The Hypothalamus in Alzheimer’s Disease

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder, characterized by irreversible decline of mental faculties, emotional and behavioral changes, loss of motor skills, and dysfunction of autonomic nervous system and disruption of circadian rhythms (CRs). We attempted to describe the morphological findings of the hypothalamus in early cases of AD, focusing our study mostly on the suprachiasmatic nucleus (SCN), the supraoptic nucleus (SON), and the paraventricular nucleus (PVN). Samples were processed for electron microscopy and silver impregnation techniques. The hypothalamic nuclei demonstrated a substantial decrease in the neuronal population, which was particularly prominent in the SCN. Marked abbreviation of dendritic arborization, in association with spinal pathology, was also seen. The SON and PVN demonstrated a substantial number of dystrophic axons and abnormal spines. Alzheimer’s pathology, such as deposits of amyloid-β peptide and neurofibrillary degeneration, was minimal. Electron microscopy revealed mitochondrial alterations in the cell body and the dendritic branches. The morphological alterations of the hypothalamic nuclei in early cases of AD may be related to the gradual alteration of CRs and the instability of autonomic regulation.


lvents using
s the performance criteria the molecular and mixture properties associated with thermodynamics, reactivity and sustainability.The proposed approach involves a fast screening stage in which numerous solvents are evaluated based on the simultaneous consideration of pure component properties.Several properties are specifically selected to represent the effects of molecular chemistry on the capture process.A few high-performing solvents are further evaluated using predictive models accounting for the very nonideal mixture behaviour.The prediction of pure component properties is supported by standard group contribution models.The solvent-water-CO2 interactions are represented within the SAFT-VR and SAFT-γ equations of state to predict accurately the mixture vapour-liquid equilibrium behaviour.The proposed developments are tested successfully on a dataset consisting of 126 potential solvent candidates.

Introduction

The identification and selection of efficient post-combustion CO2 capture solvents using chemical absorption is very challenging due to the non-ideal solvent-CO2-water chemical interactions.It requires the combined consideration of several thermodynamic, kinetic and sustainability properties as performance criteria in order to select the optimum capture solvents.Most existing works focus on developing rigorous models, which are indispensable for the accurate prediction of solvent properties prior to the selection of very few, realistically useful options.For example, the Statistical Associating Fluid Theory for potentials of variable range (SAFT-VR) has been used as an equation of state (EoS) to model accurately the fluidphase behaviour of multifunctional aqueous alkanolamines and CO2 (Rodriguez et al., 2012) based on the transfer of parameters from mixtures containing monoethanolamine (Mac Dowell et al., 2010) and alkylamines (Mac Dowell et al., 2011).A more predictive approach is the development of group parameters for these mixtures (Chremos et al., 2013) for use with the group contribution (GC) SAFT- SW (square well) approach.By the same token, continuum-solvation models have been used among other computational chemistry methods for the calculation of the amine basicity to measure molecular reactivity (Sumon et al., 2012).Detailed reaction models have also been proposed for important properties such as degradation (Lepaumier et al., 2010), associated with sustainability.Despite the usefulness of rigorous modelling, significant complexity results from the multitude of models that need to be combined to arrive at a realistic assessment, due to the large number of relevant properties and underlying physical and chemical phenomena.Increased computational cost is also incurred in some models.This prohibits the simultaneous consideration of multiple performance criteria, leading to solvents of reasonable performance only for the properties considered, limiting the selection to a few choices compar

to the very large
number of molecules and mixtures that may be considered as potential solvents.On the other hand, attempts to develop simpler models or statistical correlations that match structural molecular characteristics with pure component, mixture and/or kinetic properties associated with CO2 capture are gradually appearing in published literature.Retief (2012) considered the Hansen solubility parameter as the sole index for screening of a large number of amine-based solutions using chemical absorption.Porcheron et al. (2011) developed a statistical, neural network, model for the fast prediction of the pseudoacidity constant and the absorption isotherms for amines used in CO2 capture.The results of such efforts provide significant evidence that prior to utilizing rigorous models or lab-scale experiments, simpler models may facilitate a quick and reliable screening of CO2 capture solvents.


Proposed approach

The aim of our current work is to develop an approach to select solvents from a large pool of candidates that offer high performance simultaneously in many property criteria associated with effective chemical absorption of

O2 and solvent regeneration.We use a systematic scre
ning method, where accuracy is initially traded-off (although maintained at reasonable levels) for computational efficiency by using simple property models that have been specifically selected to identify useful capture candidates.Solvents with an infe

or performance are easily elim
nated while a few promising candidates are further evaluated using predictive models that account for non-ideal mixture behaviour.The simultaneous consideration of several properties that represent the effects of the molecular chemistry on the capture/regeneration process compensates for the initial use of simpler models and ensures the selection of fewer but more effective solvents.Selected candidates from the available pool are considered as discrete options in computations using the SAFT- SW EoS to restore accuracy in the prediction of important properties.


Property criteria for preliminary solvent screening

Numerous pure component solvent properties may be considered as preliminary screening crite ia.Several are analysed in this section based on their potential to capture important thermophysical, equilibrium, kinetic and sustainability characteristics.


Thermodynamics and reactivity

1) The Hansen solubility parameter (δ t ) represents the cohesive energy density, i.e., the energy required to overcome the solvent-solvent interactions.It may be used as an indicator of the solubility of CO2 in different solvents through the RED index (Stefanis and Panayiotou, 2008), expressed as a function of the solvent and solute dispersion, polar and hydrogen-bonding Hansen solubility parameters.Solvents with RED < 1 are generally considered to be favorable for the dissolution of the particular solute (e.g., CO2).

2) The vapour pressure (P vp ) is an important parameter for absorption/desorption CO2 capture systems because high P vp is an indicator of high volatility and potential solvent losses.

3) Higher basicity (pK a ) is associated with higher solvent rea

ivity.4) The li
uid heat capacity (C p ) of the CO2 rich stream after absorption is directly proportional to the sensible heat required in desorption (Oexman and Cather, 2010).Although the stream contains a solvent-CO2 reaction product, it may be assumed that a solvent of low C p facilitates the reduction of sensible heat.

5) The density (ρ), surface tension (σ), and viscosity (n) are solvent properties closely associated with the design and operating features of the separation columns.The density should be high to facilitate separation with reduced solvent flowrate and equipment size.Surface tension and viscosity should be low to improve the mass transfer in the packing material.6) The melting point (T m ) should be lower than the lowest process temperature to avoid solidification.The boiling point (T bp ) should be higher than the highest process temperature to avoid vaporization.


Sustainability

The solvent sustainability characteristics are evaluated using tw index categories.A cradle-to-gate Life Cycle Assessment (LCA) shows the environmental impacts of solvent production including the extraction, manufacture and disposal of raw and auxiliary materials.Safety, health and environmental hazard assessment (EHS) enables the estimation of harm potential in accidental scenarios within a CO2 capture plant in the presence of the tested solvents.7) In LCA the metrics used are the Cumulative Energy Demand (CED) in MJ-

uivalent, a resourceoriented in
icator which reflects the economic and the environmental impacts, the Global Warming Potential (GWP) in kg CO2 equivalent, which is a damage-oriented indicator reflecting the effect of greenhouse gas emissions, and the Eco-Indicator-99 -EI-99 (Goedkoop and Spriensma, 2000), providing an aggregated score for the end-point impact categories of human health, ecosystem-quality and resource consumption.

8) The EHS method (Sugiyama et al., 2008) represents an intermediate between hazards and exposure and it uses a comprehensive set of characteristic dangerous properties for each hazard category estimated from a prioritized list of physical and chemical substance properties.Substance properties are collected from material safety data sheets whenever possible and Quantitative Structure Activity Relationships (QSARs) are used to fill some of the data gaps.


Rigorous evaluation using SAFT

The SAFT-VR equation of state (Gil-Villegas et al., 1997) is a molecular-based EoS in which molecules are modelled as chains of homonuclear segments interacting via a variable range potential (here, a square-well potential), with association sites where appropriate.It provides an accurate description of the thermodynamic fluid-phase behaviour of a wide variety of compounds and their mixtures, including mixtures of CO2, water and amines (Mac Dowell et al., 2010) as well as ammonia and members of the nalkyl-1-amines homologous series (Mac Dowell et al., 2011).To increase the predictive capability of the thermodynamic description its formulation

h is required.In SAFT-γ SW (Lymperi
dis et al., 2007), molecules are formed from fused heteronuclear segments representing distinct functional groups.The underlying assumption (as in any GC approach) is that once the parameters for a particular functional group have been determined from regression to experimental data for specific molecules, they can be transferred to other molecules including that functional group without any further parameter adjustment, to predict the fluid-phase behaviour of compounds that have not yet been probed experimentally.A first step towards the description of multifunctional amines with the SAFT-γ SW EoS was presented in Lymperiadis et al. (2008) and extended in Chremos et al. (2013) to capture additional groups.This approach and the SAFT-VR models of multifunctional amines (Rodriguez et al., 2012) are used in our current work.


Implementation


Solvent database and reference set

The molecules considered as candidate CO2 capture solvents consist of 126 acyclic, aliphatic amines and hydroxyl-amines.They are obtained from an in-house database available from ETH Zurich, publicly available databases (NIST, 2014;EPA, 2014) and the commercial catalogue of Sigma-Aldrich (2014).The database is not exhaustive but provides an inclusive set of amines with different realizations of molecular structures assigned a CAS registry number.All the molecules are obtained from combinations of the following functional groups: CH3-, -CH2-, >CH-, >C<, -OH, -CH2-NH2, -CH2-NH-, -CH2-N<, >CH-NH2, >CH-NH-, CH3-NH-, CH3-N<, C-NH2.These are selected due the public availability of data required for property calculations through GC or other models.The available molecular set is separated in 3 classes: a) The reference class (R) contains 25 amines

reviously considered in the pu
lished literature as CO2 capture solvents (e.g., through determination of vapour-liquid equilibrium or reactivity properties etc.); b) The class of commercially available amines (

contains th
entire set of the 126 molecules; c) The class of commercially available hydroxyl-amines (L) contains molecules with at least a -OH group.The proposed separation is useful to investigate the possibility of identifying new molecules that might be effective as capture solvents compared to the amines in the R set.The term "commercially available amines" indicates that the molecules considered are chemical substances that have been synthesized previously (although not necessarily previously used in CO2 capture) as opposed to molecules proposed only theoretically (e.g., computationally).The reference set includes well-known solvents such as: 2-Amino-2-Methyl-1-Propanol (AMP); 2-(Methylamino)-Ethanol (MMEA); N-Methyl-1,3-Propanediamine (MAPA); 2-(Amino)-Propanol (AP); Ethylaminoethanol (EMEA); 1,2-Propanediamine (MEDA); N,N-Dimethylaminoethanol (DMMEA); 3-Amino-1-Propanol (MPA); Monoethanolamine (MEA); N,N-Diethyl-2aminoethanol (DEEA); 2-Amino-2-methyl-1,3-propanediol (AMPD); Triisopropylamine (TIPA); 4diethylamino-2-butanol (DEAB); Diethylenetriamine (DETA), N-(2-Aminoethyl)ethanolamine (AEEA).


Solvent selection and ranking

Considering the thermodynamic and reactivity properties presented in section 2.1 as performance measures, it is desired to select solvents based on the criteria set in Eq(1):
m max , min , , , , s.t. a p vp Abs bp Des pK C P n RED T T T T     (1) ,, 1 min p N i i j i j j J a x   
(2)


T Abs and T

Des are the temperatures of the absorption and desorption columns.Molecules that fail to satisfy the temperature constraints are removed from the set while the remaining molecules are rank-ordered u

ng J i to merge the con
idered properties in a unified criterion based on Eq(2).Term x i,j represents a scaled value of property J (e.g., ρ, pK a and so forth) for solvent i from the total of N p properties, and a i,j represents a unity coefficient that is positive for properties that need to be minimized and negative for those to be maximized.Solvents are ranked by minimising J i .The liquid molar density ρ is calculated as the inverse of liquid molar volume V m (Constantinou et al. 1995), σ and n from Conte et al. (2008), T m and T bp from Marrero and Gani ( 2001), P vp from the Riedel correlation (Poling et al., 2001), C p from Rayer et al. (2012), and pK a from Chemaxon Marvin (2014).After ranking the solvent set based on Eq(1) and Eq(2), the sustainability indices are calculated for each one of them.The FineChem tool (Wernet et al., 2009) is used to provide estimates (with an average relative error of 30-40 %) for LCA-based metrics using molecular descriptors as inputs.The functional unit is 1 kg of solvent produced.The Estimation Program Interface (EPI) Suite and the Toxicity Estimation Software Tool -TEST (EPA, 2014) are used for the ca

ulation of EHS metrics.An average of the dangerous properties calculated by each variation of the EHS method based on Sugiyama et al. (2008) is used to avoid missing scores in ca
e of data gaps.The EHS score is taken as the sum of the safety, health and environment scores.


Results and discussion

The average values of the index J for the top 10 molecules in each class are shown in igure 1 (left).It is apparent that the molecules in class C perform considerably better overall than those in the other classes.All classes contain molecules that may perform much better than MEA (dash marker) based on the properties considered as performance criteria.From a comparison of the C and L classes it appears that -OH has a detrimental effect on the overall performance.The close performance of the L and R classes indicates that most molecules in R are alkanolamines.The pK a and RED values are shown in Figure 1 (right) for the top 15 molecules selected based on J from the C and L classes.These are compared to the pK a and RED values of all 126 molecules.Most selected molecules indicate a simultaneously higher pK a (i.e., reaction rate) and RED (i.e,.CO2 absorption capacity) than MEA.Note that the performance (J value) of MEA is 67 % lower than the top molecule of class C. Some molecules have lower RED or pK a than molecules not selected in the top 15, but their performance is much better in the remaining properties.The top 5 molecules contained in class R are shown in Table 1, in addition to n-hexylamine (HEXA), selected from class C. MAPA exhibits a high pK a and low RED but ranks in the middle due to its high P vp .

The 20 molecules of class R mentioned in section 3.1 are ordered based on their J values, and the top 5 are reported in Table 1.Some of these molecules are outranked by those of Table 1 despite their high performance in a few property criteria.The proposed evaluation approach allows the consideration of such trade-offs.The data reported in Table 1 for several of the properties are either experimental or based on Antoine coefficients in the case of P vp .However, the molecules were rank-ordered using the GC methods reported in section 3.2 The average absolute deviation error (AADE) (from use of GC methods) is generally very low for the molecules of Table 1 compared to the standard deviation of the experimental or the Antoine correlations.For example the standard deviation in the P vp obtained from accurate Antoine models is approximately 1,100 Pa while the AADE from GC models is 340 Pa.This is an indication that the proposed approach is relatively robust to model uncertainty.Similar trends were observed for all properties.Molecules such as AMP and MAPA were considered in experimental published literature as good CO2 capture options and this is reflected in the results.HEXA is selected as an indicative example from class C. Its surface tension, viscosity, RED inde

and pK a in
icate it may be a good capture solvent.The CO2 partial pressure (Pco2) in aqueous mixtures of HEXA (SAFT-VR), MEA (SAFT-γ SW) and AMP (SAFT-VR) is shown in Figure 2 as a function of CO2 loading (Θco2) at T=313 K.A solvent composition of 30 wt % was used for MEA, 35 wt % for AMP, and 70 wt % for HEXA.MEA yields higher CO2 loading at low Pco2, while AMP performs better at higher Pco2.HEXA, which is a promising solvent (Zhang et al., 2012), has the same level of performance as MEA and AMP at Pco2 ≈ 3 MPa.The sustainability metrics calculated for the top molecules are shown in Table 2, with MEA and TIPA values for comparison.The combined EHS scores range from 1.18 to 5.07 (average of 2.86) for the 126 molecules.TIPA is the l