First two pages of discussion – that can be skipped! ARC™ Data on DTBP - Background Discussion For more than 30 years the Accelerating Rate Calorimeter (ARC™) has been commercially available. Throughout this time a solution of the liquid peroxide di-tertiary butyl peroxide (DTBP) in toluene has been used as the standard sample. Over the years other chemicals have been investigated as standard but a better alternative has never been found, 20% DTBP by weight is the concentration used. DTBP in toluene solutions have been used as a standard to all alternative adiabatic and safety calorimeters. However the ‘agreed result’ (heat of reaction) is still the subject of debate and information on this and the way to test a standard sample in the ARC is regularly requested. Results obtained by many groups and shown in scientific journals have not been determined under ‘standard’ conditions and this and no doubt other aspects has led to poor data being published. The result may vary dependent upon experimental configuration, but a ‘standard result’ that is obtained under standard conditions is what is required to verify proper operation of the ARC and to have confidence in sample test data. A standard test protocol can be described (and is described here) such that standard data can be gained. The reasons why there can be variation in result is not discussed here in detail. It has been put down to instrument or experimental inconsistencies, older poorly functioning calorimeter. However this is not really the problem.... there are many aspects that need understanding: the void volume is important for two reasons, simply the void volume is a place for material condensation or reflux resulting in heat loss and also it has been proposed that decomposition partially occurs in the gas phase. The reaction has been said to be ‘gas phase’, there may be variation of reaction on differing metal surfaces; the heat transfer between bomb and calorimeter will vary depending upon sample holder shape and size and in calculating the result the ‘phi factor’ is used. The pressure line must not have additional heavy, metal parts near the calorimeter lid that would lead to heat loss. The pressure line through the calorimeter lid must be tightly fitted. The accuracy of these aspects would mean that tests at differing phi levels will calculate to a somewhat different result. It is necessary also to understand that DTBP is available commercially from many sources, typically at 98% purity. The peroxide is sold with stabilisers (tin compounds). These inhibit low temperature slow decomposition reaction, stabilising the chemical. Since differing manufacturers will use different additives different thermal data might occur. Standardising with DTBP from Sigma Aldrich seems sensible. Toluene is available at high purity (99.9%). It is important to make up the solution accurately (2 people overseeing!) and to make a large volume that is then stored in a closed container and kept in a refrigerator. Making the sample each time and weighing into the bomb is just no good! However these variables can be eliminated with clear and precise definition of experimental protocol. During tests there should be replication of conditions between the standard sample and later sample testing. With proper use of a properly functioning ARC system and using ‘standard’ condition – consistent reliable data is obtained with the heat of reaction being determined as that expected.
A DTBP/toluene ARC calorimeter test will give full time-temperature and pressure data that may be analysed for full thermodynamic and kinetic description of the decomposition. Thermodynamic values obtained simply from the self heat rate data. Many are of interest but the ‘key value’ is the heat of reaction. (This is normally represented as Joules per gram of peroxide). Also of interest are the exotherm onset temperature (which will vary with calorimeter sensitivity) and heat generation rates at selected temperatures. The kinetic analysis parameters can also be determined by fitting of the self heat rate curve. If the kinetic modelling allows for determination of the order of reaction, the order is typically 0.7 – 0.8. However most practitioners have fixed the order to 1.0. As a first order fit, the activation energy is typically in the range 155-160 kJ/mol and this does not vary if there is poor experimentation.
With conditions standardised. the aim is to get the ‘correct’ result under these conditions and to get it reproductively and then use the ARC with confidence.
DTBP Testing Protocol and Data Use the right sample 98% DTBP (from Sigma Aldrich) made to 20.00% by weight with 99.9% toluene. A sample of 200-500 ml ought to be made with two people overseeing the sample weighing. This stock sample must be kept in an air tight bottle in a refrigerator. Use the fixed experimental protocol Use 6.00+/- 0.01g of sample in a titanium bomb (as below) Experimental conditions to be used are Ts = 90°C Heat step = 3°C Wait time = 15min Onset sensitivity = 0.02°C/min Other conditions are not important Phi is 1.30-1.33 Use the ARC properly The ARC must be calibrated, The pressure line be of standard configuration and must be filled fully with oil with the pressure transducer in the usual position. A clean titanium ‘bomb’ or sample container (Ti-LCQ ~ 5.8g) should be used (this can be re-used, to give repeatability). The calorimeter set up normally with the position of ‘bomb’ thermocouple set such that the tip is lightly pressed on the outer wall of the ‘bomb’. The through tube must be standard and clean The pressure line must be standard
A 6 gram sample mass the volume is greater and the expansion upon heating is such as to fill to bomb entirely. However the key value that heat of reaction is suggested Ts be 240-245J/g. Poor experimentation with the ARC can easily get values of 210-230 J/g. With vent sizing calorimeters this may be due to poor adiabatic control and/or heat loss through the lid part of the sample holder (as this is not thermally guarded well). Values below 240J/g indicate one or more issues.... The ARC will give values near 245J/g (+/- 5J/g) do not think the ARC or any adiabatic calorimeter will give 1-2% repeatable results! Other calorimeters have reported 250-280 J/g. This might be due to over adiabatic control. Vent sizing calorimeters typically give lower values due to poorer sensitivity and poorer adiabatic control Dewar vessel carefully used can give 240-250 J/g. Other calorimeters and users have reported values up to 280-300J/g; this may result by differing holders, over-adiabatic compensation or just poor work! The CHETAH thermodynamic computer code gives a value of 240J/g In conclusion THT believe that the 20% DTBP solution can be used as standard in the ARC and there is a clear result to be aim; this being the heat of reaction near 245J/g. If this value is determined then all other aspects of the test and results obtained will be reliable.
Data reported here... 4 DTBP tests... January 2012 In this document data is reported from four tests carried out on a THT esARC system sold in early 2012. The four tests were carried out with this new system... • • •
6 grams of 20% DTBP Ti LCE Bomb 90C Start, 3C steps, 20 min wait, 0.002C/min sensitivity, 230C End (Nothing special!).
•
Properly filled pressure line, all parts clean and the unit pre-calibrated.
In the THT esARC testing there will be 2 data sets from each test. The DAT file that contains the data from the complete test (pre-exotherm, exotherm and post-exotherm) and the EXO file that contains the data only of the exotherm – are analysed with ARCCal+ the THT software. On the following pages you can scan though the ARC Data with DTBP... as a user might do with data obtained in house.
There are other aspects of ARC testing with options etc – but these are not considered. However at the end the other often requested feature is HIGH SENSITIVITY testing. There is a brief section on using the ARC at 0.002C/min and higher to get the ultimate performance of the ARC system. The ARC will produce Word, Excel or html reports – these are not shown – only screen shots.
To illustrate how the ARC is set up to do the tests and to commence the test, THT has videos. These are available on the THT YouTube page. To illustrate how the data analysis is completed, THT has videos. These are available on the THT YouTube page. To go to these click the link below
http://www.youtube.com/user/ThermalHazardTech/videos
The four tests carried out we named (as per THT protocol) ARC120006, ARC120009, ARC12011 and ARC12012. DAT FILES First let us look at the ‘dat files’. These are not used for data analysis but show the complete experiment and so illustrate the quality of the test....
Above is the ARCCal+ temp v time curve for the complete test ARC120006 Now call in the other three dat files...and their T vs t is shown below
They look the same – which is good and expected...
Adding the pressure to one (ARC12011) it can be seen the pressure is rather high (as the void volume is low)
ARCCal+ can simply merge the files...
Similar – and we have to zoom up to see that all four files are indeed shown!
What is important in adiabatic calorimetry tests is of course quality of detection of exotherm, quality of ‘baseline’ such that the exotherm can be distinguished from the zero – and that this continues at high temperature.... and of course quality of adiabatic control. These are what can be investigated with the dat file. Focussing only on ARC12009 now... Below is the data prior to exotherm detection...
Greater zooming shows that below 100C there is no reaction (well the baseline drift is no greater than 0.002C/min) but above 100C the reaction starts, an upward drift is observed but the rate is below the threshold. At the end there is a section that shows how the ARC can indeed detect reactions down to 0.001C/min! But clearly the results are of good quality, all works well! It is very unlikely that any other adiabatic calorimeter can match this performance!
The quality of the data relates to the adiabatic tracking performance of the ARC, which is recognised to be superior to any other commercial calorimeter. This can be seen in ARCCal+ plots (though offset by the calibration offsets)
Equally important is the performance at higher temperature. Many calorimeters fall down here as they allow condensation/reflux on the lid of the sample holder. The ARC performs very well as shown below
The dat file is not used for data analysis – but to show as illustrated here that the system works to specification.
EXO FILES
Exotherm data files
The inspection and study of the results and analysis of the data can be categorised into three parts: Raw Data; Converted Data, Corrected Data, Analysed Data. The four exotherm data files are illustrated here. These four files can be added as DataSets 5-8 in the same project
Here is shown the Self Heat Rate Curve (SHR Curve) for test ARC12006. This is a Raw Data plot and others can be shown (printed, exported etc). They are not shown here – the intention of this report is not to exhaustively illustrate al features of ARCCal+ (request this from THT!)
Again merging can be done and again the similarity of the data is obvious and it can be seen by zooming that the initial part of the exotherm replicates remarkably.
The aim here though is to focus on the Thermokinetic analysis of the result - as can be achieved with ARCCal+ Software. Progressing through the Analysis portion of the software, the heat of reaction is simply determined – indeed it is directly proportional to the temperature rise so can be hand calculated in seconds.
However here this table of data is shown for all fours samples....
The heat of reaction can be seen to be between 248J/g and 252J/g.
The analysis here has shown the determination of the accepted value of the heat of reaction for DTBP and by following the guide given here all THTARC user ought to get data of this quality and reliability. The results shown here are typical of a new esARC system and results generally obtained fall within the ‘accepted’ range of values. For more information contact THT.
In addition the kinetic information shown below compares with that published and accepted. In total the report here illustrates that with normal (correct) use of the ARC data of the highest quality and reliability will be obtained.
Continuing on with the analysis... It is possible to use kinetic modelling to determine the kinetic parameters. With DTBP traditionally users have fixed the order to 1- and this is done here
Finally it is possible to make Phi Corrected plots and to convert plots to show Enthalpy, Joules, (from temperature rise, ie heat) and Power, Watts, (from the self heat rate curve)
(Note the Phi Corrected graph also shows the data prior to phi correction).
Finally alongside the traditional question of how good is ARC data the other question is how sensitive can the ARC be – what is its limits of sensitivity? The ARC was designed to get a package of data to give information on the safety of a reactive sample within the time of 24 hours... but indeed the latest instrument from THT – the esARC with the standard calorimeter and high stability electronics will produce a very stable calorimeter such that reactions of very low heat release can be studied
Ultimate Sensitivity of the THT esARC In course of testing all new esARC instruments THT carry out testing to evaluate the operation of the unit. It is clearly seen with all new units that the esARC systems when calibrated normally to a temperature of 200C, will differentiate between thermal ‘drift’ and reaction when the rate is 0.001-0.002C/min. This is seen with DTBP tests. Test 1..... 3degC steps and wait time 30 mins and 0.002C/min sensitivity
At 97C and 100C the stability of the esARC is shown and any reaction is less than 0.0001C/min! The resolution of the temperature signal of 0.001C is clearly observed
At 103C there is indication of temperature rise but the rate is 0.0001C/min. At106C, the upward temperature rise is clear – the rate is 0.0009C/min, indicating the start of reaction – and the ability to detect reaction at very low rates, potentially from 0.001C/min, twenty times that of the rate typically used in adiabatic calorimeter. This is 20 or 50 times greater sensitivity of Specifications published by most other manufactures of adiabatic calorimeters and five times more sensitive that that ever proposed as the Sensitivity by any other manufacturer. At 109C reaction is detected above the onset specified, 0.002C/min and the ARC goes into Exotherm mode.
Now with 1C steps....
If the temperature change is measured from the graph of from the data similar results are given Temp 100 101 102 103 104 105 106 107
Measured Rate (C/min) 0.0000 0.0000 0.0000 0.00002 0.00008 0.00035 0.0010 0.0019*
Calculated Rate (C/min) 0.000 0.000 0.000 -0.001 0.000 0.000 0.001 0.001* * Exotherm
The conclusion is clear – the THTesARC will work at sensitivities far above that specified and that obtainable with other instruments.
Š Thermal Hazard Technology 2012 All rights reserved.
HEAD OFFICE 1 North House, Bond Avenue, Bletchley, MK1 1SW, England. Tel: +44 1908 646800 Fax: +44 1908 645209 Email: info@thtuk.com Web: www.thtuk.com US OFFICE Tel: +1 317 222 1904 Fax: +1 317 660 2092 Email: info@thtusa.com Web: www.thtusa.com CHINA OFFICE Tel: +86 21 5836 2582 Fax: +86 21 5836 2581 Email: info@thtchina.com Web: www.thtchina.com INDIA OFFICE Tel: +91 9560 655 626 Email: info@thtindia.com Web: www.thtindia.com
Offices in ENGLAND, USA, CHINA, INDIA Representation Worldwide Offices in ENGLAND, USA, CHINA; Representation Worldwide