Etching and CVD cleaning for semiconductors

LIQUID CRYSTAL DISPLAYS, AND PHOTOVOLTAICS

Emissions vary according to the gases used in manufacturing different types of electronic devices, the process used (or more roughly, process type (e.g., CVD or etch)), the brand of process tool used, and the implementation of emission reduction technology.

The choice of methods will depend on data availability and is outlined in the decision tree, see Figure 6.1, Decision Tree for Estimation of FC Emissions from Electronics Manufacturing. Emissions from liquid FCs are estimated using Tier 1, 2 and 3 approaches and are described separately in this section.8

Continuous (in-situ) emissions monitoring is currently considered a technically and economically unviable means to estimate emissions from this industry. FC emissions are periodically measured, however, during the development of new processes and tools, and after the establishment of commercial-ready process conditions (also known as centreline process conditions).9 The industry seeks, prior to the introduction of high-volume manufacturing, centreline process designs that minimize FC emissions. However, it must be noted that FC emissions can be affected by changes in process variables (e.g., pressure, temperature, plasma power, FC gas flow, processing time). Thus, the accuracy of the methods used for estimating emissions will be affected by eventual differences between the process used in production and the reference centreline process. In addition, the efficacy of FC emission control equipment depends on operating and maintaining the equipment according to the manufacturer's specifications: Increased gas flows, improper temperature settings, and failure to perform required maintenance will individually and collectively negatively impact performance.

The accuracy of estimated emissions depends on the method used. The Tier 1 method uses default values for all parameters and does not account for the use of emission control technology. The Tier 2a method uses company-specific data on the proportion of gas used in processes with and without emission control technology, but does not distinguish between etching and cleaning, and uses default values for the other parameters. The Tier 2b method uses company-specific data on the proportion of gas used in etching versus cleaning and the proportion of gas used in processes with emission control technology, but relies on default values for some or all of the other parameters. The most rigorous method, Tier 3 method, requires a complete set of process-specific values rather than defaults.

Table 6.1 summarises the data requirements for the tiered emissions estimating methods for electronics manufacturing.

7 Where a commercial mixture is used inventory compilers will need to ensure that the conversion of the mass of the mixture to CO2 equivalents uses the appropriate conversion factors.

8 The logic depicted in Figure 6.1 does not show the possibility of combining tiers to improve estimates of emissions. For example, improved estimates of emissions might be achieved by using Tier 3 for a specific gas and process and Tier 2b for other gases and processes instead of using only the Tier 2b method. Similarly, the Tier 2a and 2b methods might be combined to produce an improved estimate compared to using only Tier 2a. However, the Tier 1 method should not be combined with any other method.

9 Centreline conditions refer to the conditions under which equipment manufacturers standardise their equipment for sale. These are nominal specifications for gas flows, chamber pressure, processing time, plasma power, etc. It is common for semiconductor manufacturers to modify these conditions to optimise for particular needs.

Table 6.1

Information sources necessary for completing the tiered emission estimating methods for

ELECTRONICS MANUFACTURING

Data

Tier 1

Tier 2a

Tier 2b

Tier 3

Process Gas Entering Tool

FCip = kg of gas i fed into specific process p or small set of common process tools (e.g., silicon nitride etch).

M

FCip = kg of gas i fed into broad process category (e.g., etching or CVD chamber cleaning).

M

M(etch) & M(CVD)

h = Fraction of gas remaining in shipping container after use (heel).

D

D

M

Process Gas Reactions and Destruction in Tool

Uip = Use rate (fraction destroyed or transformed) for each gas i and process p.

D

D(etch) & D(CVD) a

M

Emission factor for by-product emissions of CF4, C2F6, CHF3 and C3F8 respectively for gas i for each process.

D

D(etch) & D(CVD) a

M

Downstream FC Emission Control

aip = Fraction of gas i volume fed into processes with certified FC emission control technologies.

M

M

M

dip = Fraction of gas i destroyed by the emission control technology.

D

D a

M

dCF4,p, dC2F6,p, dCHF3,p and dC3F8,p = Fraction of CF4, C2F6, CHF3 and C3F8 by-products respectively destroyed by the emission control technology.b

M

Annual Production Capacity

Cd = Annual manufacturing design capacity in surface area of substrate processed (e.g., silicon, glass).

M

Cu = Fraction of annual capacity utilisation

D/M

M = measure or acquire these values. D = Use default factors from guidance.

a When available and supportable, M values may be substituted for D values for Tier 2a and 2b. See conditions in Table 6.6. b There are no default values for Tier 2a and Tier 2b because the effect of by-products has been incorporated into the D-values for dip for gas i.

Figure 6.1 Decision tree for estimation of FC emissions from electronics manufacturing

Figure 6.1 Decision tree for estimation of FC emissions from electronics manufacturing

Lockout Tagout Decision Tree

Note:

1. See Volume 1 Chapter 4, Methodological Choice and Identification of Key Categories (noting Section 4.1.2 on limited resources), for discussion of key categories and use of decision trees.

Note:

1. See Volume 1 Chapter 4, Methodological Choice and Identification of Key Categories (noting Section 4.1.2 on limited resources), for discussion of key categories and use of decision trees.

TIER 1 METHOD - DEFAULT

The Tier 1 method is the least accurate estimation method and should be used only in cases where company-specific data are not available. The Tier 1 method, unlike the Tier 2 or 3 methods, is designed to give an aggregated estimate of FC emissions although its methodology appears to produce gas-specific emissions. Estimates are made simultaneously for all gases as listed in Table 6.2 and can only be used if reported as a complete set.

The calculation of emissions relies on a fixed set of generic emissions factors. The members of the set differ depending on the sector (or class) of electronic products being manufactured (semiconductors, TFT-FPDs or PV-cells). Each member of a set, which is a gas-specific emission factor, expresses an average emissions per unit of substrate area (e.g., silicon, TFT-FPD panel or PV-cell) consumed during manufacture. For any class of electronic products, the factors (members of the set) are multiplied by the annual capacity utilisation (Cu, a fraction) and the annual manufacturing design capacity (Cd, in units of giga square meters (Gm2)) of substrate processes. The product (C • Cd) is an estimate of the quantity of substrate consumed during electronics manufacture. The result is a set of annual emissions expressed in kg of the gases that comprise the set for each class of electronic products. Because the use of FCs varies widely during PV manufacture, a third factor to account for the proportion of PV manufacture that employs FC is needed to estimate FC emissions from the PV cells manufacturing. The Tier 1 formula is shown in Equation 6.1.

Equation 6.1

Tier 1 method for estimation of the set of FC emissions

[FC, } = {EF, . Ca . Cd .[CPV . S + (1 - S)]} (i = 1,...,n)

Where:

Note: { }n denotes the set for each class of products (semiconductors, TFT-FPD or PV-cells) and n denotes the number of gases included in each set (six for semiconductors, three for TFT-FPD manufacture and two for PV-cells. See Table 6.2.). The estimates are only valid if made and reported for all members of the set using this Tier 1 methodology.

EFi = FC emission factor for gas i expressed as annual mass of emissions per square meters of substrate surface area for the product class, (mass of gas i)/m2

Cu = fraction of annual plant production capacity utilisation, fraction

Cd = annual manufacturing design capacity, Gm2of substrate processed, except for PV manufacturing which is Mm2

CPV = fraction of PV manufacture that uses FCs, fraction

S = 1 when Equation 6.1 is applied to PV industry and zero when Equation 6.1 is applied to either semiconductor or TFT-FPD industries, dimensionless

This method does not account for differences among process types (etching versus cleaning), individual processes, or tools. It also does not account for the possible use of atmospheric emission-control devices.

In using Tier 1, inventory compilers should not modify, in any way, the set of the FCs assumed in Table 6.2. Inventory compilers should not combine emissions estimated using Tier 1 method with emissions estimated using the Tier 2 or 3 methods. Neither may inventory compilers use, for example, the Tier 1 factor for CF4 to estimate the emissions of CF4 from semiconductors and combine it with the results of other FC gases from a Tier 2 or Tier 3 method. (See also Section 6.2.2.1.)

TIER 2a METHOD - PROCESS GAS-SPECIFIC PARAMETERS

This method calculates emissions for each FC used on the basis of company-specific data on gas consumption and on emission control technologies. It uses industry-wide default values for the 'heel' or fraction of the purchased gas remaining in the shipping container after use (h), the fraction of the gas 'used' (destroyed or transformed) in the semiconductor or TFT-FPD manufacturing process, and the fraction of the gas converted into CF4 or C2F6 during the process. To use the Tier 2a method inventory compilers must have direct communication with industry (e.g., annual emissions reporting) to gather data and ensure that emission control technologies are installed and in use.

Total emissions are equal to the sum of emissions from the gas FCi used in the production process plus the emissions of by-product CF4, C2F6, CHF3 and C3F8 resulting from use of the gas FCi., as shown in Equations 6.2, 6.3, 6.4, 6.5 and 6.6. Unlike the Tier 3 and 2b methods that are explained later in this section, the Tier 2a method does not distinguish between processes or process types (etching versus cleaning), individual processes or tools. The default emission factors represent weighted averages (based on expert judgments of weights), formed separately for each gas, over all etch and CVD processes.

As discussed below in the section on emission factors, the Tier 2a method uses the emission factor for the process type (CVD or etch) in which the individual FC is most frequently used in the particular electronics sector. This method reflects a current trend where individual FCs tend to be used predominantly in particular process types (CVD or etch) throughout each industry. However, in countries with companies or plants that depart significantly from the industry-wide pattern of usage (e.g., by using a gas primarily in etch while others primarily use it in CVD), inventory compilers should evaluate the potential to introduce error by using the Tier 2a method rather than the Tier 2b method.

Equation 6.2 Tier 2a method for estimation of FC emissions

Where:

Ei = emissions of gas i, kg

FCi = consumption of gas i,(e.g., CF4, C2F6, C3F8, c-C4F8, c-C4F8O, C4F6, CsF8, CHF3, CH2F2, NF3, SF6), kg h = fraction of gas remaining in shipping container (heel) after use, fraction Ui = use rate of gas i (fraction destroyed or transformed in process), fraction ai = fraction of gas i volume used in processes with emission control technologies (company- or plant-specific), fraction di = fraction of gas i destroyed by the emission control technology, fraction

Equation 6.3 By-product emissions of CF4

BPECF4,i = (1 - h)• BCF4,i • FC, • (1 - ai • dCF4)

Where:

BPECF4,i = by-product emissions of CF4 from the gas i used, kg BCF4 i = emission factor, kg CF4 created/kg gas i used dCF4 = fraction of CF4 by-product destroyed by the emission control technology, fraction

Equation 6.4 By-product emissions of C2F6

BPEC2F6i = (1 - h) • BC2F6,i • FCi • (1 - ai • dC2F6 )

Where:

BPEC2F6,i = by-product emissions of C2F6 from the gas i used, kg BC2F6,i = emission factor, kg C2F6 created/kg gas i used dC2F6 = fraction of C2F6 by-product destroyed by the emission control technology, fraction

Equation 6.5 By-product emissions of chf3

BPECHF3i = (1 - h) • BCHF3,i • FCi • (1 - ai • dCHF3 )

Where:

BPECHF3 i = by-product emissions of CHF3 from the gas i used, kg BCHF3 i = emission factor, kg CHF3 created/kg gas i used dCHF3 = fraction of CHF3 by-product destroyed by the emission control technology, fraction

Equation 6.6 By-product emissions of C3F8

BPEc3f8,i = (1 - *)• Bc3f8,i • FCt • (1 - a, • d(

Where:

BPEC3F8 i = by-product emissions of C3F8 from the gas i used, kg BC3F8,i = emission factor, kg C3F8 created/kg gas i used dC3F8 = fraction of C3F8 by-product destroyed by the emission control technology, fraction

After estimating the emission of gas i (Ei) and the CF4, C2F6, CHF3 and C3F8 by-product emissions for each gas (BPECF4i, BPEC2F6i, BPECHF3i and BPEC3F8i), inventory compilers or companies should sum these emissions across all gases to estimate the total aggregate FC emissions.

TIER 2b METHOD - PROCESS TYPE-SPECIFIC PARAMETERS

The Tier 2b method requires data on the aggregate quantities of each gas fed into all etching processes and all cleaning processes (FCi,p). Thus, it distinguishes only between broad process types (etching vs. CVD chamber cleaning), but it does not distinguish among the many possible individual processes or small sets of processes. Industry-wide default values can be used for any or all of the following:

• the fraction of the gas remaining in the shipping container after use termed the 'heel' (h);

• the fraction of the gas 'used' (destroyed or transformed) per process type (Uip);

• the emission factor for CF4 by-product emissions in the process type (BCF4ip);

• the emission factor for C2F6 by-product emissions in the process type (BC2F6ip);

• the emission factor for CHF3 by-product emissions in the process type (BCHF3ip)); and

• the emission factor for C3F8 by-product emissions in the process type (BC3F8ip).

Defaults are also presented (see Table 6.6) for the fraction of the gas destroyed by the emissions control technology by process type (di,p, dCF4,p, dC2F6,p, dCHF3p and dC3F8,p). Unless emission control technologies are installed, the default value for ai,p, the fraction of gas volume fed into processes with emission control technologies, is zero. The default values for Ui,p, BCF4ip, BC2F6ip, BCHF3ip and BC3F8 ip represent simple unweighted averages, formed separately for each gas, over all etch processes and over all CVD processes. Company or plant-specific emission factors may be substituted for default values when available. The equations account for the plant-specific use of emission-control devices, but do not account for differences among individual processes or tools or among manufacturing plants in their mix of processes and tools. Thus, Tier 2b estimates will be less accurate than Tier 3 estimates. Also, note that the Tier 2b method is applicable to semiconductor and TFT-FPD manufacture.

Emissions resulting from the use of a specific FC (FCi) consist of emissions of FCi itself plus emissions of CF4, C2F6, CHF3 and C3F8 created as by-products during use of FCi. The following calculation should be repeated for each gas for each process type:

Equation 6.7 Tier 2b method for estimation of FC emissions

Where:

Ei = emissions of gas i, kg p = process type (etching vs. CVD chamber cleaning)

FCip = mass of gas i fed into process type p (e.g., CF4, C2F6, C3F8, c-C4F8, c-C4F8O, C4F6, C5F8, CHF3, ' CH2F2, NF3, SF6), kg h = fraction of gas remaining in shipping container (heel) after use, fraction

Uip= use rate for each gas i and process type p (fraction destroyed or transformed), fraction ai,p = fraction of gas i volume fed into process type p with emission control technologies (company-or plant-specific), fraction di,p = fraction of gas i destroyed by the emission control technology used in process type p (If more than one emission control technology is used in process type p, this is the average of the fraction destroyed by those emission control technologies, where each fraction is weighted by the quantity of gas fed into tools using that technology), fraction

Equation 6.8 By-product emissions of CF4

BPECF 4,i = (l - E [-BCF 4,i, p • FCi, p • - ai, p • dCF 4, p J P

Where:

BPECF4 l = by-product emissions of CF4 converted from the gas i used, kg

BCF4,i,p = emission factor for by-product emissions of CF4 converted from gas i in process type p, kg CF4 created/kg gas i used dCF4,p = fraction of CF4 by-product destroyed by the emission control technology used in process type p (e.g., control technology type listed in Table 6.6), fraction

Where:

BPEC2F6 l = by-product emissions of C2F6 converted from the gas i used, kg

BC2F6l,p = emission factor for by-product emissions of C2F6 converted from gas i in process type p, kg C2F6 created/kg gas i used dC2F6,p = fraction of C2F6 by-product destroyed by the emission control technology used in process type p (e.g., control technology type listed in Table 6.6), fraction

Equation 6.10 By-product emissions of CHF3

BPECHF3,i = (l - h)E lBCHF3,i, p • FCi, p • (l - ai, p • dCHF3, p J p

Where:

BPECHF3,i = by-product emissions of CHF3 converted from the gas i used, kg

BCHF3,i,p = emission factor for by-product emissions of CHF3 converted from gas i in process type p, kg CHF3 created/kg gas i used dCHF3,p = fraction of CHF3 by-product destroyed by the emission control technology used in process type p (e.g., control technology type listed in Table 6.6), fraction

Equation 6.11 By-product emissions of C3F8

BPEC 3F 8,i = (l - h)^l\BC 3F 8,i, p • FCr, p • ( - ür, p • dC3F 8, p J P

Where:

BPEC3F8 i = by-product emissions of C3F8 from the gas i used, kg

BC3F8,i,p = emission factor for by-product emissions of C3F8 converted from gas i in process type p, kg C3F8 created/kg gas i used dC3F8,p = fraction of C3F8 by-product destroyed by the emission control technology used in process type p (e.g., control technology type listed in Table 6.6), fraction

Note that in certain etching or cleaning recipes, multiple FC precursors can be used concurrently and emissions of CF4, C2F6, CHF3 or C3F8 as by-products may originate from each of the individual FC precursor decomposition. In such cases, emissions of CF4, C2F6, CHF3 or C3F8 by-products should be reported as originating from the FC gas with the largest mass flow.

TIER 3 METHOD - PROCESS-SPECIFIC PARAMETERS

The Tier 3 method also uses Equations 6.7, 6.8, 6.9, 6.10 and 6.11. However, this method requires company-specific or plant-specific values for all the parameters used in these equations for each individual process or for each of small sets of processes (e.g., silicon nitride etching or plasma enhanced chemical vapour deposition (PECVD) tool chamber cleaning). Therefore, when using Equations 6.7, 6.8, 6.9, 6.10 and 6.11, inventory compilers need to interpret 'p' in these equations as a specific 'Process' (e.g., silicon nitride etching or plasma enhanced chemical vapour deposition (PECVD) tool chamber cleaning), not as 'Process type'.

For purposes of transparency and comparability, the values used for these emission parameters should be well documented (see Section 6.2.2).

CF4 formation from C-containing films during semiconductor manufacturing

The Tier 2a, Tier 2b and Tier 3 methods account for CF4 emissions formed during removal via etching of carbon-containing low dielectric constant (k) materials or cleaning CVD reactors containing low k or carbide films during semiconductor manufacture. CF4 may be formed even if the FC precursor does not contain carbon or if the FC precursor is not a greenhouse gas.

For example, cleaning low k CVD reactors with NF3 will produce CF4 as a by-product. In these cases, Equation 6.7 should be used to report NF3 emissions and the result of Equation 6.8 should be used to reflect emissions of CF4 from the process. In those situations where F2, COF2, or ClF3 is used in chamber cleaning, CF4 may also be formed. In this case, CF4 emissions are estimated using Equation 6.8 and the results added to the total CF4 emissions obtained from Equation 6.7. In both cases, BCF4ip should be measured as the fraction of the mass of CF4 produced over the mass of clean or etch gas introduced in the reactor.

After estimating emissions of each FC gas and emissions of CF4, C2F6, CHF3 and C3F8 as by-products, inventory compilers or companies should sum these emissions across all gases to arrive at an estimate of aggregate FC emissions from a particular process.

Box 6.1

Example for semiconductor manufacture

For example, if a source used NF3 (for chamber cleaning and etch), CHF3 (etch) and CF4 (etch), the total emissions, if low k films were used, are estimated using Equation 6.7 for NF3, CHF3 and CF4 and Equation 6.8 for the formation of CF4 formed when removing low k films with NF3. In equation form, the total is:

Total FC emissions = Enf3 + Eqhf3 + Ecf4 + BPEcf4,nf3

Guide to Alternative Fuels

Guide to Alternative Fuels

Your Alternative Fuel Solution for Saving Money, Reducing Oil Dependency, and Helping the Planet. Ethanol is an alternative to gasoline. The use of ethanol has been demonstrated to reduce greenhouse emissions slightly as compared to gasoline. Through this ebook, you are going to learn what you will need to know why choosing an alternative fuel may benefit you and your future.

Get My Free Ebook


Post a comment