Initial Experiments Research Articles

Exploring the Path from Instructions to Rewards with Large Language Models in Instance-Based Learning

A prominent method to model human learning is through experiential learning, where decisions are influenced by the outcomes observed in previous actions. The decisions-from-experience approach often excludes other forms of learning in humans, such as learning from descriptive information. In humans, descriptive information can enhance learning by providing a denser signal, achieved through understanding the relationship between intermediate decisions and their future outcomes, instead of relying solely on observed outcomes. To account for experiential and descriptive information, we propose the use of large language models (LLMs) to convert descriptive information into dense signals that can be used by computational models that learn from experience. Building on past work in cognitive modeling, we utilize task instructions and prompt an LLM to define and quantify the critical actions an agent must take to succeed in the task. In an initial experiment, we test this approach using an Instance-Based Learning cognitive model of experiential decisions in a gridworld task. We demonstrate how the LLM can be prompted to provide a series of actions and relative values given the task instructions, then show how these values can be used in place of sparse outcome signals to improve the model’s learning of the task significantly.

Divergent [2 + n] Heteroannulation of β-CF3-1,3-enynes with Alkyl Azides via Hydrogen Atom Transfer and Radical Substitution.

A copper-promoted divergent intermolecular [2 + n] heteroannulation of β-CF3-1,3-enynes with alkyl azides via alkyl radical-driven HAT and radical substitution (C-C bond formation) to form four- to ten-membered saturated N-heterocycles is developed. This method enables the aryl-induced or kinetically controlled site selective functionalization of the remote C(sp3)-H bonds at positions 2, 3, 4, 5, 6, 7, or 8 toward the nitrogen atom through triplet nitrene formation, radical addition across the C═C bond, HAT and radical substitution cascades, and features a broad substrate scope, excellent site selectivity, and facile late-stage derivatization of bioactive molecules. Initial deuterium-labeling and control experiments shed light on the reaction mechanism via nitrene formation and HAT.

In Vivo Two-Photon Microscopy Reveals Sensory-Evoked Serotonin (5-HT) Release in Adult Mammalian Neocortex.

The recent development of genetically encoded fluorescent neurotransmitter biosensors has opened the door to recording serotonin (5-hydroxytryptamine, 5-HT) signaling dynamics with high temporal and spatial resolution in vivo. While this represents a significant step forward for serotonin research, the utility of available 5-HT biosensors remains to be fully established under diverse in vivo conditions. Here, we used two-photon microscopy in awake mice to examine the effectiveness of specific 5-HT biosensors for monitoring 5-HT dynamics in somatosensory cortex. Initial experiments found that whisker stimulation evoked a striking change in 5-HT biosensor signal. However, similar changes were observed in controls expressing green fluorescent protein, suggesting a potential hemodynamic artifact. Subsequent use of a second control fluorophore with emission peaks separated from the 5-HT biosensor revealed a reproducible, stimulus-locked increase in 5-HT signal. Our data highlight the promise of 5-HT biosensors for in vivo application, provided measurements are carried out with appropriate optical controls.

Continuous Hydrothermal Carbonization of Olive Pomace and Orange Peels for the Production of Pellets as an Intermediate Energy Carrier

The ever-increasing volumes of food waste generated and the associated environmental issues require the development of new processing methods for these difficult waste streams. One of the technologies that can treat these waste streams directly is hydrothermal carbonization. In this work, olive pomace and orange peels were treated via a mild hydrothermal carbonization process (TORWASH®) in a continuous-flow pilot plant. For olive pomace, a solid yield of 46 wt% and a dry matter content of 58% for the solid press cakes were obtained during continuous operation for 18 days. For orange peels, the values were lower with 31 wt% solid yield and a 42% dry matter content during 28 days of continuous operation. These values corresponded fully with initial laboratory-scale batch experiments, showing the successful transformation from batch to continuous processing. The obtained hydrochar from both feedstocks showed an increase in higher heating value (HHV) and a significant reduction in ash content. Pellets produced from the solids met the requirements for industrial use, demonstrating a large increase in the deformation temperature and a significant reduction in the potassium and chlorine content compared to the original feedstock. These results indicate the excellent potential of these pellets for combustion applications.

Abstract B108: Integrated in vivo dissection of metastatic phenotypes and molecular programs in pancreatic ductal adenocarcinoma

Abstract The ability of pancreatic ductal adenocarcinoma (PDAC) cells to form tissue destructive metastases remains a critical impediment to therapy. Metastasis is a complex multistep process that is poorly understood and the molecular events and cellular states that enable cancer cells to overcome each step of the metastatic cascade remain elusive. To investigate the regulators of PDAC metastasis, we accrued and characterized a panel of 31 cell lines generated from mouse models of PDAC. Initial experiments indicated that these cell lines have an ~40-fold difference in their ability to form lung metastases after transplantation into immunocompetent recipient mice. To more accurately quantify the relationship between metastatic seeding, growth, organ tropism, and immune evasion across all cell lines in parallel, we tagged each cell line with a barcode that provides both cell line-level and subclonal-level resolution. We pooled all 31 barcoded cell lines, transplanted them to seed metastases in different organs, and performed barcode sequencing on DNA from the metastasis-bearing tissues after short (seeding) and long (expansion) time periods. Neither seeding nor metastatic outgrowth correlated with growth in culture. Metastatic seeding did not directly correlate with overall metastatic burden, indicating that these processes depend on distinct biological processes. Transplantation into syngeneic immune-competent and immune-deficient recipient mice quantified the impact of immune responses towards each cell line. Bulk and single cell RNA-sequencing uncovered unexpected intra- and inter-cell line heterogeneity. Future experiments will integrate the data on metastatic phenotypes with molecular information to uncover drivers and inhibitors of each step of the metastatic cascade. Broad-scale molecular profiling coupled with quantitative in vivo analysis of metastatic phenotypes will uncover the relationships between metastatic phenotypes, create a resource for the community, and reveal novel pathways that govern PDAC metastasis, which could have implications for developing novel therapeutic strategies. Citation Format: Saswati Karmakar, Anastasia Lyulina, Xiaochen Xiong, Yuning J. Tang, Steven Lopez, Nicolas William Hughes, Laura Attardi, Dmitri A. Petrov, Monte M. Winslow. Integrated in vivo dissection of metastatic phenotypes and molecular programs in pancreatic ductal adenocarcinoma [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr B108.

Potential use of frass from edible insect Tenebrio molitor for proteases production by solid-state fermentation

Novel, eco-saving and low-cost food sources are a global priority, such as edible insects. This work sets a valorization pathway for frass and Tenebrio molitor adult insects using solid-state fermentation (SSF) to obtain fungal proteases, which are highly used in food, beverages, cosmetics, pharmaceutical and biotechnology industries. Initial screening experiments demonstrated that Aspergillus uvarum MUM 08.01 was the most efficient proteases producer. Small-scale SSF with frass resulted in high proteases activity (65 U g-1), which increased 52% when BSG was mixed with frass (1:1, w/w). The proteases activity attained in this study using only frass or mixed with BSG was superior compared to other activities observed when using agroindustry by-products (such as rapeseed, sunflower, soybean, among others) and other species of Aspergillus genus. The optimal conditions from full factorial design using frass:BSG mixture resulted in 152 U g-1 of proteases activity, which is in full accordance with the value predicted by the model. SSF scale-up with 50 g of frass:BSG increased proteases activity up to 202.2 U g-1. This study presents an innovative utilization of insects’ by-products in the field of enzymes production, in full accordance with the circular economy guidelines.

Groundwater remediation by in-situ membrane ozonation: Removal of aliphatic 1,4-dioxane and monocyclic aromatic hydrocarbons

Groundwater contamination by widespread and persistent organic compounds requires extensive treatment efforts, for example by in-situ chemical oxidation (ISCO). In this study, we investigated ozone mass transfer and removal mechanisms of ozone-resistant monocyclic aromatic and non-aromatic compounds in a novel in-situ treatment method using ozone-permeable membranes as reactive barrier. Initial batch experiments confirmed fast depletion of ozone in presence of sub-stoichiometric benzoic acid (BA), in contrast to the non-aromatic 1,4-dioxane (DIOX), where ozone depleted much slower. Simulated in-situ membrane ozonation treatment of contaminated groundwater led to lower removal of 5 mg L−1 BA (52.7%) compared to DIOX (60.6%). Inhibited removal of BA compared to additional batch experiments could be explained by quick depletion of ozone by reactive intermediates on the membrane surface. Surprisingly, reactive porous media did not lead to substantial changes of in-situ DIOX oxidation, although a stronger impact of the media on DIOX oxidation was hypothesized. Furthermore, experimental ozone mass transfer coefficients were determined (3.94∙10−7 – 3.12∙10−6 m s−1) and compared to modeled values for different membrane types (polydimethylsiloxane and polytetrafluoroethylene). Finally, a mathematical model based on mass transfer data was developed to support upscaling efforts. We concluded that contaminant properties are crucial for the feasibility assessment of in-situ ozone membrane treatment technology.

Behavioral characterization of early nicotine withdrawal in the mouse: a potential model of acute dependence

BackgroundClinical and preclinical research have demonstrated that short-term exposure to nicotine during the initial experimentation stage can lead to early manifestation of withdrawal-like signs, indicating the state of “acute dependence”. As drug withdrawal is a major factor driving the progression toward regular drug intake, characterizing and understanding the features of early nicotine withdrawal may be important for the prevention and treatment of drug addiction. In this study, we corroborate the previous studies by showing that withdrawal-like signs can be precipitated after short-term nicotine exposure in mice, providing a potential animal model of acute dependence on nicotine.ResultsTo model nicotine exposure from light tobacco use during the initial experimentation stage, mice were treated with 0.5 mg/kg (-)-nicotine ditartrate once daily for 3 days. On the following day, the behavioral tests were conducted after implementing spontaneous or mecamylamine-precipitated withdrawal. In the open field test, precipitated nicotine withdrawal reduced locomotor activity and time spent in the center zone. In the elevated plus maze test, the mecamylamine challenge increased the time spent in the closed arm and reduced the number of entries irrespective of nicotine experience. In the examination of the somatic aspect, precipitated nicotine withdrawal enhanced the number of somatic signs. Finally, nicotine withdrawal did not affect cognitive functioning or social behavior in the passive avoidance, spatial object recognition, or social interaction test.ConclusionsCollectively, our data demonstrate that early nicotine withdrawal-like signs could be precipitated by the nicotinic antagonist mecamylamine in mice, and that early withdrawal from nicotine primarily causes physical symptoms.

Adsorption of hydrocarbon pollutants from wastewater using Cu‐ and Zn‐loaded activated carbon derived from waste tires

AbstractPetroleum refinery effluents contain hazardous aromatic and polyaromatic hydrocarbons that causes environmental degradation, that is, water, land, and air pollution; therefore, it should be properly treated before its discharge into environment. Adsorption serves an efficient route to treatment of refinery wastewater; but the high production cost and selectivity of the modern adsorbent used limits their industrial scale application. On the other hand, activated carbon due to desirable adsorption characteristics and low cost seems to hold the promise if properly modified to improve its efficiency. In this study, Cu and Zn impregnated on activated carbon (AC) derived from waste tires was used for adsorption of hydrocarbon pollutants from wastewater. AC was prepared from scrap tires through thermal pyrolysis, followed by chemical activation with KOH. Cu and Zn were loaded on AC through wet impregnation method. Initial batch mode adsorption experiments were carried out using aniline and phenol as model hydrocarbons, which indicated that maximum adsorption of aniline and phenol over Zn/AC and Cu/AC occurs in 10 min contact time, temperature 30°C, pH 6, 100 ppm initial concentration, and adsorbent dose of 0.1 g/10 mL. The removal of hydrocarbon pollutants from real refinery wastewater was also investigated under optimized conditions. Results show that the COD of the refinery wastewater decreased from 963 to 73 mg/L in case of Cu/AC (92% removal) and 78 mg/L in case of Zn/AC (91% removal), respectively. The adsorbent showed many folds reusability, and in first cycle, the removal efficiency of phenol and aniline was observed to be above 95%, whereas in second, third and fourth cycles, the removal percentage was 85%, 80%, and 70%, respectively.

Correction of Thermal Errors in Machine Tools by a Hybrid Model Approach

Thermally induced position errors are one of the main error sources on the workpiece caused by the behavior of the machine tool. In today’s industrial environment, the correction of thermal errors is usually based on simple regression approaches, where the characteristic diagrams for correction are generated experimentally. The performance of these approaches is only valid for the corresponding load regimes, which often results in insufficient correction quality in practical applications. Consequently, there is only a limited benefit or even a deterioration in machine behavior if the correction characteristic is based on an inapplicable load case compared to the initial experiment. Simulation-generated characteristic diagrams using finite element models solve this disadvantage, but do not answer the question about the choice of the right characteristic matching the current load situation, and, in addition, calculate very slowly. Structural model-based correction using reduced models, on the other hand, calculates quickly, but requires a high modeling effort for accurate correction. The approach, presented in this contribution, combines simulation-generated characteristic diagrams and a structural model-based decision algorithm for a new hybrid model in order to select the appropriate characteristic diagram for the present load situation in the control system. This paper presents the simulative characteristic diagram generation by a finite element model validated by experiments in a climate chamber and a validated structural model including the concept for the decision algorithm.

Synthetic pre-training for neural-network interatomic potentials

Machine learning (ML) based interatomic potentials have transformed the field of atomistic materials modelling. However, ML potentials depend critically on the quality and quantity of quantum-mechanical reference data with which they are trained, and therefore developing datasets and training pipelines is becoming an increasingly central challenge. Leveraging the idea of ‘synthetic’ (artificial) data that is common in other areas of ML research, we here show that synthetic atomistic data, themselves obtained at scale with an existing ML potential, constitute a useful pre-training task for neural-network (NN) interatomic potential models. Once pre-trained with a large synthetic dataset, these models can be fine-tuned on a much smaller, quantum-mechanical one, improving numerical accuracy and stability in computational practice. We demonstrate feasibility for a series of equivariant graph-NN potentials for carbon, and we carry out initial experiments to test the limits of the approach.

Quantum Applications of an Atomic Ensemble Inside a Laser Cavity

Many quantum device signals are proportional to the number of the participating atoms that take part in the detection devices. Among these are optical magnetometers, atomic clocks, quantum communications and atom interferometers. One way to enhance the signal-to-noise ratio is to introduce atom entanglement that increases the signal in a super-radiant-like effect. A coherent em field inside a laser cavity is suggested to achieve atoms’ correlation/entanglement. This may also play an important role in the basic quantum arena of many-body physics. An initial novel experiment to test the realization of atoms’ correlation is described here. A Cs optical magnetometer is used as a tool to test the operation of a cell-in-cavity laser and its characteristics. A vapor cell is inserted into an elongated external cavity of the pump laser in Littrow configuration. Higher atom polarization and reduced laser linewidth are obtained leading to better magnetometer sensitivity and signal-to-noise ratio. The Larmor frequency changes of the Free Induction Decay of optically pumped Cs atomic polarization in the ambient earth magnetic field at room temperature is measured. Temporal changes in the magnetic field of less than 10 pT/√Hz are measured. The first-order dependence of the magnetic field on temperature and temperature gradients is eliminated, important in many practical applications. Single and gradiometric magnetometer configurations are presented.

The “Effect of Marangoni Convection on Heat Transfer in Phase Change Materials” experiment: Design and performance of the cuboidal cell

The experiment “Effect of Marangoni Convection on Heat Transfer in Phase Change Materials” aims to investigate the efficacy of thermal Marangoni convection in augmenting the heat transfer rate of passive phase change materials (PCMs) that incorporate a free surface in microgravity; the so-called thermocapillary-enhanced PCMs. Compared to thermal conduction, thermocapillary flows can increase heat transport by a factor of two or more. In addition to advancing scientific understanding, the experiment seeks to evaluate the practical feasibility of using thermocapillary-enhanced PCMs as passive thermal control devices for space missions and assess possible implementation challenges. By analyzing different PCM samples, the experiment will provide crucial insights into the heat and mass transport mechanisms of thermocapillary flows during melting and solidification. The present article primarily describes the design, thermal control and ground results of the cuboidal cell of the experiment. We illustrate here the comprehensive approach towards developing space experiments adhering to rigorous and demanding scientific and technical requirements. Along with the validation of an initial experiment prototype, the evaluation of ground tests is also addressed. The hardware developed demonstrates an adequate performance, while the obtained scientific results are coherent and compliant with the established requirements; the feasibility of the proposed design is thus demonstrated.

Modifying peptide/lipid-associated nucleic acids (PLANAs) for CRISPR/Cas9 ribonucleoprotein delivery

With the first reports on the possibility of genome editing by Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and CRISPR-associated protein (Cas)9 surfacing in 2005, the enthusiasm for protein silencing via nucleic acid delivery experienced a resurgence following a period of diminished enthusiasm due to challenges in delivering small interfering RNAs (siRNA), especially in vivo. However, delivering the components necessary for this approach into the nucleus is challenging, maybe even more than the cytoplasmic delivery of siRNA. We previously reported the birth of peptide/lipid-associated nucleic acids (PLANAs) for siRNA delivery. This project was designed to investigate the efficiency of these nanoparticles for in vitro delivery of CRISPR/Cas9 ribonucleoproteins. Our initial experiments indicated higher toxicity for PLANAs with the more efficient reverse transfection method. Therefore, polyethylene glycol (PEG) was added to the composition for PEGylation of the nanoparticles by partially replacing two of the lipid components with the PEG-conjugated counterparts. The results indicated a more significant reduction in the toxicity of the nanoparticle, less compromise in encapsulation efficiency and more PEGylation of the surface of the nanoparticles using DOPE-PEG2000 at 50 % replacement of the naïve lipid. The cell internalization and transfection efficiency showed a comparable efficiency for the PEGylated and non-PEGylated PLANAs and the commercially available Lipofectamine™ CRISPRMAX™. Next Generation Sequencing of the cloned cells showed a variety of indels in the transfected cell population. Overall, our results indicate the efficiency and safety of PEGylated PLANAs for in vitro transfection with CRISPR/Cas9 ribonucleoproteins. PEGylation has been studied extensively for in vivo delivery, and PEGylated PLANAs will be candidates for future in vivo studies.

Water, Feeding Efficiencies, and Development of Dermestes maculatus DeGeer (1774)

Although dermestids have varying ecological roles, most Dermestes are generally associated with decomposition of dry animal tissues. However, Dermestes maculatus DeGeer (1774) can eat soft, hydrated tissues, so it is unclear if the apparent preference of Dermestes maculatus for dry tissue reflects physiological adaptation or competitive exclusion by blowflies and other invertebrate scavengers that arrive at carrion earlier and may have faster immature development rates. To address this question, we examined development of D. maculatus larvae on hydrated and dry food, and we determined development rates at a constant temperature of 25°C. Specifically, we conducted multiple replicated experiments in incubators. Initial experiments used with a factorial treatment arrangement of food type (wet versus dry) and larval density (1, 5, 10, 15, and 25 per 5 g of food) with 4 replications. Additionally, for dry food we determined growth, consumption, and conversion rates for 3-6 stage larvae, and consumption and conversion rates for adults. Hydrated food significantly reduced survivorship of D. maculatus larvae and adults (ca. 20% survivorship) as compared to dried food with survivorship rates of ca. 80-90%. Adults feeding on hydrated food were observed to void water, likely to overcome problems with osmotic regulation. Given dry food with little water and high conversion rates (much greater than 10%) in larvae, evidence indicated D. maculatus obtains most water metabolically and not through diet. These adaptations indicate that D. maculatus (and likely related species) have evolved physiologically to exploit dry carrion thereby avoiding competition with other carrion-feeding species. However, D. maculatus is not adapted for low humidity environments or to avoid extreme water loss, these adaptations allow use of a unique resource but do not represent adaption to xeric conditions generally. Moreover, the very adaptations that allow D. maculatus to survive on dry food, are potentially lethal when only hydrated food is available.

Single-cell fluid-based force spectroscopy reveals near lipid size nano-topography effects on neural cell adhesion.

Nano-roughness has shown great potential in enhancing high-fidelity electrogenic cell interfaces, owing to its characteristic topography comparable to proteins and lipids, which influences a wide range of cellular mechanical responses. Gaining a comprehensive understanding of how cells respond to nano-roughness at the single-cell level is not only imperative for implanted devices but also essential for tissue regeneration and interaction with complex biomaterial surfaces. In this study, we quantify cell adhesion and biomechanics of single cells to nano-roughened surfaces by measuring neural cell adhesion and biomechanics via fluidic-based single-cell force spectroscopy (SCFS). For this, we introduce nanoscale topographical features on polyimide (PI) surfaces achieving roughness up to 25 nm without chemical modifications. Initial adhesion experiments show cell-specific response to nano-roughness for neuroblastoma cells (SH-SY5Y) compared to human astrocytes (NHA) around 15 and 20 nm surface roughness. In addition, our SCFS measurements revealed a remarkable 2.5-fold increase in adhesion forces (150-164 nN) for SH-SY5Y cells cultured on roughened PI (rPI) surfaces compared to smooth surfaces (60-107 nN). Our data also shows that cells can distinguish changes in nano-roughness as small 2 nm (close to the diameter of a single lipid) and show roughness dependence adhesion while favoring 15 nm. Notably, this enhanced adhesion is accompanied by increased cell elongation upon cell detachment without any significant differences in cell area spreading. The study provides valuable insights into the interplay between nano-topography and cellular responses and offers practical implications for designing biomaterial surfaces with enhanced cellular interactions.

EXPERIMENTAL ANALYSIS OF A SOLAR DESALINATION SYSTEM WITH GRAPHENE NANOPLATELET-EMBEDDED LATENT HEAT THERMAL ENERGY STORAGE UNIT

In the present work, it is aimed to improve the yield of a single-slope solar desalination system (SDS) using graphene nanoplatelet-embedded thermal energy storage unit (TSU). In the initial stage of this work, three SDSs with different slope angles of condensation surface including 40°, 50°, and 60° are considered. According to the result of the initial experiments, the SDS with 40° gave the superior performance. In the second part of the study, three different SDS configurations with 40° slope angle including a conventional SDS, an SDS with only paraffin-containing TSU, and an SDS with graphene nanoplatelet-embedded paraffin-containing TSU have been developed and tested under the same environmental conditions. According to the findings of the second experimental stage, utilizing only paraffin-containing and graphene nanoplatelet-embedded paraffin-containing TSUs in the conventional system with 40° slope angle improved the accumulated yield of the system as 11.85% and 26.96%, respectively. Also, energy and exergy efficiencies of the SDS with 40° slope angle were improved from 16.02% to 19.69% and from 1.36% to 2.13%, respectively, using the TSU with graphene nanoplatelets. The attained findings of this work clearly present the positive influence of employing graphene nanoplatelets in TSU of a single-slope SDS.

Digital Transformation in Complex Systems

Complex systems increasingly include embedded digital technologies that interact with and are constrained by physical components and systems. Although these systems play a central role in our society, they have only been scarcely addressed in contemporary research on digital transformation and the organization of innovation. This article explores the digital transformation in complex products and systems and its consequences for organizational design. A longitudinal study of avionics development since the 1950s uncovers the application of digital technologies, first as a sequence of initial experiments, followed by the use as add-on functionality, then as an integral part of achieving critical functionality in systems, and currently combining add-on and critical functionalities enabling generativity. The findings emphasize the evolution of the intricate relationships between the systems architecture and organizational approaches when digital technology enables and enforces increased complexity, expanded functionality, increased systems integration, and continuous development. These nested dependencies are accentuated by the complexity that has emerged beyond human cognition, where increasingly sophisticated boundary objects based on modeling, simulation, and data play an important role in the organization's ability. Boundary objects relate and decouple the multifacetted dynamic relation between organization and architecture. The results also extend existing perspectives on platform strategies by outlining the importance of generativity in combination with criticality control, rather than market control. Criticality control in combination with generativity has become imperative not least as generative digital technologies have become central in achieving critical properties such as safety. Several avenues for further research are outlined.

Traditional or adaptive design of experiments? A pilot-scale comparison on wood delignification

Traditional design of experiments and response surface methodology are widely used in engineering and process development. Bayesian optimization is an alternative machine learning approach that adaptively selects successive experimental conditions based on a predefined performance measure. Here we compared the two approaches using simulations and empirical experiments on alkaline wood delignification to identify important benefits and drawbacks of Bayesian optimization in the context of design of experiments. The simulations showed that the selection of initial experiments and measurement noise influenced the convergence of the Bayesian optimization algorithm to known optimal conditions. Both methods, however, showed comparable pilot-scale results on optimal digestion conditions, where high cellulose yields were combined with acceptable kappa numbers and pulp viscosities. Bayesian optimization did not enable a decrease in the number of experiments required for reaching these conditions but provided a more accurate model in the vicinity of the optimum based on additional modelling and cross-validation. These results shed light on the practical differences between the two methodologies for process development and are an important contribution to the chemometrics and machine learning communities.

Screening Life Cycle Assessment comparing One-step and Two-step Injection Molding Compounding using Conservative and Optimistic Scenarios

One-step injection molding compounding (IMC) is an innovative process to manufacture short-fiber-reinforced polymer composites. The aim of combining compounding and injection molding into one process is to enhance component quality and minimize environmental impacts. In this study, a screening Life Cycle Assessment (LCA) is conducted to evaluate and compare the environmental impacts of the IMC process with standard two-step manufacturing. Two scenarios for the IMC are considered, each differing in terms of machinery requirements, energy consumption, and material usage. Mechanically recycled polypropylene and glass fiber are used, and considered in the LCA employing a simple cut-off approach without awarding credits for substituting (primary) materials. The functional unit is the composite produced via the respective process, assuming equal functionality. Inventory data are obtained from initial experiments, literature, and the ecoinvent database. The impact assessment method selected is ReCiPe2016. Results indicate that the environmental performance improvement achieved by the IMC compared to the reference process is minimal in the conservative scenario where energy and material usage can be reduced but machinery usage is increased. However, in an optimistic scenario, the IMC can reduce the impacts of composite manufacturing by 34 %. The contributions at the midpoint level vary, and metal usage and energy consumption are the main contributors in all scenarios. A variation of the energy source for manufacturing shows the dependency of environmental impacts of components produced in both processes on the geographical location of production and its electricity supply. Methodological choices, such as the definition of the functional unit and modeling of recycled materials, have a large influence on LCA results, and alternative options are discussed.