TW499496B - Purification, analysis and deposition of titanium amides - Google Patents

Abstract

The present invention is a process for enhancing the chemical vapor deposition of titanium nitride from a titanium containing precursor selected from the group consisting of tetrakis (dimethylamino) titrakis (diethylamino) titanium and mixtures thereof, reacted with ammonia to produce the titanium nitride on a semiconductor substrate by the addition of organic amines, such as dipropylamine, in a range of approximately 10 parts per million by weight to 10% by weight, preferably 50 parts per million by weight to 1.0 percent by weight, most preferably 100 parts per million by weight to 5000 parts per million by weight to the titanium containing precursor wherein prior to the reaction, said titanium containing precursor is subjected to a purification process to a purification process to remove hydrocarbon impurities from the titanium containing precursor. It is shown that addition of small amounts of organic amines enhance the deposition rate of titanium nitride, while the presence of hydrocarbons, such as n-decane, retard the deposition rate of titanium nitride.

Description

499496

V. Sample instructions

BACKGROUND OF THE INVENTION Titanium compounds such as titanium nitride are widely used in many applications, especially electrical materials used in the manufacture of integrated circuits for computer chips. Titanium nitride provides enhanced electrical conductivity to electronic circuits, and It has excellent barrier properties for other metal deposits on Shi Xiyi. The rapid development of the sub-device manufacturing industry generally precedes subsequent deposition of conductive metal = titaniumized human deposits as coatings. Oral literature has documented a technique for depositing titanium nitride with tetrakis (dimethylalkylamino) titanium (tdmat) and tetrakis (diethaneamino) titanium (TDEAT). For example, Wu Guo Patent No. 5, 1 39, 825 is shown in decompression and using an inert carrier gas such as nitrogen or helium at 100 to 40 (rc, preferably 15, 0 to "0 ° C, most preferably The deposition of titanium nitride is performed by reacting TDMAT or 1'1) ^ 1 'with ammonia gas under the conditions of chemical vapor deposition (CVD) of TC to 2000 to 25 (TC) to deposit titanium nitride under a condition such as heat On a silicon-containing substrate on a heated substrate.

However, the deposition rate of titanium nitride has changed greatly over the years and is not used for the repeated and accurate use in the electronics manufacturing industry. In the electronics manufacturing industry, silicon wafers are simultaneously processed 'and many batches of wafers are continuously processed.

Electrically acceptable wafers with high yields can make titanium nitride commercially acceptable to the electronics manufacturing industry. In addition, the speed of deposition is critical to provide cost-effective coatings in hundreds of process steps in individual integrated circuits from blank silicon wafers to final electrical interconnections in the electronics manufacturing industry. Wafer process with titanium nitride. The present invention overcomes the low level of titanium compounds that are removed by CV D from the carbon dioxide and the impure impurities by the addition of amine additives and /

499496 V. Description of the invention (2) The deposition rate and the non-fixed deposition rate are described in detail below. The present invention is a method for enhancing chemical vapor deposition of titanium nitride from a titanium-containing precursor, the titanium-containing precursor Selected from the group consisting of tetrakis (dimethylamino) titanium and tetrakis (monoethylamino) titanium, and mixtures thereof, the method comprising adding about 10 parts by weight to about 10 parts by weight relative to the titanium-containing precursor % By weight 'is preferably 50 parts by weight-to i. 〇 by weight, most preferably] 00 parts by weight to 5000 parts by weight of organic amines (such as dipropylamine) The titanium-containing precursor reacts with ammonia gas to generate titanium nitride. Among these reactions, when the reaction is carried out into the stomach β, the titanium-containing precursor is removed by the titanium-containing precursor to remove hydrogen impurities. The research is obvious: adding less organic amines can increase the deposition rate of titanium nitride, and for example, the deposition rate of titanium. Presence of a compound will slow nitridation Jiang Jiang :: ί: a method for determining the predicted deposition rate of a batch of titanium-containing precursors deposited on a chemical vapor: The organic amine content of the batch is used to determine the analyzed amine content; the difference between the analyzed amine content and the monoamine content standard is determined from a deposition rate standard. When the content is greater than the amine content standard, then the? Ϋ ΐ rate is greater than the deposition rate standard, and when the analyzed amine content is glutamine, the predicted deposition rate is less than the Shen Jun 0 w

\\ CHENLIN \ Lika \ PI-Α457.ptd Page 7 499496 V. Description of the invention (3) The present invention includes the compound content standard content in the deposition. When on time, the hydrocarbon deposition rate will contain titanium analysis. The amount, the difference between the comparison, the base standard, the analytical content, the predicted compound content is greater than that of the sink step, the carbon analysis of a compound sink batch, and the fraction from a precipitated hydrocarbon to a sedimentation rate is less than the product standard. It is determined that the content of a standard product of hydrocarbon condensable hydrocarbon accumulating in a matrix hydrogen compound is smaller than the precipitation hydrocarbon standard. The pre-content on the batch containing titanium is determined by the content of the compound, which is greater than the product of the standard. The content of the precursor is determined by the determination of the determination of the carbon content and the predicted carbon deposition and chlorination standard. The standard method is the chemical vapor rate method. The hydrogenation lice compound contains the hydrocarbon accumulation rate; the content of the compound is indicated as the analytical 'then the prediction_ Detailed description of the invention US Patent 5,1 39,825 is shown under reduced pressure and warmed CVD conditions The titanium-containing precursor is reacted with ammonia gas to deposit a titanium compound on a substrate, such as a stone-containing substrate, the entirety of US Patent No. 5,139,825 is incorporated herein by reference. The present invention is a method for enhancing chemical vapor deposition of a titanium-containing compound, such as titanium nitride, from a titanium-containing precursor such as tetrakis (dialkylamino) titanium. In particular, the titanium-containing precursor is selected from the group consisting of tetrakis (dioxane). A group consisting of aminoamino) titanium and tetrakis (diethylaminoamino) titanium and mixtures thereof, the method comprising adding about 10 parts by weight to 10 parts by weight relative to the titanium-containing precursor, preferably Organic amines (eg, dipropylamine) from 50 parts per million to 1.0 weight percent, most preferably from 100 parts per million to 5000 parts per million For decompression and use of an inert carrier gas such as nitrogen or helium

\\ CHENLIN \ Lika \ PI-A457.ptd Page 8 499496 V. Description of the invention (4) The body is at 100 to 400 ° C, preferably 15 (^,] 3〇 (rc, most preferably 200). Reacting the titanium-containing precursor with ammonia under chemical vapor deposition (CVD) conditions to 25 ° C to produce the titanium-containing compound on a semiconductor substrate to deposit titanium nitride on a heated substrate such as a heated silicon-containing substrate The present invention develops a prediction from, for example, tetrakis (dialkylamino) titanium (preferably tetrakis (monomethylamino) titanium) (τ with AT) and tetrakis (diethylamino) titanium (TDEAT)) A method for the performance of titanium nitride chemical vapor deposition of a titanium-containing precursor. A precursor sample is first hydrolyzed with water and then subjected to a typical solvent extraction or solid-phase microextraction (microextraction). The extract is subjected to gas Chromatograph analysis' The amount and nature of the extracted compounds can be used to predict the deposition properties of chemicals. TDM AT and TDE AT can be hydrolyzed by directly adding water or adding to ice to melt the ice. Ways of hydrolysis Depends on the amount of sample to be hydrolyzed. Flux below 0 · 1 g It is directly added to water to undergo hydrolysis, while larger amounts (0.1 to 5 grams) are hydrolyzed using ice to meet the safety considerations of the exothermic hydrolysis reaction and reduce the loss of volatile compounds. Typical methods using ice hydrolysis include 1 0 grams of water is added to a closable container (vial or centrifuge tube), the water is cooled, and the required amount of TDM AT or TDE AT is added. The container is sealed to prevent the loss of volatile compounds, and the container is placed in At room temperature, the ice is allowed to melt. During the melting of the ice, liquid water hydrolyzes TDMAT or TDEAT in a controlled manner. Extraction is performed in two modes. Solid phase microextraction (SPME) is a solvent-free technology that is increasingly being used. Used in different types of analysis. The type of fiber used depends on the type of compound involved. The general analysis described in this article uses

\\ CHENLIN \ Lika \ PI-A457.ptd Page 9 499496 V. Description of the invention (5) A lending fiber with 1 η η oxygen. Extraction of compounds. Example: The extraction efficiency of compounds based on chemical compounds has been used. Typical analysis uses volatile compounds in DMAT or TDEAT for 1 minute to perform partial chemical analysis. Longer or compound extraction or any immiscible with water found suitable extraction in one gas. These have low-boiling and high-density impurities. The concentration of the water to the object is suitable for the non-combined coating, and other types of fibers are used to increase; TDMAT and TDEat \ one ethylene rate. The nitrogen-containing accurate SPME conditions must be optimized for 2 to 4 minutes in the head space of the vial. After being hydrolyzed, and then attached to the gas layer, its adsorption. This operation is used to self-sample, & gt The main method is to desorb them into the gas layer. The shorter time may be used to reduce the analysis time. The traditional chemical solvent-based extraction and extraction compounds have two affinities. Diethyl ether, methyl isobutyl g, toluene, and dichloromethane are the most commonly used, because it has water and there is no similarity found in tdmat * tdeat extraction solvent ratio can be changed to adjust Extracted at the level of a gas chromatograph detector. SPME or solvent extracts can be analyzed by a gas chromatograph. The type of column used depends on the degree of separation desired. The following columns and gas chromatography Parameters have been successfully used to separate extractables from TDM AT and TDE AT.

\\ CHENLIN \ Lika \ PI-A457.ptd

499496 V. Description of the invention (6) Table 1 Analysis conditions 1 Phase thickness of the column (/ m) Length of the column (meters) Diameter of the column (mm) Flow rate (ml / min) Injection port temperature (° C) Injection volume (// L) Initial furnace temperature (° C) and time (minutes) Furnace temperature heating rate Final furnace temperature and time Detector detector temperature table 2 Analysis conditions 2

100% poly, 5 30 0. 53 8. 0 190 5, undivided 5 0 ° C / 5 minutes 1 0 ° C / minute 2 5 0 ° C / 5 minutes Thermal conductivity type 2 60 ° C Methylsiloxy Phase thickness of alkane column (// m) Length of column (meters) Diameter of column (mm) Flow rate (ml / min) Injection port temperature (° C) Injection volume (// L) 100% Poly 5 30 0. 53 8. 0 1901. 0 Dimethylsilane

\\ CHENLIN \ Lika \ PI-A457.ptd Page 11 499496 V. Description of the invention (7) Initial furnace temperature (° C) and time (minutes) Furnace temperature heating rate Final furnace temperature and time detector detector Temperature 5 0 ° C / 5 minutes 2 0 ° C / minute 2 5 (TC / 5 minutes Thermal conductivity type 2 60 t:

Phase thickness of the column (/ zm) Length of the column (meters) String diameter (mm) Flow rate (ml / min) Injection port temperature (° C) Injection volume (// L) Initial furnace temperature (° C ) And time (minutes) furnace temperature heating rate final furnace temperature and time detector 4 temperature detector temperature

Table 3 Analysis conditions 3 κ monofluorenylsiloxane / 5% divinylbenzene 0.5 60 0. 25 50 190 0 · 2, 1: 1 0 0 split 1 0 0 ° C / 5 minutes 1 0 ° C 2 5 0 ° C / 5 minutes Finnegan GCQ 260 ° C

Shell spectrometer analysis was used to analyze compounds found in TDMAT * TDEAT. The chromatogram obtained with a gas chromatograph using a mass spectrometer is linked to the thermal conductivity detection of the gas chromatograph to facilitate rapid classification of the extracted compounds

5. Description of the invention (8). Other detectors used in gas chromatographs The detectors listed in the tables are provided to replace the detectors listed above, nitrogen and phosphorus detectors, and photoionization detectors. Flame ionizers and atomic emission detectors can be used for plutonium; two-electron capture compounds and / or increase their sensitivity. Focusing on the analysis of specific types of chemistry Several methods can be used to advance the extracted chemistry. The standard addition method and the method used for this analysis 彳 b ° 仃 准 准 _ L Wen 1 d solid phase micro-extraction Dixie Qiu Qiu Yuyu can be used for calibration. It has been found that ^: The choice depends on the method of quantification of the compound and the standard addition method. It is possible to use LA to enclose seven types and types. Aliphatic hydrocarbons appear to be completely extracted from both of the methods described above, while Day 4 is calibrated to external standards. Because 1, and T can not be completely extracted using the compound, and the affinity of the 7 solution phase, the nitrogen-containing substance is calibrated to obtain accurate quantitative j and requires standard addition methods or based on the above-mentioned analytical methods. In the above analysis method; the person who predicts the deposition of precursors in TDEAT and TDMAT can be used to confirm the discovery of saturated amines and aliphatic hydrocarbons; the sample ff 'was found to contain inoperative . The chemical compound 5 is speculated to be derived from the synthetic compound Jian i (that is, the retention time of each chromatographic chromatography can be measured, and the preliminary thermal conductivity is detected to determine the category to which the compound belongs). j responses can be listed. Table 4 and Figures 1 and 2 are shown. The functional nitrogen heteroatom (N) was used to predict the deposition performance through the analysis method of the Institute. A compound is divided into two categories, one containing 含 and one that does not contain nitrogen heteroatom (HC), and the others

Which 496 Fifth, the description of the invention (9) could not be clearly classified in one of the two categories (unclassified). Table 4 / Oral measurement of the deposition rate from the results of extract analysis (arbitrary units) (Angstroms per minute) Ν HC Unclassified deposition rate 2829 511 131 1. 41 1265 4219 84 1. 0 2143 669 153 1. 37 954 1464 530 1.12 1331 1978 580 1.13 1355 1694 418 1.12 1481 98 62 1.19 2720 123 58 1. 42 μ μ 2 shows that there is a 3 correlation between the two types of compounds and the deposition rate performance of chemicals. However, the deposition properties of individual compounds found in the main class of JL right μ AX for the deposit number 129 135 137 138 139 140 142 144 total 3471 5568 2 9 6 5 2948 3889 3467 1641 290 1 4b. The above procedure can be used to specifically measure this = 'and thereby predict the deposition performance of TDMAT and TDEAT. (TM Erming two methods capable of reading and measuring tetrakis (dimethyl sulfenyl fluorene centi-grams) of hydrocarbon impure substances. This method can detect the detection limit of one million. From the substances, the limit This is due to the calibration of compounds in the parts per million range. The number of compounds measured makes it practical for individual compounds. The total amount of hydrocarbons is in n-decane.

\\ CHENLIN \ Lika \ PI-A457. Page 14 499496 V. Explanation of the invention (ίο) Calculation, and since all hydrocarbons are assumed to have the same sensitivity as n-decane, this value is semi-quantitative. The correctness of this method was verified by adding 680 0 and 3580 ppm by weight of n-decane to the TDMAT sample. The recoveries were 106 and 120% by weight, respectively. The analysis or measurement of molecular weight impurities of tetrakis (monofluorenylamino) titanium (TDM AT) is difficult because it is very volatile and cannot be analyzed by a gas chromatograph. Several stationary phases and temperature profiles have failed to provide reproducible results showing the overall purity of TDM AT samples. The reactivity of air and moisture prevents reverse phase liquid chromatography from being used for sigma 1 analysis. Reverse phase liquid chromatography is often used to analyze non-volatile or non-lightly volatile compounds. Normal phase liquid chromatography failed to provide analysis, which may be the result of the reaction of TDMAT with the polar position of the column packing and the residual water in the mobile phase. Current methods for evaluating TDMAT analysis include obtaining a nuclear magnetic resonance spectrum (NMR) and checking whether the spectrum has any spikes that do not belong to tdmat. Because sample preparation is not controlled and there is no standardization of any kind; calibration, so existing procedures are essentially The above is only qualitative. Since Li is able to detect impurities at a level of about 1%, "any signal present in the D-signal shows an analysis of <99%. Conversely, no spikes appear> 99 analysis (all percentages are weight percentages) ). The reactivity of TDMAT with moisture was used to develop a method to determine the amount of chlorocarbons present in TDMAT. Instead of avoiding decomposition, the entire sample was completely hydrolyzed. After hydrolysis, the sample was extracted to get the heart to the hydrolyzed Hydrocarbons and water-soluble titanium dioxide and, for example, dimethylamine 丨

499496 V. Description of the Invention (Π) '-Impurities are separated. These hydrocarbons are estimated to be derived from the solvents used to form TDMAT, as they are not by-products of the synthetic reaction. Extraction with Ik uses solid-phase microextraction. This technique uses polymethyldisilazane-coated glass fibers to adsorb compounds that have a high affinity for the coating. The polar compounds (water, dimethylamine) have a low affinity for the coating, and therefore they are less adsorbed. In this way, a hydrolyzed sample can be quickly extracted in a reproducible manner using a small amount of solvent. The adsorbed compounds are desorbed at the thermal injection port of the gas chromatograph, and analyzed by normal phase gas chromatography using appropriate detectors. A relatively small amount of water and dioxamine are delivered to this, which ensures that hydrocarbon impurities are well detected. Solid phase microextraction (SPME) was performed using a SPME apparatus from Suplc Inc Inc (Bellefonte, PA). The type of fiber used was 100 / z, fiber coated with unbonded poly (fluorenyldisilazane) (PDMS), and samples were introduced manually. A reconstituted amount of TDMAT (0.010 to 0.100 g) was injected via syringe into a closed vial with a volume of 15raL containing 5 mL of distilled deionized water, resulting in immediate solution. Because the vial has a septum and a cap, volatile compounds are retained in the bottle. After equilibrating the sample to ambient temperature for approximately ten minutes, the headspace of the vial was sampled with SPME fibers for 2.0 minutes. The fiber was transferred to the injection port of the GC, and the compound was desorbed in the 90 minute process. Figure 3 shows a typical chromatogram and GC conditions. At about 1 minute, the instrument showed a broad doublet peak,

499496 V. Description of the invention (12) This is because air enters Gc when SPME fiber is introduced. The degree of this peak depends on how long the SPME fiber is held in the Gc injection port. There is a large spike due to the water absorbed by the fiber, and a small spike due to the dimethylamine produced by the hydrolysis of TDM ^ T and the sharp peak. The instrument did not detect other highly volatile components. After a residence time of about eight or eight, a series of two, ·

peak. By using a GC-mass spectrometer and adding n-decane to the sample, positive spikes were confirmed. The GC response of the R f hydrogen compound depends on two main factors, the sensitivity of each detector to the detector, and the affinity of each compound for the fiber coating. In order to be properly calibrated, the nature of each compound and the essential factors must be confirmed. The first experiment used a male I to confirm that the impure material was taken from TDMAT. Retrieve mass spectrometry data ^; shows: most of the compounds are _〇 class and two; "f fragrant compounds. The number of impurities is found to be greater than 30. Because of%! Similar spectra 'make it clear to identify all impurities into objects. Two: = sex method, using a single-type compound i :::. The Gc residence time is marked *. Since Fu Jun II · is centered in the range of measured hydrocarbons, it is a representative compound of octahedral. Therefore, its affinity for fiber is represented by slave lice compounds. The detection rate (JCD) used is that these compounds provide approximately uniformly, ~, and power on the whole sheep ^ 疋 for all hydrocarbons, to One millionth of a million; ° = result. The calculation is performed by multiplying all V0 peaks Γ between 7 and 14 minutes.

499496 V. Description of the invention (13) Points' to obtain the total hydrocarbon spike area. No other spikes appeared after 14 minutes. Prepare an n-decane standard in isopropanol for calibration. In a similar way to hydrolyzed T D M A T, a weighted isopropanol was added to a 15 mL vial containing 5 mL of distilled water to construct a calibration curve. Calibration was also performed by adding standards to determine whether the sensitivity of the method was affected by the presence of hydrolyzed TDM AT.

SPME fibers have limited capacity. This capacity was determined by analyzing the increasing concentration criteria and changing the sample size of the TDMAT. Figures 4 and 5 show the spike area vs. sample size data. The linear range of Figure 4 shows a decane with a fiber capacity of approximately 2 5 0 // g. TDM AT samples that produce a total spike area greater than 1 200 must reduce their sample size because the fiber / detector combination response is non-linear when the fiber is overloaded. The sample size is affected by the concentration of hydrocarbons in the sample. For slaves concentrations below 1% by weight, the maximum TDM AT sample size is approximately the fiber capacity divided by the hydrocarbon concentration. (25〇 #

§ / 〇,., 〇1 = 25, 〇〇〇 // 1) ^ or 25 / ^). Figure 5 evaluates the sample size by evaluating the linearity as the sample size increases. When the sample volume is as high as 25 # [: a linear response can still be generated, showing that the fiber capacity has not been exceeded. Assuming that the total hydrocarbon impurities are approximately 1% by weight, both the n-decane calibration experiment and the TDM AT sample size show approximately the same fiber capacity. The presence of matrix effects was evaluated by preparing a calibration curve for an aqueous matrix and a hydrolyzed TDMAT, and comparing slope or sensitivity. See Table 5 below.

\\ CHENLIN \ Lika \ PI-A457.ptd Page 18 499496 V. Description of the invention (14) Table 5 Calibration data of hydrocarbons (represented by decane) in aqueous and TDMAT matrix. Water-based slope = 35.4, Water-based standard error of slope = 2.6, TDMAT slope = 36.5, Standard error of TDMAT slope = 6.5 Decane added, β g area (aqueous substrate) Area (TDMAT substrate ) 0. 0 0 2 940, 3 63 1 11. 0 526. 8 388 1, 40 02 22. 5 823 · 3, 905 · 5, 1087 3681, 4204 25 · 0 1627, 1673 5032, 5032 95% of the slope The confidence intervals overlap, showing that the presence of hydrolyzed TMDAT did not significantly affect slope sensitivity. To simplify the analysis, further calibrations were performed only on water-based substrates. Because the sensitivity mainly depends on the affinity of the fiber for hydrocarbons, and it will change with time or the change of the fiber, so the calibration should be performed regularly or when the spME fiber is replaced. 0 The calibration data shows that the accuracy may not be good, especially In TDMAT matrix. Repeated analyses and analyses performed on different days were used to assess accuracy. The relative standard deviation of repeated analyses was between 5 and 15%. The accuracy of aqueous or TDMAT-matrix samples is not different, and some unknown / uncontrolled factors seem to cause the high variability measured. One of the factors leading to poor accuracy is the small sample size. The fiber's high affinity for carbon compounds and the relatively high hydrocarbon concentration in TDMAT make small sample sizes

The value cannot be discarded using any outlier test. Arbitrarily removing this low value improves accuracy to 16606 ± 395. This result is not significantly different from the analysis on the first day.

\\ CHENLIN \ Lika \ PI-A457.ptd Page 20 499496 V. Description of the invention (16) 1 Assessment of potential hydrocarbon loss in the exothermic hydrolysis step &amp;# Add decane directly to the TDMAT sample and implement The entire analysis process. Wrong Two different amounts of decane were added to two equally divided TDM AT samples. Its # fruit summary list is in Table 7. Table 7 Calibration results of n-decay burn in TDMAT for calibration and recovery studies. All samples were analyzed twice and the amount in the table (in ppm n-decane). Uncalibrated calibration results (ppm) Fruit (ppm) Degrees (ppm) 8988 16238 6801 8988 13300 3582 Average carbon listed Hydrogen compounds containing

Difference (calibrated-not recovered) (ppm) Percentage 7250 106.6 4312 120.4 The n-decane added to TDMAT has satisfactory recovery. Samples that were initially uncalibrated were analyzed _ days earlier than those that were calibrated, showing good day-to-day reproducibility.

π A Ben Maoming has developed a compound that can detect the presence of hydrocarbons in KTDMAT, and n because only n-decane is used for calibration. This article is qualitative. Although the current accuracy is 5-15% sufficient to detect hydrocarbon impurities and between different sources &gt; The limitation of this method is high relative to confirm all impurities and calibration accuracy and semi-quantitative results. Borrow every impure or find out every impure

499496

V. Description of the invention The relative response factor of (17) to n-decane. This method can be quantitative in nature. These relative response factors must take into account the affinity of each compound for PDMS fibers. Tetrakis (diethaneamino) titanium is used in another analysis of hydrocarbon impurities and in the analysis of amines. Tetrakis (diethaneamino) titanium (TDEAT) in a 6. sacrificial film evaporator (Pope Scientific, Inc.,

Milwaukee, Wisconsin) was purified by short path distillation. Reducing many types of volatile impurities to, rotor speed, vacuum, wall temperature, branch condenser temperature, and prefractionation calcination to reduce volatile impurities to short-path distillation is a process, a low-volatile chemical May be subject to fractions or ingredients. On the inner wall of the tube heated under vacuum. Coaxial to the tube. Liquid chemicals flow down the distiller. Volatile components are coaxially located at the cold collection outlet. The resulting low-volatile liquid or residue in the condensate is removed. The original process parameters did not have an acceptable concentration. These parameters, including feed rate, rear wall temperature, main condenser temperature, and pre-bottle temperature, are the best accepted concentrations. The inventors of this case obtained the difference from the parameters of the wiping film, 'Acceptable spermatozoa can be obtained in the application of this analytical technique, and this ion is not widely exposed to heat.' Otherwise, the chemical may be decomposed and introduced into a distillation. Inside the downstream heating wall of the condenser is the inner surface of a condensing heating wall, and more condensers are subjected to volatilization and condensation. Individual liquid or distillate and remove 0 at the bottom of the still to purify TDEAT. It was originally found that a few were used at the time, and the only analysis in use at the time

499496 V. Description of the invention (18) The method is NMR. After the inventor of the present case further developed this product, the inventor of the present case developed a technique for analyzing volatile compounds such as hydrocarbons and amines. This novel analytical technique emphasizes the need for improved purification capabilities. The original parameters included an evaporation (wall) temperature of 120 ° C. This temperature at low pressure is higher than the temperature required for TDEAT evaporation. Therefore, higher boiling volatiles will also evaporate away from the wall. The original condensation (pointer) temperature was 0 ° C. This temperature is not only low enough to condense TDEAT, but it can also condense many other lower boiling volatiles. These absolute temperatures do not take into account changes in the degree of vacuum in the distillation chamber, and the degree of vacuum in the distillation chamber is a number of factors for the composition of the feed and the feed rate. Two methods are used to deal with this problem. The first method is to change the absolute temperature to the relative temperature. Considering the inherent problem of controlling vacuum, the difference between the wall temperature and the pointer temperature is more important than the absolute temperature of the wall or pointer. The contact time of the body (as a function of rotor speed) is also considered. ^ Let raw materials with different compositions be used in this system.

The second way is to focus on a more robust system. Reality of the first way = more precise temperature control is required. This includes improving the measurement of liquid on the wall, and the temperature measurement when the liquid leaves the heated part of the wall. It also needs to improve pointer temperature measurement. The temperature of the prefractionation flask was reduced to ensure a lower boiling and the volatiles did not evaporate again after being collected. Six experiments were performed to study the effect of changing parameters on purity. ^ Experiment # 1 This experiment gradually increased the pointer temperature from lor to 7 (rc.

499496 V. Description of the invention (19) The number is kept constant, and the pointer temperature is set at 10 ° C, 30 ° C, 50 t, and 70 ° C. At each setting, two liters of raw materials are processed, and a total of eight liters of raw materials are processed. One sample is processed at each pointer temperature, and the pre-concerned volume at each temperature is recorded. The hydrocarbon (HC) state of the sample was analyzed. Table 8 compares the ppm content of HC at various pointer temperatures in Experiment # 1. Because all prefractions are HC, the numbers are high. As the temperature increased, the volume of the collected material also increased, and HC showing a low boiling point did not condense on the pointer. The HC for this batch is 2630.

Temperature (° c) QASS # Raw material (8L) 1036 10 1037 30 1038 50 1039 70 1040 Total volume N / A Table 8 HC PF volume (mi) 5559 N / A 50715 5 52626 5 55101 10 47342 15 N / A 3 5 m 1 Main stream (L) Tail (L) N / AN / A 1.8 Λ 2.0 .4 1 · 8 .3 1. 2 .6 68 1. 3 Chemical experiment # 2 In this experiment, the relative temperature of the wall and pointer is set. Optimizing the operation By fixing the pointer temperature below 70 t, the wall temperature is increased to the point where the main fraction is seen flowing into the receiver. At the wall, the temperature of the needle is increased until no liquid flows in; at this temperature, the finger and the king are in the fork. Pointer temperature

\\ CHENLIN \ Lika \ PI-A457.ptd 499496 V. Description of the invention (20) The degree is reduced until the liquid flow is restored. Table 9 shows the relationship between wall temperature and pointer temperature in experiment # 2. By changing the Lofa (wall) temperature or the condensation (pointer) temperature, the flow can start and stop. These temperatures apply only to this operation. First increase the evaporation temperature until the product begins to flow. The product continues to flow until the condensation temperature increases C. At this temperature the product stops flowing. When the temperature drops to 84 ° C, flow resumes. The HC for this batch was 772 ppm.指针 Wall temperature pointer Temperature liquid flow into the receiver 70 75 70 70 No No Table 9 80 85 70 70 No No 90 90 90 90 90 7 0 75 80 85 84 Yes Yes No Yes Estimated liquid level of the thermoelectric liquid The liquid was flowed to the experimental probe, and the experimental probe was heated and left. The rotor stayed up to the bottom of the tank. Experiment # 3 on the wall was used to discuss the rotor speed and feed rate into the outer middle of the wall in degrees. The temperature of another thermocouple when it is heated. The speed is changed from 200 R P M to time. Liquid for a fixed time. Then the rotor speed observes the effect of the liquid on the wall. When the bottom of the vessel wall is a better estimate of purity, the other is added to the vessel wall at 350 RPM and the flow rate is maintained to maintain a solid overflow and keep the temperature below the effect. Another — the outer bottom on the inner wall is maintained in the tank wall with parameters to maintain a solid estimate, and the liquid pointer temperature that is estimated and the feed rate is insufficient 5 499496 V. Description of the invention (21) When the temperature has reached the optimal C 'is It is considered that the feed rate, rotor speed, and speed are reduced. Table 1 The relative between the liquid leaving the wall and the condensing temperature (pointer) in the 10-carriage shellfish U 3 (difference == brewing degree (bottom) as a function of sub-velocity, feed rate and wall temperature.). Bottom temperature is the turn time (hours) 1 Hand 80 Bottom 98 Difference 18 2 3 Table 10 4 5 84 83 84 85 82 72 72 62 2 11 12 23 6 7 8 9 83 83 84 84 66 69 80 70 17 14 4 14 Experiment # 4 This experiment examines the effect of distilling substances with a small amount of hydrocarbon impurities. Lot # 1 41 was re-steamed using the newly determined optimization technique under the parameters of Lot # 1 4 2 in Table 13 below. Table 11 is a comparison of the program parameters for lot # 141 and lot # 143. The values listed are representative but subject to change.

Table 11 Pointer Temperature Lot # 141 70 Lot # 1 43 84 Wall temperature speed Vacuum degree flow rate (liters HC (ppm) (RPM) (mTorr) / hour) 120 202 500 2.2 2630 83 275 500 2.2 33

\\ CHENLIN \ Lika \ PI-A457.ptd Page 26 499496

V. Description of the Invention (22) Experiment # 5 This experiment explores the effect of filling a prefractionation flask with dry ice to prevent any capture and re-evaporation of volatiles. The raw materials are batch materials supplied by the manufacturer, and the complex HC content is 5000 ppm. Dian Experiment # 6 This experiment explores the effect of purifying the recovered material with the optimized process parameters and customer returns. Such raw materials are normally in the range of 3000 ppm. Table 12 describes the parameters of these batches. These values are representative but will change. Table 12 Pointer middle bottom wall temperature velocity vacuum degree flow rate (HC temperature wall temperature wall temperature setting (RPM) (mTorr) liters / (ppm) point hours) Lot No. 1 44 84 87 80 1 00275 800 0.6 70 Lot No. 14584 86 72 96 275 1 000 〇 7 242 Table 13 is a summary of the experimental batch numbers. The ranges listed are high and low without considering the average. Table 13 Batch HC pointer wall temperature middle bottom vacuum flow rate velocity yield

\\ CHENLlN \ Lika \ PI-A457.ptd Page 27 499496 V. Description of the invention (23) Temperature setting wall temperature point 141 2630 10-70 12〇 one 142 1772 78-84 70-95 143 33 80-85 1〇 ° 80-91 144 70 82-84 1〇〇83-91 145 242 84 59-96 85-87 Wall temperature (liters / (RPM) 0 / 〇 (mTorr) hours)-240-950 2 · 2 202 75- 425-7001.0 350 77 66 450-600 .85200 71-8 2-3 5 0 70 800-12006 27574 -84 68 5 0 0- 1 1 〇〇 · 7 275 66 -72

Hydrocarbon impurities can be effectively removed using appropriate process parameters. Because the internal vacuum of the wiped film tank depends on many factors including the impurity concentration of the raw materials, the feed rate, the wall temperature, the vacuum pumping conditions, the temperature of the liquid nitrogen trap, and the temperature of the condenser trap in advance, so the absolute temperature Does not optimize equipment capabilities well. Relative temperature is more suitable for removing volatile impurities. The relative separation between the evaporation temperature and the degree of cold &gt; JDL is the best indicator of proper short path distillation. The absolute temperature near these two temperatures may change according to the vacuum in the tank, but the relative temperature between the two should be

Stay fixed. Condensing temperature is a fairly simple measurement. It is recommended to measure at the inlet and outlet pointer points in addition to the bath measurement. It is slightly more difficult to actually measure the liquid wiped from the wall. The easiest way to make this measurement is an unheated wall Ansang thermocouple directly below the heated wall. Non-evaporated liquid flows down the wall at this point, and the temperature at this point should be just above the condensation temperature. Such as

\\ CHENLIN \ Li ka \ PI-A457.ptd Page 28 499496

The result is lower than this efficiency. The pure substance will also be ensured to improve normality, still low, and there is also the time to reach the bottom of the tank, which is heated immediately, and evaporated to dryness. 'So distillation is not inferior in liquids. Time is also important to make the temperature of the effluent. If the flow rate of all liquids is not to be heated before the temperature to the intermediate column is completed without an extra bottom in half of the residence time, because some TDEAT does not evaporate. If it is too high above the condensation temperature, it should stay and evaporate, thus reducing There is sufficient contact between the pure wall of the main component and the liquid and high evaporation efficiency. If the wall temperature is too high, there is a danger that the liquid will decompose in "over-hot spots" and may contain some high-boiling impurities before it has evaporated. This phenomenon is achieved by matching the partial temperatures. This operation ensures that the liquid has reached the evaporation temperature, remains in the remainder, and the flow rate is high enough that the liquid reaches the tank / due to the increase in tailings and pre-distillate, the yield is increased by about 0 weight / 0. However, this technique allows modification of tails and pre-distillates to extract the remaining TDEAT, and this makes up for some lost yield. This technique seems to be applicable to any optimization of short-path distillation purification chemicals. Experiment # 7 It is suggested that adsorption is a useful technique to reduce HC impurities. Prepare a sample with a weight ratio of tdeat to adsorbent of approximately 1 ·· 1. Lot # 14〇 was selected as the T D E A T source for this experiment, while gasification, activated carbon, and p u r s p e c 2 2 2 1 (IC I) were selected as adsorbents. These samples were subjected to manual agitation 'every day and were sampled one week later to analyze HC content. There is no identifiable change in the Ti02 sample. Most of the colors in the carbon sample and some HCs have been shifted

499496

5. Description of the invention (25) Except in PurasPec sample, HC was reduced to 315 ppm, and some colors were removed. The Puraspec samples received several repeated Spme tests for HC. The experimental values were 133 and 159 ppm HC. Sample # 140VT Q20 0 6-A Item lot number 140VT Control HC 4800 4403 Table 14 Q2006-B Q2006-C Q2006-D i 02 Carbon Puraspec 4818 2646 315 Experiment # 8 夂 Repeated run # 7 to confirm the result, but the adsorbent pair The ratio of TDEAT is reduced to 1: 2 weight ratio. A 50/50 carbon and puraspec222 1 mixing ratio of the outer sample was added to determine whether color and HC could be improved in one step. At the same time, an intermediate sample (Q20i3D) was moved to ^ fhraspec 222 1 vial two days later to check the removal efficiency of both the adsorbent and the control (ΪΓ0131). Within 24 hours, five I = decreased from 5475ρρη to 2793ppm. After one week, the vials were n: t-like 'and the results are listed in the table below. The results show the same trend ^ ^ ^ σ, but the removal efficiency of both can not reach the pure pure ranchidino alone to be able to remove most colors, while Puraspec 222 1 was noticed to remove most of the colors Hc. Table 1 5

499496

Sample # 140VT Q2015-A Q2015-B Q2015-C Q2015-D Q2015-E V. Description of the Invention (26) Item Lot No. 1 40VT Ti02 Mixture Carbon Puraspec Control HC 4800 4570 1036 4010 650 4937 Experiment # 9 This experiment is to enlarge Pur a spec 222 1 program possibilities. In a glove box, a flask was filled with 209 grams of TDEAT and 204 grams of Puraspec 222 1. The ratio of this mixed TDEAT to adsorbent is 10% by weight and a volume ratio of about 4 to 1. The sealed flask was stirred for one week with a stirring rod device. The mixture was sampled on two, four and one week. The sample at one week was taken after the sorbent was filtered. Filtration was performed first using a 10 micron filter (Arbortech, Model # PolycapHD), and finally using a 0.2 micron filter (Balston, Model # AAQ). This procedure produces about 2 kg. Table 16 Sample # Q2030-A Q2030-B Q 2 0 3 3-A Q2033 Q2046 Item 2 days control 2 days adsorption 4 days control 4 days adsorption Final HC 2988 1552 932 678 441 Adsorption seems to be a viable method to remove HC from TDEAT . This method has also been shown to be effective in removing some colored impurities. Additional techniques for this procedure are also considered. These include pressure filtration, high / low temperature filtration, continuous tearing,

\\ CHENLIN \ Lika \ PI-A457.ptd Page 31 499496 V. Description of the invention (27) Chromatography, and reflux filter bed. Experiment # 10 This experiment observes the effect of a series of flask-to-flask (one plate) distillations. A 5 liter flask was charged with TDEAT (Lot 140) and subjected to distillation under dynamic vacuum. 10% by weight of the predistillate was collected, and 10% by weight of the tails were left behind. The main fraction was sampled and the procedure was repeated three times. This method seems to be another method to remove the HC content in TDE AT. This experiment shows that each distillation reduces the HC to about one-quarter of the HC at the beginning. Table 17 Sample # 140VT Q2010VT Q2032VT Q2041VT Item Batch Substance IX 2X 3X HC 4800 1677 469 92 Experiment # 11 An Older shaw distillation system was set up to check the TDE AT effect of such devices. A 5 liter reboiler was installed with a 1 "5-plate Oldershaw column, and the system was placed in a vacuum of 500mTorr (+ / 200mTorr). Two operations were performed. One of the pre-cracks was sampled and one main distillate was collected. The first operation was the reuse of lot # 1 35, and the second operation was the use of a substance from another source. The color of the main distillate was observed Much better than the original color. This technology also removes hydrocarbons

\\ CHENLIN \ Lika \ PI-A457.ptd Page 32 499496 V. Useful Method of Invention Description (28). Table 18 Sample # Q1093VT Q1097VT Q2002VT Q2014VT Project First operation First operation Second operation Second operation-raw material work-main component work-raw material work-main component HC 9019 1107 3720 1780

Based on the above results, the inventor of this case pioneered the development of a method that can analyze the amine and hydrocarbon content of titanium-containing precursors made in different batches of commercial preparations, and predict the deposition rate of titanium-containing precursors. Chitin-containing precursors are available for end use in, for example, electronics manufacturing, where yield and yield are critical to the economics of integrated circuit manufacturing. The inventor of this case has developed a method for determining the predicted deposition rate of a titanium oxide-containing compound deposited on a substrate in a chemical gas deposition.

Method: • Analyze the organic amine content in the batch to determine the amount of amines analyzed; compare the analyzed amine content with a standard for amine content (the amine content f standard may be edited from a series of amine content analysis results Average), the $ deposition rate is determined from a deposition rate standard (which may be edited by averaging a series of titanium compound deposits) in accordance with the difference between the analyzed amine content and the amine content standard; where, when the analyzed The amine content is greater than the amine content standard date :: = The predicted deposition rate is greater than the deposition rate standard, and when the analyzed ^ ΐ is less than the amine content standard, the predicted deposition rate is less than the Shen Hanzhun.

\\ CHENLIN \ Lika \ PI-A457.ptd

499496

Song =, the inventor of the present case has developed a method for determining the predicted deposition rate of a batch of titanium-containing precursors for depositing titanium-containing compounds on a substrate during chemical milk deposition, which includes: analyzing the hydrocarbon content of the batch To determine the content of sex slaves; compare the analytical hydrocarbon content with the one-carbon ^ compound 1 standard (which may be edited by analyzing a series of hydrocarbons containing i and average them), based The difference between the analytical hydrocarbon content and the hydrocarbon content standard is determined from a deposition rate standard (which may be edited by averaging a series of Chin compounds). When the analytical carbon If the content of chlorine compound is greater than the target of chlorine-containing compounds, the predicted deposition rate is smaller than the deposition rate standard, and when the content of analytical hydrocarbons is smaller than the hydrocarbon content standard, the predicted deposition rate is greater than the deposition rate standard . The inventors of this case have also developed a method to correlate the analyzed amine content in a batch of titanium-containing precursors with the analyzed hydrocarbon content to determine predictions based on the impact or impact of the amine and hydrocarbon content in the batches Deposition rate method.

The present invention provides techniques for individually determining the content of amines and hydrocarbons and / or impurities in titanium-containing precursors used for titanium compound deposition to enhance the titanium compound deposition rate. At low ppm levels, its enhanced (amine) or retarded (carbohydrate) deposition rate, it is important to have analytical techniques sensitive enough to detect 'such low levels of amine or hydrocarbon functionality. It is important to detect amine or hydrocarbon levels or to control specific amounts of amines for specific amine doping or calibration 'to enable customers who use titanium-containing precursors to deposit titanium-containing films (especially electronics manufacturing) Can accurately and accurately predict the content of commercial batches

\\ CHENLIN \ Lika \ PI-A457.ptd

499496 V. Description of the invention (30) For the deposition rate of precursors, the electronics manufacturing industry needs formulas and reproducible results to maintain high yields and accept mass production of integrated circuits and related electronic equipment. In addition, the discovery of the effects of amine content and hydrocarbon impurities on the deposition rate of titanium-containing precursors has led to the development of novel and surprising enhancements to the deposition of titanium-containing compounds (such as films) on integrated circuits and other electronic devices. Unexpected approach. Adding amines to titanium-containing precursors will not enhance the deposition rate of titanium compounds on the substrate, as for many of the better commercially viable titanium-containing precursors (such as TDMAT and TDEAT), the amine functionality is this Z precursors are by-products of these precursors when they are deposited on semiconductors by CVD. The present invention has been shown in several preferred embodiments, but the full scope of the present invention is defined in the following patent application scope. x

499496

Brief description of the diagrams Figure 1 is the relationship between nitrogen-containing compounds and sedimentation rate (Angstroms per minute) ° linear regression: slope = 67 · 2, intercept = -3334, correlation coefficient, r = 0.2924 2 is the relationship between hydrocarbons and sedimentation rate (Angstroms per minute) ° linear regression: slope = -111, intercept 29774, correlation coefficient, r =-0 · 7 9 7 〇 Figure 3 is a kind of hydrolysis A typical chromatogram of the hydrolyzed TMDAT sample taken from the headspace of the sample by Ding Xianren. GC conditions: syringe temperature: 190 ° C; furnace conditions: maintained at 40. 〇2 minutes, maintained at 25 ° c / min for 2 minutes at 2 50 ° c; Detector: Thermal conductivity type, temperature = 25.0 ° C, Guanzhu · DB-1, 30 meters long X 0.53mm ID x 5 film. Fig. 4 shows the change in peak area when a large amount of decane is added. Conditions: 5 ml of aqueous solution volume, 10 ml of headspace, 100 // m⑼ ... fiber, adsorption for 2 minutes, to absorb m minutes! Child, thermal conductivity type detector. The six data points cover 0.4 to 1 720 // g n-decay. Figure 5 is a plot of the total hydrocarbon spike area as a function of sample volume (1 eL = 100 #g). Conditions: 5 ml of aqueous solution volume, 10 ml of head space, 100 // m PDMS fiber, adsorption for 2 minutes, desorption for 1 minute, thermal conductivity type detector.

\\ CHENLIN \ Lika \ PI-A457.ptd

Claims (1)

^^ 496 -44SS2iil2〇962 VI. Patent scope 9L 2. 2¾ W years of enhanced self-titanium-containing precursors will be contained on the substrate ^ ^ method including organic amines and titanium-containing precursors and Reduce hydrocarbon impurities in Chin precursors before sedimentation. —2. The method according to item 1 of the scope of the patent application, wherein the deposition is a chemical vapor deposition. 3. The method according to item 1 of the scope of patent application, wherein the containing is removed before being deposited to remove impurities from the titanium-containing precursor. * k gas compound 4 · The method according to the first patent application, wherein 4 is an organic amine of titanium. Q Titanium precursor 5. The method according to item 4 of the patent application, wherein the titanium is tetrakis (dialkylamino) titanium. / 、 Organic amine 6 · The method according to item 5 of the patent application, wherein the tetra (alkylamine% 499496 __Case No. 891209 (^ _ year, month, date, amendment __ VI. Patent Application Base) Titanium is selected from the group consisting of tetrakis (dimethylamino) titanium, tetra (diethylamino) titanium, and tetrakis (dipropyl) A group consisting of aminoamino) titanium and tetrakis (dibutylamino) titanium and mixtures thereof. 7. The method according to item 1 of the patent application, wherein the organic amine is selected from the group consisting of an alkylamine, an alkenylamine, an alkynylamine, an arylamine, an alkylarylamine, a dialkylamine, and a diarylamine , Urea and its mixtures. 8. The method of claim 1 in which the organic amine is dipropylamine. Wherein the deposited titanium-containing material 9 · The method according to item 1 of the scope of patent application The substance is titanium nitride. The titanium-containing precursor reacts with ammonia, and in the reaction is added the method of item 10, such as the scope of the patent application, in which the chemical vapor deposition of titanium nitride is generated on the conductor base, and the base is titanium And tetrakis (diethylamino) titanium and its mixtures 11 · If the scope of the patent application, the strongest organic amine is 10% relative to titanium. 〇 method, I 丄 /, the deposition rate increases the weight of the precursor I + 10 parts per million WAdm-hos t \ 20010610 \ P11100 \ ai r 11147 \ niend01 .ptc 苐 38 2002. 02. 23.038 Wherein the organic amine 1 2 · As in the method of applying for the scope of the patent application item 丨 丨, / 4, the addition amount is from 50 parts per million by weight to the weight of 1 1 3 · as in the patent application scope of the item 12 "Ν is called, Gary is 100 to 500 parts per million by weight W'J / 7 ^, 1 of the organic amine 4. A self-titanium precursor-enhanced chemical vapor deposition method" the present The oblique earth compensating titanium nitride / 3 Chin precursors are selected from tetra (dimethylamino) amino) titanium and: M: ,, said from M '♦ and tetra (monoethyl in titanium containing A well-established ethnic group, the method includes the addition of relative and precursor weights of 100,000 ppm: titanium ΓΓ; ammonia gas reaction, so z z is produced on the semiconductor substrate, before the reaction, The titanium-containing precursor is applied to a king of clothes to remove hydrocarbon impurities from the titanium-containing precursor. Wherein the hydrocarbon is 15. The method as described in claim 14 of the patent scope: C1 to C 20 hydrocarbons. 1 6 · The method of item 15 in the scope of the patent application is C5 to C10 hydrocarbons. Among them, carbonitriding 17 · The method of item 16 in the patent scope of Shenyan is C 5 to C 10. 18 · The method according to item 4 of the patent application, wherein 2002. 02. 23. 039 499496 _ case number 89120962 _ year month__ VI. Patent application scope Purification process of removing hydrocarbon impurities is selected from distillation, adsorption, membrane separation, solvent extraction and combinations thereof Composition of ethnic groups. 1 9. A method for determining a predicted deposition rate of a batch of titanium-containing precursors for depositing a titanium-containing compound on a substrate in chemical vapor deposition, comprising: analyzing the organic amine content of the batch to determine the analyzed amine content; The analyzed amine content is compared with a monoamine content standard, and the predicted deposition rate is determined from a deposition rate standard based on a difference between the analyzed amine content and the amine content standard. 20. The method of claim 19 in the scope of patent application, wherein when the analyzed amine content is greater than the amine content standard, the predicted deposition rate is greater than the deposition rate standard, and when the analyzed amine content is less than the amine For the content standard, the predicted deposition rate is smaller than the deposition rate standard. 2 1. A method for determining a predictive deposition rate of a batch of titanium-containing precursors for depositing a titanium-containing compound on a substrate in chemical vapor deposition, comprising: analyzing the hydrocarbon content of the batch to determine the hydrocarbons to be analyzed Content; comparing the analyzed hydrocarbon content with a hydrocarbon content standard, and based on the difference between the analyzed hydrocarbon content and the hydrogen breaking compound content standard, the predicted deposition rate is determined from a deposition rate standard. 2 2. The method of claim 21, wherein when the analyzed hydrocarbon content is greater than the hydrocarbon content standard, the predicted deposition rate is less than the deposition rate standard, and when the analyzed hydrocarbon content is \\ Adm-host \ 20010610 \ P11100 \ airlll47. ^ End01.ptc page 40 2002.02.23. 040 499496 499496 \\ Adm-host \ 20010610 \ Pl 1100 \ ai rl 1147. \ mend01 .ptc Page 42 2002.02.23.042