Another Unprecedented Wieland Mechanism Confirmed: Hydrogen Formation from Hydrogen Peroxide, Formaldehyde, and Sodium Hydroxide

Article abstract of DOI:10.1002/anie.201800636

In 1923, Wieland and Wingler reported that in the molecular hydrogen producing reaction of hydrogen peroxide with formaldehyde in basic solution, free hydrogen atoms (H^.) are not involved. They postulated that bis(hydroxymethyl)peroxide, HOCH₂OOCH₂OH, is the intermediate, which decomposes to yield H₂ and formate, proposing a mechanism that would nowadays be considered as a “concerted process”. Since then, several other (conflicting) “mechanisms” have been suggested. Our NMR and Raman spectroscopic and kinetic studies, particularly the determination of the deuterium kinetic isotope effect (DKIE), now confirm that in this base-dependent reaction, both H atoms of H₂ derive from the CH₂ hydrogen atoms of formaldehyde, and not from the OH groups of HOCH₂OOCH₂OH or from water. Quantum-chemical CBS-QB3 and W1BD computations show that H₂ release proceeds through a concerted process, which is strongly accelerated by double deprotonation of HOCH₂OOCH₂OH, thereby ruling out a free radical pathway.

Full text of DOI:10.1002/anie.201800636

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Another Unprecedented Wieland Mechanism Confirmed:
Hydrogen Formation from Hydrogen Peroxide, Formaldehyde,
and Sodium Hydroxide
Authors: Robert Czochara, Grzegorz Litwinienko, Hans-Gert Korth, and
Keith U. Ingold
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201800636
Angew. Chem. 10.1002/ange.201800636
Link to VoR: http://dx.doi.org/10.1002/anie.201800636
http://dx.doi.org/10.1002/ange.201800636

10.1002/anie.201800636
Angewandte Chemie International Edition
Another Unprecedented Wieland Mechanism Confirmed:
Hydrogen Formation from Hydrogen Peroxide,
Formaldehyde, and Sodium Hydroxide
Robert Czochara, Grzegorz Litwinienko,* Hans-Gert Korth,* and Keith U. Ingold
This paper is dedicated to Alwyn Davies on the occasion of his 92st birthday.
Abstract: In 1923, Wieland and Wingler reported that in the molecular hydrogen
producing reaction of hydrogen peroxide with formaldehyde in basic solution free
hydrogen atoms (H) are not involved. They postulated that bis(hydroxymethyl)peroxide,
HOCH2OOCH2OH, is the intermediate that decomposes to yield H2 and formate,
proposing a mechanism which nowadays would be considered as a "concerted
process". Since that time several other (conflicting) ‘mechanisms’ have been suggested.
Our NMR and Raman spectroscopic and kinetic studies, particularly determination of the
deuterium kinetic isotope effect (DKIE), now prove that in this base-dependent reaction
both H atoms of H2 derive from the CH2 hydrogens of formaldehyde, not from the OH
groups of HOCH2OOCH2OH or from water. Quantum-chemical CBS-QB3 and W1BD
computations confirm that H2 release proceeds through a concerted process, strongly
accelerated by double deprotonation of HOCH2OOCH2OH, thereby ruling out a free
radical pathway. Wieland’s mechanism for H2 evolution showed truly amazing chemical
insight.
1
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
[*]
R. Czochara, Prof. G. Litwinienko,
Faculty of Chemistry, University of Warsaw
Pasteura 1, 02-093 Warsaw, Poland
E-mail: litwin@chem.uw.edu.pl
Dr. H.-G. Korth
Institut für Organische Chemie, Universität Duisburg-Essen
Universitätsstraße 7, 45117 Essen, Germany
E-mail: hans-gert.korth@uni-due.de
Dr. K. U. Ingold
National Research Council,
100 Sussex Dr., Ottawa, ON, Canada
E-mail: Keith.Ingold@nrc-cnrc.gc.ca
Supporting information and the ORCID identification numbers of the authors of this article can be found under:
https://doi.org/10.1002/anie2018xxxxx
2
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Heinrich Wieland’s mechanistic insight was unrivalled in peroxide chemistry. He was
the first to identify any free radical rearrangement, viz. Ph3CO  Ph2COPh, during
thermal decomposition of Ph3COOCPh3.[1] Decades later some of us[2] provided
unequivocal evidence favoring this (disputed) isomerisation. This stimulated us to look
for other questioned (or ignored) Wieland mechanistic proposals. We found the fast
room temperature reaction:[3,4]
H2O, 25oC
0.05 M H2O2 + 0.1 M HCHO + 0.1 M NaOH
initial 1/2 = 82 s
0.05 M H2 + 0.1 M HCOO Na (1)
Wieland[3] postulated that bis(hydroxymethyl)peroxide (H2P)[5] was the intermediate
that decomposed to give molecular hydrogen [reactions (2)–(4)], representing this
process exactly as shown in Figure 1 and concluding that free H atoms were not
involved because methylene blue or iodine were not decolorized and p-benzoquinone
was not hydrogenated.[6]
Wieland’s genius is further attested to by his pioneering description of what later
became known as the solvent cage effect.
3
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Figure 1. Mechanism of bis(hydroxymethyl)peroxide (H2P) decomposition proposed by
Wieland and Wingler.[3] (H. Wieland, A. Wingler, Über den Mechanismus der
Oxydationsvorgänge V. Zur Oxydation der Aldehyde, Justus Liebigs Ann. Chem. 1923,
431, 301–322.) Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with
permission.
Reaction (1), first reported in 1898,[7] has been the subject of considerable
mechanistic speculation, both before and after Wieland’s Figure 1 proposal, but much of
the speculation has ignored or overlooked established facts. The reaction has three
times been reported to follow second-order kinetics. In the first of these studies,[8] the
effect of added base (essential if the reaction is to proceed at a noticeable rate at or
below 25 oC) was already recognized, but the evolved gas was assumed to be O2! The
second work[9] explicitly incorporated base into a proposed mechanism. Unfortunately, it
was concluded that the H2 came from the two HO groups, viz. 2 CH(OH)O–  2
HCOO– + H2, despite an earlier study[10] in D2O in which the evolved hydrogen was
converted to water, the density of which indicated that it was H2O, proving that the
evolved hydrogen came from two non-exchangeable CH2, rather than the exchangeable
OH groups.[11] In the third study[12] H2O2 oxidation of HCHO in basic solution was
suggested to proceed via the bis(hydroxymethyl)peroxide dianion (P2–) as an
intermediate, and decomposition to H2 and formate [reaction (5)] was proposed to be
rate-limiting. However, H2 release was not investigated.
4
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
We here confirm the peroxide’s two CH2 groups as the sole source of H2 using
Raman and NMR spectroscopy, see Figure 2 and Figures S14–S25 in the Supporting
o
Information. In Ref. [9] it was also shown that the H2 formation vs time curve at 0 C
using the concentrations given in reaction (1) was identical to that for an equivalent
concentration of H2P with the appropriate [NaOH] added, reaction (4), see Figure S26.
When the reaction is fast, e.g. reaction (1) conditions, the H2 yield is 100% of theory,[3,9]
but when slow (at pH ≤ 12), additional paths (see below) may make the yield appear to
be <100%.[13]
1
Figure 2. A) H NMR spectra of the gas generated from HCHO/H2O2/NaOH (4:1:2) in
D2O (dotted) and from DCDO/H2O2/NaOH (4:1:2) in H2O (solid). Signals at 3.40 ppm:
H2O vapor. The small signal (solid) at 7.29 ppm corresponds to the HCHO impurity of
5
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
2
the applied DCDO. B) H NMR spectra of the gas generated from HCHO/H2O2/NaOH
(4:1:2) in D2O (dotted; D2 not detected; signal at 3.45 ppm: D2O vapor) and from
DCDO/H2O2/NaOH (4:1:2) in H2O (solid; D2 is detected at 7.17 ppm).
Considering all of the above (and other experimental facts) it would seem that
Wieland’s Figure 1 may be an acceptable approximation to the transition state of
reaction (1), in particular if the two hydroxyl H atoms are replaced by two negative
charges, viz. P2–. This leaves two mechanistic possibilities: either, the rate controlling
step could involve a concerted bond making/breaking (sigmatropic [2+2+2] elimination,
as implied by Figure 1), or O–O bond scission is rate controlling, being followed by two
fast in-cage C–H scissions to give a singlet pair of H atoms that combine
‘instantaneously’, reaction (6).
It seemed likely that these two mechanisms could be readily distinguished by the
HCHO/DCDO deuterium kinetic isotope effect (DKIE), which would be large for the
concerted process but small for a simple O–O scission. Determining the (possibly small)
DKIE demanded a careful kinetic study. In view of the earlier kinetic work,[9] we chose as
o
our base the conditions indicated in reaction (1), kept to 25 C and monitored H2
evolution while varying the concentration of one reagent at a time. Experimental details
and results are given in the SI, Tables S1–S18 and Figures S1–S13.
6
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Reproducibility was less than we would have liked, but was quite sufficient for our
purposes. We confirmed (approximate) second-order[14] H2 evolution with 100% yield for
fast reactions. At less than stoichiometric [HCHO] the rate was [HCHO]2 while excess
,
HCHO roughly doubled the rate (Figure 3A). The rate slightly enhances with a small
increase in [H2O2] (Figure 3B) but a stoichiometric excess of H2O2 strongly retarded the
reaction (because now hydroxymethyl hydroperoxide, HOCH2OOH, [reaction (2)] is
favored over H2P).
7
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Figure 3. Initial rates of H2 evolution. A) 0.10 M NaOH and 0.05 M H2O2 with varying
amounts of HCHO (inset: initial rate as a function of [HCHO]2). B) 0.10 M NaOH and
0.10 M HCHO with varying amounts of H2O2. C) 0.10 M HCHO with 0.05 M H2O2 with
varying amounts of NaOH (expressed as pH). The arrow indicates the stoichiometric
8
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
HCHO:H2O2:NaOH = 2:1:2 ratio. Empirical fits to the non-linear graphs are just intended
to guide an eye.
The pH dependence of the initial rate of H2 production is shown in Figure 3C. Clearly,

the pKa of the species governing H2 release must be around 13.5. Since the pKa s of
H2O2 and HOCH2OOH are 11.6 (Table S19) and 10 (estimated from computed
deprotonation free energies, see Figure S27), respectively, these observations are not
only consistent with earlier work[9,12] but also provide strong support for a rate-controlling
concerted decomposition of P2– (formed via an extensive series of equilibria, see
Scheme S1), i.e. HOCH2OH[15] and HP– (see Figure S28) are the only species present
likely to have such a high pKa. The reaction rate was the same (to within one standard
deviation) under O2 as under N2[16] and was also the same when NaOH was replaced by
CsOH (Tables S10–S18).[17] The reaction exhibited a mean HCHO/DCDO DKIE of 6.9,
which is fully consistent with a concerted elimination of H2 from P2–. Facile H2
elimination from P2– in water is consistent with a report[18] that decomposition of neutral
H2P in the gas phase at 110–200 oC yields formic acid and H2 more slowly than it yields
HCHO and HOCH2OOH, i.e. reaction (–3) is favored over H2 loss.
There are only three theoretical studies on (decomposition of) H2P.[19,20] Fleischmann
and Cioslowski[19] recognized that H2 must come from the dianion P2–. On the HF (MP2)
level they located a “stretched” conformation (–OCOO units s-trans) as the ground state
structure, and gave barriers to internal rotation. No calculations being reported on H2
loss or O–O bond cleavage. In two, essentially identical reports by Thiel (gas phase DFT
calculations),[20] a 6-center cyclic transition structure of H2P is found from which H2 can
be extruded in a concerted process. The reported structures agree with the ones found
9
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
in this study, however, it was ignored that the release of H2 in the gas phase or neutral
aqueous solution is slow and the yield is not quantitative. In order to explore possible
reaction paths of H2 formation from H2P at neutral and basic pH, we carried out
extensive gas phase and aqueous phase computations employing the composite
(PCM)CBS-QB3 method.[21] Pivotal structures were additionally computed with the high-
accuracy W1BD method, with aqueous solvation being modeled by the SMD approach.
Structures and thermochemical data are collected in Figure S29 and Tables S20–S23.
W1BD structures and activation data are displayed in Figure 4. With a few exceptions,
identical structures were found with CBS-QB3 and W1BD in the gas and aqueous
phase. Several ground state conformers and rotational transition states were located for
H2P, HP– and P2– (Table S21). The lowest-energy structure of H2P (1/1a) has its OH
groups mutually hydrogen bonded, thereby stabilizing an unfavorable arrangement for
concerted H2 elimination. Two conformeric TSs for concerted H2 evolution were found,
the lower (2/2a) of H‡(aq) = 27.8 (CBS-QB3, 27.9) kcal mol–1 activation enthalpy in
water. A further transition structure (3/3a), leading to formic acid and dihydroxymethane
(formaldehyde hydrate), was located, characterized by an elongated O–O bond
combined with a simultaneous 1,3-shift of one of the CH2 hydrogens to the remote
peroxide oxygen. The >12 kcal mol–1 higher barrier, however, renders this path
unimportant under normal conditions. Note that transition structures for simple O–O
bond homolysis were not found (for bond dissociation energies see Table S23).
10
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Figure 4. (SMD)W1BD-computed lowest-energy aqueous phase structures and
activation barriers H‡ (G‡) for bis(hydroxymethyl)peroxide (H2P) (left), its monoanion
(HP–) (middle), and its dianion (P2–) (right) and related transition structures for concerted
H2 release and intramolecular hydrogen shift. For compound labelling and bond
distances see Tables S20, S21 and Figure S29.
As expected, the ground state of the monoanion HP– (4/4a) is intramolecularly
strongly hydrogen-bonded (HHB ≤ –12 kcal mol–1, cf. 18/18a, Table S21), again
stabilizing an unfavorable conformation for concerted H2 release. The related transition
state (5/5a) is slightly lower than that of H2P (2/2a), both agreeing well with the
11
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
experiments in water,[8] see Figure S30. Interestingly, TS 6/6a for decomposition to
(initially) HCOOH and HOCH2O– now involves a 1,2-CH2 hydrogen shift to the adjacent
peroxide oxygen. This TS even lies below the barrier for concerted H2 release,
suggesting that at a pH where HP– exists, H2 production will not be quantitative but
accompanied by formation of HCOOH/HCOO– and HOCH2OH (= HCHO + H2O), as
reported by Wieland.[3]
The ground state of dianion P2– (7/7a) adopts the stretched conformation as found
earlier.[19] This conformation can reasonably be attributed to Coulomb repulsion of the
negatively charged oxygens. Two of the CH2 hydrogens are now in a favorable
orientation for concerted H2 release [H–H distance 2.82 (g)/2.55 (aq) Å]. This results in a
strongly decreased TS barrier (8/8a) with H‡(aq) = 13.1 (CBS-QB3, 16.3) kcal mol–1.
This H‡(aq) is close to the experimental “activation energy Ha = 14.5 ± 0.3 kcal mol–
1”,[9] which supports reaction (5) as the rate-determining step. The 1,2-H-shift TS 9/9a,
–
leading to (initally) HCOOH and OCH2O–, lies 7–8 kcal mol–1 above TS 8/8a. This
process thus would not compete noticeably with the concerted H2 release. Internal
reaction coordinate (IRC) computations (Figure S31) confirm the found transition states
connect with the peroxidic educts (H2P, HP–, P2–) and the products (H2 + HCOO–
/HCOOH and HOCH2OH + HCOO–/HCOOH, respectively). Further, DKIEs for H2
extrusion in the range 3.2–9.9 were computed (Table S21), in accord with the
experimental values.[22] In toto, our computational findings are congruent with a base-
induced strong acceleration and quantitative evolution of H2 from HCHO and H2O2, and
support H2 evolution from P2– in the rate-determining step.
12
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Our conclusion is that Wieland’s Figure 1 is an acceptable approximation to the
transition state. Hats off to Wieland!
Acknowledgements
We thank W. Siebrand for helpful discussions. This work was co-funded by the National
Science Centre - Poland, decision no. 2014/15/B/ST4/04835.
Conflict of interest
The authors declare no conflict of interest.
Keywords: history of science • H2O2–formaldehyde reaction • hydrogen formation •
peroxides • reaction mechanisms
13
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
[1] H. Wieland, Ber. Dtsch. Chem. Ges. 1911, 44, 2550–2556.
[2] G. A. DiLabio, K. U. Ingold, S. Lin, G. Litwinienko, O. Mozenson, P. Mulder, T. T.
Tidwell Angew. Chem. Int. Ed. 2010, 49, 5982–5985; Angew. Chem. 2010, 122,
6118–6121.
[3] H. Wieland, A. Wingler, Justus Liebigs Ann. Chem. 1923, 431, 301–322.
[4] Reaction follows second-order kinetics, so only first half-life is given.
[5] First isolated by Legler (L. Legler, Justus Liebigs Ann. Chem. 1883, 217, 381) from
lampic acid [Lampensäure] and first synthesized by Fenton [H. J. H. Fenton, Proc.
Roy. Soc. (London) 1914, A90, 492–498].
[6] The relevant sentence in Ref. [3] can be translated: The driving force for this
peculiar decay reaction is the tendency of both escaping hydrogen atoms to
combine to the molecule H2, a process which takes place before the molecular force
field [molekulare Kraftfeld] is passed. This concept was foreshadowed by Wieland et
al. (H. Wieland, E. Popper, H. Seefried, Ber. Dtsch. Chem. Ges., 1922, 55B, 1816–
1834) in a study of the Ph3CN=NPh  N2 + Ph4C reaction: … for a moment after
their bond release, atoms or groups experience a kind of gravitation [with]in the
force field of their [parent] molecules...
[7] O. Blank, H. Finkenbeiner, Ber. Dtsch. Chem. Ges. 1898, 31, 2979–2981.
[8] J. H. Kastle, A. S. Loevenhart, J. Am. Chem. Soc. 1899, 21, 262–276. Here it is
reported that HCHO consumption in the presence of KOH or NaOH is about ten
times faster than in the presence of NH4OH, see Figure S30.
14
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
[9] J. B. Jaillet, C. Ouellet, Can. J. Chem. 1951, 29, 1046–1058. Similar to the present
study, kinetics were determined from monitoring H2 release by means of a pressure-
gauge albeit at a ~110-fold higher volume.
[10]K. Wirtz, K. F. Bonhoeffer, Z. Phys. Chem. 1936, B32, 108–112.
[11]Cited [10] but mistranslated, viz.[9] …que l’oxygen dégagé provident uniquement de
la formaldehyde… “oxygen” should have been “hydrogen”.
[12]J. S. Do, C. P. Chen, Ind. Eng. Chem. Res. 1994, 33, 387–394.
[13]The Wieland mechanism has also been adopted for H2 production from HCHO and
H2O2 in acidic solution: B. L. Dunicz, D. D. Perrin, D. W. G. Style, Trans. Faraday
Soc. 1951, 47, 1210–1215. The overall process is proton-catalyzed (Scheme S2),
but H2 release from H2P is an acid-independent first-order reaction (Ea = 24.9 kcal
mol–1).
[14]The rates of H2 evolution during reaction differ slightly from pure second-order in
[HCHO], see Figures S9–S13, in agreement with Ref. [8], see Figure S30. We
neglect this small effect because of clear dependence of the initial rates on [HCHO]2
(see inset in Figure 3A).
[15]For HOCH2OH a pKa range of 12.5–14.0, focusing on 13.3, is given in the literature,
see: Y. I. Tur'yan, Croat. Chem. Acta 2000, 73, 625–641.
[16]Supporting the absence of free H atoms and an H atom chain, H + P2–  H2 +
–
HCO2– + OCH2O–  H + HCO2 .
[17]In the presence of a heavy ion like Cs+ a singlet H pair should evolve to a triplet pair
more rapidly, enhancing H escape from the cage. This free H could react rapidly
with O2 and lower the rate of H2 formation.
15
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
[18]A. D. Jenkins, D. W. G. Style, J. Chem. Soc. 1953, 2337–2340.
[19]E. D. Fleischmann, J. Cioslowski, Chem. Phys. Lett. 1990, 168, 265–268.
[20]a) W. R. Thiel, Eur. J. Org. Chem. 2004, 3108–3112; b) W. R. Thiel, J. Mol. Catal. A:
Chem. 2006, 251, 229–233.
[21]Gaussian 09, revision A.02; Gaussian, Inc.: Wallingford, CT, 2009; see SI, p. S-4 for
full references.
[22] A contribution by hydrogen-tunneling to the rate constants has to be expected for
such type of reaction, see footnotes to Tables S20, S21. Detailed investigation of the
tunneling contribution, however, is beyond the scope of our paper.
16
This article is protected by copyright. All rights reserved.

10.1002/anie.201800636
Angewandte Chemie International Edition
Entry for the Table of Contents
COMMUNICATION
Robert Czochara, Grzegorz Litwinienko,*
Hans-Gert Korth,* Keith U. Ingold
Page No. – Page No.
Another Unprecedented Wieland
Mechanism Confirmed: Hydrogen
Formation from Hydrogen Peroxide,
Formaldehyde, and Sodium
Hydroxide
Hats off to Wieland! Heinrich Wieland's 1923 foreseeing of a 'concerted'
mechanism for the release of molecular hydrogen from reaction of hydrogen
peroxide with formaldehyde in basic solution via a bis(hydroxymethyl)peroxide
intermediate has been confirmed. Kinetics, spectroscopy, and quantumchemical
computations prove both H2 hydrogens derive from formaldehyde's CH2 group.
Free radicals are not involved.
17
This article is protected by copyright. All rights reserved.