Stokes' Kiln in The Engineer

Read The Engineer at Grace's Guide.

Stokes’ rotary kiln patent of 1889 was described in detail in an article in The Engineer: Continuous Portland Cement-Making Apparatus, Vol 71, 16/01/1891, page 53. In the ensuing weeks, an extensive exchange took place on the Letters pages, beginning with some tongue-in-cheek enthusiasm from Bertram Blount, who was probably aware of the real problems being encountered by the pilot installations. He proceeded to politely emphasise these. Blount shows off his understanding of the latest developments in cement chemistry, and implies, with some justification, that Stokes – an engineer – had little understanding of the technical details of cement manufacture. All the subsequent discussion shows the enormous difficulty of understanding clinker formation without any practical means of measuring free lime - the first free lime test was developed 40 years later.

From 23/01/1891 p 76:


SIR, - I have read with much interest the account appearing in THE ENGINEER for January 16th, of the continuous Portland cement-making apparatus devised by Mr. W. Stokes. Considering the importance of the cement industry, it is certainly high time that a continuous process should displace the intermittent system still generally used. Nothing more sharply differentiates modern processes from those becoming or become obsolete than the adoption of the luminous principle of continuous working, and the inherent difficulties in the way of its application in this case appear to have been met in an adequate manner. That the economy of fuel that must result from the use of Mr. Stokes' process is considerable need scarcely be questioned, looking at the care taken to utilise the whole available heat; but one point which the description of the process leaves untouched, merits, I venture to think, a little attention.

In burning Portland cement the first reactions that occur are the expulsion of the carbonic anhydride from the chalk and the combined water from the clay, both effected at a moderate temperature, and needing no special effort for their attainment. The caustic lime derived from the chalk then acts upon the silicate of alumina constituting the clay to form with its silica the tri-calcium silicate 3 CaO SiO2 which Chatelier has shown to be the chief ingredient of cementitious value. But as the particles that are to take part in the reaction are relatively coarse, means must exist for bringing them into the intimate contact necessary for the combination of the molecules composing them. This means is found in the calcium aluminates which are simultaneously formed, and being more fusible than tri-calcium silicate, act as a flux, bringing about the union of silica and lime at a lower temperature than would otherwise be possible. In spite of the aid thus afforded, full and due combination of the raw materials is not readily attained, and the production of properly burnt clinker can only be secured by regard being had to several factors. There is little doubt that one of these is time. The lime and silica on whose union the success of the product depends, can only combine when brought into juxtaposition by the aid of the fluxing action of the aluminates mentioned above. As this is not great, and as the clinker at most sinters - never truly fuses, or it is useless as cement - the motion of the molecules it is desired to combine is necessarily slow, and in order that they may have sufficient chance of collision and ensuing union, they must be kept at a sintering temperature for a very appreciable period. This, then, is the real difficulty with which a continuous process has to contend. The dried slurry in being converted into clinker can be exposed undoubtedly to a temperature as high as that obtaining in an ordinary cement kiln, but it is by no means certain whether the time during which it is kept at that temperature is sufficient to insure the completion of the reactions I have briefly sketched. On this point, therefore, information would be very welcome. Will the clinker thus burnt prove that in it thorough combination has taken place by yielding cement of as good quality as that produced in the ordinary manner? Mr. Stokes may have in his possession evidence that will decide this question.

Laboratory, Broadway, Westminster, BERTRAM BLOUNT. January 17th.

The following week, 23/01/1891 p 85, Stokes replies, intent on showing that he has been on a Cement Technology Course:


SIR, - I have much pleasure in answering the question Mr. Bertram Blount raises in his letter, appearing in this week's ENGINEER, on my continuous cement process. So far as badly ground slurry is concerned, time is distinctly an element in the proper burning of cement, as the larger particles of chalk then take a considerable time to soak up, so to speak, the silicate of alumina. But when the slurry is properly prepared no difficulty arises in practice, and a good clinker can readily be obtained in the ten to fifteen minutes during which the cement materials are at their maximum temperature. I think Mr. Blount emphasises too much the distance through which the molecules have to act, which, when the slurry is properly "washed," is almost infinitesimal. I may mention that I have succeeded in burning a small quantity of cement by passing slurry powder through an oxyhydrogen blow-pipe flame. The amount burnt was necessarily small - only enough for a small pat - but it appeared to have all the properties of good Portland cement. In this case, of course, the time available for the reactions, referred to by Mr. Blount was only a very small fraction of a second, and yet to all appearance the necessary combinations took place.

The evidence of numerous pat and tensile tests shows that properly burnt cylinder cement is fully equal to kiln-burnt material, and that there is no evidence of an excess of free lime, which would be the case were the reactions incomplete. Microscopic examination of my clinker also confirms this, no particles of lime being visible. I am gratified to notice that Mr. Blount takes the economy in fuel as considerable. ln this he is quite correct, as the total cost of fuel, labour, and power, is only 3s. 6d. per ton of cement burnt, i.e., a saving of from 6s. to 7s. as compared with the ordinary kiln process.

WILFRID STOKES. 9, Victoria-street, S. W., January 26th.

There followed on the same page:


SIR, - In yesterday's ENGINEER Mr. B. Blount asks whether time is not an element in the calcination of Portland cement and seems to infer that therefore the continuous process of manufacture proposed by Mr. Stokes in your issue of the 16th inst. may not succeed. From trials recently made in the manufacture of a special cement, and from previous experience, it appears to me that time - not a very long time - is an essential element but with a revolving calciner there would not be much difficulty in regulating the speed of rotation and inclination of the rotator so as to retain the raw material sufficiently long under a soaking heat. However, I do not think the form of calciner, with desiccator attached, proposed by Mr. Stokes - which is, so far as I recollect identical with that patented in 1888 by Mr. Thwaite - would succeed, and that it would fail in the same way and from the same cause as Ransome's. A revolving calciner has been patented by a well-known specialist, which I am very desirous of trying; and if any of your readers feel inclined to join in a trial, I shall be happy to correspond with them on the subject.

W. SMITH. 1, Clyde-road, Dublin, January 24th.

Three weeks later, 13/02/1891 p126, Blount suggests in a roundabout way that Stokes' cement is unsound, while scoring one-upmanship points by citing a French journal:


SIR, - Remarking parenthetically, in reply to Mr. Wilfrid Stokes' remarks as to the completeness of the combination of the constituents of Portland cement produced by his process, that the distances between adjacent particles of chalk and clay even in well washed slurry are, molecularly speaking, colossal, I will, with your permission, narrow the question to the highly practical point - is the cement produced of really high quality? Mr. Stokes has put forward evidence that it is, and until quite recently the tests upon which he relies have been the best available. Specialists in Portland cement are well aware that the most conclusive of all tests are those extending over one, two, or even more years, when latent defects which are overlooked by the ordinary short-time tests often become apparent. In particular, the presence of a small proportion of free lime has been hitherto practically impossible to detect, but would nevertheless make itself felt in the long run. Most fortunately, we have now for the first time a means of ascertaining with considerable certainty, and in a short time, the quality of cement which could only be pronounced upon by older methods after the lapse of a period impracticably long.

In the Bulletin of the Société d'Encouragement pour l'Industrie Nationale M. Deval has elaborated a method of hot testing , which has proved in his hands capable of nicely discriminating between good and indifferent cements, and which will, doubtless, speedily find a place in the specifications of engineers using large quantities of Portland cement.

Had this method been generally known a short time back, Mr. Stokes would have been able to advance still more conclusive proofs of the good quality of the cement made by his process. I may say that I am so impressed with its utility, both for such experimental purposes and for supplementing the ordinary routine of cement testing, that I am taking steps to introduce it into our cement laboratory here.

BERTRAM BLOUNT. Laboratory, Broadway, Westminster, S.W., February 2nd.

On the same page, Stokes responds to the attack on the originality of his patent:

SIR, - I cannot allow Mr. Smith's remarks on my process to pass without correction, but as his information appears very incomplete, he will no doubt be glad to be set right. Mr. Thwaite's patent – the complete specification of which is dated May 19th, 1888, and which I presume is the one referred to - has no resemblance whatever to my own, except in so far as that revolving furnaces are proposed to be used in both. The gist of the proposal is to divide the burning into two stages in separate furnaces, leaving the preparation of the slurry to "a machine suitable for the purpose," and the cooling of the clinker to chance; the heating of the air being effected by regenerating chambers, &c. As regards revolving furnaces for cement burning, Mr. Smith again appears misinformed. At the present time new works are almost completed near Grays, where six revolving furnaces are to be used, their adoption being considered justified by the experience obtained with Ransome's furnace at Messrs. Gibbs' works. This does not look like the revolving furnace being a failure for cement burning. It is true that although upwards of 900 tons of cement have been burnt by Ransome's process, it has not been altogether a success, and for three very good reasons: (1) The endeavour to burn as a powder; (2) the absence of any attempt to prepare the slurry for burning; (3) the lack of means for cooling the clinker.

In my apparatus, however, those difficulties are overcome, as the slurry is automatically dried and delivered to the furnace in small sticks, which gets over the powder difficulties; and the burnt clinker-cooling cylinder not only cools the clinker, but also does away with a regenerative system of flues, which were a constant source of trouble to Mr. Ransome. With these wide differences from Ransome's process, I fail to see how Mr. Smith arrives at his conclusion that mine "would fail in the same way and from the same cause as Ransome's."

Just one word of warning to any of your readers who may feel tempted to join Mr. Smith and "the well known specialist" in a trial of the new "revolving calciner." A furnace pure and simple won't do. The apparatus must deal with the slurry as it leaves the pumps, and must carry on the operations of drying, burning, and cooling continuously, and without hand labour, otherwise a thorough commercial success cannot be obtained.

9, Victoria-street, S.W., WILFRID STOKES. February 3rd.

Also on the same page, a new critic puts forward the general view common among those with static kilns, that making clinker takes a long time. Modern rotary kilns complete the whole process in 15-20 minutes. A mention of a six-day working week hints at another concern of the many manufacturers who still insisted on shutting their plant down on Sundays.

SIR, - In your issue of 30th ult. Mr. Stokes - in reply to the letter by Mr. Blount, which appeared in your issue of 23rd ult. - states that the short time required to produce a good cement clinker by his continuous process is principally based on the assumption that the slurry is properly washed, but admits that in the case of badly-ground slurry, time is distinctly an element in the proper burning of cement, as the larger particles of chalk then take a considerable time to soak up, so to speak, the silicate of alumina.

Now, all cement manufacturers being aware of the absolute necessity of washing their slurry to the finest degree, generally provide that the slurry shall pass through a fine sifter of an approved number of meshes, so that as a rule, it is of an uniform consistency, and any "larger particles" are not allowed to leave the washmill until thoroughly dissolved, and any attempt to produce cement from slurry improperly washed would only prove disastrous in every way.

What, then, would Mr. Stokes define as properly washed slurry? for thus far his explanation is applicable to all processes of burning good cement clinker.

It would, no doubt, be interesting to many of your readers to know whether any other method superior to the washmill is adopted by Mr. Stokes for obtaining a more perfectly ground slurry than the washmill or other known methods produce.

Mr. Stokes has succeeded in obtaining satisfactory results from his apparatus on a small scale only, but if he could give some idea of the quantity of clinker, approximately, which his apparatus would be capable of turning out - say, during six working days with the amount of coal-gas likely to be consumed, your readers would be better able to form a more definite idea as to the value of such a process, as against kiln-burnt or other existing methods.

The time which Mr. Stokes allows - only ten to fifteen minutes for the thorough calcining of the slurry into cement clinker, is so very short, as compared with other methods, that this fact alone, to say nothing of the great saving in fuel effected, would, if practicable on a large scale, want no other recommendation for its adoption by all large cement manufacturers. From numerous and exhaustive tests and experiments, during several years' experience as a manufacturer of cement, with a view to determining the minimum length of time necessary to properly burn and produce good clinker from the dried slurry I regret to say that under the most favourable circumstances I have never succeeded in bringing the minimum of time under two and a-half hours and this at a temperature of about 1600 deg. Centigrade, and any attempt to hasten this process by the introduction of additional heat only proved useless, the result being that the outside crust of the slurry is simply burnt to a clinker as hard as, and similar in appearance to, the ordinary coal clinker formed on furnace bars, whilst the inside being partially protected by this coating remains almost in the same raw state in which it first entered the kiln.

PORTLAND, 20, Falsgrave-road, Scarborough, February 9th.

The next week, 20/02/1891 p 146, Stokes replies to the last comment. In justifying his patent, he provides a "calculated" heat consumption of his kiln. This works out at 5.96 MJ/kg. The lack of any empirical data shows that he has not had a production run that would in any way bear out the calculation. It is amusing to re-do the calculation in the light of modern knowledge - the heat consumption was probably at least three times that given.


SIR, - As requested by your correspondent “Portland”, I now send you the following information, trusting it may be of interest to your readers:-

(1) I consider that slurry is properly washed if on breaking a dry piece no white specks of chalk are visible.

(2) Where the old “wet” process of slurry mixing is employed , the wash mill is quite sufficient for my purpose; but in the case of the “semi-dry” process, the slurry should also - as “Portland” no doubt well knows - be passed through “wet stones” in the usual way.

(3) With respect to the consumption of coal-gas in my small apparatus, this bears such an indefinite relation to the amount of producer gas used in a large plant that it would serve no useful purpose were I to give it. I propose therefore to give figures relating to a full-size apparatus of, say, one or two tons per hour output.

The economy in fuel being the most important factor in the saving of my apparatus, and wishing to know the practical limit of possible economy, I employed a well-known specialist in such matters to report upon the total quantity of heat required for the various operations, and as his report is most interesting, I propose to give briefly the results. In my apparatus the heat expenditure may be divided into five heads:-

(a) Heat needed to decompose the calcium carbonate33.43%
(b) Heat left in the issuing clinker0.65%
(c) Heat used in drying the slurry34.80%
(d) Heat retained by the exit gases22.03%
(e) Heat lost by radiation, &c., say9.09%

Strictly speaking, some allowance should be made for the heat absorbed or evolved during the formation of the various silicates and aluminates, but no data exists as to how much or how little this should be, and it is therefore disregarded as presumably unimportant.

The following figures represent the number of B.T.U.’s required per minute, under the above-mentioned five heads, for the production of cement clinker at the rate of one ton per hour:-

(a) Decomposition of 42 lb. calcium carbonate32,000
(b) Heat left in 37⅓ lb. of clinker at 84 deg. Fah. above the air temperature622
(c) Heat required to dry 84 lb. of wet slurry under favourable conditions33,320
(d) Heat retained in exit gases if at 700 deg. Fah. above the air temperature21,087
(e) Heat lost by radiation, &c., say 10 per cent. of the above8,703
Total heat units95,732

This is equivalent to 6.6 lb. of good coal per minute, or, say, 3½ cwt. per hour. This figure compares most favourably with the 11½ cwt. of coke at present required in kiln burning. The cost of the above 3½ cwt. of coal may be taken as about 2s. 7d. The cost of labour and power in a two tons per hour plant amounts to about 1s. 2d. per ton - assuming only one apparatus at work - that is a total cost of 3s. 9d. per ton for drying, burning, and delivering cool. To “Portland”, and those of your readers who are connected with cement making, these figures will be sufficient to show that my claim for economy is a substantial one, and has a very wide margin of safety.

My own experience with revolving furnaces for cement making quite bears out the result of the foregoing calculations, as I have frequently seen cement burned at the rate of about 400 lb. of coal per ton of cement, and this in cases where the heat contained in the burnt clinker, amounting to 21 per cent. of the total heat evolved, was not utilised in any way.

(4) With regard to “Portland's” difficulty in burning quickly in a sample kiln, this is by no means uncommon, and is partly due to the proportionately large amount of heat required to heat up the kiln itself, and partly to the size of the pieces of slurry burnt. It is obvious that the volume of a piece of slurry has a very great influence on the time required in heating thoroughly during the successive operations, constantly requiring fresh supplies of heat, in driving off the contained moisture, in decomposing the calcium carbonate, and in guarding against over-burning the outside, leaving the inside under-burnt. It is also obvious that the conditions of a. revolving furnace are much more favourable to the quick development of the chemical changes necessary in cement burning than the present form of kiln.

I am much interested in Mr. Bertram Blount's account of the new hot method of cement testing, and will take an early opportunity of availing myself of its inauguration at the Broadway Testing Works. I apologise for occupying so much of your valuable space,

WILFRID STOKES. 9, Victoria-street, S. W., February 14th.

Three months later, on 22/05/1891 p 412, Stokes gives an update on the "hot test" results on his cement, having commissioned work from Blount.


SIR, - I have read with much interest the able leading article appearing under the above heading in this week's issue, and having recently had recourse to the use of the new hot test you advocate, I should like to add my testimony as to its utility.

Some of your readers may perhaps remember that a correspondence recently took place in your columns on the subject of my continuous cement process, in which the question as to the possibility of quickly burning sound cement was raised by Mr. Bertram Blount, and also by others who had not been able to obtain good results from shorter burnings than 2½ hours. Although personally quite satisfied by my own experiments in this direction, I deemed it advisable to obtain the independent testimony of an acknowledged expert, as suggested by one of your correspondents. I therefore placed myself in communication with Messrs. Stanger and Blount, feeling sure that, being already inclined to the “long time” view, their verdict would carry the more weight.

Their researches naturally occupied some time, so that I have only received their report during the last few days. It being of an exhaustive character, I may only quote the following conclusion, which, I think, sums up the case. Messrs. Stanger and Blount report “that Portland cement of high tensile strength and perfect soundness can be made, under the extremely adverse condition implied by the use of a slurry yielding a product high in lime, by exposure, in pieces of small section, to a clinkering temperature for as short a time as ten minutes.”

There can, therefore, now no longer be any question on this point. A considerable number of tests were made on clinker burnt in varying lengths of time, the shortest being nine minutes, which, oddly enough, gave abnormally high results, owing to the large proportion of lime: - five 1 in. briquettes, gauged one part cement to three of standard sand, at the end of twenty-eight days in cold water, broke at 350 lb., 375 lb., 380 lb., 425 lb., and 425 lb. per square inch respectively.

The new hot test was applied to numbers of sand briquettes and neat pats, with the satisfactory result that in all cases the cement was proved, beyond dispute, to be of perfect soundness and increasing strength. Four 1 in. sand briquettes made from a twelve minutes' burning may perhaps be taken as typical of the results obtained:-

Pounds per square inchAverage
7 days' hot test275310292.5
28 days' cold test290330310.0

These results, it will be seen, are very considerably above those required by the most stringent specifications. I would also like to mention that similar sand briquettes, made up from ordinary commercial cement burnt from the same batch of slurry as the above, and which had passed the ordinary cold neat tests, gave the following results after seven days’ hot test:- 25 lb., 45lb., and 50 lb. respectively; average, 40 lb. per square inch.

These results are most eloquent as to the utility of the new hot test in detecting faulty cement, and it is only to be regretted that the difficulty in maintaining a constant temperature in any but a properly fitted-up testing establishment will render its general adoption somewhat difficult, and in inexperienced hands somewhat uncertain.

WILFRID STOKES. 9, Victoria-street, S. W., May 19th.

However, the most eloquent evidence on the success of Stokes' kiln is the fact that Stokes never sold an ounce of cement.