Japanese pre-printed transfer tissue results

Here are results from a test of the Japanese pre-printed transfers.  The left piece was transferred onto a fully fired coupon; the surface was first prepared with gum arabic and dried.  The right piece was transferred onto a bisque fired (1925F, 25 min.) coupon; the surface was just brushed with water first.

The bisque fired piece took up the color remarkably well.  The poor lighting and focus of the photograph do not do the best job of rendering the subjective appearance of the right hand piece, though they do a good job of showing the flaws in the left hand transfer.

These pieces were fired to 2225F, held for 25 min.  They were then coated with a little clear glaze and fired at 1925F for 25 min. (alongside a bisque firing of some other pieces destined for later tests).

I subsequently did more transfers to a few bisque fired buttons and test coupons.  When I put these into a hot kiln (around 800F after a previous firing), I noticed that the entire surface of each piece rapidly turned brown.  This suggests that the tissue paper is uniformly coated with some sort of organic sizing material before it is inked.  The paper itself is entirely peeled off before firing, so the browning is from a soluble coating that has transferred alongside the underglaze image, not from the paper itself.

The material appears to be sophisticated in its refinement, though the transfer characteristics of various designs do not seem to be especially uniform.  I suspect that the material does not have an especially long shelf life, and that older material either gives irremediably poorer transfer performance, or perhaps needs more water and a longer stay before peeling.

Commercial tissue transfers (from Japan)

Today I received from Australia, several sheets of commercially produced transfer tissue from Japan: both pre-printed patterns and blank paper.

I had been working under the assumption that transfer tissue should have fairly high wet strength.  This turns out not to be the case for this paper.  Its wet strength appears to be perhaps not as poor as ordinary wrapping or craft store tissue, but not much better.  One difference is that one surface of this paper is smoother: calendered or sized.  It has that in common with sewing pattern paper.

Meanwhile, a roll of wet strength tissue arrived today from Talas.  It is very strong when wet: at least as strong as a teabag.  After comparing it to other abaca tissues that have arrived recently, I suspect that it will have fairly poor take-up of ink from intaglio printing.  So far, nothing beats sewing pattern paper.

The pre-printed patterns use a water soluble formulation, not an oil.  Online materials suggest that one can screen print onto this paper with (presumably ordinary) underglaze.  My results with screen printing of anything, glaze or not, have been abysmal.  I probably ought to give it another try.

This tissue seems very unlikely to survive intaglio printing.

Piecing together some other discussion found online, it looks like CBM cellulose might be a useful medium for formulating water soluble glaze ink.  I have experimented with CBM as a sizing for tissue, but had better results with unsized paper.  I should try making some ink with it.

 ---

Later on...

In tried transfers onto bisque fired and fully fired test coupons.  The results are promising but not immediately consistent.  I had excpected the bisqued pieces to give better results, but the transfer seems less substantial than it ought to be.  I suspect that the ware is too absorbent and pulls the water out of the transfer before the ink is able to soften and bind.  The most promising result was on a fully fired piece.  I sized the surface with gum arabic before applying and moistening the transfer.  I then allowed the transfer to dry at a floor vent from the furnace.  Next I brushed a little water onto the back of the paper then peeled before the water had too much chance to go deeper into the ink.  This gave a fairly good transfer, though with some holes in the ink.  It is firing now, to 2225F.

Decorating fired wares

I had been working under the assumption that transfer printing was done onto bisque fired wares.  There are good reasons for this not to have been the case for calico buttons.  The buttons themselves might have been produced at one location, then transported to another location for printing; my bisqued tests are too fragile to invite being dumped into bags and loaded onto horse carts.  If the same button was sold both in both plain and decorated forms, it makes especially good sense for the decoration to have been applied to fully fired wares.

My experiments with decorating fully fired wares had been dismal failures.  Ink adhesion is very poor; most of the ink comes off with the tissue.

Recently located, however, is a book by C.F. Binns: The Story of the Potter  published in 1901.  Beginning on page 217 (the book is readily available in scanned copies online), Binns gives a popularized but detailed account of the transfer printing process as used for household pottery.

Binns on transfer printing 

He notes that "a fine varnish" was applied to the wares to promote adhesion of the ink.  Taking that hint, I took a few fully fired test coupons from early body experiments, heated them, and applied a thin coating of #8 burnt plate oil.  The results are somewhat promising.  I have been able to obtain better ink adhesion, but I have increased difficulty in removing the tissue.

Another reference (specifics not remembered) suggests that a potassium soap was used in the process.  Binns states that the transfer paper is prepared with a soap and water solution before the impression is made.  Potassium soap stands to reason here, because potassium soaps tend to be liquids, whereas sodium soaps tend to be solids.

A readily available potassium liquid soap is Dr. Bronner's.  I am trying the Baby Mild (no fragrance) type; most of the fragrances contain terpenes that seem likely to act as solvents for the oil medium.  (I use one of the household citrus-based cleaners to clean ink off plates, hands, and utensils.)

I need to experiment with soap concentration to see if there might be one that allows the ink to adhere well enough to the tissue while making the impression, yet promotes release of the tissue from the ink/varnish/body when soaked after transfer.

Also, I have a few other types of tissue paper on order.  Some are sold for conservation and book repair, and claim to have good wet strength.  Thinking a bit about wet strength, I tried printing on tea bag paper: ink transfer was disappointing to the point of being useless, possibly due to porosity, though the wet strength was much better than for sewing pattern tissue.



 

Pictures of early results

Test coupons with various pattern transfers

Buttons: undecorated, and decorated with cobalt blue

Test coupons, decorated with burnt umber (iron)

These are early results.  Ink formulations, impression technique, and especially transfer technique will need a great deal of refinement.
 

Ink formulation

My typical ink consists of a flux binder, a colorant, and an oil medium.  The flux binder is Ferro 3134 frit (lower melting than 3124).  The colorant is a metal oxide pigment.  As previously mentioned, the oil medium is #8 burnt plate oil (Graphic Chemical).

A typical formulation is 0.6g 3134 frit, 3-4g colorant, and an unmeasured amount of oil.  I heat a dish on the hotplate, pick a gob of #8 from its can with a spoon, and place the spoon into the hot dish.  The oil will thin out and flow more readily into the dish.

To the oil I add portions of the dry mix, stirring with a small stainless steel spatula.  Depending upon the specific colorant, incorporation into the oil can be straightforward or more difficult.  I continue to add dry material to the ink until it feels "right".  This is regrettably difficult to communicate, and I must admit that I am still exploring what works better and what does not.  I hope eventually to be able to give a measured formula with good proportions of oil to dry ingredients.

When cooled, the ink is somewhere between sticky tar and stiffly elastic without much surface tack.  It seems to be best used when first mixed.  Unless stored under water to exclude air, it forms an oxidized skin and performs poorly when reheated.  Even when stored under water, performance seems worse when reheated.

So far, my successful pigments are all from Gamblin.  The safest and most reliable and by far least expensive are the various iron oxides, especially Burnt Umber.  Beautiful but more expensive and arguably more toxic is Cobalt Blue: my favorite.

Viridian is beautiful in the bottle but is more difficult to mix, and dehydrates during firing to much duller chrome green.  Chrome Green would likely be a better choice, but the local art store had only Viridian in stock.

Manganese Violet is beautiful but disappears when fired, working as a flux rather than colorant.  You can see the image as a subtle texture, but not as a color.

Ultramarine is formed in a kiln, but very sadly loses the essential trisulfide from its zeolite cage when fired again, losing its beautiful color.

Some workers have reported using Mason stains in oil.  I have been unable to make a usable ink from Mason stains, though they seem to work well as body colorants.  They incorporate with the oil, but seem gritty, ink the plate poorly, and give only faint density in the transfer.

A risky experiment with Deep Cadmium Red tube oil paint gave a beautifully inked tissue, but very poor transfer to the body.  The transfer and a direct painting-on to the body, disappeared when fired.  Cadmium is cardiotoxic and volatilizes, and that will be the end of my experiments with it.

Tissue transfer: paper, ink, and impression

The literature on transfer printing repeatedly mentions a now-mythical material called pottery tissue or potter's tissue.  This was a type of paper, the characteristics of which I have been unable to locate: fiber type, surface, weight, sizing or other special treatments if any were used.

I have tried numerous types of paper.  So far, I have found only one that seems usable: sewing pattern paper.  I know of no source for this as a raw material in small quantities (or any quantity).  Fortunately, pre-printed patterns can be found for $3 in fabric stores and are thus cheaper than most anything sold at an art supply store.  The print does not functionally interfere with this work, apart from visually cluttering the impression.

Classic technique used intaglio printing from a copper plate.  Exact ink compositions seem to be poorly documented, and likely were considered proprietary when used by industry.  One formulation is said to use pine tar as a base.  A worker from a UK university suggested trying #8 burnt plate oil.  That has proven to be a good suggestion.  This stuff at room temperature is at least as viscous as molasses on a cold day.  It must be heated, both when formulating ink, and when printing.  I work in small glazed baking dishes ($1.49 each from Bed, Bath, and Beyond) on a hotplate.

For plates, I use Solarplate photopolymer plates.  These are formulated on thin sheet steel, and can be cut with tin snips or heavy scissors.  My typical plate is a couple of inches on each side, cut from 8x10" raw material.  The original image is laser printed on overhead projection transparency film.  A black line on the transparency will give a groove in the plate.  The groove is filled with ink, the surface of the plate is wiped clean, and the impression is made.  A black line on the transparency gives an ink line on the final impression.  UV light hardens exposed areas of the photopolymer, leaving unexposed areas water soluble.

These plates may be exposed by sunlight or by other sources with sufficient UV light.  I use a small UV fluorescent tube in a power supply enclosure that was built for, of all people, Avon (cosmetics) as some sort of skin condition disclosing light.  This was almost certainly a vast failure, explaining the ready availability of these units, at least some years back, from surplus dealers.

Today we also have UV LED sources readily available.  I have not yet tested one if these, except for post-development hardening exposure.

Using the small fluorescent tube, my typical exposure time is 19-22 seconds with the fixture held above the plate, transparency, and clamping glass, by a stack of two tuna cans on each side.

I work with a large sodium vapor safelight originally used for traditional darkroom printing.  The exposed image can be seen faintly on the plate even before development.

Development is in a running stream or spray of tap water.  I run the water a little warmer than cool to the touch.  I also use a very soft plastic brush (sold for cleaning mushrooms) to gently scrub.  I start with a stream of water, progress to the vegetable sprayer, then the mushroom brush, more spray, and more stream.  I then dry with a hair dryer, then post-expose for 2x50 second cycles (the Avon lamp runs for around 50 seconds if you don't cut it earlier), then a but more time with a 3.6W UV LED flashlight (wearing Uvex UV blocking glasses).

For a press I use a modified hand-cranked pasta machine originally sold to work polymer clay.  Modifications include removing the cross bars and scoops from the lower side of the rollers.  I use a small square of real etching press cushion blanket (rubber with some fiber in it - got it from one of Daniel Smith or Graphic Chemical).  I fold an 8x5.5" half sheet of used copy paper in half.  Between the halves I place the plate, with dampened tissue on top then dampened newsprint on top of that.  The cushion blanket goes on the outside of the face side.  On the back side I place another folded half-sheet of copy paper, then crank this through the pasta machine.

This curls the plate: inferior to a real press, but I already had the pasta machine, and it cost less than 1/12 of what even a small press would cost.

The next challenge is to pull the tissue slowly and carefully from the plate.  With luck, it peels without ripping, and you get a richly raised ink impression on the tissue.

Inking and impression are both done with the plate hot.  This is essential to work with the thick and very tacky ink.  Thick and tacky seems to be what is needed for transfer.

The tissue I dampen by spraying either with plain water, or with soapy water.  Which works best is still unclear.  The newsprint I spray with plain water.

Next: ink formulation.

Early tissue transfer experiments: body and forming

My goal are to try to reconstruct, using readily available modern materials, the tissue transfer technique for decorating ceramics, and to investigate dry press forming of small wares.

Without yet describing exactly what I am decorating, it must be noted that my wares are not typical of mainstream personal craft.  Rather than hand-forming, extrusion, or slip casting, I use a so-called dry pressing (historically identified as "Prosser") technique.  In this technique, the ware is formed between shaped dies under pressure, from a powder mixture containing typically around 8% water by weight.  For the small pieces that I am making, the ware can be immediately fired either to bisque or to a final state, without waiting for it to dry.

My original objective was cone 10 (2350F) porcelain, but practical considerations of firing time and poor element longevity in my small test kiln, argue for modifying the body composition with frit, to give reasonably good results at 2210F.

A typical body composition starts with 73 parts G-200 feldspar and 27 parts of EPK.  On its own, the 73/27 composition when taken to 2350F and held 30 minutes (for my tiny wares), fires slightly off-white and quite translucent.  To obtain 2225F firing, I add 6-10 parts Ferro 3124 frit.  I mix the dry powders, spray on around 8% water by weight, and transfer the mixture to a mortar and pestle, where I mix until uniform.  The result is a powder with some tendency to adhere (though not tenaciously) to the walls of the mortar, and to cohere into small platelike aggregates.  When scraped from the mortar, the mixture may be readily scooped as if it were a dry powder.

For test coupons, I form between a stainless steel disk and an acrylic disk, that both fit inside a metal sleeve (bronze shoulder bushing), using an ordinary Harbor Freight hand lever action arbor press.  For shaped objects, I use upper and lower die sets transfer-molded from original objects, with the final dies being molded in epoxy resin.

The pressing procedure is simple: situate the lower die inside the sleeve, scoop in some of the powder, pack lightly with the finger or a more loosely fitting flat die, insert the top die, then press.  The object is ejected by situating the sleeve over a void (e.g. a stack of washers or a socket wrench of carefully chosen size) and pressing with something sized to push the bottom die out entirely, and the top die out far enough to eject the ware.

The ware typically emerges with a rim of loose powder around the edge, and perhaps with some small bump defects.  These can be carefully smoothed off with the smooth skin, callus, or nail of a finger.  The ware can be handled carefully, but will usually break if dropped and certainly if subjected to any bending moment.

If this is the final form desired, without decoration, it may be immediately placed in the kiln and fired.  My objects are 5/8" to 3/4" in diameter and only 2-6 mm thick, and can thus readily survive rapid heating and cooling.

Dimensional shrinkage through firing typically runs 8-13%.

For image transfer, I bisque fire to 1750F for 25 minutes.  The body still is quite fragile to bending, but has a surface condition that seems good for transfer.  Some of my forms need holes to be drilled, and drilling works fairly well at this stage.

An alternative body composition uses only feldspar and 3124 frit, with zero kaolin.  This composition requires a binder.  For reasons related to the historic origin of the process I use casein, extracted from ordinary skim milk with white vinegar, then redissolved in concentrated sodium carbonate (washing soda) solution.  The casein + alkali + remaining water (the mixture typically is a very stiff paste after storing under refrigeration) is used at 6-8% by weight, ground in the mortar with the other ingredients, then weighed and re-ground with 8% water added by spray.  This composition typically can fire at around 1950F.  I mention this more for undecorated bodies; transfer performance without kaolin seems inferior.