1088
C. GARDNER SWAIN, JAMES T. MCKNIGHT AND V. PETER KREITER
concentration of R/5.7 x = 1.3 x 10-18 M . 1 2 Alkoxide ion cannot be an intermediate in this case a t least because its Concentration, calculated from the autoprotoly-
sis constantla of ethanol (4 X large as this, oiz., 5 X
_._____
(13) B.
(12) 5.7 X I O L 2 is t h e Boltzmann conrtant times absolute temperatiire divided by t h e Plauck constant.
[CONTRIBUTION FROM
THE
VOl. 70 is on a thousandth 3s
Gutberahl a n d E. Grunwald, Tlirs J O U K N A I .75, , 571
CAMBRIDGE 39,
DEPARTMENT OF CHEMISTRY AND
MAssAciii:stins
THE LABORATORY FOR XUCLEAR INSTITUTE O F TECHNOLOGY]
SCIENCE, MASSACHUSETTS
The Mechanism of Exchange of Hydrogen between Ammonium and Hydroxyl Groups. 111-3
BY
c. GARDNERSWAIN,JAMES T. M C K N I G H T 4 A N D V. PETERKREITER RECEIVED AUGUST15, 1956
The rate of exchange of hydrogen between substituted ammonium salts and alcohols has been studied as a function of the following variables: isotope used as tracer; concentration and structure of acid, substituted ammonium salt, alcohol and solvent; and temperature. The results are consistent with the mechanism proposed in part I.
I n part I, the kinetics of exchange of ammonium bromide with methanol was examined in an inert solvent (dimethylformamide) and found to be consistent with the mechanism Solvent
+ ROH + RsNH+
ka
solvent H +
+I
kb
kc
I J _ ROH
+ R3N
kd
where I is a hydrogen-bonded alcohol-amine complex (ROH - - - NR3or RO-- - - HNR3+).
In this part, structural influences of the substituted ammonium salt, hydroxyl compound and solvent are examined in the light of this mechanism. The substituted ammonium salts were always deuterated and the alcohol undeuterated a t the start of a run, unless otherwise noted. Triethylammonium chloride (“triethylamine hydrochloride” in “Chemical Abstracts”) and the solvent methanol were used unless otherwise noted. We use substituted “ammonium” names for simplicity and uniformity with the “ammonium bromide” (“Chemical Abstracts” name) of the previous paper, although we recognize that primary, secondary and tertiary ammonium halides are all weaker electrolytes than quaternary ammonium halides as a result of hydrogen bonding. Reproducibility of the Kinetic Results.---A typical kinetic run is presented in detail in the Experimental section (Table VII) to illustrate the constancy of the experimental first-order rate constants ( k e ) within runs. The reproducibility of these rate constants is also good from one run to another considering the speed of many of the runs ( 1 ) Cf. C. G. Swain, J. T. McKnight, M. M. Labes a n d V. P. Kreiter, THISJOURNAL, 1 6 , 4243 (1954). For further details on deuterium exchanges, cf. J. T. McKnight, Ph.D. Thesis, M.I.T., September, 1953; all work with tritium was carried o u t b y V. P. K. ( 2 ) P a r t I, C . G . Swain a n d M. M . Labes, THISJ O U R N A L , 1 9 , 1084
( I 957).
(3) T h i s work was supported in p a r t b y t h e research program of t h e Atomic Energy Commission. Reproduction permitted f u r a n y porpose of t h e United S t a t e s Government. (4) Atomic Energy Commission Fellow, 1051-1953.
and the difficulty in quenching the reaction. Table I is a representative sampling of the check runs that were done and shows that the average deviation is under 5%. Factors Not Affecting the Rate.-The following variables had little or no effect on the rate (cf. Table I ) : the procedures used for preparing deuterated salt or for drying the alcohol (cf. runs 67, 80) ; the initial locus of the deuterium, whether in the salt or the alcoholic hydroxyl (Part I)2 or its percentage of the total hydrogen (58, 56); the amount of surface (84, 89); the concentration of more rapidly exchanging ammonium ions, either stronger or weaker acids (119, 89, see Discussion in section 3 below) ; and the choice of anion (runs 34,40,41,30in Table IV). Factors Affecting the Rate. 1. Isotope Used as Tracer.-Comparison of Table I1 with Tables I11 and IV shows that tritium exchanges approximately 0.8 as fast as deuterium (cf. r u n s 3 and 106, 4 and 49). If one assumes statistical distribution of deuterium between salt and methanol a t equilibrium, as was found to be the case with animonium bromide and methanol in dimethylformamide solution,2 the same isotope effect ( k ~ / k=~ 1.2) must hold for both forward and reverse rate constants. 2. Acid.-Exchange of triethylammonium chloride with methanol was complete in less than 34 sec. a t 0” when no acids were added to the methanol solution. The reaction was studied in basic media since the suggestion was made by Ogston6 that the net proton transfer from ethanol t o diethylamine a t 25’ is a slow process. I t was found that isotope exchange between deuterated diethylamine and either methanol or ethanol as solvent a t 0’ under comparable experimental conditions was complete in less than 30 sec. Thus hydrogen atoms bound to a free amine are exchanged rapidly in these hydroxylic solvents. A reasonable mechanism might involve either an ammonium ion as an intermediate or a four-center transition state with no intermediate. Higher-speed kinetic methods will have to be used t o distinguish these alternatives. ( 5 ) A G Ogston, J Chent bor , 1013 (19%)
1089
HYDROGEN EXCHANGE BETWEEN AMMONIUM AND HYDROXYL GROUPS
March 5 , 1957
TABLE I11 TABLE I EXCHANGE O F TRIETHYLAMMONIUM-d CHLORIDE I N METHAREPRODUCIBILITY OF RATE CONSTAXTS FOR EXCHANGE NOL AT 0' BETWEEN METHANOL AND 0.90 M TRIETHYLAMMONIUM-d ke IRzNHCI1,O [ H E ] , I, CHLORIDE IN METHANOL SOLUTIONS CONTAIXING HYDROGEN sec.-l X 10' Run M CHLORIDE AT 0' lH$ll,
Run
ks I see.-* X 108
Deviation from mean, %
55" 56" 57" 5Sb
0.143 .146 ,142 ,143
0.71 .61 .59 .64
11.0 4.7 7.8 0.0
67'~~ 80"*"
0.0038 .0033
0.64 13.0 12.0
4.0 4.0
81f 82/ 83d*f 8gdJ
0,424 .447 .445 .448
12.5 3.2 3.1 3.3 3.5
3.0 3.0 0.0 6.1
84' 89$ 119h
-
3.3 0.23 .24 .22 __ 0.23 3.0 3.3
0.456 ,444 .450
0.051 .052
94' 102'
4.8 4.8
TABLE I1
M
3 2
0.21 .22
B
.88 .86
r
4 1
[HCI], M
IRaNIICI],
.86
.so
n
32 .23 no054 m i .018 .082
h1EI IIA-
0" T in salt,
pc
mmole-1 sec
2 2 14 14 14 0.7
EXCHANGE OF 0.9 M
he, X 104
-1
0.80 n 83 3 3 2 7 1 7 0 7
Table I11 shows that the rate is a hyperbolic function of acidity over the range of concentration of hydrogen chloride available for study. There is very little dependence on acidity a t low concentrations of acid, but its retarding effect increases very gradually until the rate is sharply dependent on acid concentration a t high acidities. It is unlikely that this behavior is due to either weakness or dimerization of the hydrogen chloride since hydrogen chloride behaves as a simple strong electrolyte in methanol in other reactions: e.g., the
TRIETHYLAMMONIGM-d CHLORIDE I N
METHANOL Run
~ X C I I A N U P ,O F ' ~ K l ~ ~ l t Y l ~ A M M ~ z ' ClILORIDE IUM-~ IN
Run
0.048 0.226 1.3 .217 .220 0.99 ,368 .220 0.89 .51 0.32 .47 11.0 .00030 .118 ,00062 10.1 .118 9.6 .0016 .119 7.7 .0035 ,117 .0105 5.9 .119 3.0 .052 .114 .055 3.OC .110 looe 1.2 .20 96 ,116 .51 0.46 .I16 95 Total concentration of deuterated and undeuterated salt. 'Stoichiometric concentration. With 0.15 hl sodium p-toluenesulfonate added. d Data from reference 6 using triethylammonium bromide and hydrogen bromide (rather than chlorides).
TABLE IV 0.0 4.4 4.4
3.15 Salt containing 87% D. b S a l t containing 40% D , recrystallized extra time from chloroform. At 12' rather than 0'. dSalt prepared by exchange with heavy water rather than by neutralizing triethylamine with deuterium chloride (from reaction of benzoyl chloride with heavy water) as was done in runs in this table not marked by superscript d. s Methanol dried with calcium hydride (instead of magnesium turnings used in run 67). With 0.27 M water added. 0 With Pyrex glass wool added t o increase surface area by a factor of 2.6. With 0.05 M deuterated guanidinium chloride and 0.05 M deuterated anilinium chloride. With 0.12 A l rather than 0.90 M triethylammonium chloride.
KOL AT
107 106 105 53d 118 116 117 112 111 94
lHCII,a
Temp.,
M 0.35 .41 .60 .76" 1.2" 1.4 1.6
OC.
ke,
set.-* X 10'
25 6.1 27 25 5.8 43 34 25 4.3 40 25 3.9" 41 25 2.4" 25 1.8 30 25 1.3 32 25 0.33 31 3.1 25 .24 26 3.5 42 10 25 .Pol' 12 12.0 80 0.0033 79 .016 12 6.2 12 3.2 78 .@83 70 .404 12 1.n 63 .oon74 0 8.1 n 3.8 69 .0027 59 .0063 0 3.0 40 .018 n 2.3 28 ,036 n 1.5 48 .046 0 1.2 0 1.2 36 .050 56 .146 n 0.61 72 .404 n .24 0 .14 88 .694 Stoichiometric concentration. * Only approximate. Used triethylammonium p-toluenesulfonate and p-toluenesulfonic acid; no chloride ion in the exchanging solution. (I
methanolysis of @-naphthylacetate is first order in hydrogen chloride over this same range of concentrations.' The data satisfactorily.fit the following empirical relationship between ke and the concentration of strong acid [H+] Rate =
0.011 [ R s N H + ] 1 42[H+]
+
Comparison of equations 2 and 1 indicates that ka[ROH] = 0.011 and k b / k , = 42 with 0.1 111 salt (6) bf. M. Labes, Ph.D. Thesis, M.I.T., February, 1954. (7) M. Harfenist and R. Baltzly, TEISJOURNAL, 69, 362 (1947)
C. GARDNER SWAIN, JAMES T. LICKNIGHT AND IT. PETERKREITER
1090
a t 0'. Table IV shows that similar results are obtained with 0.9 M salt and a t different temperatures. At the highest acidities, above 1 JI hydrogen chloride, the repression is greater than calculated from a hyperbolic equation. This is probably a result of the large medium changes. Not only is the concentration of methanol varying significantly, but the large variations in total electrolyte concentration would change activity coefficients in an unpredictable manner in this region. The variation in rate is similar t o the variation of the activity coefficient of hydrogen chloride in water in this concentration range above 1 X , 8and it has been shown that activity coefficients of hydrogen chloride in water and methanol are similar.9 In neutral or basic solution, it appeared that the rate was a t least five times faster than in weak acid. Since equation 2 does not predict any further increase, this suggests that there may be base catalysis in basic solution. The reaction was too fast in this region to be studied successfully by our aliquot-taking procedure, but i t should be interesting to study using flow methods or other fast techniques. 3. Concentration and Structure of Ammonium Salt.-The reaction appears to be first order in amine hydrochloride. Table 111 shows that a 7.7fold increase in concentration gives only a 32% tlecrease in rate constant (runs 107, 96, 106, 105), reasonably interpreted as a salt effect. The effect of adding 0.15 AT of the strong electrolyte, sodium p-toluenesulfonate, in run 100 was also rery slight. This supports the conclusion above that the large effect of hydrogen chloride is iiiore than a simple salt effect. When salts of more basic amines are used, the rate of exchange usually decreases as shown in Table V (cf. triethyl- ZIS. trimethylamine, or ethylamine as. aniline). This is reasonable because k a and k, should be smaller relative to k b for the more basic amines. Guanidiniuiii ion is an exception. Perhaps i t cschanges by a different mechanism, utilizing its polyfunctional character. ClI,iOII
+ C(NHz)a+ + CHiOH J_ CHBOI-I~+ + EIN=C(NHs)KH3+ + CIIhO-.
.Icyclic process iiivoli.iiig only one iiiethaiiul u i d giving the saiiie intermediate but avoiding the ions from methanol is another possibility. The rapidly exchanging guanidinium and miliniuin chlorides present in small amounts in run 119 (Table I) did not affect the slower rate of exchange o f triethylammonium chloride. This indicates that their rapid exchange is not due to catalytic impurities and that the ammonium salts are not exchanging hydrogens among themselves faster than with the solvent. A4inong amines of comparable basicity, the greater the number of exchangeable hydrogens, the slower seems to be the rate of exchange ( c j . mono-, tli- and triethylamine or ammonia and trimethylamine in Table V). Thus ordinary steric hinHarned and K. W.Ehlers, THISJ O U R N A L , 66, 2179 (1033). B. B. Owen, "Physical Chemistry of Electrot i c Solutions," KeinholJ I'ubl. Curl)., X c \ v York, N. Y., 1843, p. 311,
. Ilarned and
Vol. 79
TABLE V Excaasc~TvIiii DIFFERESTAMCIILORIDES I S h f E l I I A S O L AT 0"
RELAIIVERA,I'ES01' XOSIVll
K B , ~
fie,
Run
.. 108
11194 109 107
,.
set.-! X 103
Concn., [HCI:, Amine" .M dl (NHdCSII I .0 3 :i ( C ~ H ~ I ~ X Hn . C2HaNH2 (C2HA3N ,ii .o.j? (CHd3N ,I1 ..;I €133 .I2 ,023 C ~ H ~ S H ~ 1.0 a,:+
relative
!?e,
rc1at:ve
c
103
> i n 3
0.10 0.06
0.32 0.17
3.0 11.0 l.ti
2.2 1.0 (1.00) 0.10 009 1 0 -6
(1.00) 20 0.53
c
>io3
*
L i s ain~noniuinchloride (amine 11~-drochloride). Basic
dissociation constant relative to triethylamine. \\ascomplete after 60 sec.
Exchange
drance does not seem to be a vcry important factor in this reaction. 4. Solvent.-The exchange of triethylaniinonium chloride with ethylene glycol in ethylene glycol solution containing 0.11 M hydrogen chloride a t 0" was extraordinarily slow, with an experiiiiental half-life of about an hour (Table VI). The exchange with ethanol in ethanol solution was faster than with methanol in methanol solution. Exchange between triethylamnionium picrate and 0.1 M methanol in toluene solution containing 0.01 M hydrogen chloride had a half-life of less than .5 seconds5 Table TrI shows that over this range o f solvent, rate increases with decreasing dielectric constant. T A B L E VI E X C I I A N G E O P ' r R I E T I I ~ L A M b I O N I U ~C~H L O R I D E TT.1111 FEREPI'T FOLVENTSAT
If
