基于谓词不变性的状态反馈控制在机器人遥操作中的应用

传统的基于事件遥操作系统虽然巧妙地避开了变化时延问题[8] ,但是它具有“走走停停”的特点 ,即机器人将当前状态发送出去之后直到操作者的下一个指令到来之前机器人必须保证状态没有变化。对此本文首次提出了利用基于谓词不变性的状态反馈控制来对原有的方法进行改进 ,使机器人能够对动态环境具有适应性

不同类型量表的线上调查和传统纸笔调查的比较研究

During the past 11 years, with the rapid development of the Internet, more and more psychologists began to realize and take advantage of it, which led to a growing number of psychological test administrated on the internet for data collection. But there were some controversy about the reliability and representatively of this new method. To examine the applicability of the Online Survey and how different types of scales used on the internet, we first reversed the measurement instrument, then from three different levels to investigate the equivalence of online survey and paper-and-pencil assessment, namely, sample level, scale level and item level. Both Classical Test Theory and Item Response Theory were used to analyze the invariance of different types of scale applicability on the internet. The main conclusions of this study could be drawn as follows: 1. In the sample-based study, self-select sample of the online survey was compared to the random sampled sample of paper-and-pencil assessment. The results showed there were no gender difference between them (p>0.05), but the online survey sample was characterized with high qualifications, high-income and younger features (88% of the sample with post-secondary education or above, and 71% aged 20 -29 years). There were significant differences on the scores of all scales between online survey and paper-and-pencil assessment (p<0.01). With demographic controlled, there was no significant difference on the variable of Neurotic between different surveys (p>0.05). 2. With in-group design, it was proved equivalence of the scale of BI (Attitude toward Brand Importance), BT (Attitude toword Brand Switcher), Extraversion, and Conscientiousbess in the compared study in the reliability, construct validity and average scores. 3. On the item level, the results based on the Item Response Theory analysis showed that 2PLM is appropriate for personality and attitude scales. With regard to personality scale, there emerged some items with DIF in the dimensions of Openness to the experience subscale and Agreeable subscale. However, there were no significant differences about the test function. 4. Exploring the psychometrics properties of answer formats of five-, six-, seven-, ten-points, it was showed that different measurement validity between online survey and paper-and-pencil test. It was also described the lower reliability and validity of six-point scale. In conclusion, the results support the application of personality scale online, but for attitude scale, we need to choose prudently.

Spatial Reference Frames for Object Recognition: Tuning for Rotations in Depth

The inferior temporal cortex (IT) of monkeys is thought to play an essential role in visual object recognition. Inferotemporal neurons are known to respond to complex visual stimuli, including patterns like faces, hands, or other body parts. What is the role of such neurons in object recognition? The present study examines this question in combined psychophysical and electrophysiological experiments, in which monkeys learned to classify and recognize novel visual 3D objects. A population of neurons in IT were found to respond selectively to such objects that the monkeys had recently learned to recognize. A large majority of these cells discharged maximally for one view of the object, while their response fell off gradually as the object was rotated away from the neuron"s preferred view. Most neurons exhibited orientation-dependent responses also during view-plane rotations. Some neurons were found tuned around two views of the same object, while a very small number of cells responded in a view- invariant manner. For five different objects that were extensively used during the training of the animals, and for which behavioral performance became view-independent, multiple cells were found that were tuned around different views of the same object. No selective responses were ever encountered for views that the animal systematically failed to recognize. The results of our experiments suggest that neurons in this area can develop a complex receptive field organization as a consequence of extensive training in the discrimination and recognition of objects. Simple geometric features did not appear to account for the neurons" selective responses. These findings support the idea that a population of neurons -- each tuned to a different object aspect, and each showing a certain degree of invariance to image transformations -- may, as an assembly, encode complex 3D objects. In such a system, several neurons may be active for any given vantage point, with a single unit acting like a blurred template for a limited neighborhood of a single view.

Learning Object-Independent Modes of Variation with Feature Flow Fields

We present a unifying framework in which "object-independent" modes of variation are learned from continuous-time data such as video sequences. These modes of variation can be used as "generators" to produce a manifold of images of a new object from a single example of that object. We develop the framework in the context of a well-known example: analyzing the modes of spatial deformations of a scene under camera movement. Our method learns a close approximation to the standard affine deformations that are expected from the geometry of the situation, and does so in a completely unsupervised (i.e. ignorant of the geometry of the situation) fashion. We stress that it is learning a "parameterization", not just the parameter values, of the data. We then demonstrate how we have used the same framework to derive a novel data-driven model of joint color change in images due to common lighting variations. The model is superior to previous models of color change in describing non-linear color changes due to lighting.

Time-Frequency Representations for Speech Signals

This work addresses two related questions. The first question is what joint time-frequency energy representations are most appropriate for auditory signals, in particular, for speech signals in sonorant regions. The quadratic transforms of the signal are examined, a large class that includes, for example, the spectrograms and the Wigner distribution. Quasi-stationarity is not assumed, since this would neglect dynamic regions. A set of desired properties is proposed for the representation: (1) shift-invariance, (2) positivity, (3) superposition, (4) locality, and (5) smoothness. Several relations among these properties are proved: shift-invariance and positivity imply the transform is a superposition of spectrograms; positivity and superposition are equivalent conditions when the transform is real; positivity limits the simultaneous time and frequency resolution (locality) possible for the transform, defining an uncertainty relation for joint time-frequency energy representations; and locality and smoothness tradeoff by the 2-D generalization of the classical uncertainty relation. The transform that best meets these criteria is derived, which consists of two-dimensionally smoothed Wigner distributions with (possibly oriented) 2-D guassian kernels. These transforms are then related to time-frequency filtering, a method for estimating the time-varying 'transfer function' of the vocal tract, which is somewhat analogous to ceptstral filtering generalized to the time-varying case. Natural speech examples are provided. The second question addressed is how to obtain a rich, symbolic description of the phonetically relevant features in these time-frequency energy surfaces, the so-called schematic spectrogram. Time-frequency ridges, the 2-D analog of spectral peaks, are one feature that is proposed. If non-oriented kernels are used for the energy representation, then the ridge tops can be identified, with zero-crossings in the inner product of the gradient vector and the direction of greatest downward curvature. If oriented kernels are used, the method can be generalized to give better orientation selectivity (e.g., at intersecting ridges) at the cost of poorer time-frequency locality. Many speech examples are given showing the performance for some traditionally difficult cases: semi-vowels and glides, nasalized vowels, consonant-vowel transitions, female speech, and imperfect transmission channels.

Visible Volume: a Robust Measure for Protein Structure Characterization

We propose a new characterization of protein structure based on the natural tetrahedral geometry of the β carbon and a new geometric measure of structural similarity, called visible volume. In our model, the side-chains are replaced by an ideal tetrahedron, the orientation of which is fixed with respect to the backbone and corresponds to the preferred rotamer directions. Visible volume is a measure of the non-occluded empty space surrounding each residue position after the side-chains have been removed. It is a robust, parameter-free, locally-computed quantity that accounts for many of the spatial constraints that are of relevance to the corresponding position in the native structure. When computing visible volume, we ignore the nature of both the residue observed at each site and the ones surrounding it. We focus instead on the space that, together, these residues could occupy. By doing so, we are able to quantify a new kind of invariance beyond the apparent variations in protein families, namely, the conservation of the physical space available at structurally equivalent positions for side-chain packing. Corresponding positions in native structures are likely to be of interest in protein structure prediction, protein design, and homology modeling. Visible volume is related to the degree of exposure of a residue position and to the actual rotamers in native proteins. In this article, we discuss the properties of this new measure, namely, its robustness with respect to both crystallographic uncertainties and naturally occurring variations in atomic coordinates, and the remarkable fact that it is essentially independent of the choice of the parameters used in calculating it. We also show how visible volume can be used to align protein structures, to identify structurally equivalent positions that are conserved in a family of proteins, and to single out positions in a protein that are likely to be of biological interest. These properties qualify visible volume as a powerful tool in a variety of applications, from the detailed analysis of protein structure to homology modeling, protein structural alignment, and the definition of better scoring functions for threading purposes.

Surface Reconstruction from Multiple Views using Rational B-Splines

A method for reconstructing 3D rational B-spline surfaces from multiple views is proposed. The method takes advantage of the projective invariance properties of rational B-splines. Given feature correspondences in multiple views, the 3D surface is reconstructed via a four step framework. First, corresponding features in each view are given an initial surface parameter value (s; t), and a 2D B-spline is fitted in each view. After this initialization, an iterative minimization procedure alternates between updating the 2D B-spline control points and re-estimating each feature's (s; t). Next, a non-linear minimization method is used to upgrade the 2D B-splines to 2D rational B-splines, and obtain a better fit. Finally, a factorization method is used to reconstruct the 3D B-spline surface given 2D B-splines in each view. This surface recovery method can be applied in both the perspective and orthographic case. The orthographic case allows the use of additional constraints in the recovery. Experiments with real and synthetic imagery demonstrate the efficacy of the approach for the orthographic case.

Type Inference for Variant Object Types

Existing type systems for object calculi are based on invariant subtyping. Subtyping invariance is required for soundness of static typing in the presence of method overrides, but it is often in the way of the expressive power of the type system. Flexibility of static typing can be recovered in different ways: in first-order systems, by the adoption of object types with variance annotations, in second-order systems by resorting to Self types. Type inference is known to be P-complete for first-order systems of finite and recursive object types, and NP-complete for a restricted version of Self types. The complexity of type inference for systems with variance annotations is yet unknown. This paper presents a new object type system based on the notion of Split types, a form of object types where every method is assigned two types, namely, an update type and a select type. The subtyping relation that arises for Split types is variant and, as a result, subtyping can be performed both in width and in depth. The new type system generalizes all the existing first-order type systems for objects, including systems based on variance annotations. Interestingly, the additional expressive power does not affect the complexity of the type inference problem, as we show by presenting an O(n^3) inference algorithm.

From Parallel Sequence Representations to Calligraphic Control: A Conspiracy of Neural Circuits

Calligraphic writing presents a rich set of challenges to the human movement control system. These challenges include: initial learning, and recall from memory, of prescribed stroke sequences; critical timing of stroke onsets and durations; fine control of grip and contact forces; and letter-form invariance under voluntary size scaling, which entails fine control of stroke direction and amplitude during recruitment and derecruitment of musculoskeletal degrees of freedom. Experimental and computational studies in behavioral neuroscience have made rapid progress toward explaining the learning, planning and contTOl exercised in tasks that share features with calligraphic writing and drawing. This article summarizes computational neuroscience models and related neurobiological data that reveal critical operations spanning from parallel sequence representations to fine force control. Part one addresses stroke sequencing. It treats competitive queuing (CQ) models of sequence representation, performance, learning, and recall. Part two addresses letter size scaling and motor equivalence. It treats cursive handwriting models together with models in which sensory-motor tmnsformations are performed by circuits that learn inverse differential kinematic mappings. Part three addresses fine-grained control of timing and transient forces, by treating circuit models that learn to solve inverse dynamics problems.

Adaptive Neural Networks for Control of Movement Trajectories Invariant under Speed and Force Rescaling

This article describes two neural network modules that form part of an emerging theory of how adaptive control of goal-directed sensory-motor skills is achieved by humans and other animals. The Vector-Integration-To-Endpoint (VITE) model suggests how synchronous multi-joint trajectories are generated and performed at variable speeds. The Factorization-of-LEngth-and-TEnsion (FLETE) model suggests how outflow movement commands from a VITE model may be performed at variable force levels without a loss of positional accuracy. The invariance of positional control under speed and force rescaling sheds new light upon a familiar strategy of motor skill development: Skill learning begins with performance at low speed and low limb compliance and proceeds to higher speeds and compliances. The VITE model helps to explain many neural and behavioral data about trajectory formation, including data about neural coding within the posterior parietal cortex, motor cortex, and globus pallidus, and behavioral properties such as Woodworth's Law, Fitts Law, peak acceleration as a function of movement amplitude and duration, isotonic arm movement properties before and after arm-deafferentation, central error correction properties of isometric contractions, motor priming without overt action, velocity amplification during target switching, velocity profile invariance across different movement distances, changes in velocity profile asymmetry across different movement durations, staggered onset times for controlling linear trajectories with synchronous offset times, changes in the ratio of maximum to average velocity during discrete versus serial movements, and shared properties of arm and speech articulator movements. The FLETE model provides new insights into how spina-muscular circuits process variable forces without a loss of positional control. These results explicate the size principle of motor neuron recruitment, descending co-contractive compliance signals, Renshaw cells, Ia interneurons, fast automatic reactive control by ascending feedback from muscle spindles, slow adaptive predictive control via cerebellar learning using muscle spindle error signals to train adaptive movement gains, fractured somatotopy in the opponent organization of cerebellar learning, adaptive compensation for variable moment-arms, and force feedback from Golgi tendon organs. More generally, the models provide a computational rationale for the use of nonspecific control signals in volitional control, or "acts of will", and of efference copies and opponent processing in both reactive and adaptive motor control tasks.

Axial symmetry and transverse trace-free tensors in numerical relativity

Transverse trace-free (TT) tensors play an important role in the initial conditions of numerical relativity, containing two of the component freedoms. Expressing a TT tensor entirely, by the choice of two scalar potentials, is not a trivial task however. Assuming the added condition of axial symmetry, expressions are given in both spherical and cylindrical coordinates, for TT tensors in flat space. A coordinate relation is then calculated between the scalar potentials of each coordinate system. This is extended to a non-flat space, though only one potential is found. The remaining equations are reduced to form a second order partial differential equation in two of the tensor components. With the axially symmetric flat space tensors, the choice of potentials giving Bowen-York conformal curvatures, are derived. A restriction is found for the potentials which ensure an axially symmetric TT tensor, which is regular at the origin, and conditions on the potentials, which give an axially symmetric TT tensor with a spherically symmetric scalar product, are also derived. A comparison is made of the extrinsic curvatures of the exact Kerr solution and numerical Bowen-York solution for axially symmetric black hole space-times. The Brill wave, believed to act as the difference between the Kerr and Bowen-York space-times, is also studied, with an approximate numerical solution found for a mass-factor, under different amplitudes of the metric.

Ab initio calculations of group 4 metallocene reaction mechanisms: atomic layer deposition and bond activation catalysis

Thin film dielectrics based on titanium, zirconium or hafnium oxides are being introduced to increase the permittivity of insulating layers in transistors for micro/nanoelectronics and memory devices. Atomic layer deposition (ALD) is the process of choice for fabricating these films, as it allows for high control of composition and thickness in thin, conformal films which can be deposited on substrates with high aspect-ratio features. The success of this method depends crucially on the chemical properties of the precursor molecules. A successful ALD precursor should be volatile, stable in the gas-phase, but reactive on the substrate and growing surface, leading to inert by-products. In recent years, many different ALD precursors for metal oxides have been developed, but many of them suffer from low thermal stability. Much promise is shown by group 4 metal precursors that contain cyclopentadienyl (Cp = C5H5-xRx) ligands. One of the main advantages of Cp precursors is their thermal stability. In this work ab initio calculations were carried out at the level of density functional theory (DFT) on a range of heteroleptic metallocenes [M(Cp)4-n(L)n], M = Hf/Zr/Ti, L = Me and OMe, in order to find mechanistic reasons for their observed behaviour during ALD. Based on optimized monomer structures, reactivity is analyzed with respect to ligand elimination. The order in which different ligands are eliminated during ALD follows their energetics which was in agreement with experimental measurements. Titanocene-derived precursors, TiCp*(OMe)3, do not yield TiO2 films in atomic layer deposition (ALD) with water, while Ti(OMe)4 does. DFT was used to model the ALD reaction sequence and find the reason for the difference in growth behaviour. Both precursors adsorb initially via hydrogen-bonding. The simulations reveal that the Cp* ligand of TiCp*(OMe)3 lowers the Lewis acidity of the Ti centre and prevents its coordination to surface O (densification) during both of the ALD pulses. Blocking this step hindered further ALD reactions and for that reason no ALD growth is observed from TiCp*(OMe)3 and water. The thermal stability in the gas phase of Ti, Zr and Hf precursors that contain cyclopentadienyl ligands was also considered. The reaction that was found using DFT is an intramolecular α-H transfer that produces an alkylidene complex. The analysis shows that thermal stabilities of complexes of the type MCp2(CH3)2 increase down group 4 (M = Ti, Zr and Hf) due to an increase in the HOMO-LUMO band gap of the reactants, which itself increases with the electrophilicity of the metal. The reverse reaction of α-hydrogen abstraction in ZrCp2Me2 is 1,2-addition reaction of a C-H bond to a Zr=C bond. The same mechanism is investigated to determine if it operates for 1,2 addition of the tBu C-H across Hf=N in a corresponding Hf dimer complex. The aim of this work is to understand orbital interactions, how bonds break and how new bonds form, and in what state hydrogen is transferred during the reaction. Calculations reveal two synchronous and concerted electron transfers within a four-membered cyclic transition state in the plane between the cyclopentadienyl rings, one π(M=X)-to-σ(M-C) involving metal d orbitals and the other σ(C-H)-to-σ(X-H) mediating the transfer of neutral H, where X = C or N. The reaction of the hafnium dimer complex with CO that was studied for the purpose of understanding C-H bond activation has another interesting application, namely the cleavage of an N-N bond and resulting N-C bond formation. Analysis of the orbital plots reveals repulsion between the occupied orbitals on CO and the N-N unit where CO approaches along the N-N axis. The repulsions along the N-N axis are minimized by instead forming an asymmetrical intermediate in which CO first coordinates to one Hf and then to N. This breaks the symmetry of the N-N unit and the resultant mixing of MOs allows σ(NN) to be polarized, localizing electrons on the more distant N. This allowed σ(CO) and π(CO) donation to N and back-donation of π*(Hf2N2) to CO. Improved understanding of the chemistry of metal complexes can be gained from atomic-scale modelling and this provides valuable information for the design of new ALD precursors. The information gained from the model decomposition pathway can be additionally used to understand the chemistry of molecules in the ALD process as well as in catalytic systems.

Multi-scale modelling of atomic layer deposition

High-permittivity ("high-k") dielectric materials are used in the transistor gate stack in integrated circuits. As the thickness of silicon oxide dielectric reduces below 2 nm with continued downscaling, the leakage current because of tunnelling increases, leading to high power consumption and reduced device reliability. Hence, research concentrates on finding materials with high dielectric constant that can be easily integrated into a manufacturing process and show the desired properties as a thin film. Atomic layer deposition (ALD) is used practically to deposit high-k materials like HfO2, ZrO2, and Al2O3 as gate oxides. ALD is a technique for producing conformal layers of material with nanometer-scale thickness, used commercially in non-planar electronics and increasingly in other areas of science and technology. ALD is a type of chemical vapor deposition that depends on self-limiting surface chemistry. In ALD, gaseous precursors are allowed individually into the reactor chamber in alternating pulses. Between each pulse, inert gas is admitted to prevent gas phase reactions. This thesis provides a profound understanding of the ALD of oxides such as HfO2, showing how the chemistry affects the properties of the deposited film. Using multi-scale modelling of ALD, the kinetics of reactions at the growing surface is connected to experimental data. In this thesis, we use density functional theory (DFT) method to simulate more realistic models for the growth of HfO2 from Hf(N(CH3)2)4/H2O and HfCl4/H2O and for Al2O3 from Al(CH3)3/H2O.Three major breakthroughs are discovered. First, a new reaction pathway, ’multiple proton diffusion’, is proposed for the growth of HfO2 from Hf(N(CH3)2)4/H2O.1 As a second major breakthrough, a ’cooperative’ action between adsorbed precursors is shown to play an important role in ALD. By this we mean that previously-inert fragments can become reactive once sufficient molecules adsorb in their neighbourhood during either precursor pulse. As a third breakthrough, the ALD of HfO2 from Hf(N(CH3)2)4 and H2O is implemented for the first time into 3D on-lattice kinetic Monte-Carlo (KMC).2 In this integrated approach (DFT+KMC), retaining the accuracy of the atomistic model in the higher-scale model leads to remarkable breakthroughs in our understanding. The resulting atomistic model allows direct comparison with experimental techniques such as X-ray photoelectron spectroscopy and quartz crystal microbalance.

Atomic layer deposition of copper for CMOS interconnects

This work concerns the atomic layer deposition (ALD) of copper. ALD is a technique that allows conformal coating of difficult topographies such as narrow trenches and holes or even shadowed regions. However, the deposition of pure metals has so far been less successful than the deposition of oxides except for a few exceptions. Challenges include difficulties associated with the reduction of the metal centre of the precursor at reasonable temperatures and the tendency of metals to agglomerate during the growth process. Cu is a metal of special technical interest as it is widely used for interconnects on CMOS devices. These interconnects are usually fabricated by electroplating, which requires the deposition of thin Cu seed layers onto the trenches and vias. Here, ALD is regarded as potential candidate for replacing the current PVD technique, which is expected to reach its limitations as the critical dimensions continue to shrink. This work is separated into two parts. In the first part, a laboratory-scale ALD reactor was constructed and used for the thermal ALD of Cu. In the second part, the potentials of the application of Cu ALD on industry scale fabrication were examined in a joint project with Applied Materials and Intel. Within this project precursors developed by industrial partners were evaluated on a 300 mm Applied Materials metal-ALD chamber modified with a direct RF-plasma source. A feature that makes ALD a popular technique among researchers is the possibility to produce high- level thin film coatings for micro-electronics and nano-technology with relatively simple laboratory- scale reactors. The advanced materials and surfaces group (AMSG) at Tyndall National Institute operates a range of home-built ALD reactors. In order to carry out Cu ALD experiments, modifications to the normal reactor design had to be made. For example a carrier gas mechanism was necessary to facilitate the transport of the low-volatile Cu precursors. Precursors evaluated included the readily available Cu(II)-diketonates Cu-bis(acetylacetonate), Cu-bis(2,2,6,6-tetramethyl-hepta-3,5-dionate) and Cu-bis(1,1,1,5,5,5-hexafluoacetylacetonate) as well as the Cu-ketoiminate Cu-bis(4N-ethylamino- pent-3-en-2-onate), which is also known under the trade name AbaCus (Air Liquide), and the Cu(I)- silylamide 1,3-diisopropyl-imidazolin-2-ylidene Cu(I) hexamethyldisilazide ([NHC]Cu(hmds)), which was developed at Carleton University Ottawa. Forming gas (10 % H2 in Ar) was used as reducing agent except in early experiments where formalin was used. With all precursors an extreme surface selectivity of the deposition process was observed and significant growth was only achieved on platinum-group metals. Improvements in the Cu deposition process were obtained with [NHC]Cu(hmds) compared with the Cu(II) complexes. A possible reason is the reduced oxidation state of the metal centre. Continuous Cu films were obtained on Pd and indications for saturated growth with a rate of about 0.4 Å/cycle were found for deposition at 220 °C. Deposits obtained on Ru consisted of separated islands. Although no continuous films could be obtained in this work the relatively high density of Cu islands obtained was a clear improvement as compared to the deposits grown with Cu(II) complexes. When ultra-thin Pd films were used as substrates, island growth was also observed. A likely reason for this extreme difference to the Cu films obtained on thicker Pd films is the lack of stress compensation within the thin films. The most likely source of stress compensation in the thicker Pd films is the formation of a graded interlayer between Pd and Cu by inter-diffusion. To obtain continuous Cu films on more materials, reduction of the growth temperature was required. This was achieved in the plasma assisted ALD experiments discussed in the second part of this work. The precursors evaluated included the AbaCus compound and CTA-1, an aliphatic Cu-bis(aminoalkoxide), which was supplied by Adeka Corp.. Depositions could be carried out at very low temperatures (60 °C Abacus, 30 °C CTA-1). Metallic Cu could be obtained on all substrate materials investigated, but the shape of the deposits varied significantly between the substrate materials. On most materials (Si, TaN, Al2O3, CDO) Cu grew in isolated nearly spherical islands even at temperatures as low as 30 °C. It was observed that the reason for the island formation is the coalescence of the initial islands to larger, spherical islands instead of forming a continuous film. On the other hand, the formation of nearly two-dimensional islands was observed on Ru. These islands grew together forming a conductive film after a reasonably small number of cycles. The resulting Cu films were of excellent crystal quality and had good electrical properties; e.g. a resistivity of 2.39 µΩ cm was measured for a 47 nm thick film. Moreover, conformal coating of narrow trenches (1 µm deep 100/1 aspect ratio) was demonstrated showing the feasibility of the ALD process.

Cooperation between adsorbates accounts for the activation of atomic layer deposition reactions

Atomic layer deposition (ALD) is a technique for producing conformal layers of nanometre-scale thickness, used commercially in non-planar electronics and increasingly in other high-tech industries. ALD depends on self-limiting surface chemistry but the mechanistic reasons for this are not understood in detail. Here we demonstrate, by first-principle calculations of growth of HfO2 from Hf(N(CH3)2)4–H2O and HfCl4–H2O and growth of Al2O3 from Al(CH3)3–H2O, that, for all these precursors, co-adsorption plays an important role in ALD. By this we mean that previously-inert adsorbed fragments can become reactive once sufficient numbers of molecules adsorb in their neighbourhood during either precursor pulse. Through the calculated activation energies, this ‘cooperative’ mechanism is shown to have a profound influence on proton transfer and ligand desorption, which are crucial steps in the ALD cycle. Depletion of reactive species and increasing coordination cause these reactions to self-limit during one precursor pulse, but to be re-activated via the cooperative effect in the next pulse. This explains the self-limiting nature of ALD.