Figure optionsDownload full-size imageDownload high-quality image (94 K)Download as PowerPoint slideprs.rt(\”abs_end\”);KeywordsnZVI; Milled nZVI flakes; Anaerobic corrosion; Hydrogen evolution; Groundwater remediation; Column study1. IntroductionGroundwater remediation with nanoscale zero-valent iron (nZVI) is a promising alternative to ZVI permeable reactive barriers (PRB) or conventional pump-and-treat techniques. The nZVI water slurry can be injected into the contaminated subsurface to avoid excavation of aquifer material or pumping of contaminated groundwater. In-situ remediation with nZVI is especially advantageous for poorly accessible contaminated sites (e.g. contamination underneath a building). Besides the transport behavior, the particle reactivity after its injection into the subsurface is a key factor for the successful application of nZVI.The high reactivity of nZVI is linked to its large surface area (Nurmi et?al., 2005 and Tratnyek and Johnson, 2006). Depending on the manufacturing process, suspensions of nZVI often contain particles with a core–shell structure. The (protecting) particle shells consist of a thin iron oxide layer (Fe3O4, FeO) of a few nanometers (Martin et?al., 2008 and Sarathy et?al., 2008) which encloses the Fe0 core (core–shell structure). The aging of core–shell nanoparticles is described by the shrinking of the metallic core while no changes in the particle size and surface area are evident (Liu and Lowry, 2006 and Crane and Scott, 2012).equation(1)Fe0+2H2O→Fe2++2OH−+H2Fe0+2H2O→Fe2++2OH−+H2Under TW-37 conditions zero-valent iron reacts with water to form ferrous iron (Fe2+), hydroxide ions and hydrogen. The reaction consists of two reaction steps with different reaction rates (Wang and Farrell, 2003): 1) Formation of H0-atoms adsorbed on the iron surface and 2) reaction to H2. The formation of adsorbed H0-atoms is linked to the availability of electrons from the core and is reported to be the rate controlling step in the corrosion of nZVI (Liu and Lowry, 2006). According to another kinetic model, the diffusion of reactants and products through the particle shell limits the reaction rate (Reardon et al., 2008).Under anaerobic conditions magnetite (Fe3O4) is the final corrosion product of the ZVI oxidation in de-ionised water (Ruhl et al., 2012), with ferrous hydroxide (Fe(OH)2) as its precursor (Reardon et?al., 2008 and Ruhl et?al., 2012). In complex groundwater matrices, the formation of secondary minerals depends on the concentration of inorganic groundwater solutes and the local pH value.Corrosion experiments with granular iron have indicated that the formation of mineral precipitates affect the long-term reactivity of PRB (Kober et?al., 2002, Klausen et?al., 2003, Kohn et?al., 2005, Parbs et?al., 2007 and Weber et?al., 2013). In particular, mainly iron hydroxide carbonate (chukanovite, Fe2CO3(OH)2) is formed in the presence of carbonate on the iron surface (Ruhl et al., 2011) and influences significantly the reactivity (Jeen et?al., 2006, Lee and Wilkin, 2010 and Ruhl et?al., 2012). High concentrations of inorganic carbon increase not only the dehalogenation rates of chlorinated groundwater contaminants but also the inhibition of reactive sites in the long-term (Parbs et al., 2007). Based on results of field-scale PRBs, Lee and Wilkin (2010) calculated the phase stability of siderite (FeCO3), chukanovite and ferrous hydroxide in a Fe(II)–CO2–H2O-system. The thermodynamic calculations indicated that the formation of siderite and ferrous hydroxide in anaerobic Fe0-groundwater systems is preferred either under carbonate rich (siderite) or carbonate limited (ferrous hydroxide) groundwater conditions (Lee and Wilkin, 2010).Liu et al. (2007) investigated the influence of various groundwater anions (except nitrate) on the reactivity of nZVI in well-mixed batch reactors. The presence of different groundwater solutes decelerated the dehalogenation rate of trichloroethylene (TCE), but no significant differences in the H2 evolution were found in similar batch experiments without contaminants. Solid phase characterization of nZVI particle surfaces which were oxidized in the absence and presences of common groundwater anions (except nitrate) without contaminants by Reinsch et al. (2010) supports these results. Investigation of H2 generation in batch experiments with nZVI and quartz sand indicated that the aquifer material probably decelerates the anaerobic corrosion reaction (Reardon et al., 2008).Direct remediation of a contaminant source requires the injection of high nZVI amounts into the subsurface. The influence of various nZVI loads (g iron per kg sand) on the long-term aging under realistic flow conditions in the long-term is not yet fully understood. In the present study, column experiments with different types of nZVI, including novel milled nZVI flakes, were carried out to investigate the influences of a) varying loads of iron per kg sand inside the column and of b) hydrogen carbonate and calcium on the rate of anaerobic corrosion. Köber et al. (2014) reported recently the development of novel milled nZVI flakes for groundwater remediation. In the present study H2 generation of these milled nZVI flakes was intensively investigated in column experiments.2. Material and methods2.1. ZVI nano particlesCorrosion experiments were performed with two types of nZVI: 1) Commercial nZVI particles (Nanofer25, Nanofer25S) supplied by NanoIron s.r.o., Czech republic and 2) novel milled nZVI flakes supplied by UVR-FIA GmbH, Freiberg, Germany, which were produced only for research purposes.The particles of Nanofer25 and Nanofer25S have a BET surface area of 25 m²/g and 22 m²/g (Zhuang et al., 2012) and an average particle size of 50 nm. The particles consist of a metallic ZVI core and an iron oxide (Fe3O4, FeO) shell (core-shell-structure). Particles of the type Nanofer25S were coated with polyacrylic acid by the manufacturer to support colloidal stability. The particles in suspensions of Nanofer25 and Nanofer25S contained about 80% (by weight) of Fe0.The milled particles are a new type of flake, like nZVI, which are being developed for groundwater remediation (Köber et al. (2014)). The nZVI flakes are manufactured in a two-step process by milling iron powder (ATOMED 57, Rio Tinto) in a ball mill under anaerobic conditions. During the milling process the particles are coated with additives to improve colloidal stability. A description of the manufacturing process and characterizations including reactivity and transport behavior in porous media is given by Köber et al. (2014). The nZVI flakes (lateral size several micrometer, thickness < 200 nm) have a BET surface area of 18 m²/g. The particle suspension contains agglomerates of different sizes in the range of 45 μm and 200 nm related to the milling process. A fresh particle suspension contains approximately 85% of Fe0. Before use the suspension was stored in ethylene glycol to avoid oxidation (Köber et al., 2014).Before each experiment the Fe0 content of the nZVI was determined by complete digestion under acidic conditions and quantification of the hydrogen production (triplicate analyses).2.2. Column experimentsGlass columns (inner diameter: 36 mm, lengths: 70 mm) were filled with 120 g of quartz sand (Dorsolit D8, grain size 0.3–0.8 mm) and loaded with different product types and quantities of nZVI (Table 1). The pore volumes of the sand columns were 24.8 mL, resulting in a porosity of ε = 0.36. Loadings of the sand columns were carried out using the experimental set-up illustrated in Fig. 1a). The nZVI suspension was stirred at 200 rpm in a 0.5 L gas-proof tank that was continuously flushed with nitrogen to achieve anaerobic conditions. Additionally, the pH of the particle suspension in the tank was adjusted to pH 12.5 to minimize oxidation during loading. The nZVI suspension was circulated through the sand column (filter velocity of 2.7 m/h) to deposit particles on the sand grains (filtration). After one hour of filtration time, the column was turned around, fresh nZVI particles were dosed into the tank (same quantity) and the loading was continued from the opposite side for one more hour to distribute the nZVI particles as homogeneously as possible. The quantity of nZVI inside the sand column was calculated by determining and comparing the iron concentrations in the tank before and after the loading procedure. An exemplary plot of the circulation-time dependent deposition of nZVI is shown in Fig. S1.Table 1.
Hand; Grip; Load; Muscle; Performance; Coupling1. IntroductionThe ability to manipulate objects represents a crucial motor function of daily living, while the hand per se is a frequently used model in the studies of biomechanics and motor control phenomena , ,  and . Among a number of different approaches, the force analysis of object manipulation has frequently been applied. This force analysis is typically based on a simple mechanical model of a vertically oriented handheld object (Fig. 1A). The interaction force is decomposed into the load force (LF) that originates from friction and acts in parallel to the contact surface to overcome the object\’s weight and inertia, while grip force (GF) is applied perpendicularly to the object to provide both the friction and enable the control of the object\’s position ,  and . In general, GF needs to be scaled high enough to prevent slippage, but not excessively high to cause either object deformation or muscle fatigue.Fig. 1. (A) Simple model of object manipulation. The circles illustrate the tips of all five digits applying a precision grip to produce the contact force that can be decomposed into the normal (i.e., grip force; GF) and parallel component (load force; LF). (B) Illustration of the fatigue protocol. Participants oscillated the weighted object from the shoulder to hip level, and back.Figure optionsDownload full-size imageDownload high-quality image (170 K)Download as PowerPoint slideNumerous studies, performed on a variety of static and free movement tasks, have consistently revealed a high level of GF–LF coordination through different aspects of GF control , ,  and . Among others, GF is typically scaled to provide a relatively low and stable GF–LF ratio , while continuous coupling of GF with ongoing LF changes has been observed through a high GF–LF correlation and a low GF–LF time lag , ,  and , indicating the involvement of “feed-forward” neural control mechanisms  and . However, various factors can adversely affect GF–LF coordination, such as a frequent switching of LF direction  and , an increase in the task complexity ,  and , or the presence of neural diseases ,  and . Of importance for the present study is the reduced GF–LF coordination, observed either in various tasks or in different patient populations, usually associated with impaired task performance  and . Therefore, it has been concluded that the GF–LF coordination in manipulation tasks could not be only a ‘window’ into the neural mechanisms of muscle control and movement coordination , ,  and , but also a basis for developing standard quantitative tests of hand function in various populations ,  and .Muscle fatigue represents an exercise-induced reduction in the force-generating capacity of muscle, caused by changes within both the CNS and the acting muscles . Fatigue typically decreases maximal voluntary activation of muscle , disrupts excitation–contraction coupling  and impairs both the movement coordination  and  and performance . Regarding the effects on hand function, the muscle fatigue has been shown to reduce the applied GF  and , to increase the fluctuation of GF, and to decrease the coupling between GF and LF in simple lifting tasks . The effects of fatigue on both the coordination of individual fingers producing GF and its SB 525334 have also been studied  and . However, taking into account both the role that GF–LF coordination has played in the studies of hand function and its potential importance for future neurological testing, it is surprising that the effects of fatigue on GF–LF coordination have been largely neglected. Only recently a deteriorated GF–LF coordination in a simple lifting task performed with a pinch grip has been observed .The purpose of the present study is to investigate the effects of muscle fatigue on the GF–LF coordination and performance in a variety of manipulation tasks. Based on the previously documented general effects of muscle fatigue on both task performance and movement coordination , ,  and , as well as on the positive relationship between the GF–LF coordination and performance in manipulation tasks ,  and , it is hypothesized that the applied fatiguing procedure would result in both impaired GF–LF coordination and deteriorated manipulation performance.2. MethodsFifteen healthy right-handed participants were recruited (10 males and 5 females, 20–30 years of age). They were without neurological problems and recent injuries to upper limbs. The experiment was approved by the IRB of the University of Delaware and conducted in accordance with the Declaration of Helsinki.A custom designed device, used in the previous studies of hand function  and , was utilized to record GF and LF produced by the participants (Fig. 1A). The instrumented handle used in this study consisted of two parallel grasping surfaces covered with high friction rubber and connected by a single axis force transducer (WMC-50, Interface Inc., USA). A multi-axis force transducer (Mini40, ATI, USA) was attached beneath each handle either to allow for attachment of the handle either to a fixed external support or to an added brass weight. The single-axis force transducer within the handle records the compression force exerted against the one side of the handle, while the multi-axis force transducer underneath the handle records all three components of the net force applied against the handle (see  and  for the details of GF and LF calculation). The externally fixed handle served for testing static manipulation tasks, while the other was attached to a 200 g mass (total weight 5 N) and could be freely manipulated, which was utilized for both determining maximum precision GF (i.e., the force exerted upon the handle by the tips of the fingers and the thumb) and performing a simple lift task.Prior to testing, participants washed and dried their hands. The handles of the experimental device were washed with isopropyl alcohol to eliminate any residue from prior testing sessions. The participants stood facing a table that contained the device handles and a computer screen that provided visual feedback. Their maximum GF was assessed using the free-manipulation handle. Specifically, subjects were instructed to grasp the free handle with a precision grip (i.e., the tips of all 5 digits involved; see Fig. 1) with their upper arm positioned vertically and elbow flexed 90°. With the use of verbal encouragement from the experimenter and the instruction to “squeeze as hard as possible,” the participants were given 4 s to record their maximum precision grip. This measure was taken both before and after the fatiguing protocol.Three tasks were tested using the same grip as in testing the maximum GF. Ramp-and-hold task required the participants to trace the line shown on the computer monitor by pulling up on the externally fixed handle to produce a tension force  and . The line remained constant at 0 N for 2 s, increased thereafter gradually from 0 N to 10 N at a constant rate for 4 s, and finally remained constant at 10 N for final 4 s. Oscillation task required the participant to produce an oscillating force within the range from 2 N to 10 N of tension for 12 s at a frequency of 1.5 Hz paced by a metronome ,  and . This frequency was chosen because it should be in the middle of the frequency range that allows for the comfortable execution of this type of task  and . Tracing the prescribed LF profile in the ramp-and-hold task inevitably requires ongoing feedback-based corrections; however, the relatively high frequency of the oscillation task does not allow for those adjustments . While the first 2 tasks were static (i.e., both tasks were performed against the externally fixed device), the third one was performed with the device free to move. Specifically, in the simple lift task the participants were instructed to pick up the free-manipulation handle from the table surface, lift it approximately 20 cm, hold the handle in place for a minimum of 4 s, and then replace it back  and .Two experimental sessions were conducted with 1–2 days of rest between them. Each session consisted of a familiarization procedure followed by the testing procedure. Each procedure included 2 trials of each of 3 tasks. The procedures were administered identically, with the exception of the fatiguing protocol preceding the fatigue experimental session. The sequence of both the 3 experimental tasks (i.e., ramp-and-hold, oscillation, and simple lift) and the 2 experimental sessions (non-fatigue and fatigue) were randomized across the participants. Each task was demonstrated prior to the first familiarization procedure.The fatiguing protocol resulted from pilot testing designed to reveal both the properties of the fatigue device and a procedure based on bouts of consecutive lifts, that led to a similar rate of fatigue of the ‘arm muscles’ (i.e., the LF producing muscles) and ‘hand muscles’ (GF producing muscles). The fatigue device consisted of a plastic bottle-like container filled with heavy material (diameter 6 cm, length 21 cm, net mass approximately 2.1 kg). When coated in acetate tape to reduce friction, the pilot subjects reported that the maximum number of the consecutive lifts was limited by a fatigue affecting ‘both their arm and hand muscles’.The participants’ pre-fatigue maximum precision grip was first assessed using the free-manipulation handle (see above) to be compared with the maximum post-fatigue level in order to assess the level of fatigue . The subjects then moved the fatigue device up and down from the shoulder to the hip level at a frequency of 1 Hz paced by a metronome until they were no longer able to maintain either the frequency or amplitude of the fatiguing oscillation (Fig. 1B). Thereafter, they were immediately tested for their maximum precision grip. Verbal encouragement was applied, while a stopwatch was used to track the duration of each fatigue bout. This process was repeated until their maximum GF dropped below the target level of 70% of their initial value (i.e., a 30% reduction in applied maximum GF). Participants needed between 2 and 5 bouts of lifting to complete the fatiguing procedure. The first bout lasted between 34 and 120 s, while consecutive bouts were gradually shorter. Once the target level of fatigue was reached, the participants immediately initiated the testing procedure. However, based on the pilot experiments, a brief ‘fatigue maintenance’ protocol of 15 repetitions was also applied immediately after each of the 3 tasks to maintain the fatigue conditions. Finally, note that even immediately after fatiguing procedures the maximum GF (i.e., 79 N; data averaged across the subjects) were about one order of magnitude higher than the GF forces exerted by fatigued subjects while performing the tested manipulation tasks.Raw signals from the force transducers were sampled at 200 Hz and low-pass filtered at 10 Hz with a fourth order Butterworth filter. The first out of 2 trials of each task was taken for further analysis unless it was considered unsuccessful, such as because of either a late initiation, or dropping the object, which only occurred twice. In the ramp-and-hold task, the first and final 1 s of each phase were discarded, as were the first 3 s and the last 1 s in the oscillation task. This exclusion was used to account for the potential effect of the preceding and anticipated transitions . In the simple lift task, only the middle 2 s of steady holding was analyzed . We evaluated GF–LF coordination through GF scaling and GF coupling , , , ,  and . GF scaling was calculated as a ratio between the averaged values of GF and LF. GF–LF coupling was assessed through the time lags observed from the oscillation task ,  and , as well as from the correlation coefficients observed between GF and LF in both the ramp phase of the ramp-and-hold task and oscillation task , ,  and . Due to the nature of the tested tasks, task performance was evaluated through the subjects’ ability to produce the prescribed LF profiles , ,  and . A series of dependent t-tests were used to analyze the differences between the conditions (non-fatigue vs. fatigue) on GF–LF coordination and movement performance variables (SPSS 19.0 for Windows). The level of significance was set to 0.05.3. ResultsFig. 2 illustrates typical force profiles obtained from a representative participant performing all 3 tasks. Note that the changes in GF closely reflect the changes in LF without any discernible time lag between them.Fig. 2. Profiles of the grip (GF) and load force (LF) recorded in a representative subject performing the ramp-and-hold (A), oscillation (B), and simple lift (C) task.Figure optionsDownload full-size imageDownload high-quality image (265 K)Download as PowerPoint slideFig. 3A illustrates the effect of fatigue on GF scaling through the GF/LF ratio for each of the three experimental tasks. No effect of fatigue was observed in the oscillation task (t = 0.38, p = 0.71). Regarding the ramp-and-hold task, the hold phase revealed a significant drop in GF scaling associated with fatigue (t = 2.17, p = 0.048), but not the ramp phase (t = .54, p = 0.56). In the simple lift task, GF scaling decreased significantly following fatigue (t = 2.63, p = 0.020).Fig. 3. (A) GF scaling assessed through the GF/LF ratio. (B) GF coupling observed through the Fisher-transformed GF–LF correlation coefficient. All data represent means with standard error bars (*p < 0.05).Figure optionsDownload full-size imageDownload high-quality image (161 K)Download as PowerPoint slideGF–LF coupling was assessed through the time lags observed from the oscillation task ,  and , as well as from the correlation coefficients observed between GF and LF in both the ramp phase of the ramp-and-hold task and oscillation task , ,  and . The time lags revealed 0.002 ± 0.005 s and 0.024 ± 0.022 s under the non-fatigue and fatigue conditions, respectively. Although the difference remained below the significant level (t = 1.14, p = 0.27), it should be noted that it mainly originated from two subjects who markedly increased the GF lag from virtually zero to 0.29 and 0.15 s under the fatigue conditions. However, the Fisher-transformed correlation coefficients decreased significantly following fatigue in both the oscillation task (t = 2.42, p = 0.03) and the ramp phase of the ramp-and-hold task (t = 2.63, p = 0.02; Fig. 3B). Relatively steady GF and LF over both the hold phase and the simple lift task did not allow for testing GF–LF coupling .We assessed the root mean square errors (RMSE) when tracing a line (i.e., the ramp-and-hold task) and the absolute errors (AE) of the LF maxima and minima (oscillation task), while the nature of the simple lift did not allow for testing the task performance. RMSE of the ramp phase and holding phase under the non-fatigue condition were 0.56 ± 0.05 and 0.55 ± 0.01 N (mean ± SE), respectively, while under the fatigue conditions the same values were 0.55 ± 0.06 and 0.61 ± 0.05 N. AE of the oscillation task performed under the non-fatigue and fatigue condition were 1.75 ± 0.18 and 1.97 ± 0.16 N (data averaged across the maxima and minima), respectively. None of the above presented effects of fatigue on the performance variables proved to be significant (p > 0.05; paired t-tests).4. DiscussionThe main aim of this study was to explore the effects of muscle fatigue on a variety of manipulation tasks. Regarding the hypothesized outcomes, we found a fatigue associated decrease in GF–LF coordination as seen through both a reduced GF scaling (as assessed through the GF/LF ratio) and partial decoupling of GF and LF (as assessed through GF–LF correlations). Conversely, the task performance (as assessed by the ability to exert the prescribed LF) remained unaffected by the applied fatiguing procedure.It is generally known that a deteriorated GF–LF coordination is typically associated with impaired hand function , , , , ,  and . From that perspective, the observed partial decoupling of GF and LF observed in the present study suggests that muscle fatigue could also be a factor that negatively affects GF–LF coordination. However, we also observed a reduced GF scaling that could be considered as somewhat unexpected. Namely, previous studies have typically demonstrated increased GF scaling associated with either impaired hand function ,  and , or reduced GF–LF coordination ,  and . We believe that the observed phenomenon could be interpreted by the specific effects of the applied intervention since fatigue per se impairs the muscles’ ability to exert high forces. Note that a fatigue associated decrease in GF has been already reported , although the authors neither controlled the level of the associated LF nor assessed force coordination. In addition, note also that the tasks tested in our study required GF magnitudes (i.e., typically below 10 N; see Fig. 2 for illustration) that were vastly lower than the maximum GF that fatigued muscles were able to exert. Therefore, future studies should explore whether the observed reduced GF scaling originates predominantly from the central (e.g., the CNS reduces muscle excitation and allows for a higher risk of dropping the object to slow down further fatiguing) or peripheral level (e.g., the fatigued muscles exert lower GF although the ‘central command’ remains unchanged ). Nevertheless, the observed effects of fatigue on GF–LF coordination could explain why the hand-held objects are more likely to drop when manipulated by fatigued individuals. Namely, both the decoupled GF and LG and an overall drop in GF increase the chance of having GF/LF ratio temporary below the minimally needed to prevent the slip .It is well known that muscle fatigue has adverse affects on both movement coordination and task performance , ,  and , and also negatively affect a simple lift of a hand-held object . Impairment of hand function is also typically associated with impaired ability to control LF in various manipulation tasks ,  and . Therefore, the lack of the effect of fatigue on the movement performance observed in the present study could be considered unexpected, particularly when taking into account that the fatiguing protocol was designed to induce fatigue in the GF and LF producing muscles at a similar rate. One could only speculate that the observed phenomenon could originate from either a low level of the imposed fatigue or a low level of forces required by the tested tasks. Therefore, future studies could involve both more aggressive fatiguing interventions and a larger variety of manipulation tasks regarding the magnitude of the required GF and LF.Regarding the applied methodology, we selected a fatigue procedure that closely resembles a number of repetitive daily tasks that result in muscle fatigue. The strength of our study could be the diversity of the selected static and free movement tasks that required feedback (e.g., ramp and hold) and feed-forward (e.g., oscillation) control of LF and a rather natural and a spontaneous task (e.g., simple lift) , ,  and . However, future studies should extend this line of research to distinguish among the effects of different fatiguing procedures and to explore the effects of fatigue at both the central (i.e., neural) and peripheral (i.e., muscular) level.To conclude, we found that while the ability to exert the prescribed LF pattern remained unaffected, the applied fatigue protocol resulted in both a deteriorated coupling of GF with LF and reduced GF scaling. Both effects should increase the likelihood of dropping hand-held objects over the course of manipulation performed with fatigued muscles. Taking into account the importance of hand function in everyday life and motor control research, as well as the importance of GF–LF coordination for success of manipulative actions, this line of research should be extended to different tasks and to a variety of fatiguing interventions.AcknowledgmentThe study was supported in part by a grant from the Serbian Ministry of Education, Science and Technological Development (#175037).
The types of the as-synthesized zeolites with different ratios of OH−/SiO2.OH−/SiO2Zeolite0.5 → 1.0Transformation from L to W1.0 → 1.5Crystallinity increasing of zeolite W1.5 → 2.0Transformation from W to NatroliteFull-size tableTable optionsView in workspaceDownload as CSV3.1.2. Effect of crystallization temperatureThe influence of crystallization temperature (90 °C, 120 °C, 150 °C, 180 °C and 210 °C) was investigated on the crystallinity of zeolite W with keeping the molar compositions of initial materials constant at OH−/SiO2 = 1.5, H2O/SiO2 = 14, and crystallization time is 24 h. The XRD patterns and SEM images of the materials synthesized at different crystallization temperatures are shown in Fig. 3 and Fig. 4, respectively. From the XRD patterns, the intensities of the characteristic peaks of zeolite W at 2θ = 12.5 and 27.4° were relatively low when the crystallization temperature was 90 °C. The morphology of zeolite W appeared a thin pie shape with the diameter of 2 μm ( Fig. 4a). When the crystallization temperature increased from 120 °C to 150 °C, the crystallinities and morphologies of zeolite W became more regular. The zeolite W synthesized at 150 °C possessed the highest crystallinity. However, further increase of the crystallization temperature to 180 °C, the morphology became irregular ( Fig. 4d). An impurity peak appeared at 2θ = 29.5° in the XRD pattern at 210 °C, indicating the CP-868596 from zeolite W to kalsilite proceeded. Thus, the optimal crystallization temperature for synthesis of zeolite W is 150 °C.Fig. 3. XRD patterns of the materials synthesized at different crystalline temperatures.Figure optionsDownload full-size imageDownload high-quality image (189 K)Download as PowerPoint slideFig. 4. SEM images of the materials synthesized at different crystalline temperatures. a = 90 °C; b = 120 °C; c = 150 °C; d = 180 °C; e = 210 °C.Figure optionsDownload full-size imageDownload high-quality image (706 K)Download as PowerPoint slide3.1.3. Effect of crystallization timeFig. 5 and Fig. 6 show the XRD patterns and SEM images of zeolite W synthesized with different crystallization time (2–84 h), respectively, and other synthesis conditions were kept constant as follows: OH−/SiO2 = 1.5, H2O/SiO2 = 14 and T = 150 °C. As shown in Fig. 5, it can be found that there was no crystal formed as the crystallization time before 2 h, and it appeared as flocculent structures ( Fig. 6a). Starting from the crystallization time of 4 h, the characteristic peaks of zeolite W at 2θ = 12.5° and 27.4° appeared but with low intensities. From 6 h to 24 h, the intensities of zeolite W enhanced gradually. After 24 h, the intensities of zeolite W peaks are higher, and the relative crystallinities of the samples with the crystallization time over 24 h were greater than 95% compared with that of W0 ( Fig. 5). Combining the results from XRD and SEM ( Fig. 6b–f), the optimal crystallization time is 24 h to synthesize zeolite W.Fig. 5. XRD patterns of zeolite W synthesized for different crystalline time.Figure optionsDownload full-size imageDownload high-quality image (425 K)Download as PowerPoint slideFig. 6. SEM images of zeolite W synthesized for different crystalline time. a = 2 h; b = 12 h; c = 24 h; d = 48 h; e = 72 h; f = 84 h.Figure optionsDownload full-size imageDownload high-quality image (913 K)Download as PowerPoint slide3.1.4. Effect of H2O/SiO2Fig. 7 and Fig. 8 show the XRD patterns and SEM images of zeolite W crystallized with different H2O/SiO2 ratios (10–18), respectively, and the other synthesis conditions were kept as follows: OH−/SiO2 = 1.5, T = 150 °C, and t = 24 h. It can be seen from the XRD patterns in Fig. 7 that zeolite W could be synthesized when the ratio of H2O/SiO2 was great than 14, and the characteristic peaks of zeolite W enhanced gradually with the increasing of H2O/SiO2 ratios. The crystal obtained at the H2O/SiO2 ratio of 18 exhibited the best crystallinity ( Fig. 7) and the typical morphology of zeolite W ( Fig. 8).Fig. 7. XRD patterns of the materials obtained at different ratios of H2O/SiO2.Figure optionsDownload full-size imageDownload high-quality image (235 K)Download as PowerPoint slideFig. 8. SEM images of the materials obtained at the different ratios of H2O/SiO2. a = 10; b = 12; c = 14; d = 16; e = 18; f = 20.Figure optionsDownload full-size imageDownload high-quality image (1065 K)Download as PowerPoint slideIn summary, the above researches on the synthesis conditions indicated the phase region to prepare zeolite W was relatively narrow, and the optimal synthesis conditions for zeolite W with high crystallinity >95% is: H2O/SiO2 = 18, OH−/SiO2 = 1.5, T = 150 °C, and t = 24 h.3.2. The characterization and catalytic performances of CoMo/γ-Al2O3-L/W catalysts3.2.1. XRD analyses of the synthesized CoMo catalystsThe XRD patterns of the series catalysts CoMo/γ-Al2O3-L/W are shown in Fig. 9. The XRD pattern of Cat.0 showed strong diffraction peaks located at 2θ of 47 and 73° corresponding to γ-Al2O3 framework topology. The diffraction peak at 5.5° was assigned to the LTL framework topology, and its intensity decreased with the increase of zeolite W addition in the catalysts. From Fig. S7, the XRD patterns of three typical catalysts were compared with each other to evaluate the effects of different proportions of zeolite W and L on their catalytic performances in HDS. It can be seen that the XRD pattern of Cat.3 showed the characteristic peaks of zeolite W (12.5° and 27.4°) and L (19.4° and 22.7°) concurrently, which fully demonstrated that the introduction of zeolite W and L simultaneously did not change the crystal structures of zeolites. Additionally, no obvious peaks ascribed to MoO3 crystals were observed in the XRD patterns of all the catalysts, indicating that Mo species were well dispersed on the surface of the supports or existed as small crystallites beyond the XRD detection limit of 4 nm.Fig. 9. XRD patterns of different catalysts CoMo/γ-Al2O3-L/W.Figure optionsDownload full-size imageDownload high-quality image (244 K)Download as PowerPoint slide3.2.2. Py-FTIRThe acid property of the catalyst is a crucial factor for its hydro-upgrading reaction. The Py-FTIR spectra and the quantitative calculation results of the series catalysts are shown in Fig. 10 and Table 4. The bands located at 1540 and 1450 cm−1 can be assigned to pyridine adsorbed on Brönsted (B) and Lewis (L) acid sites, respectively, and the band at 1490 cm−1 can be ascribed to pyridine co-adsorbed on both B and L acid sites . Cat.0 exhibited the highest amount of L acid sites among all catalysts, owing to the nature of thiophene having L alkalinity, the existence of L acid sites could facilitate the adsorption and conversion of thiophene molecule . However, Cat.1, Cat.2 and Cat.3 not only have L acid sites, but also possess a few amounts of weak B acid sites. The existence of B acid sites on the catalyst surface makes contribution to the cleavage of C-S bond. Comparing with other catalysts, Cat.3 has moderate acid strength and appropriate acid distribution which are expected to promote the synergistic effect between B and L acids, thus to improve the catalytic performances of HDS, hydroisomerization and aromatization.Fig. 10. FTIR spectra of pyridine adsorbed on different catalysts CoMo/γ-Al2O3-L/W after degassing at 200 °C and 350 °C.Figure optionsDownload full-size imageDownload high-quality image (345 K)Download as PowerPoint slideTable 4.