Full Papers
ing CO2 and N2, at two different proportions of H2S. The pur-
pose was to show that the overall H2S splitting can take place
without interference of these gases, which have been found to
not exhibit absorption in this spectral region (254 nm) and,
therefore, the H2S decomposition process could be applicable
in most of real gases without the need of H2S purification. The
results are summarized in Table 1. As can be seen in Table 1,
the presence of CH4 and CO2 in large concentrations with re-
spect to H2S, corresponding to compositions of real mixtures,
does not interfere in the overall H2S splitting, regardless of the
H2S concentrations (2–6%), achieving complete removal of H2S
in a few hours under our irradiation conditions, with H2S exhib-
iting the same photochemical behavior as in the absence of
these gases. The time required for complete H2S removal in-
creases along its initial concentration. This is understandable
considering that neither CH4 nor CO2 have absorption bands in
the UV region above 200 nm. Note that the reaction proceeds
to completion because elemental sulfur is deposited in the
cold parts of the photoreactor and not in the illuminated (hot)
parts that are transparent during the whole process.
Figure 3. Temporal profiles of H2 generation (solid lines) and H2S disappear-
ance (dashed lines) in mixtures containing an initial 10.8 vol% H2S concen-
tration upon irradiation with quasi monochromatic light of different wave-
lengths: 254 nm (red), 300 nm (green) and 354 nm (black). Note that the
green and black lines essentially overlap giving zero H2 production.
radiation wavelength on the photochemical overall H2S split-
ting can be easily rationalized based on the UV absorption
spectrum of H2S shown in Figure 2, which does not present
significant photon absorption beyond 270 nm.
This selectivity for overall H2S splitting in the photochemical
reaction in the presence of other gases can be easily explained
considering the absorption spectra of CH4 and CO2. It has been
widely reported that CO2 does not absorb light at wavelengths
larger than 200 nm,[8] whereas CH4 and N2 are transparent and
do not absorb UV light at all. In contrast, H2S has an absorp-
tion from 270 nm into the deep UV region. Figure 2 presents
the UV absorption spectrum of H2S gas. The sharp absorption
lines with fine structure observed are characteristic of gas-
phase molecules, whereas the broad absorption band under-
neath indicates the occurrence of some H2S aggregation at
this concentration even in the gas phase.
The quantum yield for light-induced H2 generation was cal-
culated by dividing the moles of formed H2 by the number of
photons absorbed by H2S for mixtures of 9.5 vol% H2S in N2
upon 254 nm irradiation. By determining the lamp flux using
a calibrated photodiode, it was found that 0.53 H2 molecules
were produced per photon absorbed when UV irradiation at
254 nm was used, and only 0.016 and 2.6ꢂ10ꢀ5 for 300 and
354 nm, respectively. The absolute quantum yield at 254 nm is
around the theoretical limit of 0.5. This quantum yield value
means that the formation of a H2 molecule is a process requir-
ing two photons. A simple rationalization of this quantum
yield is that each photon would break a HꢀS bond with 100%
Considering the UV absorption spectrum of H2S shown in
Figure 2, the overall H2S splitting was also tested at longer
wavelengths. Specifically, the photochemical H2S splitting was
performed at 300 and 354 nm. A mixture of 10.8 vol% H2S in
N2 was used in this study, monitoring the H2 evolution and H2S
disappearance as depicted in Figure 3.
efficiency. This hypothesis would be compatible with the prior
[6,7]
C
observation of a HS radical in the photolysis of H2S. This im-
plies that when using 254 nm, the absorption of a single
As observed in Figure 3, H2S irradiation at 300 or 354 nm did
not allow the detection of measurable H2 amounts. These neg-
ative results contrast with those observed upon 254 nm irradi-
ation, in which H2 evolution occurs. This dependence of the ir-
photon of such energy must lead to the cleavage of one SꢀH
C
bond, resulting in the generation of one H atom and HS . For-
C
mation of H2 from H would require, then, another photon,
C
C
either by splitting the HS or by reaction of H with photoexcit-
ed H2S, among other possibilities. Scheme 1 summarizes this
proposal. In any case, a quantum yield of 0.53ꢁ0.01 indicates
a very high efficiency of photochemical HꢀS bond cleavage. In
spite of this high quantum yield, an estimation of electricity
costs based on the quantum yield value at 254 nm (around
0.5), considering the photon flux of the 45 W Hg lamp (6.09ꢂ
1015 photonssꢀ1), and assuming a cost of 0.1 Ekwhꢀ1, gives
a value of 0.01 gEꢀ1 for H2 and 0.18 gEꢀ1 for S. This cost is
mainly limited by the efficiency of the electricity-to-light
conversion. Lamps with higher yield of emitted photons per
consumed W would be highly desirable.
To further investigate the reaction mechanism looking for
the detection of possible excited states or intermediates, gas-
phase transient absorption spectroscopy experiments in the
microsecond time scale were done using a Nd:YAG laser at
Figure 2. Gas-phase UV spectrum of 9.5 vol% of H2S in N2.
ChemSusChem 2017, 10, 1 – 6
3
ꢁ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ