In Situ Diffuse Reflectance Infrared Fourier‐Transform Spectroscopy Investigation of 1‐Methylcyclopropene Adsorption in Cobalt–Formate Metal–Organic Framework

Abstract

Metal–organic frameworks (MOFs) find numerous applications due to their tunable adsorption/desorption properties. This work is focused on the spectroscopic investigation of the adsorption and release of ethylene (C₂H₄) and 1‐methylcyclopropene (1‐MCP)—natural plant growth hormone and its synthetic inhibitor—in the pores of Co₃(HCOO)₆ (Co‐FA) MOF. Using in situ diffuse reflectance infrared Fourier‐transform spectroscopy (DRIFTS), we have identified the molecular‐level interactions between the adsorbed molecules and Co‐FA pores, as evidenced by the characteristic shifts of the vibrational modes. The significant confinement of 1‐MCP in Co‐FA at room temperature was demonstrated, while moderate heating enabled its temperature‐controlled release. In comparison, weaker but visible ethylene adsorption of C₂H₄ was also demonstrated, with significant desorption readily occurring at room temperature. Validation tests on bananas confirmed the superior performance of Co‐FA over alternative MOFs, providing a link between molecular‐level structure and practical applications.

In situ diffuse reflectance infrared Fourier‐transform spectroscopy spectroscopy was used to investigate 1‐MCP adsorption in the pores of Co₃(HCOO)₆ (Co‐FA) metal–organic framework (MOF). A strong confinement of 1‐MCP at room temperature and a significant release by moderate heating, proves Co‐FA as a promising material for food preservation technologies, which was validated by using this MOF to prolong the freshness of banana fruits.

1. Introduction

Metal–organic frameworks (MOFs) are porous crystalline materials with a highly tunable pore architecture [1, ²]. Such tunability is achieved by a variety of secondary building units—inorganic cornerstones represented by metal or metal–oxide clusters—that, in turn, can be connected by a variety of organic linkers. These properties of MOFs have attracted instant research attention in fields, such as catalysis [3, 4, 5, 6, 7, 8], gas separation, storage and release [9, 10, 11, 12, 13], drug delivery [14, 15, 16], and others. Moreover, due to the chemical, mechanical, and temperature stability of different MOFs, their numerous real‐life applications are being suggested [17].

Application of MOFs in food preservation technologies represents another promising and innovative direction, which is still less explored compared to other fields [18, ¹⁹]. In a series of recent works [20, ²¹], we have focused on the MOFs with open‐metal sites for covalent bonding of 1‐methylcyclopropene (1‐MCP)—a synthetic inhibitor of ethylene (C₂H₄) receptors in plants. While ethylene acts as a natural plant hormone, 1‐MCP can block ethylene adsorption sites and prevent fruit ripening [22]. Recently, a cobalt–formate framework Co₃(HCOO)₆ (Co‐FA) was reported as an attractive candidate for 1‐MCP release in fruit preservation technologies [23]. This MOF is represented by a three‐dimensional network of octahedral Co²⁺ centers bridged by formate ligands, with channel‐like pores and thermal stability up to 300°C [24]. Despite promising results for 1‐MCP and other gas storage/separation [24, 25, 26, 27], the molecular‐level understanding of the adsorption phenomenon in Co‐FA, which is crucial for the rational design of functional materials, is still lacking.

In this work, we present the first spectroscopic study of ethylene and 1‐MCP adsorption and release in Co‐FA, utilizing in situ diffuse reflectance infrared Fourier‐transform spectroscopy (DRIFTS), Fourier‐transformed infrared spectroscopy (FTIR), and mass spectrometry. Comparative ethylene and 1‐MCP experiments provide the mechanistic insights into the binding process and reveal a potential for practical applications. The latter was validated through the fruit preservation tests, connecting molecular‐level findings to real‐world applications.

2. Results and Discussion

2.1. Sorbent Choice and Characterization

The choice of 1‐MCP sorbent for practical application is based on its ability to reversibly adsorb and desorb 1‐MCP in a controllable manner at ambient pressures and temperatures. Chopra and coauthors [28] reported no release of 1‐MCP from HKUST‐1 MOF with open copper sites. In our previous works, screening of MOFs with various open‐metal sites demonstrated a strong binding of 1‐MCP, irrespective of the metal type. Co‐FA MOF was recently reported to be an attractive 1‐MCP sorbent [23]. However, no structural studies with 1‐MCP, as well as ethylene, have been performed.

Co‐FA MOF was synthesized by a solvothermal method. The experimental powder X‐ray diffraction (PXRD) data (Figure S1) were refined in P21/m space group with unit cell parameters shown in Table 1. Upon activation aimed to remove the remaining solvent, only ca. 7% weight loss was observed (Figure S2), accompanied with a slight decrease of the cell parameters and significantly increased PXRD peak width. Importantly, after sorption and desorption of 1‐MCP (vide infra), no further significant changes were observed.

TABLE 1.

Unit cell parameters and crystallite sizes of as‐synthesized and activated Co‐FA and after 1‐MCP sorption/desorption determined by PXRD.

Sample	a, Å	b, Å	c, Å	β, °	D, nma
As synthesized	11.350(2)	10.022(1)	18.448(3)	126.53(1)	174
Activated	11.283(2)	9.881(1)	18.263(3)	126.77(1)	83
After 1‐MCP	11.323(3)	9.911(2)	18.229(4)	127.16(2)	111

^a Based on the LX parameter of the angular dependency of the peak broadening.

2.2. Ethylene Adsorption in Co‐FA

For the structural investigation of the adsorbed molecules in Co‐FA, in situ DRIFTS spectroscopy was applied. The setup shown in Figure 1a was used to follow the adsorption and desorption of ethylene molecules. Under ethylene atmosphere, an increase in the infrared absorption around 950 and 1450 cm⁻¹ (Figure 1, glue line) is associated with the gas‐phase ethylene adsorption on Co‐FA (Figure 1b, gray line). This signal rapidly disappears within the first few minutes when the sample is flushed with Ar (Figure 1b, from blue to orange). However, a weak band at 957 cm⁻¹ remains after flushing, corresponding to the C–H wagging mode of the adsorbed ethylene. This band has a redshift of 6 cm⁻¹ compared to gas‐phase ethylene, typical for a weak physisorption interaction. A similar trend is observed in the region of C—H stretching vibrations (Figure S8). The weak intensity and minimal frequency shift confirm the absence of the strong ethylene adsorption sites in Co‐FA, in contrast to MOFs with open‐metal sites, such as HKUST‐1, that show pronounced ethylene binding with significant frequency shifts [11, ²⁹].

FIGURE 1.

(a) Experimental setup for C₂H₄ adsorption in Co‐FA and (b) in situ DRIFTS data collected at 30°C for the bare material activated in inert atmosphere (purple), under the flow of C₂H₄ (blue), time‐resolved data (1 scan per minute) during C₂H₄ desorption (thin lines, from blue to orange), and after 20 min of flushing with Ar (orange). DRIFTS = diffuse reflectance infrared Fourier‐transform spectroscopy.

2.3. 1‐MCP Adsorption in Co‐FA

In the next step, the experimental setup was modified (Figure 2a) to enable in situ production of 1‐MCP, which is unstable in the gas phase. An additional infrared spectrometer was used to monitor the successful 1‐MCP generation in the outgoing gas stream. Unlike in the case of ethylene, in situ DRIFTS spectra collected during exposure to 1‐MCP show no visible features of the gas‐phase vibrational pattern. Instead, two distinct bands appear at 1023 and 929 cm⁻¹, assigned to C–H wagging and C–H rocking modes, respectively, of the confined 1‐MCP molecule (Figure 2b). These bands exhibit significant shifts by ca. 10 cm⁻¹ compared to gas‐phase 1‐MCP (1031 and 919 cm⁻¹), indicating a stronger host–guest interaction compared to ethylene. Nevertheless, such shifts are still lower than those expected for a covalent bonding, suggesting that a physisorption, possibly combined with a geometric confinement, should take place.

FIGURE 2.

(a) Experimental setup for 1‐MCP adsorption in Co‐FA and (b) in situ DRIFTS data collected at 30°C for the bare material after activation in inert atmosphere (purple), under the flow saturated with 1‐MCP (blue), time‐resolved data (1 scan per minute) during 1‐MCP desorption (thin lines, from blue to orange) in the flow of Ar, and after 34 min of flushing with Ar (orange). DRIFTS = diffuse reflectance infrared Fourier‐transform spectroscopy.

The intensity of these spectroscopic features is remarkably high, confirming strong 1‐MCP uptake within Co‐FA pores. Upon flushing with Ar, a gradual decrease of 1‐MCP‐related bands is observed over about 20 min at 30°C until no further changes are observed. Even after prolonged flushing with Ar, the remaining intensity of confined 1‐MCP bands is still pronounced, which is also visible in C—H stretching region (Figure S8). Subsequently, immediate heating boosts further 1‐MCP release until complete disappearance of its bands at 105°C (Figure S4). These observations suggest the potential practical application of Co‐FA for 1‐MCP adsorption, storage, and release at close to ambient or slightly elevated temperatures.

2.4. Temperature‐Programmed 1‐MCP Release and Practical Applications

Mass spectroscopy was applied to monitor online the release of 1‐MCP from Co‐FA. Here, a preliminary activated material (30 mg) was placed in a closed cell and kept at a controlled temperature on an oil bath (Figure 3a). The cell was attached to the 1‐MCP in situ activation setup, similar to the one used for DRIFTS studies. The setup was constantly flushed with Ar at 10 mL/min. The amount of 1‐MCP source was taken in 1:1 molar ratio with respect to Co‐FA. After initial activation of 1‐MCP, a small increase of m/Z signal of 54, corresponding to the molar mass of 1‐MCP, is observed, which decreases over the first 10–30 min. This is consistent with in situ DRIFTS measurement of 1‐MCP desorption, and this signal can be attributed to 1‐MCP that was not trapped in the pores of Co‐FA. The subsequent temperature‐programmed desorption reveals a distinctive release profile starting at ca. 55°C. The integrated signal of 1‐MCP observed during heating is 9.7 times higher than that observed during 1‐MCP activation, which indicates >90% of 1‐MCP capture efficiency by Co‐FA at room temperature. Such thermal release behavior is crucial for practical applications as it demonstrates strong but reversible 1‐MCP retention with the possibility of its temperature‐induced release.

FIGURE 3.

(a) Experimental setup for 1‐MCP loading and temperature‐controlled release and (b) the signal corresponding to the ion current for m/Z = 54 (1‐MCP) registered by an online mass spectrometer.

Equipped with the spectroscopic insights, the practical application of Co‐FA was further demonstrated. The setup shown in Figure 3a was used to load three different MOFs with 1‐MCP: Co‐FA, HKUST‐1, and CPO‐27‐Cu. The vials with MOFs were placed inside the plastic containers next to the banana fruits, whose ripening was monitored over two weeks. Bananas in the container with Co‐FA maintained the yellow color and solid texture, while in the other groups, the appearance of brown spotting started earlier (already from day 2) and the final state was characterized by significantly darker color and soft composition. After day 12, all groups, including the one with Co‐FA, suffered from mold, as no additional treatment or dehumidifiers were applied. The superior performance of Co‐FA compared to other MOFs can be attributed to its balanced adsorption properties, providing sufficient 1‐MCP storage and its prolonged release at room temperature. Notably, if the vial is preheated to 55°C immediately before placing it inside the container, a similar effect is observed (Figures 4 and S7).

FIGURE 4.

Top: Suggested structures of 1‐MCP molecules adsorbed on the open‐metal sites of HKUST‐1 (left) and CPO‐27‐Cu (right) and inside the pore (middle) of Co‐FA. Bottom: Comparison of banana fruit ripening over 12 days in the control group and in the presence of 1‐MCP preloaded MOFs: Co‐FA, HKUST‐1, and CPO‐27‐Cu.

3. Conclusion

We have presented the first spectroscopic insights on ethylene and 1‐MCP adsorption and release in Co‐FA, providing a structure–function relationships by linking molecular‐level spectroscopic observations to macroscopic performance. In particular:

1. Weak ethylene adsorption in Co‐FA was evidenced by in situ DRIFTS spectroscopy.

2. A setup for in situ generation of 1‐MCP coupled with in situ spectroscopic sample environments was developed

3. A strong retention of in situ generated 1‐MCP inside Co‐FA pores was confirmed by the significant shifts of the C–H rocking and C–H wagging vibrational modes.

4. The possibility of temperature‐controlled 1‐MCP release by heating above 55°C was shown.

5. Practical validation of Co‐FA was demonstrated during fruit preservation tests over 14 days.

These findings provide useful insights into the functionality of Co‐FA for controlled‐release applications, particularly in food preservation technologies.

4. Experimental Section and Discussion

4.1. Materials

The following materials have been purchased and used as received without further purification. Cobalt (II) nitrate hexahydrate, copper(II) nitrate trihydrate, and 2,5‐dihydroxyterphthalic acid (DHTP) were obtained from Sigma‐Aldrich. N,N‐dimethylformamide (DMF), formic acid, isopropanol, and methanol were purchased from ChemMed. HKUST‐1 (Basolite C300) was purchased from Sigma‐Aldrich. A commercial source of 1‐MCP and its activator were provided by Fresh‐Forma. A 10% C₂H₄/Ar gas mixture (5.0, purity > 99,999%) was supplied by the Linde Group. High‐purity Ar (4.8, purity > 99,998%) was supplied by LLC NII KM. Banana fruits (Musa acuminata, origin: Ecuador) were provided by X5 Retail Group.

4.2. Synthesis of Co‐FA

Co‐FA MOF, Co₃(HCOO)₆, was synthesized by a solvothermal method. Cobalt (II) nitrate hexahydrate (1.309 g) was dissolved in 10 mL of DMF under ultrasonication. To the resulting solution, 1 mL of formic acid was added, and the mixture was poured into a 20 mL autoclave. The autoclave was heated and kept at 100°C for 24 h and then cooled to room temperature. The obtained precipitate was separated by centrifugation, washed three times with DMF and three times with methanol, and air‐dried. The activation procedure was carried out at 120°C to remove residual solvent from the pores.

4.3. Synthesis of CPO‐27‐Cu

CPO‐27‐Cu (Cu‐MOF‐74), used in fruit preservation tests, was synthesized by a solvothermal method. A mixture of copper (II) nitrate trihydrate (0.814 g) and DHTP (0.304 g) was dissolved in DMF (32.84 mL) under stirring. Isopropanol (1.64 mL) was then added, and the solution was transferred to a Teflon‐lined autoclave, heated at 105°C for 7 h, and cooled to room temperature. The resulting precipitate was collected by centrifugation and washed three times with DMF and three times with methanol. Obtained powder was activated at 160°C under vacuum overnight.

4.4. CoFA Characterization

PXRD measurements were performed on a Bruker D2 PHASER X‐ray diffractometer using Cu K _α radiation (λ = 1.5417 Å). The diffraction patterns were collected in the 2θ range from 5° to 60° with a step of 0.01° at room temperature. Data analysis was carried out using the Jana2006 software to extract the unit cell and peak shape parameters.

TGA was performed on TA 449 F5 Jupiter thermogravimetric analyzer with a low‐drift microbalance. The washed and overnight dried powder (4 mg) in corundum crucibles was heated under nitrogen flow from room temperature to 600°C with a rate of 10°C/min.

FTIR spectra of the as‐synthesized material were collected in attenuated total reflection (ATR) mode on a Bruker Vertex 70 spectrometer with MCT detector (HgCdTe) and a Bruker Platinum ATR accessory. IR spectra were recorded in the range between 5000and 500 cm⁻¹ with a resolution of 1 cm⁻¹, averaging 128 scans. Air was used as the reference.

4.5. In Situ Diffuse Reflectance Infrared Fourier‐Transform Spectroscopy (DRIFTS)

In situ DRIFTS experiments with ethylene and 1‐MCP gases for as‐synthesized MOF Co‐FA were carried out on the Bruker VERTEX 70 spectrometer in diffuse reflectance mode using MCT (HgCdTe) detector. Spectra were collected every 1 min with 1 cm⁻¹ resolution over the range between 5000 and 400 cm⁻¹ and were automatically converted to absorbance units using the Kubelka–Munk function. A spectrum of potassium bromide powder was used as the reference. The homogeneous powder sample was placed into a low‐temperature reaction chamber (Harrick Scientific) equipped with precise temperature control. A gas system allowed switching between the chamber and the bypass line using a three‐way valve and gas flows were controlled using mass flow controllers (EL‐FLOW, Bronkhorst High‐Tech B.V.). The sample was activated for 1 h at 120°C and then cooled to room temperature under pure Ar flow (50 mL/min). In addition to the experimental schemes, the photos of the experimental setups are shown in the Supporting Information.

In the ethylene adsorption experiment, a mixture of 10% C₂H₄/Ar was introduced to the sample at a constant flow rate of 25 mL/min until the peaks corresponding to the gas phase of ethylene were saturated (within ca. 12 min). Afterwards, the line was purged with pure Ar (50 mL/min), and the spectra were recorded continuously to follow the ethylene desorption and detect the remaining signal of ethylene adsorbed in Co‐FA.

For adsorption of 1‐MCP, the experimental setup was modified (Figure 2a) and an additional infrared spectrometer (FSM1202) equipped with a Harrick HTC‐3‐XXX high‐temperature reaction chamber was added after the sample to determine the presence of 1‐MCP in the gas phase. Gas‐phase spectra were collected every 80  s with 4 cm⁻¹ resolution over the range between 4000 and 450 cm⁻¹. Since 1‐MCP is a highly unstable gas, it was produced in situ by introducing into the gas system a glass vessel containing the commercial source of 1‐MCP (100 mg), a glass syringe with the activator dissolved in 10 mL of deionized water. Additionally, a subsequent glass vessel filled with NaOH crystals was added to remove CO₂, which is a by‐product from the commercial 1‐MCP source. Continuous stirring of the 1‐MCP source in a specific activator ensured a complete conversion. A mixture of 1‐MCP and pure Ar (10 mL/min) was sent to the activated sample until disappearance of the peak corresponding to the gas‐phase 1‐MCP at 1031 cm⁻¹, as detected by the FSM1202 spectrometer. Thereafter, the bypass line and chamber were purged with pure Ar, with the flow rate gradually increased from 10 to 50 mL/min with ongoing spectra collection.

4.6. Mass Spectroscopy (MS)

Online mass‐spectrometer experiments with 1‐MCP adsorption and desorption from obtained Co‐FA sample were carried out using Tekhmas MS7‐200 mass spectrometer using a custom‐designed experimental setup (Figure 3a). 1‐MCP was produced in situ by dissolving the commercial source of 1‐MCP (30 mg) in solution of the activator and deionized water (3 mL in total). A signal of formation and presence of 1‐MCP gas was monitored as m/Z = 54 ion current value and recorded continuously during experiments. For the adsorption process, activated sample was flushed with the produced 1‐MCP in the inert flow of Ar (10 mL/min). After reaching a plateau in the signal of 1‐MCP corresponded to zero value, the desorption process of 1‐MCP was initiated by heating from RT to 120°C. The procedure was repeated after reactivation process of the MOF sample in the flow of pure Ar.

4.7. Fruit Preservation Experiment

Using the setup described in the previous subsection, Co‐FA, HKUST‐1 and CPO‐27‐Cu MOFs were loaded with 1‐MCP and stored in a closed vials before further use. Banana fruits were put in four separate plastic containers closed with a plastic lid with holes made for ventilation. Each of the above‐mentioned MOFs was placed inside the corresponding containers next to the fruits and the vial was opened. One container was kept without any MOF as a control group. The fruits were kept under constant temperature, humidity, and light conditions over two weeks. The state of all groups was monitored continuously via time‐lapse mode using a digital camera.

Supporting Information

Additional supporting information can be found online in the Supporting Information section. The data that support the findings of this study are available in the Supporting Information of this article. Data used for the figures are openly available in Zenodo repository: https://doi.org/10.5281/zenodo.17161358. Supporting Fig. S1: PXRD data for as‐synthesized (black) and activated (gray) Co‐FA, and after 1‐MCP sorption and desorption (light gray). Theoretical patterns are shown by dashed red lines. Supporting Fig. S2: TGA data for as‐synthesized Co‐FA. Supporting Fig. S3: FTIR data measured in ATR mode for activated Co‐FA. Supporting Fig. S4: In situ DRIFTS data collected for Co‐FA saturated with 1‐MCP (dashed blue line) at 30°C, flushed with Ar at 30°C until no further changes occur in the spectra (solid orange line), and time‐resolved data (20 scans per minute) during subsequent heating from 30°C to 105°C (from orange to red). Supporting Fig. S5: Photo of the in situ DRIFTS setup. Supporting Fig. S6: Photo of the in situ setup for temperature‐programmed mass spectroscopy. Supporting Fig. S7: Validation of 1‐MCP release action of banana ripening process. Left—initial state, right—after 17 days. Supporting Fig. S8: The region of C—H stretching vibrations for in situ DRIFTS spectra. Left: Data collected at 30°C for the bare material activated in inert atmosphere (purple), under the flow of C₂H₄ (blue), time‐resolved data (1 scan per minute) during C₂H₄ desorption (thin lines, from blue to orange), and after 20 min of flushing with Ar (orange). Right: Data collected at 30°C for the bare material after activation in inert atmosphere (purple), under the flow saturated with 1‐MCP (blue), time‐resolved data (1 scan per minute) during 1‐MCP desorption (thin lines, from blue to orange) in the flow of Ar, and after 34 min of flushing with Ar (orange).

Author Contributions

Anna Yu. Pnevskaya: investigation (lead), data curation (lead), writing – original draft (lead). Andrei A. Tereshchenko: investigation (supporting). Aram Bugaev: supervision (lead), conceptualization (lead), writing – review and editing (lead).

Funding

This study was supported by the Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (TMSGI2_226131).

Conflicts of Interest

The authors declares no conflicts of interest.

Supporting information

Supplementary Material

Biographies

Anna Yu. Pnevskaya earned her PhD in Nanotechnology and Nanomaterials at the Southern Federal University in 2025. Her main research activities related with in situ characterization of MOFs including their interaction with particular guest modeling such as ethylene and 1‐methylcyclopropene (1‐MCP), relevant for further implementation of MOFs as agents for prolongated food storage technologies. She combined both experimental and theoretical approaches in her studies of MOF‐based systems in order to understand the mechanism of interaction between active sites or porous frameworks and gas molecules on the atomic level.

Andrei A. Tereshchenko got his PhD in Nanoscale structure of materials in 2023 with scientific background in in situ and operando spectroscopic techniques, including synchrotron radiation and vibrational spectroscopy to study the structural dynamics and gas adsorption behavior of metal–organic frameworks (MOFs).

Aram Bugaev is a head of SynFlow group at the Paul Scherrer Institut, which integrates expertise across chemistry, physics, engineering, and data science to advance research at large‐scale facilities. His research focuses on catalysis, materials science, and advanced characterization techniques, particularly leveraging synchrotron radiation and operando spectroscopic methods to study dynamic processes in functional materials. He has made significant contributions to understanding metal–organic frameworks (MOFs) and their interactions with gas molecules, as well as optimizing catalytic systems through computational modeling and experimental validation.

Data Availability Statement

The data that support the findings of this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.17161358, reference number 17161359.