FI91770C - Cationic tannin useful as a flocculant and composition useful as a flocculant in wastewater treatment - Google Patents

Description

91770

This invention relates to cationic tannins useful as a flocculant in coniferous bark extracts and to a composition useful as a flocculant in wastewater treatment.

The bark of conifers growing in Finland, spruce (Picea abies) 10 and manny (Pinus sylvestris), can be extracted with various aqueous and organic solvents and the obtained extract used for recovery. Under favorable conditions, 14 to 16% of spruce bark is extracted with hot water and about 6% of manny bark, based on the weight of the dry bark.

15 The hot water extract of spruce bark is very heterogeneous and often accounts for only about 50% of the water-soluble tannin (Stiasny number 50). If alkali and sulphite are present in the extraction of spruce bark with water, considerably higher extraction yields are obtained than if only 20 with hot water had been extracted (spruce about 25%, manty about 8%). Ethano-li-water (60:10% by volume) and acetone-water (50:50 by volume) dissolve about 28% and 30% of the spruce bark at 140 ° C, respectively. The extraction yields corresponding to manny bark are lower (18 - 20%). The tannins of spruce and manny bark extracts have a potential use in wood glues, especially if the extracts are fractionated by ultrafiltration, for example. In this case, however, fractions containing non-tannins are obtained, for which something other than the use of adhesives should be found.

30 It has been previously shown that extracts of spruce and mannyn bark, regardless of their tannin content, can be used to produce cationic products which flocculate colloidal silica and dredging sludge well (published by Pulkkinen E. and Seppanen R., "Preparation and testing 35 of cationic flocculants from agueous extracts of conifer tree barks ", Forest Products Research Society, 40th Annual Meeting, June 22-26, 1986, Washinton, USA, Abstract 8; 2 91770 and Seppanen R., Makela A. and Pulkkinen E ., "Use of cationic derivatives of tree bark extracts and lignins as a flocculant for wastewater treatment", Chemistry PSivMt 5-6.11.1986, Helsinki, Abstract 6.26, Kemia-Kemi 11, 5 1986). However, in the past, cationic tannins have not been used to flocculate colloidal substances in wastewater and to precipitate organic chlorine fractions in bleaching wastewater.

. The present invention demonstrates that the cationization mode and flocculation efficiency of the shell extracts 10 can be further substantially improved. Surprisingly, it has surprisingly been possible to prepare effective flocculants from fractions containing a lot of non-tannins.

The invention thus relates to a cationic tannin usable as a flocculant, characterized in that the tannin-containing softwood bark extract is cross-linked before it is cationized with glycidyltrimethylammonium chloride or N- (3-chloro-2-hydroxypropylammonium trimethyl) trimethyl. Thanks to cross-linking, the tannin flocculation capacity of the 20 cations is clearly improved.

The cationic tannins of the invention provide spruce and mannose bark extracts by modifying inexpensive polycations that neutralize the negative charges of particulate matter and flocculate inorganic and organic compounds such as organic chlorine compounds from bleaching effluents and further nitrogen and phosphorus.

Even small differences in flocculation power become clear in laboratory tests (dish test or test with a photometer-30 dispersion analyzer (PDA) when the colloidal reference is e.g. microcrystalline silica or cellulose). In large-scale wastewater treatment processes, the efficiency of cationic shell extracts in the removal of colloidal solids and bleach effluents from organic chlorine-containing compounds 3,91770 can be seen e.g. in settling, flotation and sludge reduction (dewatering).

In the invention, spruce and manna bark extracts have been cationized with the following epoxyammonium and chlorohydrin-5 ammonium salt reagents:

NaOH

theanine-OH and CH 2 CHCH 2 CH 2 ClO -V

10 glycidyltrimethylanunonum chloride GTAK reagent sannin-OCf ^ CHCH ^ tCHy ^ Cl0

OH

15 +

N «OM

rNaCl 1 h 2 °

CH2CHCH2N (CH3) 3Cl® + NaOH

Cl OH N (3-Chloro-2-hydroxypropyl) tert-methylyl sulfonyl chloride GTAK-HCl reagent (chlorohydrin reagent)

Aminomethylation of tannins with formaldehyde and dime-25 ethylamine also comes into question (Tappi Vol.

61, No. 5, May 1978). It is further apparent from the invention that GTAK reagent and GTAK-HCl reagent adhere to the shell extract in high yield (about 90%). This reduces the amount of by-products generated from the reagents and substantially improves the raw material cost per active product.

If the tannin content of the shell extract is high, it can be cross-linked before etherification (cationization) with a small amount of formaldehyde and a highly efficient cationic tannin can be obtained. If, on the other hand, the shell extract is crosslinked with epichlorohydrin before cationization, an effective flocculant can also be prepared from the shell extract or its fraction (Stiasny number e.g. 40) containing a lot of non-tannins.

The cationic tannins produced belong to the group of cationic polymers which are known to adhere, alone or with inorganic coagulants, to the surface of negatively charged particles contained in effluent, to neutralize the negative charge of the surface and thus to precipitating the effluent. If it is desired to increase the floc size of the particles neutralized with cationic tannin, a high molecular weight cationic, anionic or neutral linear polymer is added to the fine floc, whereby its long flexible molecular chains bind the contaminant particles into highly descending bundles.

It is an object of the invention to find a combination dose in a particular use situation in which the neutralizing agent is a cationic tannin and the crosslinking agent is a high molecular weight linear polymer, and to study the relative size and method of such a two-component dose to obtain the best flocculation result.

The flocculation efficiencies of cationic polyelectrolytes and cationic tannins can be compared by pre-25 tests. In the vessel test, the colloidal dispersion (silica, cellulose) is precipitated at different maximum flocculant dosages, whereby, after settling of the precipitate, the lowest fractional turbidity of the series clarifier indicates the optimum flocculation dose for the dispersion used. When applying the vessel test 30 to wastewater, it is often impossible to measure turbidity. Instead, the chemical oxygen demand (COD), phosphorus, or amount of organic chlorine in the series of brighteners, respectively, gives the optimum dose. Tests performed on the PDA indicate the relative size of the flock and also the endurance, 35 if the test is performed at different speeds. PDAs.

The 5,91770 test becomes especially valuable when applied to effluent. the effluent is titrated in a vessel of the apparatus with a solution of a certain stability of flocculant or coagulant, it is determined at what dose the formation of the floc begins, what is the optimum dose and how the strength of the floc can be affected by the crosslinking polymer.

Not only the colloidal fraction of the effluent, but also the organic acid salts it contains can be precipitated with sufficiently high doses of cationic tannin. The consumption of flocculant or coagulant is often too high to be economical, so recovery of the precipitant from the sludge will be considered.

The following examples illustrate the invention.

Example 1 15 This example details how the tannin of the hot water extract of spruce bark (Stias-ny Chapter 84) with GTAK reagent gives a cationic ether, the preparation and flocculation efficiency of which is optimized.

Table 1 shows the initial state of cationization.

20

table 1

Preparation of cationic tannin with GTAK reagent _Grams_ 25 _Mark Dry Mole Molar ratio

Spruce bark hot water extract Stiasny Chapter 84, ka. 47.5% 653 310 30 GTAK reagent, M 151.64

Activity 70%

Dry matter 79% 454 317.8 2.10 1 100% epoxy 35 40.9 (non-epoxy)

NaOH, M 40.01 15 0.37 0.18 6 91770

Significant in the starting situation is the high dry matter content (61.8%) and the molar excess of NaOH relative to the GTAK reagent. The reagent had been obtained from Raisio Oy.

Aqueous tannin for etherification (ka.

47.5%) was homogenized in a Baker-Perkins type screw mixer (volume about 1 liter), after which solid NaOH was added to the reactor and stirred at room temperature for 2 hours. The reagent was then added and stirring was continued at room temperature. Samples taken from the reaction mixture were flocculated in a vessel test in a silica dispersion (270 mg / 600 ml) after 14, 17, 40 and 52 hours. The reaction was quenched by neutralizing the reaction mixture with 2 M hydrochloric acid. The results are shown in Figure 1. The results are expressed as a percentage of forintin turbidity unit (% FTU) and the flocculant dose is expressed as a percentage by weight of silica.

It can be seen from Figure 1 that the cationization has taken place completely at room temperature in 40 hours.

The crude product thus obtained was diluted and ultrafiltered through a Diaflo YM2 membrane (Amicon Corp., exclusion limit 1000). The purified and dried product contained 4.4% nitrogen (determined by the Kjeldahl method).

Calculation of reagent yield From 100 g of ultrafiltered cationic shell extract, 25 were found: = 0.3143 x 151.64 = 47.66 g of GTAK reagent and 14 52.34 g of tannin.

GTAK reagent is bound to tannin as follows: 30 0.3143 X 310.0 = 88 6% 52.34

Flokkauskokeet

Vessel tests were performed on a (6-well) apparatus (Phipps & Bird) and the colloidal dispersion was microcrystalline silica (Min-U-Sil 5, Pennsylvania Glass and i 7 91770

Sand Corp.) / which was slurried in 1,800 g of 4 liters of water at pH 4.0 to 4.1 hours. During rapid stirring, the precipitates were added to 800 ml beakers containing 600 ml of silica dispersion (270 mg of silica) and mixed successively as follows: 20 min at 65 rpm, 20 min at 30 rpm, clarification after 30 min. Residual turbidity was measured with a Hach turbidometer from 25 ml samples taken 2 cm below the liquid surface. The results are shown in Figure 2.

It can be seen from Figure 2 that the cationic tannin purified by ultrafiltration precipitates colloidal silica considerably more efficiently than the commercial cationic polymer Fennopol 194 K, Kemira Oy (highly cationic polyacrylamide; M is about 4 million). In the tests with a photo-15 meter dispersion analyzer (PDA 2000, Rank Brothers Ltd), the cuvette was connected by a 3 mm hose to a 1.5 liter reaction vessel equipped with a motor stirrer equipped with a tachometer (stroposcope). Sampling into the cuvette took place from the bottom of the reaction vessel at a distance of 15 cm. The flow rate was obtained with an LKB 12000 Variopres peristaltic pump. The printout took place on a plotter. Sensitivity was obtained with a PDA-2000 analyzer (Gatin RMS, DC), so no front resistor was required. The flow rate under all conditions was 1.22 25 ml / min.

Test conditions:

Silica: 450 mg / 1,000 ml

Microcrystalline cellulose: 450 mg / 1000 ml pH: 4.0 - 4.1 30 Initial mixing: 65 rpm

Stir during test: 65 rpm Flow: 1.22 ml / min

Analyzer Sensitivity: RMS 1.07 Filter on DS 5.4 Limit on 8,91770

Determining the Optimal Dose with a Dispersion Analyzer (PDA) The PDA prints on the horizontal axis of the plotter the time that has elapsed since the flocculant dose was injected.

5 The vertical axis shows the relative floc size (in volts). The optimum dose at the relative flock size and the rate at which it is reached is the maximum. Based on Table 2, the optimum dose of cationic tannin, i.e. 0.058% silica, is achieved when the relative floc size is 10 3.95 V and the floc growth rate is 9.22 volts / min.

Table 2

PDA determination of the optimal dose of cationic tannin for colloidal silica 15 _

Dosage Relative Floc growth rate% of silica flock size (volts) volts / min_ 0.023 1.35 0.75 0.029 2.65 3.55 20 0.035 3.10 6.15 0.046 3.55 8.55 0.058 3.95 9.22 0.070 3.95 5.56 0.081 2.95 3.70 25 _

It is noted that the optimum dose obtained with PDA (0.058% silica) is higher than the corresponding dose obtained in the vessel test (0.035% silica).

Example 2 This example describes experiments in which tannin obtained from spruce bark by hot water extraction has been cross-linked with formaldehyde before cationization. The results show that due to cross-linking, the optimum blocking dose of the corresponding potassium 9 91770 thionic tannin with respect to colloidal silica is reduced and the resistance of the flock is increased.

The example starts with a bark extract with a lower concentration of Liu-sized tannins than the bark extract used in Example 5 (Stiasny number 70 vs. 84). The shell extract is first crosslinked with formaldehyde and then cationized with GTAK reagent.

a) In order to determine the gelation sensitivity of the shell extract, a gelation time determination was first performed.

Gel time experiment 30 g of hot water extract of spruce bark (Stiasny number 70) was dissolved in 100 ml of water containing 3.36 g of NaOH. 45.1 g of this solution (containing 10.4 g of extract) were cross-linked with 1.5 ml of 37% formaldehyde at 50 ° C in a Tecan gel time meter (GT 3 disc diameter 14 mm). The gel time was 138 min. Prior to gelation, the solution contained 22.3% dry extract. The amount of NaOH based on the dry extract was 11.5% and the CH 2 O content per dry extract was 0.178 mol / 100 g.

20

Table 3

Cross-linking and cationization of spruce bark hot water extract 25 _Grammaa_ __Mårkånå Dry Mole

Crosslinking 50 ° C / 30 h Alkaline solution of spruce bark hot water extract (Stiasny number 70) 30 16.7

Organic part 15.0

NaOH 1.67 0.042 CH 2 O 37% 1.24 0.46 10 91770

The solution contained 11.1% NaOH from the dry extract and 0.103 moles of CH 2 O / 100 g of bark extract.

_Grams_ 5 MMpark Dry_Moles

Etherification 50 ° C / 20 h

Reagent GTAK (M 151.64)

Reagent dry matter 10 was 79% and would contain epoxide 66.2% 19.1 15.1 12.6 0.083 (100% epoxide) 15 NSyte of the crude product was filtered through Diaflo YM2 (Ami-con Corp., exclusion limit M 1000 ), with 3.9% nitrogen in the purified product.

Calculation of reagent yield 100 g of ultrafiltered cationic shell extract was found = 0.2786 x 151.64 = 42.25 g of GTAK reagent and 14 57.75 g of shell extract.

25 GTAK reagents are bound to tannin as follows: 0.279 moles N_15 g bark extract_ 57.75 g bark extract 0.083 moles reagent °> 279_x 15-% = 87 3% 30 57.75 x 0.083 b) In the second experiment, the bark extract was cross-linked with 0.154 moles of formaldehyde per 100 g per tannin and then reacted with GTAK reagent at 50 ° C for 3 hours in a Berghoff autoclave to give a purified product nitrogen content of 3.5%. In a comparative experiment without I; 11,91770 crosslinks were cationized under the same conditions, giving a purified product nitrogen content of 3.6%.

Table 4

Effect of crosslinking on the molecular weight distribution of cationic tannins in gel filtration chromatograms

Cationized _Molecular mass_tannin_10,000 5,000 3,000 1,500 1,000 10 Non-crosslinked 36 56 68 82 89

Cross-linked 50 66 74 84 89 (0.103 moles of CH 2 O / 100 g) 15

The gel filtration method is described in Forss K., Kokkonen R. and Sågfors PE., "Determination of molar mass distribution of lignins by gel permeation chromatography", 1989, American Chemical Society, Symposium Series 20 397, Chapter 9, 125-133. Sephadex G-50 and 0.5 M NaOH as eluent.

The improvement in flocculation efficiency due to crosslinking is shown in Table 5.

25 Table 5

Effect of tannin crosslinking (CH 2 O) on the silica precipitation capacity of the cationic ether prepared from it in a vessel test 30 Degree of crosslinking Dosage Relative moles of CH 2 O / 100 g silica, flock size _% _ J [_% _ 100 rpm_ 0 3.6 0.17 - 0.20 1.2 0.154 3.5 0.07 - 0.11 2.8 35 _ 12 91770

Example 3

This example further describes the effect of CH 2 O crosslinking of tannin on the flocculation efficiency of the cationic ether prepared therefrom.

5 L & htOtannin is the same as in Example 1 (Stiasny Chapter 84). In the crosslinking step, 278.7 g (ca. 43.3%) of wet tannin was weighed to give 120.7 g of a crude shell extract, of which 0.46% (0.0153 moles of CH 2 O / 100 g of shell extract) was used, based on formaldehyde. .

CH 2 O was added as a 37% aqueous solution and NaOH as a 5 M solution until the pH of the reaction mixture was 12. Cross-linking took place in a 1.5 liter flat glass flask, allowing the reaction mixture to stand at room temperature for 2 hours and then at 45 ° C for 4 hours.

To the cross-linked mixture was added cationization reagent (GTAK, ca. 70%, active ingredient 60%) 80.19 g (0.529 moles calculated as active ingredient) and the reaction mixture was homogenized with a blender (Braun). The mixture was allowed to stand for 2 hours at 20 ° C and 22 hours at 40 ° C.

The stiffening effect of formaldehyde on the reaction mixture was already visible at room temperature, but the solid mass in the slurry softened to a viscous liquid during the etherification. After ultrafiltration, 3.6% of the neutralized crude product contained nitrogen.

The reagent yield calculated as follows was remarkably high: ^ = 0.257 x 151.64 = 39.0 g of GTAK reagent and 14,100 g to 39.0 g * 61.0 g of tannin.

30 GTAK reagents are bound to tannin: 0.257 x 120f7_ = g61 61.0 x 0.529 13 91770

Effect of crosslinking on flocculation efficiency

Although the crosslinked cationic tannin contained 4.4% nitrogen (Example 1) and the crosslinked tannin 3.6%, both cationizations were performed on the same bark extract, so flocculation results can be compared if the doses are calculated for cationic tannin nitrogen. towards. It is clear from the results shown in Figure 3 that crosslinking of formaldehyde reduces the optimum dose in the vessel test and increases the 10 relative floc sizes in the test performed on the PDA.

Example 4 This example demonstrates that my epichlorohydrin cross-links a high non-tannin-containing hot water extract of spruce bark (Stiasny Chapter 40), thereby significantly improving the flocculation efficiency of the cationic-15 product due to cross-linking.

2 g of spruce bark hot water extract (Stiasny number 40) dissolved in a solid NaOH solution (0.3 g of NaOH and 2 ml of H 2 O) was stirred at 50 ° C in a glass vessel for about 1 hour. Crosslinking was performed by adding 0.1 g of epichlorohydrin (ECH) to the reaction vessel and stirring was continued at the same temperature for 10 min.

To cationize, 1 g of GTAK reagent (calculated as 100% epoxy content, 6.7 moles) was added to the crosslinked reaction mixture and heating was continued at 50 ° C for 2.5 hours. The reaction was quenched by neutralizing the mixture with 2 N HCl. The crude product was dried to constant weight at 50 ° C. A portion of the dry product (1 g) was slurried at room temperature in 160 ml of 94% ethanol, filtered and the precipitate was washed with diethyl ether and dried to constant weight in a vacuum desiccator. The purified epichlorohydrin cross-linked cationic spruce bark extract was obtained by ethanol washing to 61% of the dry crude product and had a nitrogen content of 2.6%.

14 91770

Uhtfitllanne

Grams Grams in mmoles Molar ___ ratio

Bark extract 5 Stiasny chapter 40 2 2

NaOH (40) 0.3 0.3 7.5 1.12 7 ECH (92.53) 0.1 0.1 1 GTAK (151.64) 1 1.0 6.6 1 H2O _2_ _ 10 5, 4 3.4 dry matter 63% LShtO situation calculated per 100 g of shell extract: 15

Grams Grams Moles Mole _Ratio_

bark Extract

Stiasny number 40 100 100 20 NaOH (40) 15 15 0.375 1.14 7 1 ECH (92.53) 5 5 0.054 1 0.14 GTAK (151.64) 50 50 0.330 1

HjO 100 _ 270 170 dry matter 63%

Figure 4 shows the effect of crosslinking the shell extract on the silica precipitation capacity of the corresponding cationic ether 30 (Min-U-Sil-5, silica 150 mg / l) in a vessel test. The cationic tannin (2.6% nitrogen) prepared from the cross-linked bark extract gave a significantly lower optimum dose than the non-cross-linked cationic tannin (nitrogen 2.7%).

I: »9Ί770 15

Example 5 In this example, vessel tests with colloidal silica as a reference material have shown that when a small amount of Al 3+ or Fe 3+ salt is added to a solution of cationic tannin, the flocculation capacity of cationic tannin is improved. This is shown in Figure 5.

Example 6 This example describes the interaction of a cationic tannin and a linear cationic polymer on the flocculation of microcrystalline cellulose in a vessel test. The results are shown in Figure 6.

The dose of cationic tannin (0.25 ppm) selected in this experiment is not a clear cellulose dispersion alone, but when adding a number of linear cationic polymers 15 Fennopol K211, Kemira Oy (a slightly cationic polyacrylamide) is about 8 million; ppm. In this experiment, the optimum total dose was 0.38 ppm (20% ratio of cationic tannin (3.9% nitrogen) to Fennopol K2li 66:34).

Example 7 In this experiment, a combination of cationic tannin and a combination of cationic tannin and inorganic coagulant (Al sulphate, Feklor idi) was added to wastewater from a pulp mill 25 which had been biologically treated and thus intended for discharge into discharge water. The vessel test was performed with 600 ml of wastewater to which the above flocculants were added, the mixture was stirred and, after a certain time, allowed to settle. Samples from the supernatant were assayed for solids, COD, AOX (organic chlorine), color, total phosphorus, and total nitrogen. The results are shown in Table 6. It can be seen from the results, among other things, that the amount of phosphate can be significantly reduced by adding cationic tannin either alone or in combination with 35 coagulants. The results further show that cationic tannin reduces the amount of soluble aluminum when used as a precipitant in combination with Al sulfate.

i 16 91770 • η in 4J: 0 oo lo σ lo lo in Cd) «.«. * · «... o o in d) 3 EP ΙΌ IN ^ n fM (N ·« · «« i

in: 3 ^ g cn> -h ro <-η oo cn rH

• Η ί3 rH \ tn O <tn

O

rH

o

B

XI n i — i ih vo νοΓ 'm oo r ~ ro v1 «« «« ««. «. «« «« 3 O m in o ld in ro cm t-ioroi — i 3 λ; ro ro r- lo r- ~ ro lo r- oo 3 2 + + 4-> .c d)

4-> X O LO rH O O LO LO LO rO CN

3 o «« «« ««. «« «« «

ι-H x N * in N "rH ro lo O (Λ O N1 VB

i — t ¢ 1 (ro ro m oo oo ιο σ ro οι σι m d) in C 0) ind) -H lo inooro lo σ m in σ »a · 3 cn p -« «« «. *« «« «

4-> -P: 3 IN IN O ΓΟ CN ΓΟ CN LO in IN LO

3 g> r- oo oo σ σ n> σ in oo σι on X 3 • P 4-> os in> -P c #> χ σι ro lo oo oo o ro ro σ O O ««. «. «. ««

Cdi o c o lo m oo σ cn lo oo

LO d) Ή Ν 'Ν' ro m in N <r- ^ LO

p 3 -P

0 -P d) X

X -P Ό 3 X 3 -P T3Q ocmooooo loin -sr σ ro in 3 3 d) d) 0 «« ««. «. «.«. «« «« Η i — IC PiU n · · * τ o σ σ in · * τ r- ~ in ro lo 3 3 -H n * n · n * lo lo lo oo rHLor ~ oo rO Oo 3

Eh 3 I

O 3 I

X -P O

+ J + J

3 -P 30) lo i ^ cn σ d) 3 -P C I II * · ·. ^ v 4-> x: -h-η lo il in σ r- r- in X 3 r- σοοοοσ • H 3 3 d) 3 3 3 · ρ. Cn g

p: 3 ίο O o O

: 3 -P 05 IN to oo ^

Eli rH Oro. — I IN Ν 'σ O): 3 Cg o o o oo lo o o \ \ \ \ tn 3 ^ o lo o oo o o o Loinroio 3 -p <01 cn cn m lo σ σΐΝ in η ^ «

* (1 CO rH «—Η rH CN

rH γΗ

C C

• Η (D -Η 3 3 -Ρ Η II) 4-> · Ρ Ρ> 3 T3 3 0) 3 · Η 3 4-1 4Η Ρ 3: 3 · Η γ — ι Ο Η-> -1—1 3 · Η 3 Ή • Η 4J .—. ιη Αί

3 3 -Η - «4-1 · Ρ d): 3 I I

0)> 1 3 æ 3 Τ3 IP g Ή 0) in rH 3 3-HiH <Cfc (• Η rH 3 LO M-H p c #> d) ^ \

Bd) 4-> * rH O -P LO LO

04-> ^ 3 rH U0 tn «> -P -P. in X «-P N> 4-) CO 4-> · I I CN Ό 3: 3 3 2 Ή d) Ή, 3 ·· ··. XX X ~ c En ~ tn 2 2 I,

Claims (6)

1. Cationic tannin which is useful as a flocculant, characterized in that the tannin-containing bark extract from coniferous wire has been crosslinked for the cationisation thereof by a glycidyltrimethyl-ammonium chloride or N- (3-chloro-2-hydroxypropyl) trimethyl chloro trimethyl chloro 2. Cationic tannin according to claim 1, characterized in that formaldehyde is used in the cross-linking. 3. Cationic tannin according to claim 1, characterized in that epichlorohydrin is used in the cross-linking. 4. A composition which is useful as a flocculant in wastewater purification, characterized in that it contains a cationic tannin according to any one of claims 1-3. 5. A composition according to claim 4, characterized in that it additionally contains an inorganic coagulant. 6. A composition according to claim 4 or 5, characterized in that it additionally contains a high molecular weight cationic, anionic or neutral linear polymer. * li 9