface tension, are very accurately known. This latter 
 point is especially important inasmuch as a knowledge of the above con- 
 stants is necessary for the testing of our theoretical views as to the course 
 of the adsorption. 
 
 In the following measurements of the adsorption of ammonia by silica 
 gel, it will be shown that this gas is not anomalous in its behavior, either 
 from the standpoint of the extent or the rate of the adsorption, the earlier 
 discrepancies being due to the uncertain water and acid content of the 
 gel. Especial study was made of the influence of the water content of 
 the gel on the adsorption, and it was found that small differences 
 produced large variations in the adsorptive ability of the material. 
 
 Apparatus. The experimental method has been described in detail 
 by MacGavack and Patrick. 1 To give an idea of the precision attained, 
 it may be stated that the volume of ammonia was measured to within 
 0.005 cc., the pressure to within 0.03 mm., and the temperature of the 
 thermostat regulated so as not to vary more than 0.05. 
 
 Material. The ammonia used was purified liquid ammonia. Com- 
 mercial ammonia was treated with metallic sodium for one week, and 
 the accumulated gas allowed to escape at intervals of 12 hours. The 
 purified sample was tested by allowing a small stream of the gas to bubble 
 into sulfuric acid contained in a gas buret for 10 minutes; as no residue 
 was shown by this procedure it was assumed that the ammonia was free 
 from permanent gases. 
 
 The silica gel was an ordinary commercial sample that was further 
 purified by treatment with nitric acid and a thorough washing with dis- 
 tilled water. When dried in a vacuum of 1 to 5 mm. at a temperature 
 of 300 for 3 hours the gel still contained 5.21% of water, and further- 
 more an appreciable amount of nitric acid, to which reference will be 
 made later. 
 
 Procedure. The adsorption apparatus was evacuated by means of 
 a rotary oil pump and a Gaede mercury pump, connected in series. Before 
 any gel was put into the adsorption bulb, the apparatus was thoroughly 
 evacuated and then swept out with ammonia and evacuated again until 
 the MacLeod gage showed no pressure after standing under a vacuum 
 for 4S hours. The gel was weighed in the adsorption bulb, which was 
 then directly attached to the apparatus by means of a ground joint, and 
 sealed by mercury. The pumps were always started before the stopcock 
 between tin adsorption bulb and the main apparatus was opened, so that 
 the air in tin- bulb would have- k-ss rhauce to become adsorbed on the walls 
 of the apparatus. The heating of tin- adsorption bulb was started at the 
 same time, care briiijj taken not to heat the gel to a higher temperature 
 
than that at which it was prepared, so that the water content might not 
 be disturbed. The evacuation was continued until the MacLeod gage 
 showed no pressure, usually from 3 to 8 hours, at a temperature from 
 290 to 300. 
 
 In the beginning it was found almost impossible to make two determ- 
 inations that would agree. After a considerable number of measure- 
 ments had been made at 30, a few were found to check fairly well. How- 
 ever, a number of measurements made at 40 showed that the gel exhibited 
 greater adsorptive ability at this temperature than at- 30. Such in- 
 consistent and disturbing results led us to stop and thorougly inspect 
 our apparatus and method of procedure. The apparatus was first exam- 
 ined for leaks. It was thoroughly evacuated and allowed to stand for 
 one week, at the end of which the MacLeod gage showed no increase of 
 pressure. The ammonia was again examined and no foreign gases found. 
 Another series of measurements was made at 30 in which great care 
 was taken that all manipulations should be as nearly identical as possible. 
 The electric furnace, in which the adsorption bulb was heated during 
 evacuation, was kept constant and evacuation continued for exactly the 
 same length of time. Under these conditions checks could be made with 
 but little variation. This proved that the gel did not remain constant 
 
 TABLE I 
 EXPERIMENTS AT 0, 30, 40 AND 100 
 
 Expt. XX 
 
 Temp., 30; H 2 O content, 4.93%; Wt. 
 of gel., 0.5739 g.; D, 0.5939; a, 18.03; 
 I/TV, 0.2103; #,57.30; P , 874.90. 
 P X/M v 
 
 0.07 26.82 0.034 
 1.009 68.05 0.087 
 37.726 126.66 0.161 
 
 Expt. XXIII 
 
 Temp., 40; H 2 O content, 4.93%; Wt. 
 of gel, .5730 g.; D, .5769; <r, 16.70; 
 l/N, 0.2093; K, 53.53; P , 1165.80. 
 
 Expt. XXV 
 
 Temp., 0; H 2 O content, 4.93%; Wt. 
 
 of gel, .5744 g.;D,0 .6389; <r, 25 . 94; 
 
 Po, 322.10; l/N, 0.2116; K, 82.40. 
 
 Pff/P 
 
 
 p 
 
 X/M 
 
 V Pa/Po 
 
 0.001443 
 
 
 
 .153 
 
 58 
 
 .77 
 
 
 
 .070 
 
 
 
 .012321 
 
 0.020792 
 
 8 
 
 .116 
 
 123 
 
 .47 
 
 
 
 .147 
 
 
 
 .654614 
 
 0.683757 
 
 33 
 
 .494 
 
 168 
 
 .59 
 
 
 
 .200 
 
 2 
 
 .697404 
 
 
 58 
 
 .653 
 
 191 
 
 .99 
 
 
 
 .228 
 
 4 
 
 .723902 
 
 P X/M 
 0.218 36.42 
 4.026 77.03 
 20.987 101.06 
 61.632 126.25 
 
 V P*/P 
 0.050 0.003123 
 0.101 0.057672 
 0.133 0.300622 
 0.166 0.882887 
 
 P = equilibrium pressure in cm. of 
 
 mercury. 
 X/M = cc. of ammonia under 
 
 standard conditions adsorbed per 
 
 g. of gel. 
 V = cc. of liquid ammonia ad- 
 
 sorbed. 
 
 Expt. XXIX 
 
 Temp., 100; H 2 O content, 4.93%; Wt. 
 of gel, 0.5737 g.; D, 0.4589; a, 6 . 5 
 I/TV, .2885; K, 23 .07; P , 4693 .40. 
 P X/M v P<r/P 
 
 2.147 27.05 0.04473 0.00276 
 13.805 52.22 0.08636 0.01912 
 41.016 66.62 0.11017 0.05681 
 Po = vapor pressure of liquid ammonia. 
 D = density of liquid ammonia, 
 o- = surface tension of liquid ammonia. 
 
' 
 
 8 
 
 during evacuation, and especially was this true when the pumps 
 were run for different lengths of time. All the experiments recorded in 
 this paper were made under conditions that allow exact duplication. 
 
 In Table I Experiments XX, 
 XXIII, XXV and XXIX are re- 
 corded the results of the measure- 
 ments made at 0, 30, 40 and 
 100. These results are shown 
 graphically in Fig. 1. Equilibrium 
 was usually reached in from 15 to 
 30 minutes; nevertheless the am- 
 monia was allowed to remain in 
 contact with the gel for at least 2 
 hours. As more ammonia was in- 
 troduced and the pressure increased, 
 the time required to reach equil- 
 
 100 
 
 80 
 
 60 
 
 40 
 
 20 
 
 20 
 
 40 
 
 60 
 
 80 
 
 ibrium also increased. If equil- 
 ibrium was not reached within 2 
 hours after the introduction of the 
 first amount of ammonia, it served 
 as an indication that the air had 
 not been completely removed from the apparatus. It was found that 
 3 hours' evacuation with the adsorption bulb heated to 290 was suf- 
 ficient to remove all the air. All the determinations were made 
 under these conditions. After this treatment it was found that the gel 
 lost 0.28% of its water content. When this treatment was continued for 
 8 hours 0.487% of the water was removed. The gel upon analysis was 
 found to contain 0.437% of nitric acid, equivalent to 5.23 cc. of am- 
 monia. Of course as the gel lost some of its water content some of the 
 acid also was lost. 
 
 Inasmuch as our preliminary measurements showed that the water 
 content of the gel was a most important factor in determining the extent 
 of the adsorption, an effort was made to. prepare a gel of low water content. 
 It has previously been thought that the presence of the water was very 
 closely associated with the structure of the material and the removal 
 of the water below 2% was considered undesirable. The subsequent 
 results, however, indicate that such views are unsound and that it is 
 possible to obtain a gel possessing the proper adsorptive structure having 
 only 0.33% of water. 
 
 A gel containing 0.33% of water was prepared by heating it in a Pyrex 
 glass tube through which a current of dry air was passed. The tube 
 was gradually heated with a Bunsen flame for 1 hour, and then 
 in the flame of the blast lamp for 2 hours to the fusion point of Py- 
 
9 
 
 rex glass. During this process the gel became perfectly clear and color- 
 less. 
 
 After this treatment the gel was tested for its ability to adsorb ammonia. 
 It was rather surprising to find that it was still a good adsorbent for am- 
 monia, adsorbing on an average about 53 cc. less per gram than it did 
 before this drastic treatment. Two isotherms were made, one at 
 and one at 30. Fig. 3 shows the graph of these determinations as com- 
 pared with those made before dehydrating the gel. Table II, Experi- 
 ments XXXI and XXXV show the experimental results. 
 
 ACTION OF 
 Kxpt. XXXI 
 
 Temp., 30; H 2 O content, .33%; 
 of gel, 0.5739 g.; D, 0.5939; a, 18 
 l/N, 0.2487; K, 21.33; P , 874. 
 
 P X/M V 
 
 1.293 24.49 
 
 8.653 46.00 
 
 18.933 57.08 
 
 36.600 73.50 
 
 59.697 83.01 
 
 TABLE II 
 GEL AFTER HEAT TREATMENT 
 
 Expt. XXXV 
 
 Wt. Temp., 0; H 2 O content, .33%; Wt. 
 
 .03; of gel, 0. 5739 g.;D,0. 6389; <r, 24.94; 
 
 90. l/N, 0.3948; K, 28.765; P , 322.10. 
 
 Pff/P P X/M V P<r/P 
 
 0.05684 0.273091 
 
 0.07287 0.544651 
 
 0.09567 1.124572 
 
 0.15145 3.708976 
 
 0.17276 4.941874 
 
 0.03113. 0.026646 
 0.05878 0.178021 
 0.07294 0.390173 
 0.09392 0.754265 
 0.10608 1.23024 
 
 3.391 47.85 
 6.763 61.35 
 13.964 80.54 
 46.056 127.50 
 61.365 145.44 
 
 2.6 
 
 1.8 
 1.4 
 
 I 
 
 *JZ 
 
 It is necessary to mention a phenomenon that was observed with the 
 measurements made at 100 . After 33.64 cc. of ammonia had been brought 
 over 0.5741 g. of gel and the pressure had reached a value of 287.35 mm., 
 it was noticed that there was a gradual increase in the pressure. This 
 increase in pressure continued 
 until at the end of 9 days the 
 pressure in the apparatus was 
 471.40 mm. Upon cooling the 2 2 
 gel to 0, the pressure imme- 
 diately dropped to 4.53 mm., 
 while heating again to 100 
 caused an immediate rise in 
 pressure to the value of 471.40 
 mm. That this phenomenon 
 is in some way associated with 
 
 <? 
 
 1.0 1.4 .18 0.2 0.6 1.0 1.4 1.8 
 
 the water of the gel is indicated by the fact that no such behavior as 
 outlined above was exhibited with the gel containing only 0.33% of water. 
 
 Discussion of Results. 
 
 The various isotherms may be very accurately represented by the 
 ordinary empirical adsorption formula, X/M = KP 1 ' . In Fig. 2 are 
 plotted the logarithms of X/M against P, and from the straightness of 
 the lines it is apparent that the above formula accurately reflects the 
 
10 
 
 experimental data. Inasmuch as the sulfur dioxide was also well rep- 
 resented by the same relationship, the comparison of the behavior of 
 ammonia and sulfur dioxide is greatly facilitated by use of the values of 
 K at the same temperature. In the following table are given the values 
 of K for sulfur dioxide and ammonia obtained with a gel of approximately 
 the same water content, 4.8%. 
 
 
 
 30 
 
 40 
 
 100 
 
 NH 3 
 K 
 
 82.40 
 57.30 
 53.53 
 23.01 
 
 S0 2 
 
 K 
 
 29.14 
 12.93 
 
 9.75 
 
 1.12 
 
 Inasmuch as the significance of K is the number of cubic centimeters 
 of gas adsorbed by 1 g. of gel at a pressure of 1 cm., it is apparent that 
 ammonia is more strongly adsorbed than sulfur dioxide by this particular 
 sample of gel. 
 
 That this result is due to the larger solubility of ammonia in water, is 
 
 shown by the fact that the value of 
 K for ammonia at decreased from 
 82.40 to a value of 28.76 when the 
 water con tent was reduced from 4.88% 
 to 0.33%. While it is true that the 
 adsorption of sulfur dioxide is lessened 
 by a diminution of the water con- 
 tent of the gel, the effect here is not 
 so pronounced as in the case of am- 
 monia. It will be observed that 
 with a gel containing 0.33% water 
 content, the adsorption of ammonia 
 
 10 
 
 20 
 
 20 
 
 40 
 
 60 
 
 80 
 
 is lower than that of sulfur dioxide 
 in a gel containing 4.87% water. 
 It is our belief that with an anhy- 
 drous gel the adsorption of ammonia 
 will be found to be in accordance 
 with its ease of condensation as expressed by its critical temperature. 
 
 The above results for ammonia must also be corrected for the acid that 
 is present in the gel. In the beginning it was thought that prolonged 
 washing with distilled water was sufficient to remove all acid, especially 
 inasmuch as the acid diffuses readily through the gel when in the state 
 that it is when washed. Nevertheless, upon collecting the liquid that 
 is driven off from a gel containing .V21% water by heating to 800-900, 
 the same was found to require 0.653 cc. of 0.1066 7V alkali for the neutral- 
 ization per gram of gel. This acid would probably not exert an appre- 
 
11 
 
 ciable effect upon an acidic gas such as sulfur dioxide, but would mate- 
 rially affect the adsorption of ammonia. 
 
 The above facts illustrate the great need for care when generalizations 
 are drawn concerning adsorption. It is apparent that the amount of 
 ammonia that the gel is capable of taking up depends not only on the 
 capillary structure of the gel, but in addition upon the water and acid 
 content. Before generalizations as to the effect of capillarity on ad- 
 sorption may be drawn, it is 
 necessary to allow for the 1.4 
 solubility of the gas in the ~ Q 
 water and the amount of am- _ 
 monia that combines with 2 -6 
 the acid present. 
 
 If instead of expressing the 
 weight of ammonia taken up 
 
 3.4 3.8 2.2 2.6 1.0 1.4 1.8 0.2 0.6 0.1 
 
 by a unit weight of gel under a certain pressure, the volume of liquid 
 ammonia adsorbed per gram of gel under its condensation pressure is 
 used, all the isotherms fall on a smooth curve which may be expressed by 
 the equation 
 
 V=K( ) 
 
 * 
 
 These results are shown in Fig. 4, where will be found 
 
 plotted the logarithms of the volume of ammonia adsorbed against the 
 logarithms of the condensing pressures (P$/P ). 
 
 A similar result was obtained with the sulfur dioxide measurements 
 where the values of K and l/N were 0.1039 and 0.447, respectively. In 
 the case of ammonia with a gel of 4.88% water content, the equation 
 
 becomes V = 0.1679 ( ) 0.2105; while with a gel of 0.33% water content 
 
 K has the value of 0.0955 and l/N of 0.3588. 
 
 At first thought it may be considered strange that the above formula 
 should apply to the adsorption of ammonia when we consider the very 
 large portion of the ammonia that is dissolved in the water of the gel. 
 This behavior is immediately explained, however, from the fact that the 
 solubility of ammonia in water may be expressed by the same formula 
 that is used to express the results of the adsorption of ammonia by silica 
 gel. In other words, as shown by measurements made in this laboratory, 
 the solubility of ammonia in water is well represented by the formula, 
 
 V = . 
 
 It has also been shown that the solubility of sulfur dioxide in water 
 may be well represented by the above formula where K equals 0.0123 
 and l/N equals 0.91. It is therefore possible to calculate the adsorption 
 
12 
 
 due to the capillary structure alone by correcting, with the aid of the 
 above formula, for the amount of gas dissolved in the water of the gel. 
 
 /P<r\ l/N /Pff\ 1/N ' 
 
 Applying this formula, V=Kl J +0.04S#'( ) , to the measure- 
 ments of ammonia made with the gel containing 4.8% water, the value 
 of K becomes 0.1147. Doing the same with the measurements of sulfur 
 dioxide, we find that the mean value for K does not differ much from 
 0.1038. The K for the anhydrous gel is 0.0955. From these experiments 
 we see that the amount of different gases adsorbed due to capillarity is 
 in good agreement. 
 
 Summary. 
 
 1. The adsorption of ammonia by silica gel has been measured under 
 various pressures at 0, 30, 40, and 100. 
 
 2. The influence of the water content of the gel on the adsorption has 
 been studied. 
 
 3. It has been shown that the adsorption of ammonia may be satis- 
 factorily explained on the basis of capillary condensation, provided cor- 
 rections are made for the amount of the gas that dissolves in the water. 
 
BIOGRAPHY 
 
 Levi Yorgey Davidheiser was born at New Hanover, Pa, in 1885. He 
 received his elementary education in the schools of his native village, 
 Boyertown High School and Perkiomen School; from the latter institution 
 he was graduated in 1903. He then took up the teaching profession, in 
 which he continued until 1911 when he entered Ursinus College, from 
 which institution he was graduated in 1914 with the degree of A.B. He 
 again took up teaching. In 1917 he entered the graduate school of Johns 
 Hopkins University, following the subjects of Chemistry, Physical Chem- 
 istry and Mathematics. During the year 1918-19 he was employed by 
 the Aetna Explosives Co., Emporium, Pa., as head chemist of the acid 
 laboratories. In the fall of 1919 he again took up his work at Johns 
 Hopkins University, completing his course in June, 1922. 
 
THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 AN INITIAL FINE OF 25 CENTS 
 
 WILL BE ASSESSED FOR FAILURE TO RETURN 
 THIS BOOK ON THE DATE DUE. THE PENALTY 
 WILL INCREASE TO SO CENTS ON THE FOURTH 
 DAY AND TO $1.OO ON THE SEVENTH DAY 
 OVERDUE. 
 
 ' ,j ^ 
 
 
 3MSZ 
 
 
 i 
 
 
 
 
 MAR 23 1942 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 LD 21-100m-7,'40 (6936s) 
 
Pamphlet 
 
 Binder 
 Gaylord Bros. 
 
 Makers 
 Syracuse, N. Y. 
 
 PAT. JAN 21, "08 
 
 )avidheiser 
 Tha adso 
 
 imnonia by 
 
 
 
 (54 
 
 silica gel. 
 
 UNIVERSITY OF CALIFORNIA LIBRARY