Search

An extracellular chlorogenic acid esterase from Ustilago maydis (UmChlE) was purified to homogeneity by using three separation steps, including anion-exchange chromatography on a Q Sepharose FF column, preparative isoelectric focusing (IEF), and, finally, a combination of affinity chromatography and hydrophobic interaction chromatography on polyamide. SDS-PAGE analysis suggested a monomeric protein of ∼71 kDa. The purified enzyme showed maximal activity at pH 7.5 and at 37°C and was active over a wide pH range (3.5 to 9.5). Previously described chlorogenic acid esterases exhibited a comparable affinity for chlorogenic acid, but the enzyme from Ustilago was also active on typical feruloyl esterase substrates. Kinetic constants for chlorogenic acid, methyl p -coumarate, methyl caffeate, and methyl ferulate were as follows: K m values of 19.6 μM, 64.1 μM, 72.5 μM, and 101.8 μM, respectively, and k cat / K m values of 25.83 mM −1 s −1 , 7.63 mM −1 s −1 , 3.83 mM −1 s −1 and 3.75 mM −1 s −1 , respectively. UmChlE released ferulic, p -coumaric, and caffeic acids from natural substrates such as destarched wheat bran (DSWB) and coffee pulp (CP), confirming activity on complex plant biomass. The full-length gene encoding UmChlE consisted of 1,758 bp, corresponding to a protein of 585 amino acids, and was functionally produced in Pichia pastoris GS115. Sequence alignments with annotated chlorogenic acid and feruloyl esterases underlined the uniqueness of this enzyme.

Over the past two decades, a plethora of crystal structures of molybdenum enzymes has appeared in the literature providing a clearer picture of the enzymatic active sites and increasing the challenge to chemists to develop accurate models for those sites. In this minireview we discuss the most recent model studies aimed to reproduce detailed features of the pterin–dithiolene ligand, both as the uncoordinated form and as a chelate coordinated to molybdenum.

Advanced accelerator research is aimed at finding new technologies that can dramatically reduce the size and cost of future high-energy accelerators. Supercomputing is already playing a dramatic and critical role in this quest. One of the goals of the SciDAC Accelerator Modeling Project is to develop code and software that can ultimately be used to discover the underlying science of new accelerator technology and then be used to design future high-energy accelerators with a minimum amount of capital expenditure on large-scale experiments. We describe the existing hierarchy of software tools for modelling advanced accelerators, how these models have been validated against experiment, how the models are benchmarked against each other, and how these tools are being successfully used to elucidate the underlying science.

Laser-driven accelerators, in which particles are accelerated by the electric field of a plasma wave (the wakefield) driven by an intense laser, have demonstrated accelerating electric fields of hundreds of GV m-1 (refs 1–3). These fields are thousands of times greater than those achievable in conventional radio-frequency accelerators, spurring interest in laser accelerators4, 5 as compact next-generation sources of energetic electrons and radiation. To date, however, acceleration distances have been severely limited by the lack of a controllable method for extending the propagation distance of the focused laser pulse. The ensuing short acceleration distance results in low-energy beams with 100 per cent electron energy spread1, 2, 3, which limits potential applications. Here we demonstrate a laser accelerator that produces electron beams with an energy spread of a few per cent, low emittance and increased energy (more than 109 electrons above 80 MeV). Our technique involves the use of a preformed plasma density channel to guide a relativistically intense laser, resulting in a longer propagation distance. The results open the way for compact and tunable high-brightness sources of electrons and radiation.

VORPAL is a new plasma simulation code designed for maximum flexibility through use of advance C++ techniques. Through use of inheritance. VORPAL incorporates multiple models for the plasma and electromagnetic fields. The plasma models include both particle-in-cell and fluid models. Through C++ meta-template programming a single code can be used to simulate one-, two-, or three-dimensional systems with no less of performance. VORPAL can also be run in either serial or parallel, with the latter using a general domain decomposition. A new fluid algorithm that allows for regions of zero density was developed and incorporated into the code. VORPAL simulation results for the generation of laser wake fields through laser-plasma interaction are presented.

For years compressor researchers have tried to account for the fluid interaction effect of the working fluid on valve motion in displacement compressors. In recent years, the computing capacities and available CFD and FEA programs have allowed fully coupled interaction of fluids and moving structures to be modelled more comprehensively. This paper describes our experience and results from developing a model of a multi-cylinder reciprocating piston compressor with suction and discharge valve systems that are fully coupled with the pressure pulsation in the adjacent plenum. Valve dynamics are captured by the model that affects compressor performance. The results show that higher running speed causes more discharge valve delay on closing due to higher pressure pulsation in discharge plenum. The acoustic property of the discharge plenum as it relates to valve motion is studied by the developed cost-effective standalone model.

A series of molybdenum−pterin complexes produced from reactions of molybdenum and pterin reagents in various oxidation states has been investigated by X-ray photoelectron spectroscopy (XPS). Prior difficulties in making oxidation state assignments for the metal center and coordinated pterin can be resolved through comparison of Mo 3d binding energies (BE) for these new complexes with the BEs of standard molybdenum complexes. XPS analysis of molybdenum−pterin complexes produced from reactions of Mo(VI) reagents with tetrahydropterins show binding energies that are shifted by 1.5−1.8 eV to lower energies as compared to the BEs observed for the oxo-Mo(VI) reagents. The opposite shift in BE values is observed for complexes prepared from Mo(IV) chloride and fully oxidized pterins where BEs shift to higher values with respect to those for the starting Mo(IV) reagents. Remarkably, the BEs obtained for Mo−pterin complexes originating from Mo(VI)−tetrahydropterin reactions are nearly identical with those from Mo(I...

The coordination of alloxazine and pterins to molybdenum(IV) is demonstrated in this study. The synthesis of MoOCl3(pteridineH), where pteridineH is the protonated form of 1,3,7,8-tetramethylalloxazine (tmaz), 2-pivaloyl-6,7-dimethylpterin (piv-dmp), and 6,7-dimethylpterin (dmp), proceeds readily starting from Mo(IV)Cl4(acetonitrile)2 and the pteridine ligand in chloroform or methanol. X-ray crystal structures of MoOCl3(tmazH) (1) and MoOCl3(piv-dmpH) (2) show that Mo chelates each pteridine at the carbonyl oxygen and pyrazine nitrogen and that the pteridine ligand is protonated at the other nitrogen in the pyrazine ring. A third X-ray structure for MoOCl3(H3dmp) (4) is included in this work since its determination permits the comparison of metrical parameters for the oxidized and reduced forms of a pterin in identical molybdenum coordination environments. The major difference observed in the structures of 2 as compared to 4 is the Mo−N5 bond length which is significantly shorter in compound 4 containing ...

The synthesis and structure of a new molybdenum complex coordinated by a reduced pterin is reported. [MoOCl(detc)(H3dmp)]Cl (1) (where detc is diethyldithiocarbamate and H3dmp is 6,7-dimethyl-6,7,8-trihydropterin) is prepared from MoOCl2(detc)2 and H4dmp (6,7-dimethyl-5,6,7,8-tetrahydropterin). The X-ray structure determination of [MoOCl(detc)(H3dmp)]Cl·MeOH reveals an octahedral complex where the reduced pterin ligand coordinates through the carbonyl oxygen and pyrazine ring nitrogen atoms. An extensive hydrogen bonding network in the crystal lattice connects adjacent complexes, the chloride counterion, and the molecule of methanol. This hydrogen bonding persists in solution where it is identified by characteristic absorptions in the electronic spectrum. Dimethyl sulfoxide (DMSO) oxidizes [MoOCl(detc)(H3dmp)]Cl, producing 1 equiv of dimethylpterin and 1/2 equiv of oxidized dithiocarbamate, tetraethylthiuramdisulfide (TETDS), a reaction that constitutes a net five-electron oxidation of the molybdenum comp...

An extracellular feruloyl esterase (PeFaeA) from the culture supernatant of Pleurotus eryngii was purified to homogeneity using cation exchange, hydrophobic interaction, and size exclusion chromatography. The length of the complete coding sequence of PeFaeA was determined to 1668 bp corresponding to a protein of 555 amino acids. The catalytic triad of Ser-Glu-His demonstrated the uniqueness of the enzyme compared to previously published FAEs. The purified PeFaeA was a monomer with an estimated molecular mass of 67 kDa. Maximum feruloyl esterase (FAE) activity was observed at pH 5.0 and 50 °C, respectively. Metal ions (5 mM), except Hg(2+), had no significant influence on the enzyme activity. Substrate specificity profiling characterized the enzyme as a type A FAE preferring bulky natural substrates, such as feruloylated saccharides, rather than small synthetic ones. Km and kcat of the purified enzyme for methyl ferulate were 0.15 mM and 0.85 s(-1). In the presence of 3 M NaCl activity of the enzyme increased by 28 %. PeFaeA alone released only little ferulic acid from destarched wheat bran (DSWB), whereas after addition of Trichoderma viride xylanase the concentration increased more than 20 fold.

This work introduces a conformal finite-difference time domain (CFDTD) particle-in-cell (PIC) method with improved field emission algorithms to accurately and efficiently study field emission devices. The CFDTD method is based on the Dey-Mittra algorithm or cut-cell algorithm, as implemented in the VORPAL code. Locally distorted cells with edges tangential to the metallic surface are used. The fields in these distorted cells are computed using a modification of the conventional FDTD update equations. Complicated electromagnetic structures can be well described by cut-cells in the CFDTD method. For the field emission algorithm, we employ the elliptic function v(y) found by Forbes and a new fitting function t(y)2 for the Fowler-Nordheim (FN) equation. With these improved correction factors, field emission of electrons from a cathode surface is much closer to the prediction of the exact FN formula derived by Murphy and Good. In addition, divergent problems or singularities in PIC numerical simulations can be well taken care of. As the emission behavior is strongly dependent on the surface electric fields, the local fields, field enhancements around complicated field emission tips, and the corresponding electron emission can be predicted more accurately and efficiently using this new modeling method, both quasi-statically and dynamically.

We have constructed computational three‐dimensional geometry of ITER antenna modules using CAD information and component parametric descriptions. These are used in time‐domain finite‐difference simulations, done with the VORPAL software [1], to look at both vacuum and plasma loaded conditions. Simulations of a single module are presented, and progress on efforts to move to full array‐sized simulations is reported on. We diagnose field profiles, surface field strengths, and impedance. Plasma loading is provided by a cold plasma model [2], which was designed to treat linear magnetized plasma response. This model easily treats non‐axi‐symmetric equilibrium, and we look at the effect of local high‐density structures in front of the antenna. We also discuss progress in including non‐linear effects in the edge plasma model, including time‐dependant density, and sheath effects.

Superconducting radio frequency (SRF) cavities are a popular choice among researchers designing new accelerators because of the reduced power losses due to surface resistance. However, SRF cavities still have unresolved problems, including the loss of power to stray electrons. Sources of these electrons are field emission from the walls and ionization of background gas, but the predominant source is secondary emission yield (SEY) from electron impact. When the electron motion is in resonance with the cavity fields the electrons strike the cavity surface repeatedly creating a resonant build up of electrons referred to as multipacting. Cavity shaping has successfully reduced multipacting for cavities used in very high energy accelerators. However, multipacting is still a concern for the cavity power couplers, where shaping is not possible, and for cavities used to accelerate particles at moderate velocities. This Phase II project built upon existing models in the VORPAL simulation framework to allow for simulations of multipacting behavior in SRF cavities and their associated structures. The technical work involved allowed existing models of secondary electron generation to work with the complex boundary conditions needed to model the cavity structures. The types of data produced by VORPAL were also expanded to include data commonmore » used by cavity designers to evaluate cavity performance. Post-processing tools were also modified to provide information directly related to the conditions that produce multipacting. These new methods were demonstrated by running simulations of a cavity design being developed by researchers at Jefferson National Laboratory to attempt to identify the multipacting that would be an issue for the cavity design being considered. These simulations demonstrate that VORPAL now has the capabilities to assist researchers working with SRF cavities to understand and identify possible multipacting issues with their cavity designs.« less

The combination of a pterin and a transition metal in many enzymes is the motivation for exploring the chemistry of pteridine complexes in detail. Unlike other biological ligands for essential transition metals, pterin is unique in displaying multi-electron redox reactivity, an ability that resembles the redox capabilities of transition metals. It is perhaps because these two partners, metal and pterin, have this chemical similarity that their compounds defy traditional categorization by formal oxidation number. The result challenges the chemist to formulate fresh interpretations of these deceptively ordinary complexes. This review concerns reports of metal-pterin complexes that appeared from the early 1980s through 1996. In a few cases older literature is briefly mentioned to build a context for the newer work. The review comprises four sections. Section 1 introduces the pteridine family and its important contributions to biochemistry. Section 2 is devoted to studies of molybdenum (6+) complexes reacted with reduced pterins. Section 3 describes redox interactions between reduced pterins and the first row metals copper and iron. Finally, Section 4 turns to a discussion of the electronic interactions in flavin complexes of various metals. An Epilogue closes the review.

VORPAL is a simulation framework which can model electromagnetic fields and charged particle dynamics in RF structures such as those encountered in high gradient accelerating cavities. Multiple models exist in VORPAL to determine the secondary electron yield (SEY) for an impact of a primary electron on the cavity surface. Combining these models allows VORPAL to be used to study electron multipacting in radio frequency cavity structures. VORPAL also supports the tracking of individual multipacting trajectories as well as recording the impact history of multipacting electrons. We present simulation results of multipacting simulations in various RF structures including accelerating cavities as well as coupler and waveguide structures. Benchmarking results are given comparing VORPAL results from other codes as well as experimental data.

This work introduces a conformal finite difference time domain (CFDTD) method to accurately determine the dispersion relation of an A6 relativistic magnetron. The accuracy is measured by comparing with accurate SUPERFISH calculations based on finite element method. The results show that an accuracy of 99.4% can be achieved by using only 10,000 mesh points with Dey-Mittra algorithm. By comparison, a mesh number of 360,000 is needed to preserve 99% accuracy using conventional FDTD method. This suggests one can efficiently and accurately study the hot tests of microwave tubes using CFDTD particle-in-cell method instead of conventional FDTD one.

This work introduces a conformal finite difference time domain (CFDTD) method to accurately determine the dispersion relation of an A6 relativistic magnetron. The accuracy is measured by comparing with accurate SUPERFISH calculations based on finite element method. The results show that an accuracy of 99.4% can be achieved by using only 10,000 mesh points with Dey-Mittra algorithm. By comparison, a mesh number of 360,000 is needed to preserve 99% accuracy using conventional FDTD method.

Syntheses and physical properties of pterin and pteridine complexes for first-row transition metals are described. Characterization used single-crystal x-ray diffraction, spectroscopic, and microanalytical methods. These data all indicate that pterin and pteridine ligands chelate via oxygen and nitrogen atoms. Three copper complexes have been structurally determined. Cu(tppb)(pterin) [tppb − =tris(3-phenylpyrazolyl)hydroborate] crystallizes in the triclinic space group P1

The preparation and characterization of new model complexes for the molybdenum cofactor are reported. The new models are distinctive for the inclusion of pterin-substituted dithiolene chelates and have the formulation Tp ∗ MoX(pterin-R-dithiolene) (Tp ∗ = tris(3,5,-dimethylpyrazolyl)borate), X = O, S, R = aryl. Syntheses of Mo(4+) and (5+) complexes of two pterin-dithiolene derivatives as both oxo and sulfido compounds, and improved syntheses for pterinyl alkynes and [Et 4 N][Tp ∗ Mo IV (S)S 4 ] reagents are described. Characterization methods include electrospray ionization mass spectrometry, electrochemistry, infrared spectroscopy, electron paramagnetic resonance and magnetic circular dichroism. Cyclic voltammetry reveals that the Mo(5+/4+) reduction potential is intermediate between that for dithiolenes with electron-withdrawing substituents and simple dithiolates chelates. Electron paramagnetic resonance and magnetic circular dichroism of Mo(5+) complexes where X = O, R = aryl indicates that the molybdenum environment in the new models is electronically similar to that in Tp ∗ MoO(benzenedithiolate).

Summary form only given. Recent advances and implementations of new "cut-cell" algorithms now permits very accurate 2nd-order geometry representation in modeling tools which have the required self-consistent treatment of space-charge and mode fields, necessary for design of the high-space-charge beam-to-RF conversion subcomponents. This development significantly reduces design and modeling effort, while simultaneously improving precision. We demonstrate this new capability in the VORPAL software framework for devices which convert bunched beam kinetic energy to RF energy, including output cavity components of klystrons and IOT's in accelerator upgrade programs. In addition, we test the ability of this new modeling capability to accurately predict the energy recovery process in a superconducting rf cavity. In this latter design calculation, we require an order-or-magnitude greater precision than typically required for similar calculations of RF sources. In all situations, we also require accurate information of the induced particle spread and velocity shears on exit from the cavities. Time permitting, we also investigate the feasibility of the modeling capability to be applied to other complex-geometry aspects of RF sources, such as the gun and collectors.

Multipacting continues to be an important issue in Superconducting Radio Frequency (SRF) cavities, particularly near waveguide couplers. Most modern simulations of multipacting are not self-consistent, using the fields from a purely electromagnetic simulation to drive the motion of multipacting electrons. This approach works well for the onset on multipacting but as the electron density increases in the cavity it can have an effect on the cavity mode. Recently VORPAL has demonstrated its ability to mode the electrodynamics of SRF cavities using finite difference time domain (FDTD) algorithms coupled with cut- cell treatments of the cavity boundaries based on the Dey- Mittra method. The FDTD approach allows us to easily incorporate multipacting electrons as PIC particles in the simulations. To allow multipacting simulations to be done with EM- PIC we have been developing particle boundaries for the cut-cells. Recently we have added particle removal boundaries at the particle sinks which will correct the unphysi- cal build up of image charge at the boundaries. We have also modified the secondary electron routines in VORPAL so they are now aware of the cut-cell boundaries and use the normal to the boundary surface when emitting secondaries. Using these new particle boundary conditions we have started preliminary simulations of multipacting electrons in a SRF cavity.

The combination of a pterin and a transition metal in many enzymes is the motivation for exploring the chemistry of pteridine complexes in detail. Unlike other biological ligands for essential transition metals, pterin is unique in displaying multi-electron redox reactivity, an ability that resembles the redox capabilities of transition metals. It is perhaps because these two partners, metal and pterin, have this chemical similarity that their compounds defy traditional categorization by formal oxidation number. The result challenges the chemist to formulate fresh interpretations of these deceptively ordinary complexes. This review concerns reports of metal-pterin complexes that appeared from the early 1980s through 1996. In a few cases older literature is briefly mentioned to build a context for the newer work. The review comprises four sections. Section 1 introduces the pteridine family and its important contributions to biochemistry. Section 2 is devoted to studies of molybdenum (6+) complexes reacted with reduced pterins. Section 3 describes redox interactions between reduced pterins and the first row metals copper and iron. Finally, Section 4 turns to a discussion of the electronic interactions in flavin complexes of various metals. An Epilogue closes the review.

A laser driven wakefield accelerator has been tuned to produce high energy electron bunches with low emittance and energy spread by extending the interaction length using a plasma channel. Wakefield accelerators support gradients thousands of times those achievable in RF accelerators, but short acceleration distance, limited by diffraction, has resulted in low energy beams with 100% electron energy spread. In the present experiments on the L’OASIS laser, the relativistically intense drive pulse was guided over 10 diffraction ranges by a plasma channel. At a drive pulse power of 9 TW, electrons were trapped from the plasma and beams of percent energy spread containing > 200 pC charge above 80 MeV and with normalized emittance estimated at < 2 π-mm-mrad were produced. Data and simulations (VORPAL code) show the high quality bunch was formed when beam loading turned off injection after initial trapping, and when the particles were extracted as they dephased from the wake. Up to 4 TW was guided without trapping, potentially providing a platform for controlled injection. The plasma channel technique forms the basis of a new class of accelerators, with high gradients and high beam quality.

Considerable resources are required to run three dimensional simulations of multipacting in superconducting radio frequency cavities. Three dimensional simulations are needed to understand the possible roles of non-axisymmetric features such as the power couplers. Such simulations require the ability to run in parallel. We consider the versatile plasma simulation code VORPAL as a possible platform to study such effects. VORPAL has a general 3D domain decomposition and can run in any physical dimension. VORPAL uses the CMEE library to model the secondary emission of electrons from metal surfaces. We present a three dimensional simulation of a simple pillbox RF cavity to demonstrate the potential of VORPAL to be a major simulation tool for superconducting radio frequency (SRF) cavities.

High quality electron beams with hundreds of picoCoulombs of charge in percent energy spread above 80 MeV were produced for the first time in high gradient laser wakefield accelerators by guiding the drive laser pulse

In the Phase I project we concentrated on three technical objectives to demonstrate the feasibility of the Phase II project: (1) the development of a parallel MDSplus data handler, (2) the parallelization of existing fusion data analysis packages, and (3) the development of techniques to automatically generate parallelized code using pre-compiler directives. We summarize the results of the Phase I research for each of these objectives below. We also describe below additional accomplishments related to the development of the TaskDL and mpiDL parallelization packages.

Summary form only given. Researchers are interested in rapid switching high-current diodes for applications including accelerators, microwave sources, and flat panel displays. Gas-filled diodes, such as the pseudospark, are one class of candidates for such high-current applications. To yield a better understanding of their operation, numerical simulations of these devices must include a number of physical processes: gas ionization, secondary electron emission and ion-induced electron emission, and radiation. We are enhancing the VORPAL code to model these physical effects, and we present here electromagnetic particle-in-cell (PIC) simulation results for a generic one-dimensional diode. These simulations include models of some of the processes mentioned above, as well as the effects of the self-consistent electro-magnetic fields of the particle beams. We discuss the numerical models used, including techniques to allow users to use these models from a variety of languages, including Java and C++. We examine the case of the planar diode for a variety of conditions and discuss the several physical regimes obtained in terms of analytical models. This study provides an excellent test-bed for verification and validation of both numerical and physical models

Summary form only given. Ion acceleration via laser plasma interaction is an area of research that attracted a lot of attention over the past few years. The general experimental setup is a laser pulse hitting an over- or underdense foil target. Several mechanisms responsible for the ion acceleration have been identified, including sheath normal acceleration or shock acceleration at the foil interior. For underdense or slightly overdense targets, the ion acceleration is mainly in forward direction, parallel to the initial laser propagation. For some applications it might be of interest to accelerate ions in the backward direction. One concept to accomplish this is to send the laser pulse through an underdense, low-Z coating before it hits the overdense, high-Z target. The target is spherically shaped and acts as a focusing mirror. The incident laser intensity is chosen low enough that the ions are not accelerated via the ponderomotive force. We are currently investigating this concept using the fully kinetic plasma simulation code VORPAL and present results of these simulations

Summary form only given. A direct simulation Monte Carlo (DSMC) method is being added to VORPAL to model dilute (large Knudsen number) gases. VORPAL is a multi-dimensional, parallel particle-in-cell (PIC) plasma simulation code that supports a variety of models for both plasma and electromagnetic fields. The addition of the DSMC capabilities will allow VORPAL to run hybrid simulations of gas/plasma mixtures using the current plasma simulation capabilities of VORPAL along with the newly added DSMC capabilities. In particular, the resulting version of VORPAL can be used in various applications, including space propulsion (plasma thrusters) and plasma manufacturing. We will present a general overview of VORPAL's current capabilities along with a more detailed discussion of the DSMC model. VORPAL's DSMC simulation model was built on the existing PIC model by adding a Monte-Carlo collision model and adding the capability to do ensemble averaging. Comparisons of VORPAL simulations with the well-established DSMC code MONACO will be presented. These will include simulations of 1D high speed Couette flow, 1D thermal Couette flow, and 2D high speed flow around a flat plate

A proposed luminosity upgrade to RHIC includes a novel electron cooling section, which would use ∼55 MeV electrons to cool fully‐ionized 100 GeV/nucleon gold ions. High‐current bunched electron beams are required for the RHIC cooler, resulting in very high transverse temperatures and relatively low values for the magnetized cooling logarithm. The accuracy of analytical formulae in this regime requires careful examination. Simulations of the friction coefficient, using the VORPAL code, for single gold ions passing once through the interaction region, are compared with theoretical calculations. Charged particles are advanced using a fourth‐order Hermite predictor‐corrector algorithm. The fields in the beam frame are obtained from direct calculation of Coulomb’s law, which is more efficient than multipole‐type algorithms for less than ∼106 particles. Because the interaction time is so short, it is necessary to suppress the diffusive aspect of the ion dynamics through the careful use of positrons in the simulations.

High-quality electron beams, with a few 109 electrons within a few percent of the same energy above 80 MeV, were produced in a laser wakefield accelerator by matching the acceleration length to the length over which electrons were accelerated and outran (dephased from) the wake. A plasma channel guided the drive laser over long distances, resulting in production of the high-energy, high-quality beams. Unchanneled experiments varying the length of the target plasma indicated that the high-quality bunches are produced near the dephasing length and demonstrated that channel guiding was more stable and efficient than relativistic self-guiding. Consistent with these data, particle-in-cell simulations indicate production of high-quality electron beams when trapping of an initial bunch of electrons suppresses further injection by loading the wake. The injected electron bunch is then compressed in energy by dephasing, when the front of the bunch begins to decelerate while the tail is still accelerated.