This invention relates to a stable lithium diiso¬ propylamide (LDA) composition, and a method of its preparation- Lithium diisopropylamide (LDA) is widely used as a reagent in the preparation of pharmaceuticals and specialty chemicals. LDA has a low solubility and irreversibly precipitates in hydrocarbon solvents. Consequently, LDA is not commercially available in solution form. The only commercially available LDA has been as a pyrophoric solid or slurry in hydrocar¬ bon, and this has been available only in low volumes. The safety hazard in the shipment and handling of the pyrophoric LDA solid or slurry, as well as the diffi¬ culty in using these forms of LDA in reactions, severely limit the usefulness of these LDA forms in commercial applications.

Users of LDA generally prefer an LDA solution. Although LDA is soluble in ethers, it is quite unsta¬ ble in this medium and quickly decomposes at room temperatures. Therefore, large volume users of LDA must synthesize their own LDA as it is needed, typically by reacting n-butyl-lithium with diisopro- pylamine in cold tetrahydrofuran (THF). This reaction is fairly easy to carry out, but may present safety hazards for those users unfamiliar in the handling of pyrophoric n-butyllithium.

Accordingly, the need exists for a stable and nonpyrophoric form of lithium diisopropylamide which could be produced and shipped in quantity and which presents fewer handling problems than the currently available forms of lithium diisopropylamide.

In accordance with the present invention, there is provided a composition comprising lithium diiso-

propylamide and a limited amount of tetrahydrofuran that forms a lithium diisopropylamide composition which is thermally stable at mild temperatures for several months. Furthermore, the composition is non- pyrophoric and thus fewer precautions are required to ensure safe shipping and handling.

Tetrahydrofuran (THF) is the preferred liquid ether for use in the present invention, since LDA is quite soluble in THF. However, as earlier noted, the currently available forms of LDA are unstable in the presence of THF.

It has been found that by limiting the amount of THF to no more than about 1 mole of THF for each mole of LDA present, an LDA/THF composition is obtained which exhibits enhanced thermal stability. Prefer¬ ably, the mole ratio of THF to LDA is within the range of 0.5:1 to 1:1, and most desirably within the range of 0.8:1 to 1:1.

Surprisingly, while not critical to the present invention, the presence in the LDA composition of small amounts of excess diisopropylamine, one of the reactants in the preparation of the LDA, and/or the presence of other amines, such as non-metalable tri- ethylamine (TEA) for example, has a significant stabilizing effect on the decomposition of the LDA. As little as one mole percent excess amine has a stabilizing effect but an excess of 4 or more mole percent is preferred. Excesses of up to 100 mole percent can be utilized. The most preferred molar excess of amine is about 10 to 11 moles of amine per mole of lithium diisopropylamide.

Thus, in accordance with one broad aspect of the present invention, there is provided a stable, non¬ pyrophoric form of lithium diisopropylamide comprising lithium diisopropylamide, and tetrahydrofuran in an

amount not exceeding about one mole of tetrahydrofuran per mole of lithium diisopropylamide. For added stability, the composition may also include at least one C- to C. _g amine, preferably a tertiary amine. In its preferred and more limited aspects, the present invention provides a stable, nonpyrophoric lithium diisopropylamide solution which comprises a 1 to 3 molar solution of lithium diisopropylamide in a mixture of tetrahydrofuran, at least one amine selected from the group consisting of diisopropylamine and triethylamine, and an inert liquid hydrocarbon cosolvent, the tetrahydrofuran being present in an amount not exceeding about 1 mole of tetrahydrofuran per mole of lithium diisopropylamide. While the soluble, stable LDA composition of the present invention can be produced by any of several different methods, including the conventional method involving the reaction of n-butyllithium with diiso¬ propylamine, as well as by reacting other organo- lithiums and organodilithiums , such as methyllithium, ethyllithium, phenyllithium, cyclohexyllithium, dilithiobutane for example, the preferred method of preparation in accordance with the present invention involves the preparation of lithium diisopropylamide directly from lithium metal. This approach has very significant economic advantage over the traditional method of preparation, since one mole of lithium metal will produce one mole of LDA, while two moles of lithium metal are required when producing LDA from an alkyllithium compound.

In accordance with the preferred method of the present invention, the lithium diisopropylamide is produced directly from lithium metal with the use of an electron carrier such as styrene or isoprene. The reaction is carried out in tetrahydrofuran (THF), with

the amount of the THF being limited to no more than two moles per mole of electron carrier. Limiting the amount of THF results in a stable product, as earlier noted, and also advantageously avoids sticking together of the lithium metal during the reaction. An inert liquid hydrocarbon cosolvent may also be employed to adjust the final LDA concentration as desired. The stable LDA composition produced by this method will also include as a part of the hydrocarbon cosolvent system, the reduced electron carrier, e.g. ethylbenzene (b.p. = 136°C) where styrene was used as the electron carrier and 2-methyl-2-butene (b.p. = 36°C) where isoprene was used. Since the boiling points of these two byproducts differ widely, the particular electron carrier which is used may be selected to facilitate recovery of the user's end product from the reduced electron carrier. Thus, for example, where the end product is a liquid, if the user is preparing a high boiling compound, an LDA solution containing the low boiling 2-methyl-2-butene would be selected, whereas the choice for a low boil¬ ing compound would be an LDA solution containing ethylbenzene. For those instances where the butene or ethylbenzene are not suitable, other electron carriers may be used, such as butadiene, divinylbenzene, and naphthalene, for example depending upon the method of separation to be employed.

In accordance with an alternative method in accordance with the present invention, LDA was also prepared for the first time in a non-ethereal sol¬ vent. In accordance with this aspect of the present invention, LDA is prepared directly from lithium metal and an electron carrier such as styrene or isoprene, in the presence of diisopropylamine in a hydrocarbon solvent, producing a stable but relatively insoluble

mass which may be later rendered soluble if desired by addition of tetrahydrofuran.

The present invention will be understood more fully from the description which follows, and from the 5 accompanying examples, in which particular embodiments of the invention are shown. It is to be understood at the outset, however, that persons of skill in the appropriate arts may modify the invention here described while still achieving the favorable results

10 of this invention. Accordingly, the description and examples which follow are to be understood as being a broad teaching disclosure directed to persons of skill in the appropriate arts.

I. Thermal Stability of LDA Solutions

1.5 Tests were conducted to determine the thermal stability of various LDA solutions as a function of several variables, and the results are summarized in Table 1 below. These tests illustrate how LDA com¬ positions in accordance with the present invention

20 exhibit superior thermal stability as compared to LDA compositions not in accordance with the invention. The LDA solutions were prepared either from lithium metal in the presence of THF and diisopropylamine through the use of a styrene or isoprene electron

25 carrier in accordance with the procedures described in Example 1 below, or by the conventional route by reacting n-butyl-lithium (NBL) with diisopropylamine in THF. The decomposition of the LDA was determined from gas chromatography methods based upon the free

30 diisopropylamine (DIPA) resulting from the metallation of tetrahydrofuran by lithium diisopropylamide using active (non-hydrolized) injection techniques. The decomposition of LDA is understood to occur as follows:

+ LiN ( CH( CH ₃ ) ) ₂ — >HN( CH( CH ₃ ) ₂ ) ₂ +

THF LDA DIPA ETALATED

THF

Table I shows that at even relatively mild tem¬ peratures (20.5°C +_ 2.5°C) LDA solutions containing more than one mole equivalent of THF decompose rapidly, liberating free DIPA via metalation of uncom- plexed THF. For example, LDA solutions containing 2 and 6.4 mole equivalents THF/LDA (Samples D and C) , lose 15% and 1001 activity after 30 days and 7 days, respectively. Conversely, LDA solutions containing 1 or less mole equivalents THF/LDA were found to be thermally stable.

At elevated temperatures (27.5°C _+ 7.5°C) all LDA solutions tested degraded. LDA solutions with limited THF (THF/LDA ^~ 0.5 to 1) decomposed slightly (6% loss in 30 days) while an LDA solution containing only a slight excess of THF (THF/LDA = 1.13) depicted an accelerated rate of decomposition (10% loss in 30 days) . Refrigerated samples (0°C) containing 1 equiva¬ lent of THF or less remained clear yellow and stable. However, a small amount of decomposition (1.5% after 30 days) was noted for a refrigerated sample (see Sample E above) which contained a slight excess of THF (THF/LDA = 1.13). Finally, an LDA sample (F above) containing 0.54 mole equivalents THF deposited a white crystalline precipitate on the walls and bottom of the sample bottle after 50 days at room temperature and 0°C. Additional THF redissolved the precipitate,

which was identified as LDA. Samples containing 0.82 to 0.95 mole equivalents THF/LDA remained clear (no pptn. ) at 0°C for at least 90 days indicating no solubility or stability problem (see samples B, G, H, I and J) .

In summary, the use of excess THF (THF/LDA 1) results in LDA solutions which rapidly decompose at room temperature. LDA solutions containing THF in an amount below the preferred minimum level (THF/LDA = 0.5) slowly precipitate product. The preferred operating range to produce a thermally stable and soluble LDA product is from 0.5 to 1, and most desir¬ ably from 0.8 to 1.

During the course of thermal stability studies it was surprising to learn that the presence of small amounts of excess diisopropylamine (DIPA) or even non-metalable triethylamine (TEA) had a stabilizing effect on the decomposition of solutions of LDA in THF/cyclohexane. The rate of decomposition of an LDA solution containing no stabilizer (Sample G) was twice (11.7% loss/mo.) that of comparable LDA solutions (6% loss/mo.) containing excess DIPA or TEA (4 and 5 mole % - Sample F, J and A). LDA solutions containing more stabilizer (10 %) degraded even less (4% loss/mo. Sample I and B). At 40°C all LDA samples tested degraded. However, with increasing stabilizer the degradation rate/day signi icantly decreased from 1.49 to 0.81 %/day (Samples M, N, 0 and P) . Interestingly, an LDA solution containing 13 mole % excess THF and 18 mole % stabilizer (see above Table Sample E) thermally degraded at all three test temperatures. Also, LDA solutions (Sample D and K) containing 100% excess THF degraded at significantly different rates (0.49 and 0.17%/day) due to the use of different stabilizers, indicating that TEA may be superior to excess DIPA.

Thus, it was apparent from the above data that the presence of excess DIPA plays an important role as a stabilizer in the preparation of an LDA solution, and the preferred DIPA level appears to be about 4 to 100 mole percent, with the optimum level about 10 mole percent.

Thermal stability of LDA solutions of various compositions were tested under simulated spring and summer plant temperatures (18 to 35°C) and also under refrigerated conditions (0°C). The data is presented in the summary Table I. At 0°C LDA solutions, ideal in terms of THF and stabilizer (discussed above) remained clear light yellow with no precipitation and stable for three months (see Samples A, B, G, H, and I). However, samples with excess ether (THF or t-butylmethylether(TBME)) underwent some thermal decomposition at 0°C depending on how much excess ether was present (see Sample E and L). At the inter¬ mediate temperature level (20.5 +_ 2.5°C) an ideal LDA sample (Sample B) was thermally stable (no increase in free DIPA), but deepened in color (It. orange) and became quite hazy. Another sample (Sample F) showed no decomposition at 20.5 _+ 2.5°C, but was metastable in that precipitation of LDA occurred because of insufficient THF. Other samples decomposed, deepened in color and became hazy with solids to. varying extents, depending on the amount of excess Lewis base (Samples, C, E, D, , and L). At elevated temperature (27.5 _+ 7.5°C) all samples degraded, deepened in color and generated insoluble particulate matter, again, depending upon the major variables (see summary Table I and Fig. I, II, and III). The most stable LDA solution (0.13% loss/day) at 27.5 +_ 7.5% (Sample I) contained 10.6 mole % stabilizer and less than 1 equivalent THF (THF/LDA = 0.95).

In summary, all LDA solutions degrade and produce insoluble degradation products at elevated tempera¬ tures (27.5 +_ 7.5°C and 40°C). At intermediate temperatures (20.5 _+ 2.5°C) LDA samples containing optimum amounts of THF and stabilizer do not decompose according to GLC methods, but do deepen in color and form slight haziness. At 0°C LDA solution containing slightly less than 1 equivalent of ether are stable and remain clear yellow. The thermal stability of the LDA-THF composition appears to be not significantly affected by the LDA concentration, and 1:1 LDA-THF solutions as high as 3.6M may be prepared directly from lithium metal by the electron carrier route in the absence of hydro- carbon cosolvent. However, at concentrations greater than 2.3M LDA, the composition will precipitate at lower temperatures, but can be redissolved at room temperature. While the solid or slurry form may be useful in certain applications, the solution form is most conveniently handled. Accordingly, to avoid precipitation during storage and shipment under re rigeration where the solution form is required, it is desirable that the concentration be no greater than 2.3M. A preferred range of concentrations for practical commercial quantities of the LDA solution is from about 1 to about 2.3M.

B. Preparation of Stable LDA Solutions The preferred method of preparation in accordance with the present invention involves reacting lithium metal particles in an ether medium, preferably tetra¬ hydrofuran, wherein the amount of ether does not exceed about 2 moles of ether per mole of electron carrier, with an electron carrier and with slightly more than 2 moles of diisopropylamine per mole of electron carrier to thereby produce a soluble, stable

lithium diisopropylamide composition. In this reac¬ tion, the electron carrier is the limiting reagent. The electron carrier is characterized by its ability to readily receive an electron from the lithium and form a radical anion. Conjugated unsaturated hydro¬ carbons are noted for this ability, as shown by Ziegler et al., Liebigs Annalen der Chemie, 511, .64 (1934), and particularly preferred for use as an electron carrier in the present invention are styrene and isoprene. Other suitable electron carriers may include butadiene, divinylbenzene and napthelene for example.

The production of lithium diisopropylamide directly from lithium metal through the use of an electron carrier, namely styrene, has been previously described, see M.T. Reetz and W.F. Maier, Liebigs Annalen der Chemie, 10, 1471 (1980). In this proce¬ dure, LDA was prepared by adding a styrene/ethyl ether feed to a refluxing slurry of granulated lithium and diisopropylamine in ethyl ether. The formation of LDA is promoted by styrene which functions as an electron carrier by accepting an electron from the metal to form a radical anion. The radical anion is rapidly quenched by diisopropylamine to form LDA and a neutrally charged radical. The electron carrier then receives a second electron from another lithium atom to form what is believed to be alpha-methylbenzyl- lithium which subsequently is quenched by diisopropyl¬ amine to form a second lithium diisopropylamide. However, the resulting LDA rapidly metalates the ethyl ether, forming the expected decomposition products ethylene and lithium ethoxide, as identified by NMR.

Having discovered in accordance with the present invention that enhanced stability is achieved by limiting the amount of THF, applicants repeated the

Reetz et al procedure limiting the amount of ethyl ether to 1 mole equivalent per mole equivalent of LDA. However, it was found that the resulting com¬ position is not sufficiently soluble to be of commer- cial value. The subsequent addition of ethyl ether to give a 1.7 ether to LDA ratio gave a soluble solu¬ tion. However, the thermal stability of the resulting composition, even in the presence of a large excess of stabilizer (DIPA) was considerably less than the composition of this invention (see Table I, Sample

Q). It is surprising that this enhanced stability is achieved using THF, since it is well recognized that THF if a much more labile ether than ethyl ether. In preparing the lithium diisopropylamide in accordance with the preferred method of this inven¬ tion, it has been found that significant improvements in yield are achieved by combining the majority of the total THF requirement with the electron carrier feed. The reaction vessel thus contains only the diisopro- pylamine, the hydrocarbon solvent, the lithium metal, and the remaining balance of the THF requirement. For example, yields as high as 95% have been achieved by combining 90% of the total THF requirement with the electron carrier feed, and adding the THF and electron carrier (e.g. styrene or isoprene) dropwise to the reaction vessel. The addition of most of the required THF along with the electron carrier feed permits the formation of LDA in the absence of excess free THF, and thus limits the probability of metalation of the THF by in situ formed LDA.

It has also been found to be desirable to ini¬ tially heat the reaction mixture to about 35 to 40°C prior to addition of the electron carrier feed so as to accelerate initiation of the reaction. Once the reaction is initiated, it is sufficiently exothermic

that cooling is generally required to maintain an optimum reaction temperature of about 30 to 45°C. The rate of dropwise addition of the THF in electron carrier mixture is limited by the efficiency in cool- ing the reaction vessel to maintain the desired reac¬ tion temperature. Excessive reaction temperatures are to be avoided, as this encourages the formation of side products resulting from metalation of THF, ethyl benzene and/or loss of LiH from various lithium species.

It has also been discovered in accordance with the present invention that it is possible to prepare LDA in the absence of an ether solvent. LDA has been prepared in the absence of THF by adding styrene dropwise to a stirred dispersion containing lithium sand in diisopropylamine and cyclohexane. The result¬ ing product mass was ve ^' ry viscous, contained solids and appeared to be unfilterable. The addition of 100% excess DIPA failed to dissolve the product. However, the addition of one equivalent of THF dissolved the solids and thinned the product mass to a filterable solution. While the yield from this procedure is relatively low (65%), a significant yield improvement (86%) was observed upon repeating the described ether- less preparation of LDA in the presence of about one equivalent triethylamine. The resulting composition was viscous, but again could be thinned by addition of THF. See Examples 6 and 16 below.

In the practice of the present invention, pure or essentially pure lithium metal or a commercial source of lithium metal is used. While lithium in the form of rod, shot and powder can be used, the preferred form of lithium metal for use in the present inven¬ tion, for reasons of handling and good reactivity, is particulate lithium of a particle diameter of about

500 to about 5000 microns (about 12 to about 100 mesh) known as lithium sand. It has been found desirable to use at least a 10% excess of lithium for good yields. The diisopropylamine (DIPA) and tetrahydro uran (THF) and electron carriers (e.g. styrene and iso¬ prene) are used as received from the supplier. Typi¬ cally, these reagents are colorless liquids, 99+ per¬ cent pure. As much as a total of 0.165 percent water has been present in these materials without any noticeable adverse effects in the yields of LDA. It is also desirable to use a slight excess of diisopro¬ pylamine.

Other amines which may be useful in the composi¬ tion as stabilizers include amines ranging from C. to C. - such as diethylamine, ethylamine, di-n-pro- pylamine, dibutylamines, diarylamines, dihexylamines , dicyclohexylamine, hexamethyldisilazane, butylamine, hexylamine and cyclic amines, such as 'derivatives of pyrrolidine and piperidine and even diamines such as tetramethylethyleriediamine , and other tertiary amines such as tri-n-propylamine, tri -n-butylamine , tri-n- hexylamine.

The preferred cosolvent for use in the present invention is cyclohexane. Other cosolvents which may be used to customize the product for particular appli¬ cations include toluene, xylene, n-heptane, n-hexane, benzene, and cyclic or straight or branched chain C-. to C. _π hydrocarbons such as pentanes, hexanes, and heptanes, or mixtures of paraffinic hydrocarbons such as petroleum ether.

Example 1 Preparation of LDA Complex An oven dry 500 ml three-neck, round bottomed flask equipped with a pressure equalizing dropping funnel, mechanical stirrer, Y-tube, thermometer well

and cold finger condenser was assembled while hot and purged with argon until cool. The reaction vessel was charged with 0.70 mole of lithium metal, 0.65 mole diisopropylamine and 0.056 mole dry THF. Target molarity of the LDA solution may be controlled by prereaction addition of cyclohexane. For example, add 103 ml of cyclohexane for a theoretical 2.3M solution or 251 ml of cyclohexane for a 1.5M solution. Next, the reaction mixture was heated to about 35°C using a hot air gun or heating mantle, followed by dropwise addition of a solution of 0.32 mole styrene in 0.52 mole dry THF over a thirty minute period while main¬ taining a reaction temperature of 35-40°C by cooling in a dry ice/hexane bath. The reaction was complete in about three hours. The resulting dark grey mixture was filtered to remove a small excess of lithium metal and other insolubles (e.g. LiOH) , and yielded a pale yellow to amber colored solution of LDA.

Examples 2 - 16 Preparation of LDA from Lithium Metal

Samples of LDA were prepared using a procedure similar to that described in Example 1, and employing varying amounts of lithium, DIPA, THF, and either styrene ^~ τ isoprene as the electron donor, all as set forth in Table II below. The actual concentrations are shown in the table, and the actual yield as deter¬ mined by the amount of ethyl benzene or 2-methyl-2- butene generated. Yields were also confirmed by wet analysis (total and active lithium). The above examples present several experiments showing various preparations of LDA using styrene or isoprene as an electron carrier in ether or etherless solvent and hydrocarbon cosolvent. Yield optimization was observed by limiting the amount of THF. For instance when the mole ratio of THF was greater than one

(Examples 2 and 3) the yields were 86 and 85%, respec¬ tively. When no THF (Example 6) was employed the yield dropped to 65%; however, this yield was improved to 86% by using triethylamine as a cosolvent in Example 16. A noticeable yield improvement to 93% was observed in Examples 4 and 5 where the mole ratio of THF was limited to 1 and 0.5, respectively. Still further yield improvement was achieved by combining the THF with the electron carrier feed.. This optimi- zation in procedure is clearly shown by Examples 8 through 15 where the yields are greater than 94%.

Comparative Example A A method of preparation of lithium diisopropyl¬ amide as published by F. Gaudemar-Bardone et al "A Convenient Preparation and Utilization of Lithium

Dialkylamides", June 1979, pp 463-465, was repeated as closely as possible in view of the terse description. All equipment was oven dried and purged cool with argon. The reaction vessel was a 100 ml, three neck flask (14/20 joints) equipped with a teflon coated magnetic bar stirrer, thermometer well, and argon gas inlet with back bubbler. To the reaction flask was added 20 ml (0.24 mole) THF, 20 ml (0.19 mole) anhy¬ drous grade diethylether, 0.72 g (0.10 mole) lithium sand and 6.2 g (0.052 mole) isopropenylbenzene (more commonly called alpha-methylstyrene) . The reaction mixture was stirred with the magnetic stirrer while cooling with a water bath to about 18°C. A deep red color developed about 30 seconds after the addition of the isopropenylbenzene. To the burgundy red mixture was then added slowly via syringe 10.8 g (0.107 mole) diisopropylamine (DIPA) while cooling with the water bath and maintaining a reaction temperature of about 30°C which required an eight minute addition time. The initial burgundy color gave way to an orange color

near the end of the addition. After three hours the stirring action was stopped and the solids allowed to settle. Gas chromatograph (GC) analysis of a sample of the "live" supernatant liquid indicated a very low 15% yield. The reaction mixture was then filtered through a "C" frit glass filter under argon in the usual manner to give a pale yellow solution. GC analysis of a sample of the LDA solution hydrolyzed with amyl alcohol indicated a 35% yield. A Watson- Eastham type active R-Li titration of this solution indicated a 49% yield of product (1.7M LDA in theory.) The results are presented in Table III.

Comparative Example B Example 1 of this invention was repeated using the ingredients and proportions disclosed by F.

Gaudemar-Bardone et al. To an oven dried argon purged 500 ml three-neck flask equipped with a 250 ml addi¬ tion funnel, mechanical stirrer, Y-tube, thermowell and cold finger condenser was added 3.5 g (0.504 mole) lithium sand, 53.1 g (0.525 mole) DIPA and 100 ml

(0.944 mole) diethylether (which was considered the replacement for cyclohexane in the LCA prep). Next to the addition funnel was added 100 ml (1.18 mole) THF of which 10 ml was drained into the reaction mixture. To the THF in the addition funnel was then added

31.2 g (0.264 mole) isopropenylbenzene. Meanwhile, the reaction mixture was heated to about 35°C with a heat gun. Then the isopropenylbenzene/THF solution was added dropwise to the metal/DIPA mixture over a 45 minute period while cooling with a dry ice/hexane bath and maintaining a reaction temperature of 35-40°C. The peach colored reaction mixture was stirred for three hours and then filtered in the usual manner under an argon atmosphere to yield a yellow solution. GC analysis indicated a 100% yield while titration

indicated a 98% yield (theory 1.7M LDA). The results are presented in Table III.

A very high yield of LDA was obtained, but the product exhibited poor thermal stability (1.6% loss of LDA per day). The high yield was not surprising, since we have previously seen such high yields when large amounts of ether were employed. Unfortunately, the resulting material has no practical shelf life and must be used immediately. Example 17

Example 1 was repeated but employing isopropenyl- benzene as the electron carrier in place of styrene or isoprene as described herein. The ingredients used were as follows: 6.8 g (0.98 mole) lithium sand, 68.27 g (0.675 mole) DIPA, 100 ml cyclohexane, 42.82 g (0.594 mole) THF, and 37.88 g (0.3205 mole) isopro¬ penylbenzene. The yield of LDA was calculated by two methods, GLC analysis and titration (theory 2.3M LDA), and the average percent loss of LDA per day was cal- culated. The results are presented in Table III.

The Comparison Example A was repeated a number of times and a wide range of yields of LDA (15-49%) was obtained. The stability was poor as evidenced by the loss of 2.4% LDA per day. The Comparison Example B gave high yields which were not surprising in view of the large amounts of ether present. Unfortunately, the product did not have an acceptable shelf life.

The Example 17, an example of this invention, in comparison, had good shelf life and showed a 0.0% loss of LDA.

TABLE I

LDA THERMAL STABILITY

THF/LDA

Sample LDA Cone. (Mole LDA % Loss/Day at 1 Various °C (1)

No. Route (N.) Ratio) 0°C 20.5- ^, -2.5 ^o C 27.5 ⁺ 7.S ⁶ C 40°C Stabilizer (Mole %)

A NBLC ⁶ ) 1.50 1.00 0( ) 0.06( ³ ) 0.20 TEA (5.0)

B Styrene 1.50 0.92 0(4) 0(3) 0.16 DIPA (14.2)

C Styrene 0.75 6.40 15.80 DIPA (14.2)

D NBL( ⁶ ) 1.50 2.00 0.49 DIPA (7.2) 00

I

E Styrene 2.92 1.13 0.05 O.OόC ³ 0.32 DIPA (18.0)

F Styrene 1.43 0.54 (5) 0(5) 0.19 DIPA (4.0)

G Styrene 1.60 0.85 (4) 0.39 None (0)

H Isoprene 1.54 0.82 (4) 0.20 DIPA (7.0)

I Styrene 2.29 0.95 (4) 0.13 DIPA (10.6)

J Isoprene 2.39 0.90 (4) 0.21 DIPA (4.0)

K NBL( ⁶ ) 1.50 2.00 0.17 TEA (5.6)

L NBLC ⁶ ) 1.50 3.2θ 2) 0.18 2.9 DIPA (7.0)

M Styrene 2.76 0.87 1.49 DIPA (0)

THF/LDA

Sample LDA Cone. (Mole LDA % Loss/Day at Various °C (1) No. Route (N.) Ratio) 0°C 20.5 ⁺ 2.5°C 27.5^7.5°C 40°C Stabilizer (Mole %)

Styrene 1.75 0.87 1.32 DIPA (2.7)

0 Styrene 2.76 0.89 1.16 DIPA (5.4) P Styrene 1.76 0.89 0.81 DIPA (11.8) Styrene 1.86 1.7l(7) 1.05 7.50 DIPA (27.0)

1) TEA = triethylamine; DIPA = diisopropylamine and is the amount used in excess. 2) Used t-butylmethylether (TBME) instead of THF (TBME/LDA - 3.20) o 3) A ter 30 days sample slightly turbid. I 4) After 30 days sample clear yellow containing no pptn, thus no degradation. 5) After 30 days LDA precipitated from solution. 6) NBL = n-butyllithium 7) Used ethyl ether instead of THF (Et2θ/LDA = 1.71)

0891W30087Wmd

≠J

TABLE II. LDA PREPARED USING LITHIUM METAL

Mole Ratio of Materials

Avg Temp

Example LL ¹ DIPA Styren THF (°C) Yield % ^■

2 1.00 1.00 0.500* 2.53 37 86 ^κ

3 1.20 1.00* 0.539 2.53 35 85 ^κ

4 1.00* 1.05 0.504 1.01 36 93

5 1.10 1.00* 0.527 0.523 32 93 IoV)

I

6 1.00* 1.29 0.532 θ 38 ^L 65

7 1.62 1.00* 0.513 0.105 ^B 37 ^L 89

8 1.18 1.00* ^J 0.517 0.875 ^c 35D 98

9 1.14 1.00* ^J - 0.528 0.869 ^c 36D 100

10 1.10 1.02 ^J 0.500* 0.887 ^c 38D 100

11 1.12 1.04 ^J - 0.500* 0.907 ^c 37D 99

12 1.12 1.00 0.500* 0.874 ^C 35D 95

13 ^E 1.12 1.00 0.500* 0.874 ^C 35D 95

Mole Ratio of Materials

Avg Temp

Example ^L I ¹ DIPA Styrene/Isoprene THF (°C) Yield %'

14 1.15 1.03 0.500* 0.904 ^c 35D 94

15* 1.15 1.03 0.500* 0.904 ^c 35D 94

16 1.24 1.00* 0.528 OG 35D 86

(A) LDA prepared as very viscous mixture; 0.094m THF was added to solubilize.

(B) Product became very viscous near end of reaction, so added 0.35 eq THF IV) to thin enough to filter. I

(C) Approx. 10% of THF in the reaction mixture; the other 90% added with electron carrier feed.

(D) Heated reaction mixture to about 35°C before beginning electron carrier feed.

(E) Prepared by diluting Example 12 with cyclohexane.

(F) Prepared by diluting Example 14 with cyclohexane.

(G) No THF in the reaction step (used 1.06 eq triethylamine (TEA) as a substitute); however, the final reaction mixture was very viscous with LDA precipitating, so added 0.45 mole eq THF to solubilize.

(H) Yield based on amount of ethylbenzene or 2-methyl-2-butene generated.

(I) Lithium metal sand.

(J) DIPA contained 0.165% H ₂ 0.

(K) Yield based on Watson-Eastham (active base) titration of the LDA solution.

(L) The reaction mixture required heating to approx. 35°C before the exothermic reaction would begin. * Limiting reagent

TABLE III. ANALYTICAL RESULTS FOR COMPARISON OF LDA PREPARED BY DIFFERENT PROCEDURES

Average (%) Loss

Example Yield (%) of LDA by GLC Analysis ¹ LDA by Titration ⁴ of LDA/Day ⁵

Molarity Molarity Yield

IPB Present ² DIPA Generated ³ Actual Theory (%)

Comparison A 15 35 0.82 1.67 49 2.4

I l\> Comparison B 100 100 1.73 1.77 98 1.6 I 17 79 77 2.13 2.29 92 0.0

1. Described in detail in LCA Technology Report T-21.

2. Yield of LDA calculated by amount of isopropenylbenzene present in "live ¹¹ sample versus theoretical amount if 100% yield.

3. Yield of LDA calculated by amount of diisopropylamine generated on hydrolysis.

4. A Watson Eastham type active R2N- i titration with s-butanol.

5. Titration values were used to calculate the loss of LDA per day due to instability while stored 4 days at room temperature (RT=25°C).

0895W30087Wjd