U. S. Department of Commerce Research RP1809 National Bureau of Standards Volume 39, July 1947 Part of the Journal of Research of the Nationa l Burea u of Standards

Development of Standards for Analytical Filter By Bourdon W. Scribner and William K. Wilson

Filter papers used for ordiuar.v q uali tative and q uantitative analyses were treated to obtain data for recomnlencted standards of Cjuality. The papers comprised the produ cts of two domestic and three foreign manufacturers. T ests were made for retention of fine precipitates, time of flow of water, ash, and wet bursting strength by m odi fi ed methods developed at the Bureau. Density, alpha-cellulose, copper number, and acidity were also determined. Instances of filter paper becoming too brittle for use through deteri oration h ave been observed. Stable papers are characteri zed by a high content of alpha-cellulo e, a low copper n umber, and low acidity. Furthermore, the degraded cellulose of deteriorated fil ter papers may cause erroneous analytical results. The ash, bursting strength, den sity, and time of fl ow of II'ater varied considerably for papers made for the same staled use. The alpha ce llulose con tent and copper number also varied considerably in papers made by different manufacturers. The acidity of all the papers was low. The papers showed a correlation between density and time of flow of water. Data are given showing the much greater decrease in alpha-cellulose and increase in copper Ilumber of papers of poor condition of cellulose t h an of papers of good Cjuality of cellulose when subjected to accelerated aging. Data are also included on the poor condition of cellulose in deteriorated paper as compared with paper tlutt had not deteriorated. De­ graded cellulose is soluble ill alkaline solu t ions and to some ex tent in water, and examples are given of how this may affect analyses. R ecommended standards are given for the properties considered of the most impor­ tance: a lpha-cellulose, copper number, acidity, ash, bursting strength, and t ime offiolY of water.

1. Introduction of the quality of foreign and domestic products. In a short time the Bureau was able to report that An iU'v-estigation directed toward the develop­ the new American prod ucts compared very fa­ ment of recommended standards for analytical vorably with those made abroad and , therefore, filter papers was undertaken by the National that no shortage of satisfactory filter paper was Bureau of Standards in 1942. The investigation anticipated. was prompted by the lack of adequate testing The study of the testing methods was continued methods for the valuation of such papers and the until they were considered suitable to serve as nonexistance of recognized standards of quality. standard testing procedures. The results of this The need for such information was acute at the work have been published. In order to develop time because the war had caused curtailment of standards of quality, these methods were then upply of foreign-made papers. Thus induced used to make extensive tests of filter papers of American manufacture of the higher grades not both foreign and domestic manufacture, the heretofore produced in this country, and in lieu results of wllich are reported in this paper. of the extensive experience with the foreign pro­ ducts, means had to be found quickly to evaluate II. Papers Tested the new domestic products. Testing methods Newly made papers, representative of the were involved that were adequate for comparison various kinds and grades used for ordinary analyt-

Standards for Filter Papers 21 icnl work, were tested. They were the products suggested TAP PI method for the evaluation of of the following companies: Domestic: Carl filter papers.3 chleicher & Schuell Co. ("S & S"); Eaton-Dike­ An accelerated aging test, used as a m easure of man Co. ("ED"). Foreign: W . & R. Balston, deterioration of the papers that might occur in L td ., England ("Whatman"); J. Barcham Green storage, was made by the TAPPI m ethod T453m, & Son, England ("J . Green"); Grycksbo Papers­ heat-test for relative stability of paper. The bruck A. B., Sweden ("Munktell"). T ests were test is made by heating the paper for 72 Ill' at also made on some old papers that had deterio­ 100 0 C. rated , relative to the correlation of degraded cellu­ lose with aging quality and to the eff ects of de­ IV. Test Data for Newly Made Papers graded cellulose on chemical analyses in which T est da ta fol' newly made analytical filter deteriorated papers are used . papers are contained in tables 1, 2, 3, and 4. The In the tables of test data, the papers are identi­ papers are classified in the tables according to fied only with respect as to whether they are of the manufacturers' descriptions. The quality of domestic or of foreign manufacture. The manu­ the majority of the papers, with r espect to the facturers of the papers are not identified , because condition of the cellulose and the acidity, was these data may not be representative of their very good. Stable papers are characterized by a later products. The domestic papers were ma­ high content of alpha-cellulose and low acidity. chine-made and the foreign papers were hand­ Studies at the Bureau of the deterioration of made. papers 4 have shown that the conten t of alpha­ cellulose and the copper number are good criteria III. Testing Methods of the condition of the cellulose of paper relative to the stability of papers. Alpha-cellulose T ests were made by improved m ethods devel­ is that part of the cellulose that is insoluble in a oped at the Bureau for retention of barium sulfate, strong solll tion of sodium hydroxide under defi­ time of flow of water through the papers, and con­ nitely prescribed conditions. It is a measure of th e tent of ash. 1 2 The bursting strength of wet unmodified fraction of the cellulose. The copper paper was determined by a modification of the nUJ;nber designates the amount of copper precipi­ standard m ethod T 403m of the T echnical Asso­ ta ted as cuprous oxide by cellulose under definitely ciat ion of the and Paper Industry (TAP PI) specified conditions from a solution of copper sul­ for dry paper . The followi11 g m ethods of the fate. It is a measure of the amount of unstable association were used for the other tests made: modified forms of cellulose present. An alpha­ Thickness (T411m) and weight in grams per cellulose content of not less than 95 percent, a square m eter (T410m). D ensity was calculated copper number of not more than l.0, and an from these values. Alpha-cellulose (T429m ), acidity equivalent to a pH value of not less than 5. 0, are required in the Federal Government copper number (T430m) and acidity as pH specifications for permanent record papers. The (T435m ). These are related to the condition of effects of degraded cellulose in filter paper on the cellulose and the stability of paper . chemical analysis are discussed in the following The foregoing m ethods served as a basis for a section.

1 H . Bogatyand }' . T. Carson, M easuremell t of rate of fl owofwater tli ro ugh 3 T esting analytical filter papers, T471 sm. filter paper, 1. Research N BS 33, 353 (1944) RP1613. • A. E. K im berly and B. W. Scribner, Summary report of National BW'cau ' B. W. Scri bner a nd W. K . Wilson, Methods [ or the evaluation of filter of Standards research on preservation of rccorels, N B S M isc. Pub. M154 papers, 1. Reseat'ch N BS H , 453 (1945) RPI653. (1937).

22 Journal ~£ Research TA RLt} I.- T est data 0'11 unwashed I anal!lt'icol papers

~ I ' im c of flow of water A sh per Bursting .\Ipha- Copper 1)11, hOL Thick------Sample 2 designation Ash I I-em strength, \I'eighL Density ce llulose number extration circle wet ness Average 1 Range I PAPERS ron COARSE pREC [PI TA'I' ES

Percent Percent my Points y/m' in. y/em' sec sec AI. ______9{ 0,2 7,3 0, 035 0,45 7,5 140, 0 0, 0138 0,40 14 11 to 16, A 2 ______97 ,3 7,3 , 040 , 35 3,0 83,5 , 0086 ,38 11 9 to 13, IlL ______97 , 4 7, 1 , 04 5 , 40 4,2t 87,5 ,0072 , 48 11 10 to 13, 8ti I l.i n.7 ,24 1,9 2,5 85, ° , 0104 ,32 14 10 to 17, ;:::-::: ::::_:::::::::::::'::: --: -::::: I 97 0,2 7,4 , 075 0,50 5.25 no , 0074 ,39 23 20 to 29, 1 ,4 i .O , 040 , 25 6,25 68, I ,0064 , 42 21 18 to 24, ~~~ ~: ::: :::: ::::: :::: ::::: -::-::-:::-:: I 9597 , I 7, 5 ,075 , 55 3,5 84,9 , 0074 , 45 20 17t023,

1 PAPEns FOR MEDWM PRECIPl'l' ATES

A3 ______--- 9{ 0,2 7, 4 0,045 0,35 84,0 0, 0080 0,4 1 34 25 to 43, B2 ______3 5 1 --- 97 ,3 { , O , 04U ,30 81, 8 , 0066 , 49 29 25 to 32, B3 ______--- 97 ,2 {,O , 035 ,30 95, 8 ,0072 ,52 48 44 to 5!i. B4 ______--- 97 ,0 {,3 ,to I, 7 ::'7.25 I 180,0 ,0146 , 49 41 34 to 48, B5 ______- --, 97 ,2 8,9 ,15 1,1 3,0 84,2 , 0074 . 45 19 74 to 9 1, E2 ______-- - 88 1,2 6,7 ,055 0,30 3, 75 58,7 , 0046 , 50 22 16t030, E3 ______--- 87 1,1 6,5 , 005 ,35 2,25 59.7 , 0056 , 42 45 35 to 5

97 0,3 1 7, I 0, 030 0,25 4, 25 85,6 0, 0068 0,49 79 73 to 84, ~~ _:_:::::::::_:::::::::::::: :::::::::1 97 , 2 7. (i , 10 ,9 5, 0 98,6 ,0078 ,50 81 74 to 91 , 89 ,7 8-8 , 29 1, 8 5, ° 69,3 , 0058 ,47 (il 5J to 71, 90 , 9 8,5 ,49 4,2 5,75 93, ° ,0090 , 41 60

! i\ot washed with acid . 2 P apers designated wilh the sa me \eUer arc of the same manufacturer. '}'he papers designated A. and F 1:11'(' of domestic manufacturers.

T ABLE 2,- Test dala on single acid-washed I anolytiral filter papeTs - 'rime of flow of water Ash per Bursting 1\ 1pha- Copper plJ , hot Thick- - Ash ll ·cm strength, \V ci~ ht D ens ity Famplc designation cellulose extraction number circle wet ness A\"o r'age I Ra nge I I I PAPERS FOR COARSE pnECIPITATES

Percent Percent mo Poin t" o/m' in. ojcm' sec sec PI L ______97 0,3 6,4 0,020 0,18 3,5 90,5 0,0086 0,41 8 to 10 Ell ______92 ,8 6,5 ,055 , 45 3,5 88,4 ,0096 ,36 17 13 to 25,

PAPEHS EOn !\

9- I 4,5 31 to 41 0, 2 1 o 014 1 0, 13 1 0, 0078 1 0.44 1 ~:~ - -- ::: :::::::::::::::::::::::::::::: I 8~ I 1,4 646,8 , 000 ,35 2,5 87560, I , 0066 , 36 3824 1610 36, 1 I 1 1 PAPER FOn FIN." PREC1PlTATES

B1L ______1 96 1 0, 01 5 1 °11 I 6 75 1 88 0 I 0, 0068 1 0 51 I 120 I 107 to J49'

~ __~ _____L ____ ~ ____L_ __ ~ ____~ _____ ------~--- I Papers washed with hydrochloric acid ,

Standards for Filter Papers TABLE 3.-T est data on do uble acid-washed I analytical filt er papers

~ l' im e of flow of water Ash per Bursting Sample d esignation . \ Ipha- Copper I pH, hot Ash ll-cm strength , W eight Thick- D ensity cellu lose numher Iextraction circle wet ness o\ vcrage I Range

PAPERS FOR COARSE PRECI PITATES

Percent Percent m y Points y/m' in. g/m' sec sec 1.2 6.8 0.006 11. 0 152.5 0.0142 0.42 10 A21 ~ ------~ ------93 0.08 15 to 18 A22 ______97 0.3 6. 5 . 005 .03 19. 5 64.5 . 0068 . 37 10 9 to 11 A23 ______96 . 5 6.1 . 007 .05 5.0 91. 7 .0088 . 41 16 15 to 17 B 2L ______97 . 2 6. 3 . 012 . 10 3.5 92. 0 .0092 . 39 9 8 to 10 C2L ______88 1.2 6.7 .020 . 15 6. 0 75.8 . 0074 . 40 29 19 to 36

PAPERS FOR MEDIUlvl PRECI PITATES

97 0.4 6.2 0.010 0.08 4.75 84. 1 0.0074 0.45 46 39 to 55 ~~~ ::::::::::::::::::::::::::::::::::::I 97 .2 6.5 . 010 . 09 3. 75 90.0 . 0072 .49 34 32 to 37 022 ______81 1. 8 6.6 . 022 . 18 4. 5 89.5 . 0078 . 45 24 16 to 31 E2L ______78 2.4 5.6 . 007 . 06 5.0 99.7 . 0080 .49 31 20 to 45 En ______79 1.9 5.7 . 008 . 05 3.5 65.3 .0048 . 54 56 46 to 64 I PAPER S FOR FINE PRECIPITATES

A25 ______96 0.5 6.0 0.008 0.06 5.0 88.4 0.0064 0.54 220 165 to 288 A26 ______96 . 3 6.3 .007 . 06 4.0 86.7 . 0066 .52 145 125 to 159 B23 ______96 . 4 6.3 .011 . 10 6.0 97.5 . 0088 .44 137 125 to 150 B24. ______97 .3 6. 1 . 011 . 08 4.25 75.2 .0068 .43 126 106 to 160 C23 ______. ______92 . 7 6.3 . 040 . 30 5. 0 89.9 . 0072 .49 79 55 to III E23 ______80 1.9 6. 3 . 020 .13 2.25 62. 7 . 0048 . 51 71 62 to 82

1 W ashed with hydrochloric and h ydrofl uoric acid .

T A BLE 4.- Test data on hurdened 1 analytical filler papers

Time of flow of water Ash per Burst- I Alpha- Copper pH,hot ing Thick- Sample designation Ash Jl-cm \Veight, D cn sity cellulose number extraction circle strength, ness wet Average I Range , I PAPERS FOR C OARSE PRECIPITATES

Percent Percent my Points y/m' in. y/em' SeC sec B3L ______96 0.6 6.5 0. 005 0. 04 31. 0 90.2 0. 0066 0.54 9 8 to 11 B34, ______97 . 3 6.2 .012 .10 49.5 91. 0 . 0062 .58 17 16 to 20 C3L ______84 1. 5 7. 9 . 22 2.6 30. 0 123.5 . 01 22 . 40 14 11 to 17 }'33 ______96 0. 6 8. 9 .17 1.2 49. 25 72.7 . 0062 . 46 16 13 to 20

PAPERS FOR MEDIUM PREC IPITATES

B33 ______97 0.4 6. 3 0. 018 0. 16 47. 5 96.8 0.0064 0. 59 36 28 to 44 F3L _____ . ______96 .3 9.0 . 74 6. 5 68.75 95. 7 . 0070 . 54 40 32 to 46 F 32 ______97 . 2 9. 3 . 25 1.8 49. 0 SO. 0 , 0070 . 45 21 16 to 27

PAPERS FOR FINE PHECI PI'l'A'I' E S

A31 ______97 0. 2 6. 7 0.020 15 139. 0 SO. 8 0. 0038 0.84 259 236 to 284 B 32 ______97 . 4 6. 4 . Oll 0. 10 1 61. 25 97. 4 , 004 6 . 83 132 115 to 189 I I I

1 H ardened by trcatment with acid or oLh er special treatment.

24 Journal of Research The ash, bursting strength, weight, thickness, TABLE 5. - Effect of acrelemted aging I on cellulose of filter density, and time of flow of water varied con ider­ papers 2 ably for papers made for the same stated use. Alpha-eo llulose Beta-cellulose Gamma-cellu- Copper num- 'With few exceptions, paper made for fine precip­ lose ber pH, ------hot itates retained completely a barium ulfate precip­ extrac- origi- l Change tion Original I Change °1~~1i- Change nal O~~i - Change itate prepared under specified condition .5 ------A relation between time of flow of wa tel' and Percent Percent PeT c~nt density of paper is to be expected. 95.8 -3. 0 3.4 3.3 0.8 - 0. 3 0.5 0.2 6. 0 The data show that density has an important 96.3 -3. 1 3.1 3.3 .6 -. 2 .3 . 1 6.3 95.8 - 1.5 3.3 1.8 .9 -.3 . 4 . 1 6.3 bearing on the time of flow , but that it is not 97.3 - 1.9 2.1 1.6 .6 .3 .3 . 1 6. 1 always a dominating factor. 89.0 - 7.5 10.2 7.3 .8 .2 1.4 .4 6.8 77.8 - 8,3 21. 4 8. 2 .8 . 1 2.4 .3 5. 6 80. 5 -8. 4 18.4 8.'> 1. 1 - . J 2.0 ., 6.6 v. Effects of Degraded Cellulose in 86. 2 - 12.0 12.7 12. 1 1.1 -. 1 1. 5 .5 6. 7 Chemical Analysis 1 Heated 72 h I' at lO5° C. , All acid-washed. Each pai r of " alues are for paJjcrs of difTcrcnt manu- As mentioned in section VI, when a sample fadurers. of cellulose is treated in a definitely prescribed manner with a strong solution of sodium hydroxide T AB L E 6.- Conriition of cellulose in nondeteriomted and and filtered, the insoluble portion is defined as deter-iomted filter papers 1 alpha-cellulose. The filtrate contains the so­ Dcteriorated Nondc­ called beta- and gamma-fractions. The portion Co ndition of I,aper terioraterl that precipitates when the filtrate is acidified with Sam pIc 1 Sa mple 2 ------1----- a strong mineral acid is defined as the beta-frac­ Alpha-cellulose, %. ______97.3 68.8 65.3 tion, and that which remains in solution is called Beta-cellulose, %______1. 7 30. 7 34. 1 Gamma-cellulose, %______1.0 0.5 0.6 the gamma-fraction. Although the test is an Co pper number ______empirical one, it gives some very practical infor­ 0.2 2. 8 3. 2 mation about cellulose. 1 J)apcrs of Lhe same kind and manufaciurer. For the purpose of this discussion, degraded cellulose is defined as the beta- and gamma-frac­ damaged by improper treatment during the manu­ tions and contains oxidized and/or hydrolyzed facturing proces. The papers that have been cellulose. The beta-fraction is almost always washed with acid to reduce the ash content may larger than the gamma, and in any process that suffer if the acid treatment is not carefully con­ results in an increase of degraded cellulose, the trolled, or if the acid is not carefully washed out increase is predominantly in the beta-fraction. of the fibers after treatment. This is indicated by the data in table 5 for some With the above information as a basis, filter filter papers that had been subjected to accelerated paper call be discussed wi th respect to the effect aging by heating for 72 hI' at 105 0 C. This same of degraded cellulose in chemical analysis, a sub­ increase in beta-cellulose rather than in gamma­ ject that heretofore ha.s been given httle attention. cellulose has been noted at the Bureau in work D egraded cellulose in filter papers may cause on the accelerated aging and also natural aging erroneous analytical results because it is partially of other papers of known history. Some data on soluble ill any alkaline solution, and to a lesser the condition of the cellulose of nondeteriorated extent, in hot water, in cold wator, and dilute . I and deteriorated filter papers that had not been acids. If a filtrate contained beta-cellulose, this subjected to accelerated aging are contained in fraction would be precipitated. if the solution were table 6. The deteriorated papers were so brittle acidified. If an analytical precipitation were as to be unfit for use. subseql.lClltly carried out in acid medium and the Bigh-grade analytical filter papers arc usually precipitate dried and weighed without ignition, mack from fibers having a high alpha-cellulose the weigh t would be too high. If a filtrate con­ content, generally from cotton or linen cuttings of taining soluble cellulose were evaporated to dry­ textile mills. The cellulose , however, may be ness and the res idue weighed, the weight would , Soe footnote 2. include the soluble cellulose.

Standards for Filter Papers 25 The following are instances experienced at the chemical nature of cellulose is so complex that its Bureau, where fil ter papers containing large degree of solubility is not a simple function of the I amounts of degraded cellulose have interfered with amount of degraded cellulose present, except under specifi c analytical determinations: the conditions of the method for the determination 111 atel'ial extracted from paper of poor cellulosic of alphi -cellulose. The cellulose in the filtrates quality somet.imes causes incomplete precipitation was determined by oxidation with potassium di­ of R Z0 3• The rea,ction is probably like that of chTOmate in acid solution. cer tain organi c hydro).:y acids that are sometimes An interesting phenomenon concerning cellulose used in chemical analysis to prevent the precipi­ solubility was observed by passing a second 50-m1 tation of iron and aluminum hydroxides. portion of solvent through the same circle of filter The extracted material seriously interferes in paper in the a.bove procedure. One-fourth to one­ oxidimetry by acting as a reducing agent. The half as much cellulose was dissolved by the second determination of calcium oxalate with permanga­ 50-ml portion as by the first. This slight but per­ nate, and cases where potassium dichromate is sistent solubility of cellulose, even after several , used as an oxidizing agent arc examples. washings, has been obser ved by Strachan.7 D egraded cellulose sometimes causes erroneous It is quite clear that filter paper cannot be con­ results in colorimetry. For example, an inferior sidered as an inert material even toward water. paper may produce a yellow color in the colOl'i­ metric method for the determination of chromium, VI. Recommended Standards even though chromium is absent. As filter papers containing degraded cellulose arc 1II ention should be made also that even cellulose usually unstable in storage and may cause errors of good quality is slightly soluble.6 This is shown in analytical work, the necessity for cellulose of for fil ter paper by the data in table 7 for two high quality is indicated. Quality of cellulose is papers having alpha-cellulose contents of 65 per­ defined satisfactorily by alpha-cellulose content cen t and 97 percent, respectively. Fifty milli­ and copper number. It is likewise known by liters of the solvents indicated in table 7 were experience with papers in general that if the passed at room temperature through 7 -em circles acidity is too high , on aging they will become fitted in ordinary 60-degre e . This pro­ brittle and the cellulose will become degraded. cedure cannot be consi dered as a quantitg,tive Acidity expressed as-the pH value of an aqueous method for the determination of the solubility of extra~t of the paper is a satisfactory safeguard in cellulose, but was selected because it simulated this respect. actual filtering conditions. These data are merely From an inspection of the test data, it would illustrative of what migh t occur, because the appear that an alpha-cellulose co ntent of not less than 95 percent, a copper number of not more than TARLF. 7.--Solubility of cellulose 0/ filter papers 0.5, and a pH valu e of not less than 6, could be

SAMPLE I, A L PHA-CE LL ULOSE ~ 65 PERCENT req uired. The other important proper ties for satisfactory performance of the papers are con­ Sulfuric Sod ium sidered to be ash content, strength, time of flow of Solvent 1 W ater acid. h ~ dr ox- 0.2 N O,~eN water, and retentiveness of precipitates . ------1------With regard to the ash content of the double Milliliters of 0.1835 N potassium dichromate acid-washed papers, only one showed an ash con­ requi red to oxidize cellulose in filtrate ______0.7 0.6 1. 3 Milligrams cellulose in fil trate ______. 9 . 7 1.6 tent as great as 0. 3 mg per ll-cm circle, and only three others an ash content above 0.1 mg per

SAMPLE 2, ALPHA-CELLULOSE ~9 7 PERCE"'!' ll-cm circle. It appears that 0.1 mg of ash in an ll-cm circle should be a reasonable maximum lVIilliliters of 0.1835 N potassi um d ich ro mate requirement. required to ox idi ze cellulose in fil trate. ______0.2 0.2 0.6 Milligrams cellulose in fi ltrate ______. 2 . 2 . 7 Qualitative papers usuall y are not acid-washed and therefore con tain more mineral matter than

' 50 Illl passed through 7-cm circles. the acid-washed papers. The data in table 1

' J a m es Strachan, Solubility of cellulose in water, Natme 141,332 (1938). 7 See footnote 6.

26 Journal of Research show that the highest ash content for any of tbe precipitates, 85 seconds for unwashed papers, and unwashed papers was 4.2 mg per 1] -cm circle, and 150 seconds for acid-washed papers. only 6 of the 27 papers had an a 11 con ten t greater W ith regard to the bursting strength, it is than 1 mg per ll-cm circle. A reasonable value recommended that all fil ter papers have a mini­ for maximum ash conten t for qualitative paper mum wet-bursting strength of at least 3.0 and should be 1.0 mg per ll-Cll1 circle. preferably 4.0 points. It is true that some of It is difficult to make defini te recommendation the papers tested had a lower bursting strength, for time of fl ow and reten tion of prec ipi tates, as but this is not desirable, e pecially if the papers reten tion is not a simple function of time of flow , are to be used in a Buechner . If papers and values for time of flow show such great varia­ of unusual strength are desired, attention is tion, as shown in tables I , 2, 3, and 4. Factors direc ted to the qualitative and quantitat ive papers that may influence the time of flow and retentive­ that have been "hardened" to increase their ness are the physical condition of the fib ers, strength. density, thickness, formation of the sheet, and A type of nonuniformity that j very important probably others not men tioned. In view of the is the presence of pinholes in fil ter paper. As many variables it is very unlikely that l'etentive- this is something that cannot always be dis­ I ness should bear a simple relation to time of flow. covered in the testing of random amples, the The data on time of flow of water andl'etention analyst should inspect each circle by transmitted of barium sulfate bear out this point. In an light before use. extreme case the retention of barium sulfate wa The recommended tandal'ds are nece arily incomplete for a paper with an average tim e of based on the present quality of fil ter papers. The flow of 0 seconds, and complete for a paper, usual experience of the Bureau has been that made by another manufacturer, wi th an average when an investigaton of this kind shows a large time of flow of 14 seconds. vari ation in des irable properties of materials, a A recommended procedure for selecting a general improvement in their quali ty results. fil ter paper of the propel' speed and retentiveness Improvement of the condi t ion of the cellulose in is to find by trial and en ol' the fastes t paper that filter papers and in their time of flow are par tic­ retains the material in qu estion. Inasmuch as ularly des irable, and a comparison of the test mamlfactUl'ers usually err on the side of safety data for paper of diffe.cen t manufacturers indi­ in their recommendations in this respect, it is cates that such improvemen t is feasible. often advisable to experimen t with faster papers than recommended if 'the number of The assistance of Thelma L. O'Brian, Ali ce J. wal'l'an ts the trou ble. As a general guide for Padgett, and V. Wor thington in performing the filter papers for ordinary use, maximum times testing reported herein is gratefully acknowledge d of flow may be set as follows: papers for coarse by the authors. precipitates, 20 seconds; papers for medium precipitates, 40 seconds; and papers for fine VVASHINGTON, September 11 , 1946.

Standards for Filter Papers 27