Monthly Archives: May 2016

Renewable Energy and Electric Vehicles

“Why do we not have electric vehicles yet? A 2006 documentary film explores the creation, limited commercialization, and subsequent destruction of the electric vehicle in the United States (specifically the General Motors EV1 of the mid 1990s). The film explores the roles of automobile manufacturers, the oil industry, the US government, the Californian government, batteries, hydrogen vehicles, and consumers in limiting the development and adoption of this technology. Why this recalcitrance on the part of our political leaders? One word describes it all, “”GREED””.

Wikipedia describes the film as dealing with the history of the electric car, its development, and commercialization. The film focuses primarily on the General Motors EV1, which was made available for lease mainly in Southern California, after the California Air Resources Board passed the ZEV mandate in 1990. Also discussed are the implications of the events depicted for air pollution, environmentalism, Middle East politics, and global warming.

The film details the California Air Resources Board’s reversal of the mandate after suits from automobile manufacturers, the oil industry, and the George W. Bush administration. It points out that Bush’s chief influencers, Dick Cheney, Condoleezza Rice, and Andrew Card, are all former executives and board members of oil and auto companies. The EV1 was eliminated from the GM Line in 1999 but now they are back with the “”Volt”” portraying themselves now as the saviors of the country.

A large part of the film details GM’s efforts to demonstrate to California that there was no demand for their product, and then to terminate the leases on every EV1 and dispose of them. A few were disabled and given to museums and universities, but almost all were found to have been crushed; GM never responded to the EV drivers’ offer to pay the residual lease value ($1.9 million was offered for the remaining 78 cars in Burbank before they were crushed). Several activists, including actress Alexandra Paul, are shown being arrested in the protest that attempted to block the GM car carriers taking the remaining EV1s off to be crushed.

The film explores some of the reasons that the auto and oil industries worked to kill off the electric car. Wally Rippel is shown explaining that the oil companies were afraid of losing out on trillions of dollars in potential profit from their transportation fuel monopoly over the coming decades, while the auto companies were afraid of losses over the next six months of EV production. Others explained the killing differently. GM spokesman Dave Barthmuss argued it was lack of consumer interest due to the maximum range of 80–100 miles per charge, and the relatively high price.

The film also showed the failed attempts by electric car enthusiasts trying to combat the cancellation of EV1 and the surviving vehicles. Towards the end of the film, a deactivated EV1 car #99 was found in the garage of Petersen Automotive Museum, with former EV sales representative, Chelsea Sexton, invited for a visit.

This was a complicity involving “”leaders”” in the highest levels of corporate and political America and again motivated by pure “”GREED”” and plays directly into the hands of our nation’s enemies who wish to see us crippled economically.  ~ Theodore Sheldon SUmrall

Dissertation on Insensitive Highly Energetic Materials (Section 2.4)

Article Summary: This article contains Section: 2.4 (Ingredient Selection and Theory) of the “Dissertation on the Study of Insensitive Highly Energetic Materials” as part of the Doctoral requirements for Theodore S. Sumrall at The University of Tokyo, March of 1998. Theodore S. Sumrall was awarded a Doctorate Degree from the Department of Chemical Systems Engineering in April of 1998 as a result of his research, development testing and dissertation presentation.

2.4 Ingredient Selection and Theory

The overall logic for selection of raw materials is detailed in Figure 2.4-1. Once ingredients were selected, scale up to 450 gram mixes occurred. Ingredients were selected to help ensure that project objectives (Table 1-3) were achieved.

Binder Screening and Selection

Theoretical calculations for non AP containing PBX compositions which utilize an HTPB binder system revealed that significant amounts of un-reacted carbon was being generated in both the burn and detonation reactions. The code predicted that, while the majority of the aluminum was oxidized to Al2O3, the majority of the binder (a hydrocarbon) was un-oxidized.

Theoretical Calculation Logic Flow

Figure 2.3-1 Theoretical Calculation Logic Flow

It was postulated, and supported by thermochemical code output, that replacement of the non-oxygenated HTPB hydrocarbon binder with an oxygenated binder might allow the oxygen to enter into the detonation and/or deflagration reaction. This approach, if successful, would yield a dual benefit. First, oxygen from the binder could be utilized to burn the residual hydrocarbon and secondly, the more oxygenated (and more sensitive) ingredients (such as RDX) could be reduced in content. This would hopefully allow the PBX to pass tests which PBX-109 fails, namely Slow Cook-Off (SCO), Fragment Impact (FI) and Sympathetic Detonation (SD). In the process of oxygenated binder evaluation, binders were considered that not only had relatively high oxygen content, but which would also adequately wet the solids (to ensure low viscosity (Objective #4)) and which had a demonstrated capability of rapid cure (Objective #5). Both of these characteristics would be desirable from production cost standpoints.

Finally, only binder ingredients which were commercially available in large quantities and at a relatively low cost were chosen (Objective #8).

After a thorough evaluation of a number of potential binder candidates, three oxygenated, curable binders were selected for further evaluation, polyethyleneglycol (PEG), polypropyleneglycol (PPG), and ethyleneoxide-propyleneoxide (EOPO). The plasticizer chosen was triacetin (TA). In order to meet rapid cure goals, two types of cure catalysts were chosen for evaluation. The first cure catalyst evaluated was ferric-acetylacetonate (FeAA). The second cure catalyst evaluated was di-butyltin di-laurate (DBTDL) in conjunction with a cure delay/cure stimulation catalytic system consisting of minute concentrations of triphenyl bismuth (TPB) and maleic anhydride (MA) .

A comparison of the theoretical improvement which was predicted to occur by switching from an HTPB binder to a PPG binder for a well characterized explosive (PBX-109) is detailed in Table 2.4-1.

Table 2.4-1

Comparison of Theoretical Performance Improvement with Oxygenated Binders


HTPB Binder

PPG Binder

Density (g/cm3)



Detonation Velocity (m/sec)



Detonation Pressure (MPa)



Percent Un-reacted Carbon



Temperature (K)



As shown later, PPG was down selected as the final polymer of choice due to superior processing, curing, and physical property characteristics.

Oxidizer Screening and Selection

Theoretical calculations showed a direct correlation between energetic material density and oxygen content with detonation velocity, detonation pressure and Impulse Density. It was predicted by the TIGER and NASA/Lewis codes that the inclusion of oxidizers such as Ammonium Nitrate (AN), Ammonium Perchlorate (AP), Potassium Nitrate (KN), etc., would improve detonation velocity, detonation pressure and blast pressure impulse.

The addition of oxidizers such as: Ammonium Nitrate (AN); Ammonium Perchlorate (AP); Potassium Nitrate (KN); and etc., have been proven beneficial to the more efficient combustion of fuels. A higher degree of fuel combustion will result in higher temperatures and therefore higher blast pressures. AP has long been the oxidizer of choice for solid rocket propulsion such as the Space Shuttle SRB and NASDA H-II Boosters. However, at the time that this project was initiated, AP availability had decreased (and cost had increased) due to an incident at one of only two major AP producers in the US. Also testing by other researchers revealed a correlation between AP content and Slow Cook-Off (SCO) test failure.

Criteria other than cost and availability which guided oxidizer selection were: high oxygen balance; non-hygroscopic character, and high-moderate density. Potassium Nitrate (KN), for example, met all of the screening criteria. At a crystal density of 2.1 g/cm3, and despite having 67% condensed products, KN reacts to form one half more mole of free O2 than AP as indicated by the following equations.

2KNO3 à K2O + N2 + 2 1/2 O2 Equation 2.4-1

2NH4ClO4 à N2 + 3H2O + 2HCl +2½ O2 Equation 2.4-2

Lead Nitrate (PbN) and Barium Nitrate (BaN) also have high densities (4.53 g/cm3 and 3.24 g/cm3 respectively) are non hygroscopic, and react to form three moles and 2.5 moles of excess O2 respectively according to Equations 2.4-3 and 2.4-4.

Pb(NO3)2 à 2Pb + N2 + 3O2 Equation 2.4-3

Ba(NO3)2à BaO + N2 + 2 1/2 O2 Equation 2.4-4

KN and PbN were selected as the oxidizers of choice during this phase of research. KN was eventually chosen over PbN due to superior sensitivity and environmental characteristics.

Molecular Explosive Screening and Selection

To achieve performance goals, it was determined that at least some molecular explosives would be required. Molecular explosives are defined as explosives which have the fuel and oxidizer segments linked via chemical bond. Common examples of molecular explosives are TNT, RDX, and HMX. The chemical structure of these three molecular explosives is found in Annex-A. Molecular explosives are therefore unlike other types of explosives, such as emulsion explosives or composite explosives where, although the fuel and oxidizer are in relatively close proximity to each other, they are not linked via chemical bond. Molecular explosive type was screened for incorporation into the insensitive energetic design matrix as outlined in Figure 2.3-1.

As a result of this analysis, from theoretical insensitivity, performance, and cost standpoints, it was determined that the only well characterized, economically attractive, insensitive, high performance explosive which was readily available at the time was nitroguanidine (NQ). Other low sensitivity, high performance explosives existed, but their cost and availability were not comparable to NQ. Tri-Amino Tri-Nitro Benzene (TATB), for example, while highly insensitive has a cost of » ¥7700/kg.

NQ Availability and Selection Rational

Four crystalline configurations of NQ were available at the time this research was being conducted from domestic and international producers. Figure 2.4-2 details the production differences of these four crystalline types. Low Bulk Density Nitroguanidine (LBDNQ), has a very high length to diameter (L/D) ratio. The typical diameter is approximately 5m however, the length can exceed 100m (Figure 2.4-3). The LBDNQ is very fibrous with a consistency of cotton and occasionally, the needles are hollow. The bulk density of the LBDNQ is » 0.17g/cm3. The LBDNQ, although inexpensive, is processable only in small quantities (<6%). Additionally, due to the potential of entrained air in the hollow needles, the potential for “hot spot” formation exists which could make the energetic material quite sensitive.

By dissolving LBDNQ in a solution of water and methyl cellulose, followed by re-crystallization, a much larger (»150m – 300m) and more processable crystalline form of NQ is produced. This material is referred to as “Cubical Nitroguanidine” (CNQ) or “Un-pulverized Nitroguanidine” with a bulk density of » 0.9g/c m3. (Figure 2.4-4)

CNQ is subsequently pulverized to a particle size of » 40m-120m and after this process is termed “High Bulk Density Nitroguanidine” (HBDNQ) or “Pulverized Nitroguanidine” (PNQ). This process increases the bulk density of the NQ to a bulk density of »0.4g/cm3 for HBDNQ. Depending on the manufacturer, the particle sizes of HBDNQ vary.

Figure 2.4-5 shows PNQ which was ground in plant and Figure 2.4-6 show PNQ which was ground at a facility in Md.

The fourth crystalline form of NQ is termed “Spherical Nitroguanidine” (SNQ) (Figure 2.4-7). This material is manufactured in a manner similar to CNQ, however, the re-crystallization solvent is an organic solvent rather than water.

Theodore Sheldon Sumrall

Dissertation on Insensitive Highly Energetic Materials (Sections 2.3)

Article Summary: This article contains Section 2.3 (Theoretical Calculations) of the “Dissertation on the Study of Insensitive Highly Energetic Materials” as part of the Doctoral requirements for Theodore S. Sumrall at The University of Tokyo, March of 1998. Theodore S. Sumrall was awarded a Doctorate Degree from the Department of Chemical Systems Engineering in April of 1998 as a result of his research, development testing and dissertation presentation.

2.3 Theoretical Calculations

Prior to any small scale mixing, a theoretical evaluation of a proposed composition, (from safety, performance and cost standpoints) permitted the elimination of a large number of ingredient candidates.

A number of thermochemical computer codes were evaluated to determine which codes would be considered the most accurate for this research project. Additionally, other considerations (such as cost and proposed raw material availability) were taken into consideration. Also, the predicted rheological characteristics of the proposed composition were calculated, based upon past experiences with these materials, and a geometric analysis code. A code was written for the purpose of permitting maximum solids packing so that appropriate particle sizes would be chosen for the purpose of achieving maximum density and minimum viscosity. The upper viscosity limit goal was 2kP as measured by the Brookfield viscometer. This value was required to meet the goal of TNT processing equipment compatibility (Objective #4). TNT processing equipment has very low shear capability and therefore, a very thin composition was required in order to meet these processing goals.

A number of thermochemical codes were evaluated to determine which codes would be considered the most accurate for this research project. At the time that this research was underway, the most reliable codes were the NASA/Lewis [4,5,6] burn code and the TIGER [7] detonation code. A burn code as well as a detonation code were chosen for this effort, because in an aluminized composition both detonation and deflagration reactions occur. First, a detonation occurs through the composition where in the molecular explosive (i.e. TNT, RDX, etc.) detonates. The detonation reaction heats and disperses the other non-molecular explosive ingredients (i.e. Al and AP). The effect is essentially an artificial “fuel/air” explosive where the Al is the fuel and the AP provides the oxygen to burn the Al powder. Within the blast zone, little atmospheric oxygen is present, because the blast wave pushes the air away from the reaction site. However, due to the presence of oxygen generated by AP, there is sufficient oxygen available for the super-heated aluminum powder to react and undergo a rapid deflagration or explosion. This imparts tremendous heaving properties to the explosive and results in greater rock breakage.

First the NASA/Lewis code was employed to determine the Theoretical Maximum Density (TMD) for the formulation. The input information required to run the NASA/Lewis code are: empirical formula; densities; ingredient heats of formation; and weight percentages of each ingredient. The first NASA/Lewis run is calculated at a pressure of 1000 psi. Additional information of interest from the NASA/Lewis program is a value known as the Impulse Density or ID. The ID is the product of the specific impulse (Isp) and the density. The data of primary interest from NASA/Lewis is a factor known as “Impulse Density” (ID). Impulse density is calculated by multiplying the composition’s Specific Impulse (Isp) by the density. The specific impulse a measure of force per unit weight of explosive consumed in a burn reaction (since the post detonation reaction is essentially a burning reaction) and is a measure of the theoretical efficiency of the explosive’s blast potential. ID is therefore the product of Isp times density. The ID therefore is a relative measurement of the theoretical blast impulse potential of an explosive. While calculation of the ID at pressures predicted by the TIGER code would have been more realistic, the NASA/Lewis code was not capable of consistently generating ID values at pressures representative of a detonation. However, relative rankings were possible even at the lower pressures.

Once the overall formulation TMD is determined, then this is input into the TIGER code. Information needed to run the TIGER code is essentially the same as for the NASA/Lewis code. Once the TIGER output is obtained, then the detonation pressures and velocities (along with the ID from NASA/Lewis) are input into a three dimensional statistical optimization program (called MATLAB [8]) to optimize performance as a function of ingredient content (within processing limitations).

The overall process for these theoretical calculations is detailed in Figure 2.3-1. After conducting theoretical calculations, the literature was reviewed once again to help ensure that initial theoretical calculations were within realistic parameters.

A number of theoretical calculations were conducted to determine the effect on performance (Detonation Velocity, Detonation Pressure, Impulse Density) when the HTPB non-oxygenated binder was replaced with an oxygenated binder as well as the effects of replacing the nitramine (RDX) with a perchlorate salt (AP). The baseline for this study was TP-H8299. Additionally, the effect on aluminum addition to these performance characteristics was also calculated.

Figure 2.3-2 shows the predicted improvement anticipated in the detonation velocity of TP-H8299 as HTPB based binder is replaced with a PEG based binder at various levels of RDX (replacing AP). An approximately 5% performance improvement is anticipated by virtue of binder replacement alone.

Figure 2.3-3 shows the predicted change in detonation velocity as RDX is replaced with AP for both HTPB and PEG based versions of TP-H8299 at various concentrations of aluminum. While at lower levels of aluminum concentration, the replacement of RDX with AP appears to have a negative effect on detonation velocity, as aluminum concentration increases, this gap narrows. Indeed, it was later calculated that 20% aluminum is about the ideal concentration. Also of note, as aluminum concentration increases, detonation pressure decreases.

Figure 2.3-4 shows the predicted change in detonation pressure as RDX is replaced with AP for both HTPB and PEG based versions of TP-H8299 at various concentrations of aluminum. Again, at lower concentrations of aluminum, the replacement of RDX with AP appears to have a negative effect on detonation pressure until a concentration of approximately 20% aluminum is reached where the points converge.

Figure 2.3-5 shows the predicted improvement in Impulse Density as a function of aluminum content, binder type and AP/RDX Ratio. Generally, as aluminum concentration increases, Impulse Density increases. Of major note is that the composition containing a 50/50 ratio of AP and RDX (with a PEG binder and at a level of 20% aluminum) had the highest ID of all other compositions, including the baseline.

These theoretical calculations substantiated the belief that replacement of the non-oxygenated binder with an oxygenated version would improve performance.

Additionally, the belief that replacement of RDX with AP could be accomplished while actually improving performance.

~ Theodore Sheldon Sumrall

Dissertation on Insensitive Highly Energetic Materials (Sections 2.1-2.2)

Article Summary: This article contains the heading for Chapter-2 (Plastic Bonded Insensitive Explosive Development) and Sections: 2.1 (Introduction); and 2.2 (Mechanistic Discussion) of the “Dissertation on the Study of Insensitive Highly Energetic Materials” as part of the Doctoral requirements for Theodore S. Sumrall at The University of Tokyo, March of 1998. Theodore S. Sumrall was awarded a Doctorate Degree from the Department of Chemical Systems Engineering in April of 1998 as a result of his research, development testing and dissertation presentation.


2.1        Introduction

2.1.1        Background

As previously stated, prior researchers believed that the physical properties associated with explosives (such as TNT) help contribute to a poor response to impact stimulation.  TNT is a very hard material with little strain capability.  Stress values for TNT rarely exceed 700 kPa prior to breakage in a JANNAF dog-bone specimen.  Plastic Explosives were developed by the British approximately 50 years ago [1].  It was theorized that by blending sensitive molecular explosives (such as PETN or RDX) with a soft, hydrocarbon binder, the molecular explosives might become less susceptible to accidental initiation by what would be considered normal activities such as dropping the energetic material.  Plastic Explosives, therefore, have physical properties on the opposite end of the spectrum relative to TNT.  While TNT has no strain capability and good stress capability, Plastic Explosives have no stress capability and excellent elasticity (high strain) capability.  These Plastic Explosives were very pliable and could be easily deformed by hand.

However, for explosives which required both stress and strain capability (such as shape charges for perforating oil well casing to start oil flow), it was postulated that manufacturing an explosive composition with physical properties similar to solid rocket propellant would allow the energetic material to absorb and evenly distribute shock input (due to impact) while maintaining sufficient physical properties for specific needed applications.  These compositions were referred to as “Plastic Bonded Explosives (PBX) because the binder material typically underwent a catalyzed reaction which caused one end of the binder molecule (i.e. HTPB) to react and bond with an adjacent binder molecule.  Unfortunately, while PBXs absorb shock at lower levels of shock input, at higher shock levels (such as the adjacent detonation of another energetic material as far as 15 meters distant) the binder becomes in-compressible and assumes properties similar once more to TNT and the advantages associated with a soft, pliable binder are minimized.  Therefore, PBX-109 for example, fails the sympathetic detonation test.  Additionally, since HTPB is actually a rocket fuel (and is therefore designed to burn) PBX-109 fails the Slow Cook-Off test.  Typical PBX compositions also suffer from a number of other inadequacies.  Due to the low density and inert nature of HTPB, an associated loss of explosive performance is observed.  Also, due to the much higher viscosities associated with typical PBX compositions they cannot be manufactured in traditional TNT type equipment (Figure 2.1) and it becomes necessary for PBX compositions to be manufactured in expensive solid rocket propellant manufacturing equipment.  Also, by virtue of the cure chemistry involved (to cross-link the HTPB polymer), a minimum of two days oven cure time is typically required.  Significant additional expenditure in terms of processing and curing equipment cause many commercial explosive manufacturers to avoid manufacturing PBX compositions in spite of the improved safety associated with these compositions.

25kg Explosives Melt Kettle

Figure 2.1 25kg Explosives Melt Kettle

2.1.2        Originality

The original aspects of this phase of research was as follows:

  1. This was the first PBX developed (to the researcher’s knowledge*) which incorporated both an oxygenated binder and a solid oxidizer. Theoretical calculations indicated that the replacement of non-oxygenated binder (HTPB based) with oxygenated binder (PPG and PEG based with TA or EOPO plasticizer) and replacement of nitramine with less sensitive oxygen containing salts (such as KN) would result in a decreased sensitivity without loss of performance (Objectives #1 and #2).

  1. This was the first PBX developed (to the researcher’s knowledge*) which was processable in steam kettles (<2kP viscosity (Objective #4)).

  1. This was the first PBX developed (to the researcher’s knowledge*) which was cured within 24 hours at moderate temperatures (Objective #5).

  1. The PBX compositions developed during this phase of research were among the lowest cost compositions available (to the researcher’s knowledge (Objective #10)).

* Based upon an extensive literature search.

2.2 Mechanistic Discussion

A number of PBX formulations have been developed in the US and abroad. The most widely produced General Purpose (GP) aluminized formulation is termed “PBX-109”. PBX-109 is a much improved insensitive energetic material relative to aluminized TNT based formulas. PBX-109 has fragment velocity and air blast characteristics comparable to the very high performance TNT based explosive (H-6), however, PBX-109 has much improved sensitivity characteristics relative to H-6. The formulational characteristics of PBX-109 and H-6 are detailed in Table 2.2.

Table 2.2 H-6 and PBX-109 Formulation Characteristics







TNT/Wax Binder



HTPB Based Binder



Aluminum Powder






PBX-109 passes both Fast Cook-Off (FCO) and Projectile Impact (PI) tests where H-6 fails both of these tests. It is theorized that because PBX-109 does not contain any oxygen generating inorganic salts (such as TNT or AP), PBX-109 results in a much milder and cooler burn. Thus when PBX-109 ignites, the energy is released over a much longer period of time. Also, due to the softer nature of the HTPB based binder, PBX-109 is able to absorb the shock associated with a projectile impact and allow this energy to dissipate over a larger area without detonating. Even if the PBX-109 ignites, again, due to the lack of oxygen containing inorganic salts, the burn is relatively mild [2]. The FCO and PI tests are described in more detail in Chapter-4.

However, despite this reduction in sensitivity to both mechanical and thermal shock, PBX-109 is still quite sensitive (>200 Cards) and is subject to accidental detonation due to the high concentration of nitramine (RDX). In fact, PBX-109 is only slightly less sensitive than H-6 with regards to card gap testing at 3.7cm.

It was postulated, and supported by thermochemical code output, that replacement of the non-oxygenated HTPB hydrocarbon binder with an oxygenated binder might allow the oxygen to enter into the detonation and/or blast reaction. This approach, if successful, would yield a dual benefit. First, oxygen from an oxygenated binder could be utilized to burn the residual hydrocarbon (which normally does not react in HTPB based PBXs) and secondly, the more sensitive oxygen containing ingredients (such as RDX) could be reduced in content, hopefully, allowing the PBX to pass tests which PBX-109 failed, namely Slow Cook-Off (SCO), Fragment Impact (FI) and Sympathetic Detonation (SD).

Also, additional oxygen could be incorporated into a PBX composition by addition of nitrate or perchlorate salts such as potassium nitrate (KN) or ammonium perchlorate (AP). These salts are typically much less sensitive than nitramines such as RDX. Typically the more energetic AP will not detonate at particle diameters below 10m except at very large diameters. Therefore, the replacement of additional nitramine with oxygenated salts should still permit oxygen to become available for the detonation/combustion reaction while decreasing sensitivity to shock. Also, this would result in a composition cost decrease due to the fact that nitrate and perchlorate salts are less expensive than nitramines.

Also State of the Art (SOTA) PBX formulas were not steam kettle processable. A reformulation (with a solids redistribution) would permit a viscosity reduction and not only a raw material cost reduction but also a tremendous cost reduction in processing costs by virtue of permitting steam kettle processing. Also utilization of more reactive cure agents and catalysts would permit traditional cure times to be decreased  50%.

While the addition of nitrate or perchlorate salts would cause a composition to become more sensitive to an adverse slow cook-off type situation, other ingredients would be added to passivate this response. The ingredient of choice would be nitroguanidine (NQ) which has not only the benefit of severely depressing a burning type reaction but also is a very insensitive high explosive itself, but not sufficiently powerful by itself and also to difficult to initiate.

Figure 2.2 illustrates the mechanics and logic behind this approach.

With the addition of nitrate or perchlorate salts to PBX-109, the composition would now begin to become more characteristic of a solid rocket propellant, similar to that used for large boosters such as the propellant utilized in the Castor IV-A booster motors flown on the Delta-II rockets [3]. Therefore, trade studies were conducted with the Castor IV-A propellant (TP-H8299) as a baseline since TP-H8299 was classified as a Class 1.3 material (i.e. will not detonate above 70 cards) while PBX-109 was a Class 1.1 material.

~ Theodore Sheldon Sumrall