skip to main content


Title: pH-Controlled forms of 1-amino-1-hydrazino-2,2-dinitroethylene (HFOX): selective reactivity of amine and hydrazinyl groups with aldehydes or ketones
1-Amino-1-hydrazino-2,2-dinitroethylene (HFOX) is a potential reactive intermediate for a new class of energetic materials. Now we describe its condensation with various carbonyl compounds in the presence of acidic and basic catalysts. Condensation of HFOX with α-diones and β-diones gives products of much interest. α-Diones undergo cyclization in the presence of base to form six-membered ring products, while β-diones cyclize to five-membered ring products in the presence of acid. One of the exciting reactions is the formation of ammonium (5,6-dimethyl-1,2,4-triazin-3-yl)dinitromethanide salt, 5c, which was isolated by using aqueous ammonia as a nucleophilic base. All new compounds were fully characterized by advanced spectroscopic techniques. The structures of 5, 5c, 5e, 9, 11, and 12a are supported by single crystal X-ray diffraction analysis. Most of the new six membered ring compounds have good thermostabilities (>200 °C), while the fluorinated five membered ring compound, 12b, has the highest density of 2.04 g cm −3 at 25 °C. Heats of formation and detonation properties were calculated by using Gaussian 03 and EXPLO5 software programs. Nearly all new compounds have very good detonation properties, especially, triazine salts, 5e ( D v = 7513 m s −1 ; P = 24.45 G P a), and 5f ( D v = 7948 m s −1 ; P = 26.27 G P a) as well as azide derivative 11 ( D v = 8166 m s −1 ; P = 25.48 G P a), which are superior to TNT ( D v = 6824 m s −1 ; P = 19.40 G P a). These findings provide a new perspective for the synthesis of novel high performing energetic materials.  more » « less
Award ID(s):
1919565
NSF-PAR ID:
10346972
Author(s) / Creator(s):
; ;
Date Published:
Journal Name:
Materials Advances
Volume:
3
Issue:
10
ISSN:
2633-5409
Page Range / eLocation ID:
4289 to 4294
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Nitrogen-rich heterocycles are essential for designing novel energetic green materials with the combination of high explosive performance and acceptable mechanical sensitivities. In this work, two sets of high nitrogen-azoles, derived from tetrazoles and triazole assemblies with N -trinitromethane, 5,5′-(2-(trinitromethyl)-2 H -1,2,3-triazole-4,5-diyl)bis(1 H -tetrazole) (TBTN) and N -methylene tetrazole, 5,5′-(2-((1 H -tetrazol-5-yl)methyl)-2 H -1,2,3-triazole-4,5-diyl)bis(1 H -tetrazole) (TBTT) are described. Their molecular structures were confirmed using multinuclear ( 1 H, 13 C, and 15 N) NMR spectra and single-crystal X-ray diffraction analysis. These molecules are attention attracting results emanating from methodologies utilized to access a unique class of tri-ionic salts in reaction with nitrogen-rich bases. The thermostabilities, mechanical sensitivities, and detonation properties of all new compounds were determined. Surprisingly, the nitro-based tri-cationic salts, 5b (Dv = 9376 m s −1 ) and 5c (Dv = 9418 m s −1 ), have excellent detonation velocities relative to HMX (Dv = 9144 m s −1 ), while those of the nitro-free tri-cationic salts, 8b·H2O (Dv = 8998 m s −1 ) and 8c·0.5H2O (Dv = 9058 m s −1 ), are superior to that of RDX (Dv = 8795 m s −1 ) and approach HMX values. Additionally, nearly all new compounds are insensitive to mechanical stimuli because of the high percentage of hydrogen bond interactions (HBs) between the anions and cations, which are evaluated using two-dimensional (2D) fingerprint and Hirshfeld surface analyses. It is believed that the work presented here is the first example of high-performing and insensitive tri-cationic energetic salts, which may establish a discovery platform for the “green” synthesis of future energetic materials. 
    more » « less
  2. Energetic properties of bistetrazole derivatives are improved by the step-by-step introduction of functionalities which improve heat of formation, density, and oxygen content. The incorporation of unsaturation between bis(1 H -tetrazol-5-yl) and bis(1 H -tetrazol-1-ol) derivatives leads to planarity which enhances the density of the final product. In this manuscript, we have synthesized compounds 1,2-di(1 H -tetrazol-5-yl)ethane (4), ( E )-1,2-di(1 H -tetrazol-5-yl)ethene (5), and ( E )-5,5′-(ethene-1,2-diyl)bis(1 H -tetrazol-1-ol), (6) using readily available starting materials. Their corresponding dihydroxylammonium salts 7, 8 and 9 are obtained by reacting two equivalents of hydroxylamine (50% in water). New compounds are analyzed using IR, EA, DSC and multinuclear NMR spectroscopy ( 1 H, 13 C and 15 N). The solid-state structures of compounds 6, 7, 8 and 9 are confirmed by single-crystal X-ray diffraction. The energetic performances are calculated using the EXPLO5 (v6.06.02) code and the sensitivities towards external stimuli such as friction and impact are determined according to BAM standard. Compound 6 {( E )-5,5′-(ethene-1,2-diyl)bis(1 H -tetrazol-1-ol)} exhibits a surprisingly high density of 1.91 g cm −3 at 100 K (1.86 g cm −3 at 298 K). Its detonation velocity (9017 m s −1 ) is considerably superior to those of RDX (8795 m s −1 ), which suggests it is a competitive high-energy-density material. 
    more » « less
  3. Obeid, Iyad Selesnick (Ed.)
    Electroencephalography (EEG) is a popular clinical monitoring tool used for diagnosing brain-related disorders such as epilepsy [1]. As monitoring EEGs in a critical-care setting is an expensive and tedious task, there is a great interest in developing real-time EEG monitoring tools to improve patient care quality and efficiency [2]. However, clinicians require automatic seizure detection tools that provide decisions with at least 75% sensitivity and less than 1 false alarm (FA) per 24 hours [3]. Some commercial tools recently claim to reach such performance levels, including the Olympic Brainz Monitor [4] and Persyst 14 [5]. In this abstract, we describe our efforts to transform a high-performance offline seizure detection system [3] into a low latency real-time or online seizure detection system. An overview of the system is shown in Figure 1. The main difference between an online versus offline system is that an online system should always be causal and has minimum latency which is often defined by domain experts. The offline system, shown in Figure 2, uses two phases of deep learning models with postprocessing [3]. The channel-based long short term memory (LSTM) model (Phase 1 or P1) processes linear frequency cepstral coefficients (LFCC) [6] features from each EEG channel separately. We use the hypotheses generated by the P1 model and create additional features that carry information about the detected events and their confidence. The P2 model uses these additional features and the LFCC features to learn the temporal and spatial aspects of the EEG signals using a hybrid convolutional neural network (CNN) and LSTM model. Finally, Phase 3 aggregates the results from both P1 and P2 before applying a final postprocessing step. The online system implements Phase 1 by taking advantage of the Linux piping mechanism, multithreading techniques, and multi-core processors. To convert Phase 1 into an online system, we divide the system into five major modules: signal preprocessor, feature extractor, event decoder, postprocessor, and visualizer. The system reads 0.1-second frames from each EEG channel and sends them to the feature extractor and the visualizer. The feature extractor generates LFCC features in real time from the streaming EEG signal. Next, the system computes seizure and background probabilities using a channel-based LSTM model and applies a postprocessor to aggregate the detected events across channels. The system then displays the EEG signal and the decisions simultaneously using a visualization module. The online system uses C++, Python, TensorFlow, and PyQtGraph in its implementation. The online system accepts streamed EEG data sampled at 250 Hz as input. The system begins processing the EEG signal by applying a TCP montage [8]. Depending on the type of the montage, the EEG signal can have either 22 or 20 channels. To enable the online operation, we send 0.1-second (25 samples) length frames from each channel of the streamed EEG signal to the feature extractor and the visualizer. Feature extraction is performed sequentially on each channel. The signal preprocessor writes the sample frames into two streams to facilitate these modules. In the first stream, the feature extractor receives the signals using stdin. In parallel, as a second stream, the visualizer shares a user-defined file with the signal preprocessor. This user-defined file holds raw signal information as a buffer for the visualizer. The signal preprocessor writes into the file while the visualizer reads from it. Reading and writing into the same file poses a challenge. The visualizer can start reading while the signal preprocessor is writing into it. To resolve this issue, we utilize a file locking mechanism in the signal preprocessor and visualizer. Each of the processes temporarily locks the file, performs its operation, releases the lock, and tries to obtain the lock after a waiting period. The file locking mechanism ensures that only one process can access the file by prohibiting other processes from reading or writing while one process is modifying the file [9]. The feature extractor uses circular buffers to save 0.3 seconds or 75 samples from each channel for extracting 0.2-second or 50-sample long center-aligned windows. The module generates 8 absolute LFCC features where the zeroth cepstral coefficient is replaced by a temporal domain energy term. For extracting the rest of the features, three pipelines are used. The differential energy feature is calculated in a 0.9-second absolute feature window with a frame size of 0.1 seconds. The difference between the maximum and minimum temporal energy terms is calculated in this range. Then, the first derivative or the delta features are calculated using another 0.9-second window. Finally, the second derivative or delta-delta features are calculated using a 0.3-second window [6]. The differential energy for the delta-delta features is not included. In total, we extract 26 features from the raw sample windows which add 1.1 seconds of delay to the system. We used the Temple University Hospital Seizure Database (TUSZ) v1.2.1 for developing the online system [10]. The statistics for this dataset are shown in Table 1. A channel-based LSTM model was trained using the features derived from the train set using the online feature extractor module. A window-based normalization technique was applied to those features. In the offline model, we scale features by normalizing using the maximum absolute value of a channel [11] before applying a sliding window approach. Since the online system has access to a limited amount of data, we normalize based on the observed window. The model uses the feature vectors with a frame size of 1 second and a window size of 7 seconds. We evaluated the model using the offline P1 postprocessor to determine the efficacy of the delayed features and the window-based normalization technique. As shown by the results of experiments 1 and 4 in Table 2, these changes give us a comparable performance to the offline model. The online event decoder module utilizes this trained model for computing probabilities for the seizure and background classes. These posteriors are then postprocessed to remove spurious detections. The online postprocessor receives and saves 8 seconds of class posteriors in a buffer for further processing. It applies multiple heuristic filters (e.g., probability threshold) to make an overall decision by combining events across the channels. These filters evaluate the average confidence, the duration of a seizure, and the channels where the seizures were observed. The postprocessor delivers the label and confidence to the visualizer. The visualizer starts to display the signal as soon as it gets access to the signal file, as shown in Figure 1 using the “Signal File” and “Visualizer” blocks. Once the visualizer receives the label and confidence for the latest epoch from the postprocessor, it overlays the decision and color codes that epoch. The visualizer uses red for seizure with the label SEIZ and green for the background class with the label BCKG. Once the streaming finishes, the system saves three files: a signal file in which the sample frames are saved in the order they were streamed, a time segmented event (TSE) file with the overall decisions and confidences, and a hypotheses (HYP) file that saves the label and confidence for each epoch. The user can plot the signal and decisions using the signal and HYP files with only the visualizer by enabling appropriate options. For comparing the performance of different stages of development, we used the test set of TUSZ v1.2.1 database. It contains 1015 EEG records of varying duration. The any-overlap performance [12] of the overall system shown in Figure 2 is 40.29% sensitivity with 5.77 FAs per 24 hours. For comparison, the previous state-of-the-art model developed on this database performed at 30.71% sensitivity with 6.77 FAs per 24 hours [3]. The individual performances of the deep learning phases are as follows: Phase 1’s (P1) performance is 39.46% sensitivity and 11.62 FAs per 24 hours, and Phase 2 detects seizures with 41.16% sensitivity and 11.69 FAs per 24 hours. We trained an LSTM model with the delayed features and the window-based normalization technique for developing the online system. Using the offline decoder and postprocessor, the model performed at 36.23% sensitivity with 9.52 FAs per 24 hours. The trained model was then evaluated with the online modules. The current performance of the overall online system is 45.80% sensitivity with 28.14 FAs per 24 hours. Table 2 summarizes the performances of these systems. The performance of the online system deviates from the offline P1 model because the online postprocessor fails to combine the events as the seizure probability fluctuates during an event. The modules in the online system add a total of 11.1 seconds of delay for processing each second of the data, as shown in Figure 3. In practice, we also count the time for loading the model and starting the visualizer block. When we consider these facts, the system consumes 15 seconds to display the first hypothesis. The system detects seizure onsets with an average latency of 15 seconds. Implementing an automatic seizure detection model in real time is not trivial. We used a variety of techniques such as the file locking mechanism, multithreading, circular buffers, real-time event decoding, and signal-decision plotting to realize the system. A video demonstrating the system is available at: https://www.isip.piconepress.com/projects/nsf_pfi_tt/resources/videos/realtime_eeg_analysis/v2.5.1/video_2.5.1.mp4. The final conference submission will include a more detailed analysis of the online performance of each module. ACKNOWLEDGMENTS Research reported in this publication was most recently supported by the National Science Foundation Partnership for Innovation award number IIP-1827565 and the Pennsylvania Commonwealth Universal Research Enhancement Program (PA CURE). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the official views of any of these organizations. REFERENCES [1] A. Craik, Y. He, and J. L. Contreras-Vidal, “Deep learning for electroencephalogram (EEG) classification tasks: a review,” J. Neural Eng., vol. 16, no. 3, p. 031001, 2019. https://doi.org/10.1088/1741-2552/ab0ab5. [2] A. C. Bridi, T. Q. Louro, and R. C. L. Da Silva, “Clinical Alarms in intensive care: implications of alarm fatigue for the safety of patients,” Rev. Lat. Am. Enfermagem, vol. 22, no. 6, p. 1034, 2014. https://doi.org/10.1590/0104-1169.3488.2513. [3] M. Golmohammadi, V. Shah, I. Obeid, and J. Picone, “Deep Learning Approaches for Automatic Seizure Detection from Scalp Electroencephalograms,” in Signal Processing in Medicine and Biology: Emerging Trends in Research and Applications, 1st ed., I. Obeid, I. Selesnick, and J. Picone, Eds. New York, New York, USA: Springer, 2020, pp. 233–274. https://doi.org/10.1007/978-3-030-36844-9_8. [4] “CFM Olympic Brainz Monitor.” [Online]. Available: https://newborncare.natus.com/products-services/newborn-care-products/newborn-brain-injury/cfm-olympic-brainz-monitor. [Accessed: 17-Jul-2020]. [5] M. L. Scheuer, S. B. Wilson, A. Antony, G. Ghearing, A. Urban, and A. I. Bagic, “Seizure Detection: Interreader Agreement and Detection Algorithm Assessments Using a Large Dataset,” J. Clin. Neurophysiol., 2020. https://doi.org/10.1097/WNP.0000000000000709. [6] A. Harati, M. Golmohammadi, S. Lopez, I. Obeid, and J. Picone, “Improved EEG Event Classification Using Differential Energy,” in Proceedings of the IEEE Signal Processing in Medicine and Biology Symposium, 2015, pp. 1–4. https://doi.org/10.1109/SPMB.2015.7405421. [7] V. Shah, C. Campbell, I. Obeid, and J. Picone, “Improved Spatio-Temporal Modeling in Automated Seizure Detection using Channel-Dependent Posteriors,” Neurocomputing, 2021. [8] W. Tatum, A. Husain, S. Benbadis, and P. Kaplan, Handbook of EEG Interpretation. New York City, New York, USA: Demos Medical Publishing, 2007. [9] D. P. Bovet and C. Marco, Understanding the Linux Kernel, 3rd ed. O’Reilly Media, Inc., 2005. https://www.oreilly.com/library/view/understanding-the-linux/0596005652/. [10] V. Shah et al., “The Temple University Hospital Seizure Detection Corpus,” Front. Neuroinform., vol. 12, pp. 1–6, 2018. https://doi.org/10.3389/fninf.2018.00083. [11] F. Pedregosa et al., “Scikit-learn: Machine Learning in Python,” J. Mach. Learn. Res., vol. 12, pp. 2825–2830, 2011. https://dl.acm.org/doi/10.5555/1953048.2078195. [12] J. Gotman, D. Flanagan, J. Zhang, and B. Rosenblatt, “Automatic seizure detection in the newborn: Methods and initial evaluation,” Electroencephalogr. Clin. Neurophysiol., vol. 103, no. 3, pp. 356–362, 1997. https://doi.org/10.1016/S0013-4694(97)00003-9. 
    more » « less
  4. The gyroscope like dichloride complexes trans -Pt(Cl) 2 (P((CH 2 ) n ) 3 P) ( trans -2; n = c, 14; e, 18; g, 22) and MeLi (2 equiv.) react to yield the parachute like dimethyl complexes cis -Pt(Me) 2 (P((CH 2 ) n ) 3 P) ( cis -4c,e,g, 70–91%). HCl (1 equiv.) and cis -4c react to give cis -Pt(Cl)(Me)(P((CH 2 ) 14 ) 3 P) ( cis -5c, 83%), which upon stirring with silica gel or crystallization affords trans -5c (89%). Similar reactions of HCl and cis -4e,g give cis / trans -5e,g mixtures that upon stirring with silica gel yield trans -5e,g. A parallel sequence with trans -2c/EtLi gives cis -Pt(Et) 2 (P((CH 2 ) 14 ) 3 P) ( cis -6c, 85%) but subsequent reaction with HCl affords trans -Pt(Cl)(Et)(P((CH 2 ) 14 ) 3 P) ( trans -7c, 45%) directly. When previously reported cis -Pt(Ph) 2 (P((CH 2 ) 14 ) 3 P) is treated with HCl (1 equiv.), cis - and trans -Pt(Cl)(Ph)(P((CH 2 ) 14 ) 3 P) are isolated (44%, 29%), with the former converting to the latter at 100 °C. Reactions of trans -5c and LiBr or NaI afford the halide complexes trans -Pt(X)(Me)(P((CH 2 ) 14 ) 3 P) ( trans -9c, 88%; trans -10c, 87%). Thermolyses and DFT calculations that include acyclic model compounds establish trans > cis stabilities for all except the dialkyl complexes, for which energies can be closely spaced. The σ donor strengths of the non-phosphine ligands are assigned key roles in the trends. The crystal structures of cis -4c, trans -5c, trans -7c, and trans -10c are determined and analyzed together with the computed structures. 
    more » « less
  5. null (Ed.)
    Heat of formation and density have a significant influence on the detonation performance of a compound and are greatly influenced by the nitrogen and oxygen content of a material. Here a family of new materials with high oxygen and nitrogen content was synthesized and characterized. Compound 1 has a nitrogen and oxygen content of 85.3%, with a high density (1.93 g cm −3 ) and high detonation properties (detonation velocity v D = 9503 m s −1 ; detonation pressure P = 41 GPa). The ammonium salt 3 has a nitrogen and oxygen content of 84.9%, a density of 1.86 g cm −3 , a detonation velocity of 9317 m s −1 , and acceptable sensitivities (8 J, 120 N) which are similar to those of HMX . The potassium salt ( 5 ) was characterized by single-crystal X-ray diffraction. 
    more » « less