Oct., 1945
1815
HYDROLYSIS OF CYANAMIDE IN ACIDSOLUTION
[CONTRIBUTION FROM
THE
DEPARTMENT OF CHEMISTRY ,AND CHEMICAL ENGINEERING OF THE UNIVERSITY OF
PENNSYLVANIA]
The Hydrolysis of Cyanamide in Acid S o l u t i o n l a * b BY MILLERJ. SULLIVAN2 AND MARYL. KILPATRICK Cyanamide hydrolyzes, in an aqueous acid Our results indicate that the hydrolysis of cysolution, to give urea; the identity of the prod- anamide in acid solution is a two-stage protolysis, uct was established by Hetherington and Bra- the rate-determining step being the reaction beham,a who also showed the reaction to be of the tween the cation of cyanamide, NH2CN-H+, and first order with respect to cyanamide. The only any bases present. Before the experimental work extensive investigation of the effect of acids, is presented, it should be mentioned that our progprior t o the present one, is that of Grube. Grube ress was retarded by the specificity of action of and M o ~ zand , ~ Grube and Schmid,s studied the the halogen acids, which appear to form unhyreaction in nitric acid solution, the acid concen- drolpzable, or difficultly hydrolyzable, complexes tration ranging from 0.05 to 5 N . They found the with Cyanamide. rate to increase steadily with increase in acid conExperimental centration, their results conforming to the equation The hydrolysis of cyanamide to urea is accompanied log (kob./"No*) A -k BNENO by a contraction (ca. 9 cc. per mole of cyanamide) which we made use of to follow the reaction dilatometrically. where kobs. is the first-order constant, L ~ H N O the , Cyanamide.-Through the courtesy of the American normality of the nitric acid, and A and B are con- Cyanamid Company a carefully purified sample of cyanstants. At the lowest acid concentrations Grube amide was obtained for which the following analysis was and Motz detected a decrease in kobs. with time. reported: This they attributed t o the fact that urea is a Anal. Calcd. for NHpCN: N, 66.64. Found: N (tistronger base than cyanamide, in consequence of tration), 05.38, N (Kjeldahl), 64.95,N (dicyanodiamide), which the hydrogen-ion concentration would de- 0.09. Since experiments performed with the above sample, crease as the reaction progresses. They deter- and with a sample prepared by us, gave consistent results, mined the equilibrium constant the above sample was used without further purification. During the course of the work the m. p. remained 43-44O, K1 a [NH2CN*H+]/[NHzCN][Hf] and checks upon the analysis by the silver titration con(the symbol H + is used throughout this paper, for tinued in good agreement. Dilatometers.-The experiments with a low initial conbrevity, t o denote the solvated hydrogen ion) and centration of cyanamide (0.045M o r less) were performed as they expected, found i t smaller than the equilib- in dilatometers like that described by Bronsted, Kilpatrick rium constant and Kilpatrick? These are called the "large" dilatometers. All other experiments were carried out in sealed' Ks [ ( ~ z ) z C O ~ ~ + l / ~ ( ~[H+l z~zCOl dilatometers. With acid in excess over cyanamide, Grube and Nitric Acid as Catalyst Motz detected no trend in kobs.. Grube and In these experiments the initial cyanamide conSchmid studied the effect of various nitrates, the acid concentration being kept constant a t 0.25 N . centration was 0.25 M , twice that used by Grube, and a t low acid concentrations kobs. decreased as In every case, the salt accelerated the reaction. Grube concluded from these findings that the the reaction progressed. To utilize the results hydrolysis of c y a n 9 i d e is an example of hydro- a t the low acid concentrations, i t was necessary gen-ion catalysis, with a large linear salt effect. to consider the effect upon the rate of reaction of This conclusion was accepted by Hammett and the basic nature of cyanamide and urea. Calculation of Rate.-Let it be assumed the Paul: who used the hydrolysis of cyanamide as an illustration of specific hydrogen-ion catalysis, activity-coefficient ratios fN&CN.H+/fNHsCNfH+ and their criterion being conformity to the relationship f(NHI),CO,H+!f(NHt)zC4fH' E!main unity over the range of acid concentration employed. Let i t be Ho log Rob. = constant assumed that eauilibrium is maintained between .. __._ where Ho is the acidity function of Hammett. NH&N and NH2CN.H+, and between (NH2)*(1) (a) Abstracted Irom the dissertation of Miller J. Sullivan preCO a d (NH2)2CO.H+. sented t o the Faculty of the Graduate School of the University of If NHzCN is the reactive species, and the ratePennsylvania in partial fulfilment of the requirement for the degree determining step is the reactibn between NHzCN of Doctor of Philosophy, November, 1943. (b) This work was and H+, one has originally directed by Martin Kilpatrick, who since January, 1945,
+
has been on leave of absence from the University of Pennsylvania. (c) Presented a t the Spring Meeting of the Philadelphia Section of the American Chemical Society, June 13, 1945. (2) Present address: The Socony-Vacuum Oil Company, Paulsboro, N. J. (3) Hetherington and Braham, Tms JOURNAL, 46,824 (1923). (4) G r u b and Motz, 2. physih. Chnn., 116, 14s (1925). (6) Grube urd Schmid, 0.5, it can be shown that ride, NH+2N.2HCll or chlorofonnamidine hydro- be the case when NHCI will c Ibe positive.16 ~hloride,~' is a stable compound in the solid state. ~ E A I ~ N H In pure hydrochloric acid solution k o b s / N H c l In dilute solution it must be largely broken up, however, since Hantzsch and Vagt16 found the first increases, then decreases, with increase in M solution to be prac- "CI, molar conductance of a while in the acid-salt mixtures it decreases steadily (cf. Tables I11 and IV); kobs., on the other tically that of two moles of hydrochloric acid. The kinetic data lend support to this hypothe- hand, exhibits a maximum in both cases. This is sis. Let it be assumed that there are present in readily explained provided one makes certain ashydrochloric acid solution the unhydrolyzable sumptions concerning the salt effects. Let it be assumed that in pure acid solution complexes NH*CN.HCI.H+ and NH2CNeHCI; possibly higher complexes, such as NH2CN.HCI. the effect of hydrochloric acid upon the secondorder constant is the same as that of nitric acid, 2Hf+, are also present. Let i. e., let it be assunled that a t 30' (7b) is appli(Ka)" = [ NHiCN HC1.H +] / cable. Let i t be assumed that in the acidsalt [ NH&Pi.H 1' [H1' [ C l - ] f ~ fcl+ = Ka/f2 (9a) mixtures
+
+
+
KC)^ = [NH&N.HCl]/[NH2CN] [H+][Cl-]fa+fcl- =
K d f Z (9b)
where the activity coefficients of NH2CN.HCI.H + and NH2CN.H+, and of NH2CN.HCl and NH2CN, are taken as equal. If the concentration of acid is high relative to that of cyanamide, so that [H+] and [Cl-] may be considered constant dur(11) Krieble and Holst, Tars JOURNAL, 60, 2976 (1938). (12) Benrath, 2. anotg. allgem. Chcm., 161, 53 (1926). (13) Taylor, J . Chem Soc., 2741 (1930). ( 1 4 ) J o h n w n and Spr.igue. Tars JOURNAL, 61, 176 (1939) I i) r i i n t , . r h ~ n l%i ' l i t . t n n 314* 366 (ison)
+
+
=& +
+
+
$1
+
- log k = 3.8600 0.0278 NACI (12) (16) In this connection, it is of iuterest to consider the decrease i n E A with increase in Ntfci observed in the hydrolysis of propionitrile by Ravinovitch, Winkler and Stewart (Con. 1.Research, BOO, 121 (1942)), aho concluded thnt "for reactions in which real changes in the parameters of the Arrhenius equation occur with change in catalyst concentration, apparent relations between reaction rate and other factors such at mean ion activity, acidity function, salt effect, etc., probably do not represent fundamental analysis of the factors influencing the reaction." The constancy of E A with nitric acid RS catalyst, in the hydrolysis of cyanamide, and its increase with increase in acid concentration with hydrochloric acid, or hydrobromic, as catalyst, is an argument against relegating the "other factnr-" to n minor position
HYDROLYSIS OF CYANAMIDE
Oct., 1945
where 3.8600 is the intercept of the line of Fig. 3, and 0.0278 is the slope found by D u ~ o u x who ,~~ measured the rate of inversion of sucrose in sodium chloride-hydrochloric acid mixtures 3.5 N in chloride. When the acid is present in large excess over cyanamide [H+] = Naci = x and one may write Ink = a
+ bx
where b is larger for the pure acid case, and smaller for the case of the mixtures, and positive for both cases. Change in K1 with change in medium will be neglected; the change in Ks and K I is determined by the change in the mean activity coefficient of the ions of hydrochloric acid. Since dk = b k , a n d Z z o ( x dKa x ) or ( x dK4 x ) =0
IN
ACIDSOLUTION
1821
two most dilute points, it is 95.X The experimental slope is thus of the expected sign and magnitude. For the mixtures, K and b are obtained from (12). Putting in the numerical values, one has
(d (kohl% dx
)x-o
725 X 10-'(2.303 (0.0278)
- 0.244 - 4K4I
I n Fig. 7 kobo/NHCl for the mixtures, a t 30°, is plotted as a function of NHCI.The upper broken line of Fig. 7 has the slope computed from the equation above with Kh set equal to zero, - 130 X 10-6. From the steeper slope of the curve i t
r-----
i t follows from (10) that (d-)
+PO
(d %;Ix)) x=o
= k(b
- Kt)for the pure acid
= k(b
- K I - 4K4) for the mixtures
(13a)
(13b)
For the pure acid case, K is the second-order constant at infinite dilution, and is obtained from equation (7b); b is the slope from (7b). Putting in the numerical values, one has
(qgjxe0 9
311 X 10-612.303 (0.2042)
- 0.2444) = 70 X
10-6
In Fig. 6, k o b s l N H C 1 for the pure acid solutions, a t 30°, is plotted as a function of " ~ 1 . The slope of the broken line is 75 X In drawing the curve, the most dilute point was disregarded, since here the cyanamide concentration was half the acid concentration ; however, if the initial slope is taken as the slope of the line through the 400
2 2ou 100
Fig. 6.-The
1
\
\
i-
I 0
I
1 .o
Norm. HCI. effect of HCI upon solutions at 30'.
2.0 koba./iVHCl
(17) Duboux. Hclo. Chim. A d a . Pi, 236 (1938).
1.0 2.0 3.0 4.0 Norm. HCI. Fig.7.-The effect of HCI upon k o b a . / N H c I in HCI-NaCI solutiolis at 30".
follows that, if the assumptions made are good, 4 Kq cannot be neglected in comparison to K1. The curve drawn has an initial slope of -580 X introducing this into the equation above, and solving for Kh, one obtains Kh = 0.16 in 4 N sodium chloride solution, and (K&, = K4!f: .= 0.06.18 For the acid-salt mixtures, then, the initial slope of the curve is of the expected sign, and its magnitude indicates either that Kq is not negligibly small or that K I is as large as 3.5 (0.244) in 4 N sodium chloride solution. Turning now to the first-order constant, one obtains from (10)
I'
'
0
3.0 in HCI
and upon proceeding as before, one finds k = 311 X k = 725 X
for the pure acid solutions and
for the mixtures
(18) By interpolation of the data of Hawkins (THISJ O U R N A L ,64, 4480 (1982)) it was found that logf 0.1958 0.03223 N H C Iin solu. tions 4 N in chloride, at 25".
+
MILLERJ. SULLIVAN AND MARYL. KILPATRICK
1822
Vol. 67
and VIIIB, the catalytic constant for the trichloroacetic acid molecule was found about twice as great as that for the hydrogen ion, which a t once cast doubt upon the mechanism suggested. In addition, the stability of an aqueous solution of cyanamide argues against the existence of general acid catalysis. Buchanan and Barsky20 found that solutions of cyanamide prepared from calcium cyanamide and sulfuric acid, and just acid to methyl red, suffered practically no change in concentration over a period of several months, and Grube and Kruger2' found no change, over the 159 X 10-0for a solution 0 25 -V in HNOs 383 X low6for a solution 0.23 n' in " 0 3 and 4 A- in period of observation of five days, in the concenh'aSOa tration of an aqueous solution of cyanamide kept Thus the increase in rate caused by the presence a t 50'. On the assumption that the anion of the acid of sodium chloride is of the same magnitude as HX catalyzes the hydrolysis, as does the water that caused by the presence of sodium nitrate. molecule functioning as a base, and on the asAn estimate of the magnitude of K , may be sumption that the rate-determining step is the made in the following fashion. By rearrangement reaction between the cyanamide cation and the of (lo), one has catalyst, one has In both cases it is-expected that the initial slope ~ will be positive; it is also of the kobs vs. N H Ccurve expected that the initial slope will be greater in the case of the mixtures. The upper broken line of Fig. 3 is drawn with a slope of 311 X low5,the and one sees lower, with a slope of 725 X that the initial slopes observed agree well with the calculated. It is of interest to compare the effect of 4 N sodium chloride upon the reaction with the effect of 4 N sodium nitrate. Grube and Schmid found the following values of k o b s l / i ~ ~ N oa8t 25'
( K I ) o Kl(Kt)o[H+I
~/fZ(k/kobs[C1-] (1 K [ H + l ) / [ H + l[Cl-l} =
+
and upon plotting u vs. [H+]expects a line of slope K1(Ka)oand intercept (K& provided no higher complexes are formed, and provided k and K1 are correctly taken. The points for the acid-salt mixtures, up to 2 N acid, were found to lie on a line of slope 0.45 and intercept 0.04 and, at higher acid concentrations, to lie above it. The experiments were not designed for the evaluation of equilibrium constants, but the results do indicate that K3 is larger than K4, and that higher complexes are probably present in the strongly acid solutions. (2) The hydrolysis of cyanamide in trichloroacetic acid solution was found to be more rapid than expected for a reaction catalyzed specifically by the hydrogen ion, since [H+]HNo~/[H+I CClsCOOCH = 1.7 at 0 5 N, and 2 0 at 1 N acid19
and (kobs)HNOs/(koba)CCliCOOH
