JOURNAL OF CHEMICAL RESEARCH 2011 DECEMBER, 675–677 RESEARCH PAPER 675

The mechanism of the reaction of the Tollens reagent

William E. Benet, Gabriella S. Lewis, Louise Z. Yang and D. E. Peter Hughes*
Westminster School, 17 Dean’s Yard, London SW1P 3PB, UK

The Tollens test for aldehydes has been used for over 100 years but no reason has been given for adding sodium
hydroxide to the silver nitrate before the solution is cleared with ammonia. In this paper, a mechanism is proposed
which explains why the addition of alkali makes the test much more sensitive.

Keywords: Tollens reagent, aldehyde, gem diol, free radical

Bernhard Tollens1 introduced his silver test for aldehydes when established so rapidly that equation (1) cannot have a signifi-
he was working on carbohydrates in the late ninteenth century. cant effect on the rate of the overall reaction, even with less
This test involves the reduction of ammoniacal silver nitrate by reactive aldehydes such as propanal. All gem diols behave as
an aldehyde to give a silver mirror. If a few drops of sodium very weak acids and at pH <10 they would be essentially
hydroxide are initially added to the silver nitrate, the test unionised. However, at higher pH ionisation would occur, that
becomes much more sensitive though no explanation has been for chloral being most pronounced as it is the strongest acid
given for this. There are two possible reasons. The higher pH (pKa 10.0)4 because of the electron-withdrawing effect of the
could increase the yield of the reaction and/or it could increase chlorine atoms. On the basis of this evidence it is likely that the
the rate of reaction. second step in the reaction is as follows.
RCH(OH)2 + OH− → R(CHOH)O− + H2O (2)
Results and discussion
The overall reaction may be written as follows. Oxidation of an aldehyde to an acid involves the transfer of
two electrons and must need two Ag+ ions, via a free radical.
RCHO + 2Ag(NH3)2+ +3OH− → RCO2− + 2 Ag + 2H2O + 4NH3 A possible mechanism for these two electron transfers is as
If the reaction is reversible, an increase in pH would increase follows.
the yield. This suggestion was tested by setting up two electro- R(CHOH)O− + Ag+ → R(CHOH)OAg (3)
chemical cells. The first cell was a Ag/Ag(NH3)2+ half-cell
against a Ag/AgCl half-cell and the second a glucose/sodium R(CHOH)OAg → R(HCOH)O• + Ag(0) (4)
gluconate half cell against a Ag/AgCl chloride half cell. The R(CHOH)O• + Ag(NH ) → RCO2H + NH4
+ +
3 2
addition of sodium hydroxide changed the emfs of these cells + NH3 + Ag(0) (5)
only to a very small degree. Their values suggest that, in the
absence of side reactions, the reaction would go to completion. RCO2H + OH → RCO2 + H2O
− −
(6)
The fact that the Tollens test is more sensitive when the pH is +
In equation (3) the Ag ion may be either hydrated or in the
increased and when the temperature is raised can only be
form of the ammine.
explained by looking at the individual steps of the reaction.
A study of the kinetics of the reaction was carried out in
These kinetic factors will be considered next.
order to test the proposed mechanism. The first set of results
A mechanism has been proposed for the reaction of ammo-
was to establish the rate of the reaction with respect to the
niacal silver nitrate with several aldehydes.2 The authors did
concentration of the aldehyde. The rate of the reaction was
not record the pH at which they conducted their experiments
found using a colorimeter to measure the turbidity produced
but it can never have been >10 as they did not add any sodium
by the precipitated silver. A solution containing 0.025 mol dm−3
hydroxide to the mixture. Under these conditions, they found
of Ag(NH3)2+ at pH 10.5 was prepared. A 2.0 cm3 sample of
that the rate did not change when the pH was varied. They sug- this solution was placed in a cuvette and the colorimeter was
gested that reaction took place by the following mechanism. zeroed at 520 nm. A measured sample of 0.10 mol dm−3 chloral
RCHO + H2O → RCH(OH)2 (1) was injected, the solution mixed and the times for absorbance
readings of 0.1, 0.2, 0.3 etc. recorded. The results gave a
RCH(OH)2 + Ag → RC (OH)2 + H + Ag(0)
+ • +
reasonable straight line whose slope was a measure of the rate
RC•(OH)2 + Ag+ → RC(OH)2+ + Ag(0) → RCO2H + H+ of the reaction, Fig. 1.
A calibration graph was set up for absorbance against silver
While this mechanism may be operating at pH <10, it does not concentration so that the rates could be expressed with units
explain why the reaction goes much faster if the pH>10. mol dm−3 s−1, Fig. 2. It was found that silver at a concentration
of 0.00042 mol dm−3 in a 1.0 cm cuvette gave one unit
It is known that some aldehydes (for example methanal and
of absorbance. All experiments were carried out at room
chloral) very quickly give a positive reaction with Tollens
temperature which was 20 ± 1 °C.
reagent at room temperature. Ethanal (and the higher alde-
The concentration of chloral was varied and it was shown
hydes) react much more slowly. Chloral and methanal rapidly
that rate was directly proportional to the chloral concentration,
hydrate to give gem diols3 and it seems likely that the initial
Fig. 3.
hydration, as proposed by Prasad et al.2 is an essential first
A similar set of experiments was carried out, keeping the
step. This hydration is a reversible reaction whose position of
concentration of the chloral constant and varying the concen-
equilibrium varies with the aldehyde under investigation. The
tration of the Ag(NH3)2+. The concentration of the Ag(NH3)2+
equation shows that the position of equilibrium is unaffected
was always in an excess (≥1.0 × 10−3 mol dm−3), larger than
by alkali. Alkali does, however, act as a catalyst and equilib-
that of the chloral (5.0 × 10−4 mol dm−3). Under these condi-
rium is established very rapidly at pH>10. This equilibrium is
tions, when the concentration of Ag(NH3)2+ was varied, there
was no significant change in rate, Table 1.
It was not possible to measure the rate at very low concen-
* Correspondent. E-mail: [email protected] trations of Ag(NH3)2+ because the amount of silver precipitated
676 JOURNAL OF CHEMICAL RESEARCH 2011

Table 1 Rate against silver concentration

Rate/mmol dm−3 s−1 0.03 0.03 0.04 0.04
[Ag(NH3)2+]/mol dm−3 0.001 0.01 0.05 0.1

would reduce the amount of free aldehyde available for reac-

tion with the ammoniacal silver nitrate and would make the
test less sensitive. A similar reduction might also take place if
too high a concentration of alkali was used. Under strongly
alkaline conditions, aldehydes may undergo two reactions, the
aldol reaction and the Cannizzaro reaction. These reactions
take place fairly slowly and it was unlikely that they caused a
significant reduction in the amount of free aldehyde present
during the time that the Tollens test was carried out. In order
to minimise these side reactions, it was necessary to prepare
Fig. 1 Typical absorbance against time graph for chloral with
Tollens reagent with care, making sure that the ammonia was
Ag(NH3)2+. not in an excess and that the pH was not excessively high.
Because the reaction involved single electron transfers, it
must be a free radical reaction. A free radical mechanism can
be confirmed by adding a small sample of TEMPO which,
being itself a free radical, slows down the reaction.5 This was
found to be the case because the addition of TEMPO slowed
the reaction down by a factor of five.
If the proposed mechanism is correct, then at low pH, the
rate should decrease because the equilibrium for the ionisation
of the gem diol, equation (2), would lie to the left. With chloral,
whose gem diol has a pKa 10.0,4 the rate should decrease below
pH 11, while for ethanal (pKa 13.0), the rate should increase
only above pH 13. Such changes in rate were observed as
shown in Table 2.
Finally Tollens reagent was tried out on a variety of different
aldehydes. The results, Table 3, correlate with the ease of
formation of the gem diol of the aldehyde.
The time for the test to be positive (solution going grey) was
Fig. 2 Calibration of absorbance against [Ag(NH3)2+]/mol dm−3.
recorded as follows : < 5 s very rapid; 5–10 s rapid; > 1 minute
slow; >5 minutes very slow. The warmed solutions were placed
in a water bath at 80 °C. Benzaldehyde, salicylaldehyde and
sucrose gave negative results at pH 10.5 but they gave a slight
positive reaction at pH 13 when the solution was warmed.
The difficulty in getting a positive response with benzalde-
hyde explains why there has been disagreement as to whether
it gives a positive reaction to the Tollens test or not.6,7 The
slow positive reaction with sucrose was presumably due to
hydrolysis when the solution was warmed.

Table 2 Changes in rate with pH for chloral and ethanal

Choral pH 7.9 11 11.6 11.8
Rate/ mmol dm−3 s−1 0.003 0.003 0.02 0.03
Ethanal pH 12.8 13.0 13.2 13.3
Rate/mmol dm−3 s−1 0.0005 0.0005 0.003 0.003

Fig. 3 Rate against [chloral]/mol dm−3. Table 3 Qualitative results for Tollens’ reagent with different
aldehydes at pH 10.5
Aldehyde Observation
was too small for the absorption to be measured. Thus, because Methanal Very rapid at room temperature
the Ag(NH3) concentration was always in excess, the rate was Chloral Very rapid at room temperature
limited only by the concentration of the chloral (pseudo first Ethanal Rapid at room temperature
Propanal Rapid at room temperature
order conditions). Because the rate was unaffected when the 4-Nitrobenzaldehyde Slow when warmed
Ag(NH3)2+ was increased, it is likely that equation (4) and not 2,4-Dinitrobenzaldehyde Rapid when warmed
equation (5) was the rate determining step. Benzaldehyde Nil
When making Tollens reagent, care must be taken not to Salicylaldehyde Nil
have excess ammonia which makes the test much less sensi- Glucose Slow at room temperature
tive. It is known that ammonia readily reacts with aldehydes to Fructose Slow at room temperature
Sucrose Nil
give a variety of addition products. These addition reactions
 JOURNAL OF CHEMICAL RESEARCH 2011 677

Conclusion to become grey. If, after several minutes, there was no obvious reac-
The addition of alkali makes the Tollens test much more sensi- tion at room temperature, the experiment was repeated with the tube
tive because the rate is much faster. The key feature seems to in a water bath at 80 °C. Because the higher aldehydes are sparingly
be the formation of the anion of the gem diol. The ease of for- soluble in water, they were initially dissolved in a 50/50 by volume
propanone/water mixture.
mation of this anion varies with the aldehyde and correlates
well with how rapidly they respond to Tollens test.
CAUTION: Sodium hydroxide solution must be handled with
care, especially at high pH. Tollens reagent was not kept for longer
Experimental than an hour as it may deposit explosive side products.
All chemicals were supplied by Aldrich and used without further puri-
fication. Tollens reagent of the required concentration and pH was
made up immediately before use. Typically 0.5 cm3 of 1.0 mol dm−3 Received 22 October 2011; accepted 2 November 2011
NaOH was added 20 cm3 of 0.050 mol dm−3 AgNO3. The mixture was Paper 1100833 doi: 10.3184/174751911X13206824040536
stirred magnetically and 1.0 mol dm−3 ammonia slowly added until a Published online: 27 December 2011
faint precipitate remained. The solution was then filtered and the
filtrate made up to 40 cm3. This typically produced 0.025 mol dm−3 References
Ag(NH3)2+ at pH 10.5. A 2.0 cm3 sample of this solution was put in a
1 B. Tollens, Landwirt. versuchs-station., 1887, 29, 392.
1.0 cm plastic cuvette which was placed in a Sherwood 260 colori-
2 T.V. Prasad, B.S. Ram and T. Rao, Curr. Sci., 1982, 51, 749.
meter fitted with a 520 nm filter. The colorimeter was zeroed with this 3 R.P. Bell, Advances in physical chemistry Academic Press, New York,
solution. A measured volume of 0.1 mol dm−3 of chloral or ethanol 1966, Vol. 4, p. 8.
was injected from a syringe, the mixture shaken and times recorded 4 R.P. Bell, Advances in physical chemistry Academic Press, New York,
for absorbance readings of 0.1, 0.2, 0.3, etc. The pH was measured at 1966, Vol. 4, p. 16.
the end of the run. For the qualitative tests on aldehydes, a solution of 5 www.en.wikipedia.org/Radical_(chemistry).
Ag(NH3)2+ was made up with the above specifications. To approxi- 6 F.G. Mann and B.C. Saunders, Practical organic chemistry, Longmans,
mately 2 cm3 of this solution, enough aldehyde was added to make its London, 1960, p. 409.
concentration ≈0.1 mol dm−3. The time was recorded for the solution 7 www.ask.com/questions-about/Tollens-Reagent-with-benzaldehyde.