Open Access
Add to BibSonomy
Abstract
The measurement of the hydrogen desorption properties of sodium alanate (${\rm{NaAl}}{{\rm{H}}_4}$) powder compacts presents a particular challenge due to various material changes. Therefore, a parallel measurement method for both spectral and gravimetrical data acquisition has been developed. The optical tracking of such chemical reactions with gaseous exhalation requires a special apparatus, since hydrogen absorption/desorption alters the granularity and volume of the sample. The technique combines a commercial Fourier transform infrared spectrometer with a self-developed heatable attenuated total reflection (ATR) cuvette that maintains the indispensable close contact between the ATR medium and sample. In particular, strongly absorbing or scattering samples can be characterized with respect to their transmission changes and their mass loss, due to the evanescent field’s penetration depth of a few micrometers into the bulk of the sample. In this work, the first and second desorption step of 2 mol. % ${\rm{CeC}}{{\rm{l}}_3}$- and ${\rm{TiC}}{{\rm{l}}_3}$-doped ${\rm{NaAl}}{{\rm{H}}_4}$ compacts has been investigated. It was shown that the technique is able to monitor the hydrogen content in an alanate compact by means of the purely optically measured signal. In addition, a detailed analysis of the individual phase transformations was carried out. While the first reaction from the tetragonal to the monoclinic phase (${\rm{NaAl}}{{\rm{H}}_4} = \gt {\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$) showed a conversion nearly linear in time, the following reaction from the monoclinic to the hydride phase (${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6} = \gt {\rm{NaH}}$) showed a nonlinear desorption process. The phase changes were accompanied by discontinuous intensity changes of certain absorption bands, as well as by nonconstant changes over time. Therefore, the second-phase conversion could be divided into two regions with different desorption behaviors. The optical and gravimetric measurements were supported by x-ray diffraction measurements to verify the composition of the materials after hydrogen desorption. For the technical application of this method, e.g., for use as a level sensor, these special optical features are essential and must be taken into account.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. INTRODUCTION
For the elucidation or monitoring of chemical reactions, a variety of physical and chemical methods is used. Spectroscopic and chromatographic methods are established in many laboratories, as well as thermogravimetry (TGA) or differential scanning calorimetry (DSC) [1–4]. A more specific spectroscopic method is the technique of attenuated total reflection (ATR) in combination with Fourier transform infrared (FTIR) spectroscopy [5]. The instrumentation and interpretation of the spectra is more complex compared to classical transmission spectroscopy, but offers several advantages.
ATR spectroscopy is mainly used in the mid-infrared region (MIR; ${{400 {-} 4000}}\;{\rm{c}}{{\rm{m}}^{- 1}}$), where most of the fundamental vibrations of typical molecular groups are found [6]. For measurement, a highly refractive ATR unit is brought into tight contact with the sample surface. Typically, the technique is used to analyze fluids and their chemical composition. Furthermore, thin or monomolecular layers that are formed directly on the surface of the ATR unit can be studied [7]. The investigation of solids is more sophisticated than other methods because a profound contact between the ATR unit and the sample material is required. Even distances in the submicrometer region may lead to an insufficient, or even no, penetration of the evanescent field of the infrared light into the sample material. Especially in the case of a powder, the contact is incomplete. To overcome this, the powder material is compacted into pellets to form an optically smooth surface for the ATR unit. In contrast with transmission spectroscopy, the powder does not need to be incorporated into a potassium bromide (KBr) pellet. This allows the reactivity of the powdery grains to be studied because they are not isolated by a surrounding KBr matrix [8]. Moreover, strongly absorbing materials are also accessible for ATR measuring. The penetration depth of the evanescent field is in the micrometer region, which represents an adequate “sample thickness.”
Chemical reactions with gas release are not only characterized by their stoichiometry, but also by their starting and end products. Besides that, several phases can occur during the reaction simultaneously, or other intermediate products may emerge. The course of the reactions, as well as the respective phase composition, can be traced in the infrared range if molecular vibrational bands are present, which correspond to species to be transformed. For metal hydrides like sodium alanate, it was necessary to find out whether the analysis of the vibrational bands of the optical signal would be sufficient to determine the existing phase composition. Since the hydrogen content of the material scales with the phase composition, precise information about the amount of hydrogen bound within the sample can be provided. The idea was that in the best case, the evaluation of only a few absorption peaks would allow the remaining hydrogen content to be monitored during a desorption process. Consequently, the technique could be used to develop a filling-level sensor for metal hydride tanks.
For sodium alanate, the hydrogen desorption takes place by heating in a three-step decomposition reaction with a theoretical hydrogen capacity of 7.4 wt. % (pure material) [9,10],
For this work, we investigated the hydrogen desorption of catalyst-doped sodium alanate pellets by FTIR-ATR spectroscopy parallel to gravimetry. The optical and mass signals showed a correlation with each other that can be used as a measurement principle for a purely optics-based filling-level sensor. Furthermore, we showed the ability of the measurement setup to follow the hydrogen release at the solid–solid phase changes during the desorption process that cannot be revealed by the discrete single measurement of the optical or gravimetrical signal alone.
2. EXPERIMENTAL
A. Sample Preparation
Sodium alanate (${\rm{NaAl}}{{\rm{H}}_4}$—hydrogen storage grade) and catalysts (${\rm{TiC}}{{\rm{l}}_3}$, ${\rm{CeC}}{{\rm{l}}_3}$) were used as received from Sigma Aldrich. For each milling process, 3 g of ${\rm{NaAl}}{{\rm{H}}_4}$ with 2 mol. % of catalyst were weighed into a milling pot with 20 milling balls (10 mm) made of hardened steel that is hermetically sealed. All the steps were performed in a glovebox (MBraun) with nitrogen atmosphere (${{\rm{O}}_2} \le {{20}}\;{\rm{ppm}}$, ${{\rm{H}}_2}{\rm{O}} = {0.5}\;{\rm{ppm}}$), as the material is highly reactive to oxygen or humidity. The ${{\rm{TiCl}}_3}$-doped powder was milled for 2 h and the ${\rm{CeC}}{{\rm{l}}_3}$-doped material for 4 h, with a cooling interval of 0.5 h after 0.5 h of milling. The rotational speed was set to 350 rpm in a planetary ball mill “S1000” from Retsch. After milling, the powder material was placed in a hydrogen atmosphere of 100 bar at 120°C for 2 h to ensure that only ${\rm{NaAlH}}_4$ is present. For the sample preparation, 500 mg of the milled powder material was uniaxially compacted into a cylindrical mold to form a pellet with 10 mm diameter and approximately 5 mm height. For the compaction, a pressure of 250 MPa was applied by using an “Atlas” hydraulic press from Specac. Polished press shoes were used for the compaction so that the surface of the pellet became smooth and visually reflective (see Fig. 1). This is substantial for a good signal-to-noise ratio of the FTIR-ATR spectroscopic measurements.
Fig. 1. Optical image of a freshly manufactured ${\rm{TiC}}{{\rm{l}}_3}$-doped ${\rm{NaAl}}{{\rm{H}}_4}$ pellet (500 mg; Ø10 mm), compacted at 250 MPa. The shimmery parts represent the reflections of the smooth surface.
B. Measurement Setup
Figure 2 gives a schematic illustration of the experimental setup. The central component is a gas-tight ATR cuvette (for details, see Section 2.C) with a gas inlet/outlet. A wireless temperature sensor is used to control the contactless radiative heating system.
Fig. 2. Schematic illustration of the parallel FTIR-ATR spectroscopic and gravimetric measurement setup.
Fig. 3. (a) FTIR setup with the ATR cuvette mounted on the precision balance and (b) optical beam path within the sample chamber of the spectrometer. For details of the ATR cuvette, see Fig. 5.
The setup consists of an FTIR spectrometer (Spectrum65) from Perkin Elmer for the measurement of the spectral range from ${{400 {-} 4000}}\;{\rm{c}}{{\rm{m}}^{- 1}}$. The spectrometer is equipped with Harrick’s reflective optics Seagull for light coupling into the ATR cuvette and collecting the reflected light. Figure 3(a) shows the measuring site (FTIR spectrometer with inserted ATR cuvette and precision balance), and Fig. 3(b) explains the beam path.
The ATR method makes use of the evanescent wave that penetrates into the medium with the lower index of refraction while the light is totally internally reflected at the interface between the sample and the ATR unit, as schematically presented in Fig. 4. The penetration depth is determined by ${{1/{\rm{e}}}}$ for the amplitude of the wave function and depends on the wavelength of light, the angle of incidence, and the indices of refractions of the sample material and the ATR unit (for more details, see Ref. [5]).
The radiation heating system in the sample chamber consists of four halogen lamps of 33 W each, controlled via a 230 V solid-state relay and a PID controller CN7500 from Omega. The four lamps are arranged symmetrically around the cuvette and enclosed by a radiation shield made of aluminum foil. A Cubis MSE524P-100-DU balance from Sartorius with a modified wind shield is used for the gravimetric measurement. The resolution of this balance is 0.1 mg, with a maximum load of 520 g.
C. ATR Cuvette
The requirements placed on the cuvette by the process parameters are operating temperatures up to 200°C, pressures up to 100 bar (10 MPa), and a total weight below the maximum load of the balance. On the optical side, a transmission range of 5–12 µm has to be provided. In addition, an MIR transparent separating layer between the sample and the ATR element is required as protection against mechanical exposure and corrosive chemical reactions [21]. Furthermore, a gas inlet and outlet providing separation from the surrounding atmosphere is needed. A detailed presentation of the cuvette structure is given in Fig. 5.
Fig. 5. Schematic representation of the measuring cuvette with detailed views of the sample chamber and tube fitting. Cuvette head (upper part) and cuvette body (lower part) can be separated from each other via screw connections for inserting and removing the sample. The seals separating the sample chamber from the atmosphere are placed between the screwable parts.
The material for the cuvette body is aluminum of type EN-AW-6082, which is light, stable, easy to machine, insensitive to embrittlement by ${{\rm H}_2}$, and therefore also used for tank liners [22]. For the ATR unit, a hemisphere made of zinc selenide (ZnSe) was chosen. This material offers a wide transmission range of 0.48–20.5 µm, is thermally stable up to 300°C, nonhygroscopic, and sufficiently stable to withstand the mechanical strain [23,24]. Moreover, the thermal expansion of ZnSe is below that of aluminum, which prevents additional stresses during heating. With a base diameter of 25 mm, the ZnSe hemisphere is well adapted to the standard Harrick Seagull optics for ATR measurement.
The head section of the cuvette is provided with a hemispherical bore and openings for the beam path of the spectrometer. The bore is loaded with a thin protective insert made of silicone rubber to accommodate the ATR element. In order to avoid solid–solid reactions, a CVD diamond platelet from Diamond Materials GmbH is used as a separating layer between the powder compact and the ATR unit. The plate dimensions are a diameter of 15 mm and a thickness of 100 µm with optically polished surfaces. Diamond is used because it is chemically inert, hard, and has a high transmission in the MIR and a similar refractive index to ZnSe [25].
The cylindrical sample chamber underneath the ATR element serves to hold a pellet of 10 mm diameter and a maximum height of 10 mm. The pellets are guided through a sample sleeve with gas channels to simplify sample handling and to protect the sample chamber. For a continuous optical contact between the sample and the ATR element, a spring is used. The contact spring presses the pellet against the diamond plate with a pressure force of up to 740 N (adjustable by the spring screw). In addition, the spring is able to compensate thermal and/or structural volume changes as they are present for the ${\rm{NaAl}}{{\rm{H}}_4}$ material during the phase transformations at hydrogen absorption and desorption [10].
The supply and discharge of gaseous reactants is realized by pipe fittings (PEEK ferrules and nuts from Machery Nagel) and a conical bore in the cuvette body. The connection to the gas inlet takes place via a quick coupler, and a backpressure valve (7 kPa) is placed at the gas outlet (Swagelok). In this way, a backflow of the surrounding atmosphere into the cuvette is prevented.
For the seals, fluorocarbon rubber o-rings are used (placed between the screwable parts in Fig. 5) because they are temperature-resistant up to 200°C and almost impermeable to hydrogen.
D. Measurement Procedure
The overall experiment is conducted via a LabVIEW routine for time-dependent control and acquisition of the temperature, weight, and spectral values.
The temperature is controlled by a temperature sensor (type PT100) integrated in the pressure plunger below the sample. The data transmission takes place by a self-developed Bluetooth radio transmitter every 5 s with a resolution of 0.5°C. Wireless data transfer is required because the cuvette has to be positioned force-free on the analytical balance for the gravimetrical measurement.
The sample pellet is inserted into the ATR cuvette inside a glove box. To obtain an isothermal measurement, the cell is pressurized with 40 bar (4 MPa) nitrogen to prevent premature hydrogen release during heating to the desorption temperature. Once the desorption temperature is reached, the backpressure valve is opened to release the nitrogen pressure, and the desorption process begins. At that point, the ATR cuvette is quickly repositioned and time-controlled data recording is started for the spectral and weight measurements.
In order to obtain a fast optical data acquisition (${\lt}{{10}}\;{\rm{s}}$), each spectroscopic measurement was taken by one single scan with a resolution of ${{4}}\;{\rm{c}}{{\rm{m}}^{- 1}}$. For the reflection spectrum $R(\tilde \nu)$ at each wavenumber $\tilde \nu$, the reflectance of the sample ${R_s}(\tilde \nu)$ and the background reflectance ${R_b}(\tilde \nu)$ must be calculated as follows:
The background spectrum is obtained by taking the sample from the cuvette and replacing it with a high-reflecting gold plate that is pressed against the diamond. Otherwise, total reflection could already occur at the transition from the ATR unit to the diamond plate. In this case, the spectral fingerprint of the diamond would not be contained in the background spectrum and would appear in the sample spectrum.The conversion of the reflection spectrum into the absorbance $A(\tilde \nu)$ follows the Lambert–Beer law,
3. RESULTS AND DISCUSSION
A. Correlated Hydrogen Desorption Measurements of ${\rm{NaAl}}{{\rm{H}}_4}$
In the following, the measurement results are explained on the basis of a spectral and gravimetrical measurement series for the ${\rm{TiC}}{{\rm{l}}_3}$-doped material.
Figure 6 shows the ATR spectra at the hydrogen desorption of the ${\rm{NaAl}}{{\rm{H}}_4}$ phase of the first cycle at 130°C. The spectral data acquisition took place every 2 min. The lower spectra correspond to the starting point (0 min) of the desorption process and a fully hydrogen loaded pellet. During desorption, the intensity of the absorption band of the ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ phase (peak at ${{1256}}\;{\rm{c}}{{\rm{m}}^{- 1}}$) gradually increases. A more detailed description of the individual spectral areas of the sodium alanate material can be found in a previous work [21]. The upper curve (18 min) indicates the end of the phase transformation. Within the optical measurements, the data acquisition shows a high reproducibility of the absorption intensities, with a spectral accuracy of ${0.5}\;{\rm{c}}{{\rm{m}}^{- 1}}$. It is noticeable that within the first minutes, a deviation in the range of the ${\rm{NaAl}}{{\rm{H}}_4}$ phase (peak at ${{1630}}\;{\rm{c}}{{\rm{m}}^{- 1}}$) is present. The intensity of the absorption signal rises from the lower (0 min) to the dashed curve (6 min) before it starts to decrease. Typically, a continuous decrease would be expected during the phase transformation. The origin of the observed effect has not been completely clarified yet, but it is probably due to the measurement procedure and not to a material-specific property. On that assumption, the initially rising absorption values result from a relaxation process along which the contacting of the sample to the ATR element improves after the nitrogen backpressure has been released. This automatically leads to the question why this effect is not visible for the whole spectral range. First, shorter wavelengths have a reduced penetration depth of the evanescent wave, which makes them more sensitive for contacting compared to the penetration depth of higher wavelengths. Second, at the beginning of the desorption process, only a small portion of the second phase, and thus also of the associated absorption band, has formed. Therefore, the effect is most obvious at the more pronounced absorption band of the ${\rm{NaAl}}{{\rm{H}}_4}$ phase at the beginning of the desorption process. The effect was not observed in comparative measurements using a fiber-bound probe prototype with a flat ATR element from art photonics GmbH [26]. This may confirm the assumption that the observed effect occurs through a relaxation process because the relatively small ATR element (${{\emptyset}}{{2}}\;{\rm{mm}}$) of the probe allows a much stronger contact pressure to be produced. But due to the fiber connection, no simultaneous gravimetric data acquisition can be performed.
Fig. 6. FTIR-ATR spectra of hydrogen desorption of a ${\rm{NaAl}}{{\rm{H}}_4}$ pellet (containing 2 mol. % ${\rm{TiC}}{{\rm{l}}_3}$ as catalyst) at 130°C and a backpressure of 7 kPa (data acquisition interval: 2 min).
The appearance of a well-formed isosbestic point (point of constant intensity) between the two phase peaks proves the assumption that there is no further species involved in the decomposition from ${\rm{NaAl}}{{\rm{H}}_4}$ to ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$.
Figure 7 shows the associated gravimetrical measurement to the spectral data acquisition. Each point represents one spectral measurement. The measurement accuracy is ${{\pm}}{0.5}\;{\rm{mg}}$ based on a regression analysis, which corresponds to a measurement error of about ${{\pm}}{0.1}\%$ by weight. The measurement shows a desorbed hydrogen quantity of approximately 3.3 wt. % in the course of time. The theoretical value of 3.7 wt. % is slightly higher, but the use of the catalyst also reduces the storage capacity of the material $({\approx} {3.5}\;{\rm{wt}}.\;\%)$. Moreover, an inhomogeneous distribution of the reacting solid phases and/or an incomplete synthesis route can affect the efficiency of the overall reaction. This may lead to the formation of inert species that do not participate in the hydrogen storage process, reducing the storage capacity, respectively [27].
Fig. 7. Gravimetric measurement of hydrogen desorption during phase conversion associated with the spectral data of Fig. 6.
For both spectral (Fig. 6) and gravimetrical (Fig. 7) measurements, a completed desorption is reached within 18 min. In order to show the correlation between the parallel measurement techniques, a normalization was done. Figure 8 shows the percentage phase transformation of the normalized values from the gravimetrical and spectral measurements over time.
Fig. 8. Percentage phase transformation of the normalized values from spectral and gravimetric measurements of Fig. 6 and Fig. 7, respectively. The spectral normalization was applied for the ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ phase at a wavenumber of ${{1256}}\;{\rm{c}}{{\rm{m}}^{- 1}}$.
The gravimetrical normalization is based on the maximum amount of hydrogen desorbed. For the spectral normalization, the intensity changes of the ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ phase at a wavenumber of ${{1256}}\;{\rm{c}}{{\rm{m}}^{- 1}}$ were compared to each other. Both signals correlate very well with each other, which enables the use of the purely optical signal as the basis of a reliable filling-level sensor.
B. Measurement of the ${\rm{NaAl}}{{\rm{H}}_4}$ and ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ Desorption Process
For the study of the hydrogen desorption of ${\rm{NaAl}}{{\rm{H}}_4}$ and ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$, the individual reaction steps should be regarded. This can either be done by two different temperatures or different pressure levels to suppress the second desorption step [27–29]. Here we can use the parallel in situ measurement of the FTIR-ATR spectra and the gravimetrical data as a powerful tool to define the start and end point of each reaction step and to trace the reaction processes in detail.
Figure 9 shows the result of the gravimetric measurement during the hydrogen desorption of ${\rm{CeC}}{{\rm{l}}_3}$-doped (2 mol. %) alanate performed at a desorption temperature of 150°C. The determined regions were defined by the analysis of the related spectral data given in Figs. 10 and 11. The first region (I) shows the decomposition of ${\rm{NaAl}}{{\rm{H}}_4}$ into ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ corresponding to Eq. (1) with fast hydrogen desorption. During desorption of the first phase, approximately 3.0 wt. % of hydrogen is released from the sample. The second region (II) shows the transformation of ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ to the hydride (NaH) phase as given in Eq. (2). The hydrogen release of approximately 1.3 wt. % takes place at a much slower desorption rate and shows a nonlinear slope. This continues by the third region (III), with another hydrogen release of 0.9 wt. %. The overall desorption process yields a total hydrogen amount of 5.2 wt. % at a desorption time of 560 min. Due to the remaining slope at the end of the curve, it is assumed that a small amount of hydrogen is left inside the sample. For a fast release of the remaining hydrogen, a higher temperature or prolonged desorption interval is needed.
Fig. 9. Gravimetric measurement of hydrogen desorption from ${\rm{CeC}}{{\rm{l}}_3}$-doped (2 mol. %) ${\rm{NaAl}}{{\rm{H}}_4}$ at 150°C and a backpressure of 7 kPa.
Fig. 10. ATR spectra of ${\rm{CeC}}{{\rm{l}}_3}$-doped (2 mol. %) ${\rm{NaAl}}{{\rm{H}}_4}$ showing the transformation to the ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ phase according to region I of Fig. 9. The time interval between the spectra is 2 min.
Fig. 11. ATR spectra of ${\rm{CeC}}{{\rm{l}}_3}$-doped (2 mol. %) ${\rm{NaAl}}{{\rm{H}}_4}$ showing the decomposition reaction of ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ according to region II and III of Fig. 9.
Figure 10 depicts the FTIR-ATR spectra for the first region to the gravimetrical data of Fig. 9. In the course of time, the intensity changes of the absorption band for the ${\rm{NaAl}}{{\rm{H}}_4}$ and ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ phases took place for the ${\rm{CeC}}{{\rm{l}}_3}$-doped material with similar behavior to that described in the previous section for the ${\rm{TiC}}{{\rm{l}}_3}$-doped material. The initial nonlinearity for the spectral change of the ${\rm{NaAl}}{{\rm{H}}_4}$ phase is present as well. This also corroborates the assumption that this effect is more likely to arise from the measurement procedure than from a material property.
After the first desorption step, the FTIR-ATR spectral changes took place as presented in Fig. 11. With the course of time, a transition from the lower to the dashed graph is obvious, according to 16 and 160 min, respectively. This time sequence corresponds to the second region (II) of Fig. 9. During this transformation, the peak of the ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ phase shows only small changes in intensity. The residual spectral fractions show a gradual shift towards higher absorption levels, especially in the range of ${{1000}}\;{\rm{c}}{{\rm{m}}^{- 1}}$ and ${{1500}}\;{\rm{c}}{{\rm{m}}^{- 1}}$. After the state of the dashed graph is reached, only minor spectral changes take place over a comparatively long period of time. The upper graph marks the final spectral acquisition at the end of the desorption measurement at 560 min.
For ${\rm{TiC}}{{\rm{l}}_3}$-doped (2 mol. %) material, similar results were observed. Figure 12 shows the gravimetric measurement at a desorption temperature of 150°C. As with ${\rm{CeC}}{{\rm{l}}_3}$-doped material, the hydrogen desorption process may be divided into three regions. However, these are also apparent without the inclusion of the optical measurement data, since the curve bends off significantly at the transition from II to III. Furthermore, the ${\rm{TiC}}{{\rm{l}}_3}$-doped material shows a much faster reaction rate. Typically, the ${\rm{CeC}}{{\rm{l}}_3}$ catalyst is reported to show superior kinetics compared to ${\rm{TiC}}{{\rm{l}}_3}$ for the powdery material [16]. But the synthesis route and material preparation, especially compaction, may have a significant influence on this behavior. The transition from ${\rm{NaAl}}{{\rm{H}}_4}$ to ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ (${\rm{I}} = \gt {\rm{II}}$) is completed within the first 4 min. Due to the high desorption rate, a part of the hydrogen release has already taken place in the short period between the backpressure release and the repositioning of the cuvette and is therefore not included in the gravimetric measurement. This results in a measured hydrogen release of approximately 2.3 wt. % for the first region (I).
Fig. 12. Gravimetric measurement of hydrogen desorption from ${\rm{TiC}}{{\rm{l}}_3}$-doped (2 mol. %) ${\rm{NaAl}}{{\rm{H}}_4}$ at 150°C and a backpressure of 7 kPa.
However, the weighing of the pellet sample before and after desorption shows in total an equal amount of hydrogen released as with the ${\rm{CeC}}{{\rm{l}}_3}$-doped material. Therefore, the hydrogen release of the first desorption step is in the same range for both materials. This continues by a similar hydrogen release of the second region (II) with 1.4 wt. %, followed by 0.8 wt. % of the third region (III).
Figure 13 shows the spectral changes for regions II and III of the gravimetric data of Fig. 12. Since desorption proceeded at such a high rate, the relaxation effect is still visible in the decomposition reaction of ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$, even after the phase transformation of ${\rm{NaAl}}{{\rm{H}}_4}$ from the first region (I) is completed. This shows in the intensity jump from the lower curve (4 min) to the following spectral measurement. Basically, the measurement showed only moderate expression of the spectra, which can be explained by a poor contact with the ATR element. Comparative measurements of the ${\rm{TiC}}{{\rm{l}}_3}$ samples (not shown here) show spectral characteristics equivalent to those of the ${\rm{CeC}}{{\rm{l}}_3}$-doped pellets. Nevertheless, the progression from the lower to the dashed curve (40 min) and then to the upper curve (240 min) is of comparable quality to the curves obtained when using the ${\rm{CeC}}{{\rm{l}}_3}$ material.
Fig. 13. ATR spectra of ${\rm{TiC}}{{\rm{l}}_3}$-doped (2 mol. %) ${\rm{NaAl}}{{\rm{H}}_4}$ showing the decomposition reaction of ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ according to regions II and III of Fig. 12.
Looking at Eqs. (1)–(3), one could assume that the assigned regions reflect the three-step decomposition process. However, even with an added catalyst, the final decomposition process only takes place at significantly higher temperatures (375°C and higher) [30,31]. This was also confirmed by x-ray powder diffraction (XRD) measurements to analyze the composition of the sample material after the desorption process. The measurements were performed in Debye-Scherrer geometry (device: Stadi P, STOE & Cie. GmbH; Ge[111] monochromator, ${\rm{C}}{{\rm{u}}_{{\rm K} \alpha 1}}$ radiation, $\lambda = {1.54060}\;{\unicode{x212b}}$; detector: Mythen1K) with the samples filled in sealed capillaries (under argon). Quantitative phase analyses (Rietveld method) of the measurements were performed using the program package GSAS-II [32]. The results are given in Table 1 (wt. %).
Table 1. Mass Distribution of the Detected Components by Quantitative Phase Analysis of the X-Ray Powder Diffractograms
The results show the presence of four main components. The major part is deposited aluminum (54% and 56%), followed by NaH (29% and 33%) and ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ (10% and 4%) for the ${\rm{CeC}}{{\rm{l}}_3}$ and ${\rm{TiC}}{{\rm{l}}_3}$ catalysts, respectively. Moreover, 7% of sodium chloride was detected. This results from the reaction of excess sodium with the chlorine content of the catalysts. Sodium alanate in its initial state was not detected, which indicates a complete phase transformation of the first step. The low residual content of the second-phase (${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$) supports the finding that the hydrogen desorption of the sample was not completely finished. As the third phase (NaH) is still present in large quantities, the hydrogen release according to Eq. (3) can be excluded.
The assigned regions II and III both belong to the process of the second-phase transformation. Nevertheless, the spectral measurements suggest a change in the reaction behavior. With increasing reaction progress and thus increasing presence of reaction products, the reaction-limiting step could change. Emerging hydrogen must first pass through areas that have already reacted. This could also lead to the development of different (local) reaction areas, since material close to the surface provides easier hydrogen release than internal material. The exact understanding of these effects requires further investigations of the reaction kinetics on compacted samples. Moreover, the observed optical signal changes must be taken into account with the second-phase conversion for use as a level sensor, as the signal intensity no longer scales linearly with the level.
4. CONCLUSION
By parallel gravimetric and spectroscopic in situ measurement, the hydrogen desorption process of ${\rm{TiC}}{{\rm{l}}_3}$- and ${\rm{CeC}}{{\rm{l}}_3}$-doped ${\rm{NaAl}}{{\rm{H}}_4}$ compacts has been investigated. The technique shows the possibility of monitoring the hydrogen level in an alanate compact by the pure optically measured signal, as the percentage-phase transformation of the normalized values from spectral and gravimetric measurements correlate linearly with each other. However, the optical observation showed a discontinuous change of the absorption intensity at ${{1630}}\;{\rm{c}}{{\rm{m}}^{- 1}}$ for ${\rm{NaAl}}{{\rm{H}}_4}$ at the beginning of desorption. This can be attributed to the improved contact of the ATR element to the sample surface after releasing the backpressure. Especially for technical frameworks with fluctuating pressure levels, e.g., in the case of cycling the material with hydrogen in a tank, this can be of high relevance and should therefore be taken into account and examined in detail, depending on the application. The discontinuity was present independently of the catalyst used. The same applies to the basic reaction process for the first decomposition step with similar shapes of the spectral and gravimetric curves for both materials.
In a continuous measurement at an elevated temperature of 150°C, the decomposition reaction of the first- and second-phase transformations was observed. For the ${\rm{CeC}}{{\rm{l}}_3}$-doped material, the gravimetrical data showed two obvious regions with a linear decomposition of ${\rm{NaAl}}{{\rm{H}}_4} = \gt {\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ followed by a nonlinear hydrogen desorption. In contrast, the spectral data showed three reacting regions as the nonlinear decomposition was further divided into two regions with varying degrees of spectral change in absorption intensity over time. For the ${\rm{TiC}}{{\rm{l}}_3}$-doped material, the three regions were directly obvious in the gravimetrical data. With the use of x-ray diffraction measurements, the composition of the samples were investigated after the desorption process. The results showed that both of the optically assigned regions (II and III) for the nonlinear hydrogen release belong to the decomposition of ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6} = \gt \;{\rm{NaH}}$ of the second decomposition step. Due to the nonlinear shape of the desorption curve and the different accompanying regions in the optical data, it is assumed that the reaction path during the second desorption process could change. A further aspect results from the consideration of the thermal decomposition reaction of pristine ${\rm{NaAl}}{{\rm{H}}_4}$. A two-stage transformation of the second phase from monoclinic $\alpha - {\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ to cubic $\beta - {\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ takes place before it decomposes further to NaH [33,34]. With doped material, however, the decomposition typically takes place in one step. Here the question arises about whether the single-stage decomposition for the doped material at the second phase can be suppressed by compaction back to a two-stage process. Alternatively, the two-stage transformation of ${\rm{N}}{{\rm{a}}_3}{\rm{Al}}{{\rm{H}}_6}$ could just be made visible by compaction. For example, if the energy barriers for the decomposition are very close to each other and the process regularly takes place very quickly for the powdery material, it would appear as a one-step process. By compaction, the individual phase transitions could be inhibited, accompanied by spread energy barriers. This in turn would also explain why for the ${\rm{TiC}}{{\rm{l}}_3}$ sample, the two regions (II and III) are also found in the gravimetric measurement. With faster kinetics, conversion processes are more strongly separated due to the temperature-dependent Arrhenius behavior. The temperature range applied then also has a decisive influence on the reaction process. Consequently, for ${\rm{CeC}}{{\rm{l}}_3}$, a comparable picture to ${\rm{TiC}}{{\rm{l}}_3}$ could be obtained at higher temperatures for the desorption process, or for ${\rm{TiC}}{{\rm{l}}_3}$, an equivalent picture to ${\rm{CeC}}{{\rm{l}}_3}$ at lower temperatures. In order to clarify this phenomenon, the reaction mechanism should be investigated within the scope of a kinetic analysis study.
Furthermore, it needs to be clarified whether the significantly slower reaction of the second process can still reflect the total hydrogen loading of the sample, since the ATR method is a surface-sensitive method.
Therefore, the nonlinear behavior of the second-phase transformation needs to be taken into account for use as a level sensor. In this case, a more sophisticated calibration needs to be done. In addition, a high sensitivity is required as the intensity changes in absorption become relatively small. On the other hand, whether the increased temperature and the significantly slower hydrogen release of the second phase decomposition justify this additional expenditure at all is a question in its own right, or whether the technical application as a level sensor should be limited to the first desorption step.
Funding
Bundesministerium für Bildung und Forschung (ProfUnt grant no. 13FH063PX5); HUBER+SUHNER Cube Optics AG; Anleg GmbH.
Acknowledgment
We gratefully acknowledge the financial support for this research from the Federal Ministry of Education and Research, HUBER+SUHNER Cube Optics AG and Anleg GmbH.We thank H. J. Beck for the manufacturing of the mechanical cuvette components. Special thanks go to art photonics GmbH for the first spectroscopic measurements with the flat ATR probe. In addition to that, we want to thank C. Hess from the Technical University of Darmstadt for the discussions.
Disclosures
The authors declare no conflicts of interest.
REFERENCES
1. W. M. Groenewoud, “Thermogravimetry,” in Characterisation of Polymers by Thermal Analysis, W. M. Groenewoud, ed. (Elsevier, 2003), pp. 61–76.
2. W. M. Groenewoud, “Differential scanning calorimetry,” in Characterisation of Polymers by Thermal Analysis, W. M. Groenewoud, ed. (Elsevier, 2003), pp. 10–60.
3. E. Stauffer, J. A. Dolan, and R. Newman, eds., Fire Debris Analysis (Academic, EBSCO Industries, 2008).
4. J. N. Rayner, “Spectral analysis,” in Gender and Reproductive Health, M. Sieverding, ed. (Universitätsbibliothek Heidelberg, 2001), pp. 14861–14864.
5. F. M. Mirabella, ed., Modern Techniques in Applied Molecular Spectroscopy (Wiley, 1998).
6. A. S. Gilbert, “IR spectral group frequencies of organic compounds,” in Encyclopedia of Spectroscopy and Spectrometry, J. C. Lindon, ed. (Elsevier, 2000), pp. 1035–1048.
7. M. Punzet, D. Baurecht, F. Varga, H. Karlic, and C. Heitzinger, “Determination of surface concentrations of individual molecule-layers used in nanoscale biosensors by in situ ATR-FTIR spectroscopy,” Nanoscale 4, 2431–2438 (2012). [CrossRef]
8. X. Cao and G. Fischer, “Infrared spectra of monomeric l-alanine and l-alanine-N-d3 zwitterions isolated in a KBr matrix,” Chem. Phys. 255, 195–204 (2000). [CrossRef]
9. B. Sakintuna, F. Lamari-Darkrim, and M. Hirscher, “Metal hydride materials for solid hydrogen storage: a review⋆,” Int. J. Hydrogen Energy 32, 1121–1140 (2007). [CrossRef]
10. A. Züttel, ed., Hydrogen as a Future Energy Carrier (Wiley-VCH, 2004), pp. 216–231.
11. J. O. Jensen, Q. Li, R. He, C. Pan, and N. J. Bjerrum, “100–200°C polymer fuel cells for use with NaAlH4,” J. Alloys Compd. 404–406, 653–656 (2005). [CrossRef]
12. P. Pfeifer, C. Wall, O. Jensen, H. Hahn, and M. Fichtner, “Thermal coupling of a high temperature PEM fuel cell with a complex hydride tank,” Int. J. Hydrogen Energy 34, 3457–3466 (2009). [CrossRef]
13. B. Bogdanović and M. Schwickardi, “Ti-doped alkali metal aluminium hydrides as potential novel reversible hydrogen storage materials,” J. Alloys Compd. 253–254, 1–9 (1997). [CrossRef]
14. D. L. Anton, “Hydrogen desorption kinetics in transition metal modified NaAlH4,” J. Alloys Compd. 356–357, 400–404 (2003). [CrossRef]
15. G. Sandrock, K. Gross, and G. Thomas, “Effect of Ti-catalyst content on the reversible hydrogen storage properties of the sodium alanates,” J. Alloys Compd. 339, 299–308 (2002). [CrossRef]
16. B. Bogdanović, M. Felderhoff, A. Pommerin, F. Schüth, and N. Spielkamp, “Advanced hydrogen-storage materials based on Sc-, Ce-, and Pr-doped NaAlH4,” Adv. Mater. 18, 1198–1201 (2006). [CrossRef]
17. C. Pohlmann, L. Röntzsch, J. Hu, T. Weißgärber, B. Kieback, and M. Fichtner, “Tailored heat transfer characteristics of pelletized LiNH2–MgH2 and NaAlH4 hydrogen storage materials,” J. Power Sources 205, 173–179 (2012). [CrossRef]
18. G. A. Lozano, C. N. Ranong, J. M. Bellosta von Colbe, R. Bormann, J. Hapke, G. Fieg, T. Klassen, and M. Dornheim, “Optimization of hydrogen storage tubular tanks based on light weight hydrides,” Int. J. Hydrogen Energy 37, 2825–2834 (2012). [CrossRef]
19. Y. F. Khalil, S. M. Opalka, and B. L. Laube, “Experimental and theoretical investigations for mitigating NaAlH4 reactivity risks during postulated accident scenarios involving exposure to air or water,” Process Saf. Environ. Prot. 91, 463–475 (2013). [CrossRef]
20. G. A. Lozano, J. M. Bellosta von Colbe, R. Bormann, T. Klassen, and M. Dornheim, “Enhanced volumetric hydrogen density in sodium alanate by compaction,” J. Power Sources 196, 9254–9259 (2011). [CrossRef]
21. I. Franke, O. Hentschel, D. Nitsche, M. Stops, H.-D. Bauer, and B. Scheppat, “Parallel FTIR-ATR and gravimetrical in-situ measurements on sodium alanate powder samples during hydrogen desorption,” Int. J. Hydrogen Energy 38, 16161–16167 (2013). [CrossRef]
22. R. Urbanczyk, K. Peinecke, M. Felderhoff, K. Hauschild, W. Kersten, S. Peil, and D. Bathen, “Aluminium alloy based hydrogen storage tank operated with sodium aluminium hexahydride Na3AlH6,” Int. J. Hydrogen Energy 39, 17118–17128 (2014). [CrossRef]
23. H. Qi, X. Zhang, M. Jiang, C. Liu, Q. Wang, and D. Li, “Optical properties of zinc selenide slabs at 373 and 423 K in the wavelength 2–15 µm,” Opt. 127, 5576–5584 (2016). [CrossRef]
24. R. L. C. Wu, A. W. McCormick, P. P. Pronko, and J. Keeley, “Surface hardening of ZnSe, ZnS, and ZnS/ZnSe optical materials by implantation of 1 MeV hydrogen ions,” Nucl. Instrum. Methods Phys. Res. B 59–60, 1232–1235 (1991). [CrossRef]
25. R. S. Balmer, J. R. Brandon, S. L. Clewes, H. K. Dhillon, J. M. Dodson, I. Friel, P. N. Inglis, T. D. Madgwick, M. L. Markham, T. P. Mollart, N. Perkins, G. A. Scarsbrook, D. J. Twitchen, A. J. Whitehead, J. J. Wilman, and S. M. Woollard, “Chemical vapour deposition synthetic diamond: materials, technology and applications,” J. Phys. Condens. Matter 21, 364221 (2009). [CrossRef]
26. “Fiber Optics Company,” 2020, https://artphotonics.com/.
27. G. A. Lozano, C. N. Ranong, J. M. Bellosta von Colbe, R. Bormann, G. Fieg, J. Hapke, and M. Dornheim, “Empirical kinetic model of sodium alanate reacting system (I). Hydrogen absorption,” Int. J. Hydrogen Energy 35, 6763–6772 (2010). [CrossRef]
28. G. A. Lozano, C. N. Ranong, J. M. Bellosta von Colbe, R. Bormann, G. Fieg, J. Hapke, and M. Dornheim, “Empirical kinetic model of sodium alanate reacting system (II). Hydrogen desorption,” Int. J. Hydrogen Energy 35, 7539–7546 (2010). [CrossRef]
29. X. Fan, X. Xiao, L. Chen, L. Han, S. Li, H. Ge, and Q. Wang, “Thermodynamics, kinetics, and modeling investigation on the dehydrogenation of CeAl4 -doped NaAlH4 hydrogen storage material,” J. Phys. Chem. C 115, 22680–22687 (2011). [CrossRef]
30. J. Šubrt and K. Tobola, “DTA study of the kinetics of sodium hydride decomposition,” J. Therm. Anal. 10, 5–12 (1976). [CrossRef]
31. K. Suárez-Alcántara, J. R. Tena-Garcia, and R. Guerrero-Ortiz, “Alanates, a comprehensive review,” Mater. 12, 2724 (2019). [CrossRef]
32. B. H. Toby and R. B. von Dreele, “GSAS-II. The genesis of a modern open-source all purpose crystallography software package,” J. Appl. Crystallogr. 46, 544–549 (2013). [CrossRef]
33. A. Léon, Hydrogen Technology (Springer, 2008), p. 112.
34. C. Weidenthaler, A. Pommerin, M. Felderhoff, W. Schmidt, B. Bogdanović, and F. Schüth, “Evidence for the existence of β-Na3AlH6: monitoring the phase transformation from α-Na3AlH6 by in situ methods,” J. Alloys Compd. 398, 228–234 (2005). [CrossRef]
