THERM: a computer code for estimating thermodynamic properties for

Aug 1, 1991 - THERM: a computer code for estimating thermodynamic properties for species important to combustion and reaction modeling. Edward R. Ritt...
8 downloads 13 Views 991KB Size
400

J. Chem. If. Comput. Sci. 1991, 31, 400-408

THERM: A Computer Code for Estimating Thermodynamic Properties for Species Important to Combustion and Reaction Modeling EDWARD R. RITTER’ Department of Chemical Engineering, Chemistry, and Environmental Science, New Jersey Institute of Technology, Newark, New Jersey 07 102 Received January 14, 1991 A computer package has been developed called THERM, an acronym for THermodynamic property Estimation for Radicals and Molecules. THERM is a versatile computer code designed to automate the estimation of ideal gas phase thermodynamic properties for radicals and molecules important to combustion and reaction-modeling studies. Thermodynamic properties calculated include heat of formation and entropies at 298 K and heat capacities from 300 to 1500 K. Heat capacity estimates are then extrapolated to above 5000 K, and NASA format polynomial thermodynamic property representations valid from 298 to 5000 K are generated. This code is written in Microsoft Fortran version 5.0 for use on machines running under MSDOS. THERM uses group additivity principles of Benson and current best values for bond strengths, changes in entropy, and loss of vibrational degrees of freedom to estimate properties for radical species from parent molecules. This ensemble of computer programs can be used to input literature data, estimate data when not available, and review, update, and revise entries to reflect improvements and modifications to the group contribution and bond dissociation databases. All input and output files are ASCII so that they can be easily edited, updated, or expanded. In addition, heats of reaction, entropy changes, Gibbs free-energy changes, and equilibrium constants can be calculated as functions of temperature from a NASA format polynomial database.

INTRODUCTION Detailed reaction mechanisms are widely used in simulations of processes such as combustion and chemical vapor deposition (CVD). Such models, based upon fundamental thermodynamic and kinetic principles, can offer insight into the controlling chemistry of such complex phenomena. In developing detailed reaction mechanisms, researchers must supply accurate thermodynamic data over a wide temperature range for all stable and radical species considered in a mechanism. The THERM computer code was developed to aid in estimating this data. THERM relies upon the principles of group additivity as developed by Benson and co-workers1 to estimate thermodynamic properties for species where literature data are not available. Benson’s group additivity is referred to as a second-order estimation technique2 since it incorporates non next nearest neighbor corrections and steric effects. Less sophisticated estimation techniques include first-order or bond additivity methods and zero-order or atomic contribution methods. It should be noted that with the exception of a few properties such as the molecular weight, zero-order atomic contributions are of little value. First-order methods work well for simple molecules such as normal hydrocarbons. These, however, are less accurate for more complex molecules. Second-order group contribution techniques incorporate important corrections for cyclization, gauche interactions, steric effects, repulsive and attractive effects for aromatic substituents, etc. In principle, there is no limit to the number of interaction groups which can be included, nor to the accuracy which can be obtained when these effects are taken into consideration. There are two limitations to this approach, however. First, there are only limited thermodynamic data available to determine the interaction contributions. Secondly, one must recognize the interactions of importance a priori, or resulting estimates will be less accurate than anticipated. Nevertheless, this method of group contributions has been embraced as the best all-around method for estimating ideal gas thermodynamic proper tie^.^-^ THERM includes groups for hydrocarbons, oxygen, nitrogen, halogen, and sulfur-containing species. Ring correction ‘Present address: Departmentof Chemical Engineering,Villanova University, Villanova, PA 19085. 0095-2338/91/1631-0400$02.50/0

and other interaction groups are also included to correct for such things as optical isomers. One aspect that distinguishes THERM from other computer implementations of group additivity is the bond dissociation group data file. These data are included for the estimation of radical species important to high-temperature reaction-modeling and combustion studies. Groups include bond dissociation energy, entropy change, and heat capacity change reflecting loss of an H atom from a parent stable molecule. have been developed to use Benson’s Other computer group additivity method. Most notable is the CHETAH program distributed by the ASTM.*,’,* CHETAH is designed specifically for use in chemical reaction hazard assessment, which is not addressed by THERM. THERM was designed specifically to automate several routine tasks encountered in reactionmodeling studies. THERM outputs data directly in a format that can be used with such industry-standard modeling codes as CHEMKIN,9 HTC,l0 and STANJAN.” In addition, THERM automates the update of molecules to reflect changes in group contributions. It should be emphasized that some group contributions have been estimated based upon other group contributions due to the lack of available thermodynamic property data. It is desirable to easily recalculate thermodynamic property estimafes to allow new data to be incorporated as it becomes available. This is especially true for bond-dissociation energies, where estimates and measurements are continually being refined. Group Contribution Database. Group contributions were derived by Benson and co-workers and various other res e a r c h e r ~ ’ * ~by * ’ breaking ~ - ~ ~ similar groups of molecules with known thermodynamic properties into their constituent groups and then performing multivariable linear regression to find group contributions that gave the best fit to available experimental data. Sources for the group database are presented elsewhere.5 Group data files contain group identification (name) followed by contributions to AHrO and So at 298 K and Cp’s at 300, 400,500,600,800,1O00, and I500 K (if available). This data is arranged in a tabular format, with a reference included for all groups not directly from Benson. The reference information is on the same line as the group data, following the C,,entry for 1500 K. THERM allows an on-line review of the group database including references entries. 0 1991 American Chemical Society

J . Chem. If. Comput. Sci.. Vol. 31, No. 3, 1991 401

THERM Program F l o w c h a r t Contrlbutlon

M o I . c u ~ ~ / R ~ ~ I c ~ I dm8crl~tion

9 THERM

L I S T FILE

Reaction Descr

l o tion

c 1nDut

to

CHEMKIN

Tharmodynemlc Analysis

tor

RaactIons

Figure 1. Functional diagram for the THERM ensemble of programs. Emphasized boxes are used to distinguish programs from file input and output. THERM consists of four programs: THERM, THERMFIT, THERMLST, and THERMRXN.

The group database is subdivided into seven files, currently containing more than 300 entries. A user may choose to change the number of files or the file names by changing information in the configuration file, THERM.CFG. When THERM is running, all group data, including reference information, are memory resident. As a result, searches for group information and calculation of properties are nearly instantaneous. This large memory requirement limits the number of groups in memory to 400. If a database is expanded beyond 400 entries, the program will prompt the user to choose which files are needed for the current session when the program is first started. Group data files supplied with THERM contain primarily, group values of Benson' and Stein and Fahr.'' Additional groups that were derived as part of this work have been made consistent with Benson group values. Other hydrocarbon group data have recently been published by researchers such as Cohen12 and D ~ m a l s k ihowever, ;~ these do not form a comprehensive database. Although the revised hydrocarbon group values are based upon more recent experimental data, Benson group values have been selected for distribution with THERM. Mixing revised hydrocarbon groups with other Benson groups, such as those for halogens, may result in erroneous estimates. This may occur since halogen and other Benson groups were derived with Benson's hydrocarbon group values and not the revised hydrocarbon group values. Group data files can be edited, however, so that a user can employ any groups he chooses. Group data files, like all of THERM'S files, are ASCII and can be easily altered by using any convenient text editor to change group values or add new groups to the data files. T H E T H E R M COMPUTER PACKAGE: S T R U C T U R E A N D FUNCTIONS

Figure 1 presents a functional diagram which shows the relation of the THERM ensemble of computer programs to one another. Individual executable computer programs are

emphasized with heavy boxes to distinguish them from interactive or file input and output. The parent program is THERM, which uses a group contribution database in addition to a molecule/radical description to estimate the thermodynamic properties for species of interest. THERM generates both a tabular listing or "LIST" file and a documentation or "DOC" file which contain the information about the species that were estimated or entered to the database. The LIST file contains the heat of formation (AHfo)and entropy ( S O ) at 298 K (AH:), heat capacities (C,)between 300 and 1500 K, creation data, phase (gas), number of internal rotors, and elemental composition in a 132-column table format (see Table I). The DOC file contains this information along with the groups which were considered, species molecular formula, and symmetry correction to the entropy. DOC file entries for radical species include the bond-dissociation energy and entropy correction for electron spin in addition. Examples of this format are presented in Figures 3b and 4b. As shown in Figure 1, the molecule description either can be input from the keyboard by the user or can be read from a DOC file. An entire DOC file can be processed automatically, allowing database entries (DOC and LIST files) to be recalculated to reflect changes to the group contribution database. This function is particularly useful when one wishes to incorporate new literature data for bond energies or to examine the effect of uncertainties in group values. This allows the database to remain flexible to new input, eliminating a significant barrier to the use of new data when it becomes available. Data in Table I and Figure 3b are shown with units of kJ/mol (Up) and J/(mol.K) (Sand C ); however, units can easily be toggled between these units and tcal/mol (AH:) and cal/(mol-K) (So and Cp). THERM can read, write, and translate files in either set of units. The THERM parent programs can spawn the three child processes THERMFIT, THERMLST, and THERMRXN. The subdivision of the package into these programs was necessitated by memory limitations encountered under MSDOS.

402 J . Chem. Inf. Comput. Sci., Vol. 31, No. 3, 1991

RITTER

oooooooooooooooooo--Noooo0

OOOOOOOOOOOOUOOOOODOOOOOOO 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

000-00000000M0--0000OOOOOO

0 -

N

-

-

N

ZLL

Lr -

-

O

-

~

N

N

-

N

N

-

N

~

W

~

~

O

~

~

"

J. Chem. InJ Comput. Sci., Vol. 31, No. 3, 1991 403

THERM THERM0 300., lOOO., 5000. C6H50 3/20/89 THERMC 6H 50 1 OG 1.573247643+01 1.54394760E-02-5.31383756E-06 -1.76799222E+03-6.20052317E+01-3.73072513E+OO 2.61335718E-08-4.61968583E-12 4.318275643+03 CH30CH3 3/20/89 THERMC 2H 60 1 OG 8.436311623+00 1.30235626E-02-4.45885307E-06 -2.62777965E+O4-2.24530764E+Ol 1.532632003+00 -5.294519293-10 5.96605679E-13-2.32831259E+04 END

300.000 5000.000 1404.000 8.28591116E-10-4.82238305E-14 6.71840729E-02-5.88121307E-05 4.022133343+01 300.000 5000.000 1364.000 6.93233262E-10-4.0275754OE-14 2.37669572E-02-8.1417755OE-06 1.663408853+01

01 2 3 4 21 2 3 4

column numbers: 1

10

20

30

40

50

60

70

80

I23456789 1123456789 1123456789 1123456789 I123456789I123456789 I1234567891123456789







Each of these child programs fulfills a specific function. They

can be run either from within the THERM main program or directly from the DOS prompt. THERMFIT is used to extrapolate C, data contained in a LIST file from IO00 to 5000 K and create NASA format polynomial^^^^^ for direct use by CHEMKIN and other codes which use this format. The resulting polynomial representations of thermodynamic data are valid from 298 to 5000 K and are written to a file that contains polynomial coefficients. Polynomial representations for heat capacity, enthalpy, and entropy functions use the following equations: heat capacity

c,(T) = R(Ul + u ~ T+ U 3 P + U 4 p + U 5 p ) C,(T) = R(U8

+ agT + UloP + U l l p + U l 2 T ' )

T > Tbk T < Tbk (1)

enthalpy function

HO(T)= R [ u ~ + T ( ~ 2 / 2 ) P+ ( ~ 3 / 3 ) P+ ( ~ 4 / 4 ) P+ (a5/4)p Ho(T)= R[QT

+ 061

T > Tbk

+ ( U 9 / 2 ) P + (alo/3)p + (U11/4)p + ( a I d 5 ) p + 0131

T

< Tbk

(2)

entropy So(?")= R[al In ( T )

+ a2T + ( a 3 / 2 ) P + (a4/3)7' + (05/4)p

So(T) = R [ u In ~ (T)

+ 071

T > Tbk

+ agT + (ulo/2)P + (uI1/3)7' + (ai2/4)p

aid

T

< Tbk

(3)

where R is the gas constant, Tis the temperature in Kelvin, ul-u14are the polynomial coefficients described in Figure 2, and Tbk is the break point temperature (a point of forced tangency between the two polynomials). THERMFIT accepts input directly from the keyboard or from a LIST file and creates NASA format polynomials similar to that shown in Figure 2. The NASA format includes four lines for each species. The first contains the species name, creation date/reference, elemental composition, phase, lower and upper temperature limit, break point temperature, number of internal rotors in the species, and the index number 1. The